Conceptualizing Music:

Cognitive Structure,

Theory, and Analysis

Lawrence M. Zbikowski

OXFORD UNIVERSITY PRESS

CONCEPTUALIZING MUSIC

AMS Studies in Music

Lawrence F. Bernstein, General Editor
Editorial Board
Joseph Auner
Scott Bur nham
Richard Crawford
Walter Fr isch, Chair
Sarah Fuller
Robert Judd
Janet Levy
Jessie Ann Owens
Kerala Snyder
Judith Tick
Gary Tomlinson

CONCEPTUALIZING

MUSIC

Cognitive Structure, Theory, and Analysis

Lawrence M. Zbikowski

2002

Oxford New York

Auckland Bangkok Buenos Aires Cape Town Chennai
Dar es Salaam Delhi Hong Kong Istanbul Karachi Kolkata
Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi
So Paulo Shanghai Singapore Taipei Tokyo Toronto

Copyright 2002 by Oxford University Press, Inc.

Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York 10016

www.oup.com

Oxford is a registered trademark of Oxford University Press

All rights reserved. No part of this publication may be reproduced,

stored in a retrieval system, or transmitted, in any form or by any means,

electronic, mechanical, photocopying, recording, or otherwise,

without the pr ior permission of Oxford University Press.

Library of Congress Cataloging-in-Publication Data

Zbikowski, Lawrence Michael.

Conceptualizing music : cognitive structure, theory, and analysis / Lawrence M. Zbikowski.

p. cm. (AMS studies in music)

Includes bibliographical references (p. ) and index.

ISBN 0-19-514023-0

1. Musical perception. 2. Musical analysis. 3. Cognition. I. Title. II. Series.

ML3838 .Z25 2002

781'.11 dc21
2001058756

Printed in the United States of America

on acid-free paper

dem andenken meiner Mutter:

Anneliese Margerite Zbikowski, 1926 1999

This page intentionally left blank

preface

n picking up a book with the title Conceptualizing Music: Cognitive Structure,

Theory, and Analysis, one might reasonably assume that it deals with music cognition and how our knowledge of that discipline can be applied to music theory and
analysis. This book does not do that, or at least not in a simple way. To begin, it does
not have much to say about the fairly large body of research usually placed under the
rubric music cognition. This work, having been developed out of music psychology and infor med by recent research in the brain sciences and mind sciences, proceeds by carefully crafted exper iments, which are subjected to closely argued statistical and logical analysis. As practiced by such eminent and able researchers as Carol
Krumhansl, John Sloboda, and David Huron, the study of music cognition has told
us much about how humans process sonic and musical infor mation.
But this book proceeds in a somewhat different way. Drawing on the same body
of research from the brain sciences and mind sciences that shaped studies in music
cognition, it explores how basic cognitive capacities are specied for understanding music. The project takes inspiration from recent work in linguistics and rhetor ic
by researchers like Ronald Langacker, Gilles Fauconnier, Mark Turner, and George
Lakoff, and it is based on the assumption that musical understanding relies not on
specialized capacities unique to the processing of patterned sound but on the specialized use of general capacities that humans use to structure their understanding
of the everyday world. The methodology, in consequence, relies not on exper imental design and data analysis but on using a broad and quite extensive body of
research to inter pret recurrent tropes of musical understanding. These tropes
involve such things as the importance to musical understanding of relatively small
and compact musical phenomena like motives, themes, and chords; the use of
terms grounded in nonmusical domains terms like space and depth to character ize musical events; and the reliance on patter ns of logical inference to reason
about music.
The result of this investigation is a theoretical perspective on musical organization, but one rather different from what usually counts as music theory. To make
sense of this claim requires a bit of explanation about the intellectual context of
music theory, for music theory is, within the rolling seas of humanistic studies, a

viii

p re fac e

rather strange sh. Put bluntly, it is clear that much of what music theory does, as a
discipline, does not count as any sort of theory in moder n scholarship. This is exemplied by each of the two distinct but related and intertwined strands that make up
contemporary music theory. One strand is occupied with pedagogy, the other with
speculative and highly systematic approaches to musical organization.
Music theor y, as it is presented in the classroom, is most often engaged with a
careful and often relentless explication of what, for want of a better ter m, we can call
musical grammar. Consider the following, from Edward Aldwell and Carl Schachters
Harmony and Voice Leading:
Like VII,V has 2 as its bass.V , in fact, resembles VII6 so closely that they are almost
interchangeable chords. The bass of V is a more neutral tone than that of V (or, as
you will see,V ) and can move convincingly either to 1 or to 3. Consequently,V , like
VII, forms a natural connection between I and I and appears very frequently as a
passing chord within an extended tonic.1

The prose and ter minology are impressively dense. But one should not be misled
into thinking that the authors are concer ned only with abstruse compositional
techniques, for immediately after this excer pt Aldwell and Schachter refer to a passage from an impromptu by Franz Schubert that illustrates the niceties of voice
leading with which they are concer ned. Their assumption is that the reader is familiar with the music and counts it as typical, and it is this familiar ity that provides a
phenomenological anchor for what might appear to be rather thick jargon. If you
know Schuberts impromptu, or (better yet) can summon it in your sonic imagination when reading the example in score, Aldwell and Schachters point about the
harmonization of the second scale-step in the bass is not just clear but even obvious.
At the heart of pedagogical music theor y are familiar or typical examples of
music, the mysteries of which are revealed by a music theor ist (or theor ists) eager to
share the secrets and wonder of this music with others. As elegant and persuasive as
this approach might seem, it is, within our cur rent cultural climate, more than a little unrealistic: music by Schubert and his contemporar ies is often unfamiliar to the
students who read Aldwell and Schachters text (or any of a number of similar texts)
and is not typical of the music that resonates through these students digitized and
hypercommercialized environments. That this should be so is often regarded as
symptomatic of an illness of the late twentieth centur y, an illness that leads to an
insufcient engagement with the g reat traditions of Western culture. For Classical
music (as it is so styled), the antidote is music theory. Music theory, with its careful
explication of the musical grammar of Mozart and Beethoven and Schubert, thus
becomes the last redoubt against the dissolution of Western culture represented by
a dwindling interest in the music of eighteenth- and nineteenth-centur y Europe.
If, for a moment, we step back from Aldwell and Schachters text and generalize
its intent beyond the specic repertoire relative to which it is framed, we might be
able to avoid this rather sanctimonious stance.We could argue that music is a highly
complex and idiosyncratic mode of human communication and that having a knowl1. Edward Aldwell and Carl Schachter, Harmony and Voice Leading, 2nd ed. (San Diego: Harcourt
Brace Jovanovich, 1989), 112.

p re fac e
edge of the grammar of this mode of communication is essential to its deeper
appreciation, no matter what for m music might take. The argument is a familiar
one to me, not the least because I often nd myself making it. And yet, something
rings hollow. The grammar that music theor y teaches is unavoidably tied to the
repertoire to which it refers, and just how this is generalized to apply to other repertoires is not immediately apparent: I know of no theory text that explains how the
grammar of Schuberts musical discourse is manifested in the music of, say, John
Coltrane or Pr ince. Given its isolation from contemporary culture, the music theory of the classroom appears to be little more than a ghost that haunts the echoing
halls of a crumbling cultural empire.
The second strand of music theory partakes of the systematic quality inherent
in grammars but generalizes it away from natural language and toward a free-standing intellectual construct. As an example of this sort of theory, consider the following brief passage from David Lewins analysis of a section of the opening of Claude
Debussys piano prelude Reets dans leau. In this excer pt, X, Y, and Z represent
specic collections of musical notes, RI refers to the compound operations retrograde and inversion, RICH is a function that effects ser ial transfor mations, and T
refers to transposition:
In measure 10 the music of measure 9 is repeated and extended. The crescendo recurs.
In the melody the repetition gives rise to a rotated for m of Z1, marked rot Z1 on
gure 10.10. Rot Z1 is Bb-Ab-F-Eb; it embeds ser ially the or iginal for m of Y, Ab-FEb, and precedes this Y by its overlapping inverse-RI-chained for m Bb-Ab-F. (Bb-AbF is RICH1(Ab-F-Eb).) This relationship is more or less inherent in the der ivations
of X,Y, T(X), their repetitions, and Z1.2

It is, of course, something of a challenge to evaluate this passage in isolation. Not

only is it just one part of a larger analysis, but also it comes late in a book occupied
with various and sundry applications of formal algebra and mathematical mapping
theory to music. Nonetheless, what should be clear is that more than familiar ity
with Debussys prelude is required to make sense of Lewins inter pretation of the
passage. The reader must also be familiar with a style of abstract thought that is
bound to appear cabalistic to the uninitiated, one in which the transfor mation of
musical entities is at least as important as the entities themselves. For some, the inaccessibility of this mode of thought is one of its char ms. For others, it is proof of the
hermeticism and ir relevance of music theory.
Before continuing, I should make clear that I have deep respect for the theor ists
whose work I have cited. I use this work in my teaching and continue to be intr igued
and stimulated by it. I also want to emphasize that these excer pts by no means
reect all that there is to music theory. I take them as representative of two strands
of thought that are replicated and woven together in all sorts of different ways to
create the texture of contemporar y theoretical practice.What is important for my
purposes here is that the practice of contemporary music theory is not like that of
contemporary cultural or social theory. Instead of probing the cultural or histor ical
2. David Lewin, Generalized Musical Intervals and Transformations (New Haven, Conn.:Yale University Press, 1987), 234.

ix

p re fac e

context for musical utterances, or the complex networks of social interaction that
give rise to musical behavior, music theory continues to focus on details of musical discourse with an obsessiveness that is both maddening and quixotic to cultural
and social theor ists.
Given the impressive traditions of music theory and their inuence on my own
thought, I cannot guarantee that what I offer here is a great improvement on this
situation. One of the things I want to do in the chapters that follow, however, is to
develop a somewhat different view of music theory one that sees music theory as
a response to a problem. The problem is that of musical understanding: how it is
that we can make sense of sequences of nonlinguistic patter ned sound, that we can
do so with amazing rapidity, and that (often as not) we can return to these or similar sound sequences and nd continued reward. I will argue that our understanding of music relies on a play of concepts and conceptual structures that emerges
from training basic cognitive capacities on musical phenomena and that music theory and music analysis der ive from this play.
This mode of inquiry is not one common in the discipline of music theor y,
despite its similar ities to work done by Leonard B. Meyer, Eugene Nar mour, and
Robert Gjerdingen. It does, however, share features with the approach to music evident in Susan McClarys recent Conventional Wisdom. That this should be so reveals
a debt on my part, for it was Professor McClar y who rst suggested that I read
Mark Johnsons The Body in the Mind back in the late 1980s, and this had a profound
effect on my research. In her book, McClary explores the cultural and histor ical
forces that have shaped genres like opera and the blues and compositional practices
like tonality. My focus is on a somewhat different set of forces: those that shape the
way humans think. It seems inevitable that these forces are in some sort of grand,
if ill-dened, dialogue with cultural and histor ical agencies: it is, after all, human
cognitive processes and human culture and histor y about which we are talking.
That cultural and social theor y have turned a deaf ear to this dialogue is not surprising: cognitive theory has had little room for and less patience with culture, and
the detail of its investigative method is no match for the epic sweep of high theoretical practice at its best. And yet it seems we must, at some point, come to ter ms
with cognitive structure, for if we do not develop an understanding of how cognitive processes shape the basic mater ials of thought, we risk accepting these mater ials as things g iven by nature, just as culture and histor y and music, for that matter were once assumed to be given by nature.
A glimpse of the problem can be seen in McClarys compelling analysis of Robert
Johnsons 1936 Cross Road Blues. McClary contrasts Johnsons blues with those of
Bessie Smith, noting that the inuence of Johnsons music on white British blues players of the 1960s was due in part to a misconception: the Br itish musicians believed
that the idiosyncrasies of Johnsons blues style represented authentic blues practice.
Describing the unique sound of Johnsons Cross Road Blues, McClary writes,
An affect of dread and entrapment pervades this tune partly the result of his strangulated, falsetto vocals and his uncanny replication of that timbre on the guitar.
Moreover, Johnsons percussive guitar pulse, which locks in at the eighth-note level,
allows almost no sensual movement: even though Johnsons singing constantly strains

p re fac e
against that beat, the listeners body is regulated by those short, aggressively articulated
units. The guitar thus seems to represent simultaneously both oppressive outside forces
and a desperate subjectivity ghting vainly for escape.3

I have no quar rel with McClar ys analysis indeed, in this short passage, she has
captured a number of the essential features of Johnsons performance style. The
difculty comes with the ultimate justication for the affect produced by the song.
Why do Johnsons vocals and guitar work yield dread and entrapment rather than
joyful anticipation and a feeling of liberation? Clearly, the rhythmic framework is
important, but why is it that the r igidity of Johnsons beat constrains us rather than
providing a secure foundation from which we can coolly regard his plaint? These
are not easy questions, all the more so because of the relatively unique character of
Johnsons recording when compared with other blues recordings of the per iod, and
because hear ing Cross Road Blues as something other than a moving, haunting
song is to misunderstand it rather thoroughly. Cultural, social, and histor ical context
cannot, by themselves, explain the or igin of our affective response to the song, for
our broad agreement on the effect of Johnsons music transcends these implements
of high theory (even if they have a profound inuence on what we do with Cross
Road Blues once we have heard it). I propose that explor ing the way cognitive
structure infor ms our understanding of music gives us a way to account for the
source of our broad agreement on the affect that pervades Johnsons blues and can
help us understand better the ways culture, society, and history reshape musical
practice.
Again, the way I want to accomplish this is by reconceptualizing what it means
to theor ize about music. This can be done by approaching music theor y from the
perspective provided by recent work in cognitive science. My intellectual guideposts include not only the wealth of work done in the mind sciences and the brain
sciences but also contemporar y and histor ical ways of theorizing about music.
These theor ies of music now with theory understood in somewhat more traditional ter ms capture important aspects of how it is we structure our understanding of music. At their best, they represent a technical and systematic articulation of an accord on what matters in music that is similar in kind, if not ter minology,
to our accord on how Robert Johnsons blues move us: we agree on how music is
put together and we agree on what music means because both are structured by
basic cognitive processes through which we organize our understanding of the
world. And so music theor y this rather strange sh in the seas of humanistic
scholarship may yet tell us some quite interesting things about the cultural and
social construction of music. Understanding the way music theory instantiates cognitive processes will also help explain its continuing value. It will offer such help
whether that theor y be prosaic as within the classroom or within cr itical discourse (since even the most radical of cultural or sociological theor ists inevitably
makes recourse to basic music-theoretical constructs) or poetic, as with the elusive and allusive constructs of abstract theory.
3. Susan McClary, Conventional Wisdom: The Content of Musical Form (Berkeley: University of California Press, 2000), 51.

xi

xii

p re fac e

As should be apparent from the preceding, my intent here is to address not only
music theor ists but also musicologists and ethnomusicologists who nd the challenges of theorizing about music intr iguing.What follows will also be of interest to
those with either a professional or avocational interest in cognitive science, for
music presents a number of interesting problems for cognitive processing, not the
least of which are its embeddedness in culture and the demands it places on realtime processing. In entertaining the thought of such an audience, however, I should
briey clar ify a distinction I draw between sound and music and the cognitive abilities related to each, which is based on three suppositions. First, not all sound is
music. Second, an account of how humans process sound is not the same thing as
how they understand music. Third, phenomena relevant to musical understanding
exist at a conceptual level that is, at a level of cognitive activity at least potentially
accessible to conscious thought. I should emphasize that I regard the conceptual
level as occupying only a small part of our total cognitive activity, and I am not at all
opposed to efforts by music psychologists and others who try to explain how structures at the preconceptual level connect with and motivate structures on the conceptual level. For me, however, it is at the conceptual level that I nd the most profoundly interesting questions, for concepts are the tools that allow us to construct
the complex notions essential to musical understanding. From this perspective, conceptualizing music is fundamental to inquir ies about music, whether those be from
the perspective of music cognition, or ethnomusicology, or musicology, or theory.

acknowledgments

n the course of writing this book I enjoyed the assistance of many people, all of
whom improved the product immeasurably. Among the many colleagues who
read drafts of material or responded to presentations I have given, are the following:
Ko Agawu, Holly Aksnes, Jeanne Bamberger, Larry Barsalou, Philip Bohlman,
Candace Brower, Scott Bur nham, Clif Callender, Thomas Chr istensen, Martin
Clayton, Rick Cohn, Nick Cook, Arnie Cox, Bob Gjerdingen, Robert Hatten, Bob
Holzer, Brian Hyer, Rich Janda, Carol Krumhansl, the late Jim McCawley, Marc
Perlman, Anne Robertson, John Rothgeb, Janna Saslaw, Martin Stokes, and Mark
Turner. Their comments, suggestions, and quer ies helped me strengthen and reformulate my arguments. Ben Brinner, Gilles Fauconnier, Douglas Hofstadter, Travis
Jackson, Bobby Short, Sumarsam, and Susan Youens were generous in their response
to questions I posed to them and provided me with infor mation and insight I simply could not have gained elsewhere. Students in my classes and seminars at the
University of Chicago listened to many of the ideas presented in this book, to
which they responded with questions, challenges, and comments. Although it
would be impractical to acknowledge all of them here, they contributed immensely
to the shape those ideas have taken in this book.
Deborah Gillaspie of the Chicago Jazz Archive and Regenstein Library patiently
endured my requests for obscure recordings and was indispensable in helping me
nd mater ials that I needed. The musical examples were expertly prepared by Jrgen Selk of Music Graphics Inter national. Philipp Goedicke and Heinr ich Jaeger
helped me with the translation of difcult passages in Ger man, and my brother
Gene Zbikowski assisted in decipher ing some of the more recalcitrant clauses in
Rameaus and Prousts French. Jos Antnio Martins and Aron Topielski served as
my research assistants, ably tracking things down, organizing my sprawling les, and
getting to know the photocopier, perhaps better than they wanted to. Maribeth
Payne, when she served as Executive Editor for Music at Oxford University Press,
listened to my ideas for the book (at least once at the r isk of missing her train) and
then provided valuable encouragement and assistance in the process of turning a
proposal into a nal manuscript.
Institutional support came through a yearlong fellowship at the Franke Institute

xiv

ac k nowle dg m e nt s

for the Humanities (for merly the Chicago Humanities Institute) at the University
of Chicago and from the Department of Music of the University of Chicago. The
latter provided not only a home for my research but nancial support as well. Some
of the recordings discussed in chapter 5 were made available by the Archives for
Traditional Music in Bloomington, Indiana, and by the Bowling Green Sound
Recordings Archives, Bowling Green, Ohio.
Finally, production of this book was facilitated by subventions from two scholarly societies. The Society for Music Theor y provided a subvention to defray the
costs of producing the musical examples, and the result of this assistance can be seen
in the number and appearance of the examples I was able to include. The Publications Committee of the American Musicological Society was equally enthusiastic
about the book and ag reed to include it in its ser ies, AMS Studies in Music. This
decision not only made the book more affordable but also gave me an excellent and
endlessly patient editor in Lawrence Ber nstein.
My family Vicky Long, Anna Katia, and Andrei Nicolai had to put up with
sharing me with this project, and I am afraid they got the short end of the deal more
than once.Without their support, understanding, and uncr itical companionship, this
book would have been much the poorer.
My musical examples generally adhere to the readings of the following editions: the
new Beethoven Werke; the Johannes Brahms Smtliche Werke; the Neue Mozart Ausgabe; Casimir is edition of the works of Palestr ina; the Neue Schubert Ausgabe; Clara
Schumanns edition of the Lieder of Robert Schumann; the new Richard Wagner:
Smtliche Werke (in conjunction with the C. F. Peters score of Tristan); and the Collected Works of Giaches Wert for Corpus mensurabilis musicae. Some indications of
editor ial intiative in those editions generally with respect to dynamic markings
added on the basis of parallel readings have been realized tacitly in my examples
for the sake of clarity of appearance. The setting of Bernhard Kleins Trockne Blumen was prepared from the 1822 edition by E. H. G. Christiani in Berlin.
When dealing with histor ical sources wr itten in languages other than English, I
have generally followed one of two practices. Where there is no translation of the
work, or where I think existing translations are not wholly satisfactory, I provide my
own translation, with the or iginal in a footnote.Where satisfactory translations exist,
I use these and do not include the or iginal in the footnote. In a few cases, I have
thought readers might like to make reference to both the or iginal and a translation
and consequently have included citations for both in the footnotes.
Some of the mater ial that follows has been previously published elsewhere and
is used here by kind per mission of the journals in which it or iginally appeared. This
material includes the following: Conceptual Models and Cross-Domain Mapping:
New Perspectives on Theor ies of Music and Hierarchy, Journal of Music Theory
41/2 ( 1997, Journal of Music Theory); Musical Coherence, Motive, and Categorization, Music Perception 17/1 ( 1999 by The Regents of the University of California); and The Blossoms of Trockne Blumen: Music and Text in the Early
Nineteenth Century, Music Analysis 18/3 ( 1999, Blackwell Publishers, Ltd.).

contents

Introduction: Conceptualizing Music 3

part i. aspects of cognitive structure

1. Categorization 23

2. Cross-Domain Mapping 63

3. Conceptual Models and Theor ies 96

part ii. analysis and theory

4. Categor ization, Compositional Strategy, and Musical Syntax 137

5. Cultural Knowledge and Musical Ontology

201

6.Words, Music, and Song: The Nineteenth-Century Lied 243

7. Competing Models of Music: Theories of Musical Form and Hierarchy 287

Conclusion: Cognitive Structure, Theory, and Analysis

Bibliography 335

Index 353

325

This page intentionally left blank

CONCEPTUALIZING MUSIC

This page intentionally left blank

introduction

conceptualizing music

arly in the opening volume of A la recherche du temps perdu, Marcel Proust presents the rst of a number of memorable accounts of listening to music. He
describes Charles Swanns initial encounter with the andante of Vinteuils sonata for
violin and piano:

Doubtless the notes which we hear at such moments tend, according to their pitch
and volume, to spread out before our eyes over surfaces of varying dimensions, to trace
arabesques, to give us the sensation of breadth or tenuity, stability or capr ice. But the
notes themselves have vanished before these sensations have developed sufciently to
escape submersion under those which the succeeding or even simultaneous notes have
already begun to awaken in us. And this impression would continue to envelop in its
liquidity, its ceaseless overlapping, the motifs which from time to time emerge, barely
discernible, to plunge again and disappear and drown, recognized only by the particular kind of pleasure which they instill, impossible to descr ibe, to recollect, to name,
ineffable did not our memor y, like a laborer who toils at the laying down of rm
foundations beneath the tumult of the waves, by fashioning for us facsimiles of those
fugitive phrases, enable us to compare and to contrast them with those that follow.
And so, scarcely had the exquisite sensation which Swann had exper ienced died away,
before his memory had furnished him with an immediate transcr ipt, sketchy, it is true,
and provisional, which he had been able to glance at while the piece continued, so
that, when the same impression suddenly returned, it was no longer impossible to
grasp.1

What Proust summons in this lyr ical, enchanted vignette is the awakening and initial consolidation of musical understanding. Swanns rst impressions of Vinteuils
sonata are vague and unfor med, his mind simultaneously struggling with and savoring the ineffability of the music. But then, with the aid of memory, patterns emerge.
Although these are incomplete and subject to revision, they offer him a way to
make sense of the music, even as it continues to play.
Conceptualizing Music provides an exploration of the process of musical under1. Marcel Proust, Swanns Way (vol. 1 of Remembrance of Things Past), trans. C. K. Scott Moncr ieff
and Terence Kilmartin (New York: Vintage Books, 1981), 227.

i nt roduc t i on

standing that is, the process through which those liquid impressions spoken of by
Proust are transfor med into structures that make it possible to grasp music. In what
follows, I argue that Swanns or anyones understanding of music draws on the
same cognitive processes that humans use to organize their understanding of the
world as a whole. Confronted with musical sound, these processes create musical
concepts, the things that enabled Swann to gain a g rasp of the music. The act of
conceptualizing music is the beg inning of a whole chain of cognitive events that
allow us to theor ize about music and to analyze the things that populate our aural
past, present, and future.
The notion that Swanns musings might give rise to musical concepts demands
some further consideration. Concepts are often thought of as highly stable cognitive
structures of considerable complexity, a view hardly commensurate with the ephemera
attended to or produced by Swann. Recent work in the brain sciences and the mind
sciences, however, has changed how we view concepts. There are now persuasive
arguments that concepts are quite uid, that they are not ir revocably wedded to
words or to concrete representations, and that they are not even unique to our
species.2 In consequence, the provisional replicas of musical phrases that make it
possible for Swann to secure a foothold in the unfamiliar ter rain of Vinteuils sonata
need not automatically be excluded from the conceptual domain. In fact, they are
very much like the concepts we use to structure our understanding of the everyday world.
This same work in the brain sciences and the mind sciences suggests a way to
account for the apparent simplicity and immediacy of musical understanding, which
seems incommensurate with the complexity of musical structure. For instance, the
sonata by the ctional Vinteuil is intended to be a relatively complex contemporary
work that has captured the fancy of the musical elite of the Paris salons. Nonetheless, Swann, whose connoisseurship does not extend to music, is able to gain a grasp
of the work almost immediately. That he is able to do so is not simply novelistic
license but is, in fact, thoroughly plausible: almost everyone has had the exper ience
of listening to an unusual composition or exotic repertoire and being able to make
something of it. This possibility has suggested to some a latent musicality in humans
comparable to the sort of competence for language proposed by Noam Chomsky.3
Competencies of this sort raise as many questions as they answer, however, particularly where cultural entities such as music are concer ned. It seems more promising
to follow the path of researchers who have rejected linguistic competence as a given
2. See, for instance, Douglas R. Hofstadter and the Fluid Analogies Research Group, Fluid Concepts
and Creative Analogies: Computer Models of the Fundamental Mechanisms of Thought (New York: Basic Books,
1995), chaps. 5 6, 8 10; Gerald M. Edelman, The Remembered Present: A Biological Theory of Consciousness
(New York: Basic Books, 1989), chap. 8; and Donald R. Grifn, Animal Minds (Chicago: University of
Chicago Press, 1992), chap. 6.
3. The notion of musical competence has generated a range of commentar y and scholarship. See
John Blacking, Music, Culture, and Experience, in Music, Culture, and Experience: Selected Papers of John
Blacking, ed. Reginald Byron, with a foreword by Bruno Nettl (Chicago: University of Chicago Press,
1995), 228 31, on musical competence and culture; Steven Pinker, How the Mind Works (New York:
W.W. Norton, 1997), 528 38, on musical competence and its relation to other competencies; and Allan
Keiler, The Origins of Schenkers Thought: How Man Is Musical, Journal of Music Theory 33 (1989):
273 98, on a nineteenth-century conception of musical faculties akin to musical competence.

conc e p tual i z i ng mu s i c
5
and who have set about explor ing the cognitive foundations of language.4 Their
task has been to discover what processes are basic to human cognition and then to
determine how they are specied for language. For my part, I would like to explore
how some of these same general cognitive processes are specied for music. By this
means I hope to account for the apparent ease and real rapidity with which we can
conceptualize a highly complex, completely unfamiliar music on our rst encounter,
without having to postulate the faculty of musical competence.
At a bit more of a remove, but no less important for a complete account of cognitive processing, is the way concepts come to be organized into the more extended
cognitive structures with which our thought is usually occupied. This is where theories come into play, for theor ies are the cognitive tools that guide the way we reason about the things we exper ience. At rst, this might seem to be a rather specialized use of the notion of theory, for the theory with which much current literature
is occupied within and without music scholarship is hardly the stuff of everyday exper ience. Recent research has countered this view of theory by demonstrating that the elegant and abstract theor ies of science have much in common with the
tools for reasoning used by very young children.5 Theories are the basic means by
which we make our exper ience coherent and guide further action. The rough-andready transcr ipt that guides Swanns listening thus has something in common with
the more fully articulated and systematic structures we usually associate with the
idea of music theory.
That music theory might have alliances with everyday thought processes is a
provocative claim. On the one hand, music theory often manifests itself as a relentlessly practical discipline: a codication of the scales, chords, and grammatical rules
proper to a highly circumscr ibed portion of musical discourse, assembled with the
intent of rendering music comprehensible to those who would become musically
literate. On the other hand, music theory can reach into the far cor ners of abstraction to embrace complicated mathematical concepts or the arcane symbologies of
voice-leading graphs, as any reader of the Journal of Music Theory, Music Theory Spectrum, or Music Analysis will quickly discover. Nonetheless, I want to argue that music
theory, in all its diverse forms, reects the same basic processes that guide our understanding of the everyday world. Theorizing about music is an activity specialized
only in its domain, not in the cognitive processes it involves.
What might these cognitive processes be, and how would they manifest themselves? To answer these questions, let us begin at a beginning, with two theor ies of
4. For general introductions to some of the working assumptions of cognitive linguists, see George
Lakoff, The Invariance Hypothesis: Is Abstract Reason Based on Image-Schemas? Cognitive Linguistics
1 (1990): 39 74; and Michael Tomasello, Introduction: A Cognitive-Functional Perspective on Language Structure, in The New Psychology of Language: Cognitive and Functional Approaches to Language Structure, ed. Michael Tomasello (Mahwah, N.J.: Erlbaum Associates, 1998), viixxiii. In the eld of cognitive
linguistics, Ron Langackers work is particularly notable for its thoroughness and its systematic approach.
See Langacker, Theoretical Prerequisites, and Descriptive Application, vols. 1 and 2, respectively, of Foundations
of Cognitive Grammar (Stanford, Calif.: Stanford University Press, 1987, 1992); idem, Grammar and Conceptualization (Berlin: Mouton de Gruyter, 2000).
5. See Alison Gopnik and Andrew N. Meltzoff, Words, Thoughts, and Theories (Cambridge, Mass.:
MIT Press, 1997); and Alison Gopnik, Andrew N. Meltzoff, and Patricia K. Kuhl, The Scientist in the Crib:
Minds, Brains, and How Children Learn (New York:William Morrow, 1999).

i nt roduc t i on

music from Greek antiquity. These theor ies, and the music to which they refer, are so
unfamiliar that even many who make music theory the focus of their research have
only passing knowledge of what they involve. But there is an advantage in this unfamiliar ity, for the disor ienting effect it can have also serves to loosen our notions about
what a theory of music, or a theoretical construct, should be. Despite connections
with music theory as it was practiced in Europe and (to a lesser extent) in the Arabic
world, Greek music theory of antiquity is the theory of an alien society. Nonetheless,
this theory was a beginning one of the starting points for accounts of musical organization in the Western tradition. We can discern within it, therefore, the theor ists of
antiquity grappling with basic constructs equivalent to, if still different from, the sort
of constructs taught in beginning music theory classes today. Both aspects will allow
us to see the role basic cognitive structures play in our understanding of music.

ancient music theory and

modern cognitive science
Those who wrote on Greek musical practice in antiquity concer ned themselves
with a wide variety of topics, including the place of music in society, musical aesthetics, the construction and nature of musical instruments, and the organization of
pitch mater ials. Specic discussions of pitch mater ials the usual topic of disquisitions more directly or iented to music theory centered around the set of pitch
relationships that has come to be called the Greater Perfect System.6 As shown in
gure I.1, this consisted of a set of four tetrachords (hypaton, meson, diezeugmenon,
and hyperbolaion), which, together with one additional note (called Proslambanomenos), provided the framework for a two-octave system of pitches basic to Greek
music. The end points of the tetrachords (the notes Hypate hypaton, Hypate meson,
Mese, and so on), together with Proslambanomenos, were regarded as stable, unmovable pitches. There was, however, no xed standard for tuning the note-names
given in gure I.1 are simply for the pur poses of illustration.Within the boundar ies
marked by each tetrachord were two other pitches, whose placement varied according to which of three different genera was understood to be in play. (In the tetrachord hypaton, for instance, the variable notes above Hypate hypaton were Parhypate
hypaton and Lichanos hypaton.) The diatonic genus located the movable pitches in a
manner analogous to moder n diatonic scales (for instance, given the reference
pitches on g. I.1, Parhypate hypaton in the diatonic genus would be equivalent to
F3, and Lichanos hypaton would be equivalent to G3), but the chromatic and enharmonic genera situated the pitches in ways that have no comfortable analogue in
modern scale construction. The result was a system in which the placement of
movable pitches could vary widely and in which intervals between successive
pitches could be smaller than a half step and larger than a whole step.
6. For a more detailed discussion of the Greater Perfect System and its place in Greek theor y of
antiquity, see Andrew Barker, ed. and trans., introduction to Harmonic and Acoustic Theory (vol. 2 of Greek
Musical Writings), Cambridge Readings in the Literature of Music (Cambr idge: Cambridge University
Press, 1989), 12 13. For a more general overview of Greek music theor y and a thorough discussion of
the sources, see Thomas J. Mathiesen, Apollos Lyre: Greek Music and Music Theory in Antiquity and the
Middle Ages (Lincoln: University of Nebraska Press, 2000), part 4.

conc e p tual i z i ng mu s i c
7
C5

Nete hyperbolaion
Tetrachord hyperbolaion

G4

Nete diezeugmenon
Tetrachord diezeugmenon

D4

Paramese

C4

Mese
Tetrachord meson

G3

Hypate meson
Tetrachord hypaton

figure I.1

D3

Hypate hypaton

C3

Proslambanomenos

Diagram of the Greater Perfect System of Greek music theory

There were two main theoretical approaches to presenting the elements of the
Greater Perfect System. The Pythagorean approach der ives from a metaphysics associated with Pythagoras of Samos, who lived during the sixth centur y b.c. Its hallmarks are a persistent interest in number and the deployment of numer ical conceptions in cosmological contexts. In contrast to this, the Aristoxenian approach,
associated with the Peripatetic school of the fourth century b.c., places little reliance
on number, trusting instead in observation and reason as the means to knowledge
about music.

Pythagoras and the Blacksmiths

Nicomachus of Gerasa, a mathematician and har monist wr iting around the beginning of the second century a.d., described Pythagorass discovery of the basic pr inciples of music theory as follows:
He [Pythagoras] was plunged one day in thought and intense reasoning, to see if he
could devise some instrumental aid for the hear ing which would be consistent and
not prone to er ror, in the way that sight is assisted by the compasses, the measur ing rod
and the dioptra, and touch by the balance and by the devising of measures; and happening by some heaven-sent chance to walk by a blacksmiths workshop, he heard the
hammers beating iron on the anvil and giving out sounds fully concordant in combination with one another, with the exception of one pairing; and he recognized among
them the consonance of the octave and those of the fth and the fourth. He noticed
that what lay between the fourth and the fth was itself discordant, but was essential
in lling out the greater of these intervals. Overjoyed at the way his project had come,
with gods help, to fulllment, he ran into the smithy, and through a great variety of
experiments he discovered that what stood in direct relation to the difference in the

i nt roduc t i on

sound was the weight of the hammers, not the force of the str ikers or the shapes of the
hammer-heads or the alteration of the iron which was being beaten. He weighed
them accurately, and took away for his own use pieces of metal exactly equal in weight
to the hammers.7

Nicomachus continues the story by descr ibing how Pythagoras used the weights to
conduct further exper iments. After suspending the weights from identical str ings,
Pythagoras plucked pairs of strings and discovered the same concords as he had
heard produced by the blacksmiths. He further discovered that the interval of an
octave was produced by weights in a 2:1 ratio, that of the fth by weights in a 3:2
ratio, and that of the fourth by weights in a 4:3 ratio, as shown in gure I.2a. The
one discordant interval that of a second sounded by the middle two weights
was the product of a 9:8 ratio. Additional exper imentation showed that the smallest weight sounded a fourth, with the next to smallest weight (8:6
4:3) and a fth
with the next to largest (9:6
3:2; see g. I.2b). The octave could thus be viewed
as the product of either a fourth plus a fth (12:9:6; g. I.2c) or of a fth plus a
fourth (12:8:6; g. I.2d). According to Nicomachus, Pythagoras also discovered that
these ratios held constant throughout the musical domain. It made no difference
whether the constituent notes of the intervals were produced through string tension, string division, beating on pots, or blowing on tubes the relationships
between these notes always reduced to the self-same ratios.8
The Pythagorean view of music outlined by Nicomachus assumes that music has
its origins in the natural world and that the natural world has a basic (if often
unseen) order that can be expressed through number. It is thus important to Nicomachuss story that the r inging of the hammers is accidental and not contr ived: their
harmony has everything to do with the inherent order of the world and almost
nothing to do with the blacksmiths. It is also signicant that the concordant intervals are immediately apparent to Pythagoras and that he can discer n them even
amid the discordant clang of the major second. Not only is the basis of musical
order natural, but also it is manifest to all who have ears to hear. The association of
these intervals with the pounding hammers provides the computational tool for
7. Nicomachus, Enchiridion, in Barker, Greek Musical Writings, 2: 256 57. The dioptra was a rod that
was used for the indirect measurement of the height of tall objects.
8. Nicomachus, Enchiridion, 258. It should be noted that Nicomachuss story is a complete ction.
There is no evidence whatsoever that Pythagoras ever conducted any empirical research on the acoustic
origins of harmonic relationships, with or without blacksmiths. Perhaps more important, the ratios
described by Nicomachus simply do not work. To sound the intervals descr ibed in the story, the values
of the weights must be squared that is, the weights must be in the ratio 4:1 to produce the octave, 9:4
to produce the fth, and 16:9 to produce the fourth. The ratios given by Nicomachus only work when
used to segment a str ing into different sounding lengths. One-half the length of a str ing will sound an
octave with the entire length of the string; two-thirds the length of the str ing will sound a fth with the
entire length of the str ing; and three-fourths the length of a str ing will sound a fourth with the entire
length of the str ing. The importance of Nicomachuss story lies in its inuence: in the for m Boethius
gave it in the sixth century, it became the standard account of basic Pythagorean principles for the Middle Ages and Renaissance; see Ancius Manlius Severinus Boethius, Fundamentals of Music, ed. Claude V.
Palisca, trans. Calvin Bower, Music Theory Translation Ser ies (New Haven, Conn.:Yale University Press,
1989), 17 19. A somewhat different perspective on Nichomachuss story can be found in Mathiesen,
Apollos Lyre, 399.

conc e p tual i z i ng mu s i c
9
a.

12

9
4:3

3:2
2:1

b.

12

3:2

4:3

2:1

c.

12

9
4:3

d.

12

3:2

8
3:2

figure I.2

4:3

Ratios of the octave, fth, and fourth from Pythagorean legend

which Pythagoras had searched, for it allowed him to translate the constituent notes
of the har monic intervals into magnitudes that is, into number.9 These numbers
then provide further proof of the order of nature, for the ratios of the concordant
intervals (the octave, fth, and fourth) are all simple (1:2, 2:3, and 3:4), while the
ratio of the discordant second is relatively complex (8:9). Finally, the numbers
involved in the ratios of the concords1, 2, 3, and 4 were also those of the tetraktys of the decad, which Pythagoreans regarded as the fount and root of everowing nature.10
The account of musical organization presented by the story of Pythagoras and
the blacksmiths is a model of concision. At its core are but three intervals: the
octave, fourth, and fth. The notes that make up these intervals are assigned magnitudes, the cor respondence of which yields numerical ratios. These ratios come
to stand for the intervals the ratio 1:2 is the octave and can also be used to
describe relationships between them, leading to a precise account of the composi9. The notion of a computational tool I employ here derives from the work of Edwin Hutchins,
especially that presented in the second chapter of his Cognition in the Wild (Cambridge, Mass.: MIT Press,
1995). Hutchins proposes that various of the navigational tools used by seafarers are in fact computational
tools in that they facilitate computation by transfor ming analog information into digital infor mation.
This is exactly what the hammers did for Pythagoras: they transfor med the analog information of sound
into the digital infor mation of hammer weights (with an inter mediary stage occupied by the computational tool of a scale for measur ing the weights).
10. Barker, Greek Musical Writings, 2: 30. A tetraktys is any coordinated group of four items; those of
the tetraktys of the decad sum to 10, which is the basis of the base-10 number ser ies used by the Greeks.

10

i nt roduc t i on

tion of intervals. The interval between the Mese and Proslambanomenos of gure I.1
is an octave, a 1:2 ratio. This octave is made up of a fourth (from Mese to Hypate
meson, a 3:4 ratio) and a fth (from Hypate meson to Proslambanomenos, a 2:3 ratio).
The fth is in tur n made up of a fourth (from Hypate meson to Hypate hypaton,
another 3:4 ratio) and a tone (from Hypate hypaton to Proslambanomenos, an 8:9
ratio). With these components in place, a Pythagorean theor ist could descr ibe,
through number, relationships between any of the xed notes of the Greater Perfect
System and eventually character ize the relationships that obtained among the
movable notes of the various genera. The order of the cosmos, which was for
Pythagoreans the order of number, thus found sounding expression in the domain
of music.

Aristoxenus and Aristotelianism

Although the Pythagorean perspective on musical order was an inuential one
for instance, it infor med Platos and Aristotles writings on music it was not the
only one available to antiquity. The alter native offered by Aristoxenus in his Elementa harmonica (most likely written toward the end of the fourth century b.c.) starts
with a denition of the science of harmonics:
It is to be understood as the science which deals with all melody, and inquires how the
voice naturally places intervals as it is tensed and relaxed. For we assert that the voice
has a natural way of moving, and does not place intervals haphazardly.We try to give
these matters demonstrations which confor m to the appearances, not in the manner
of our predecessors, some of whom used arguments quite extraneous to the subject,
dismissing perception as inaccurate and inventing theoretical explanations, and saying
that it is in ratios of numbers and relative speeds that the high and the low come
about. Their accounts are altogether extraneous, and totally in conict with the appearances. Others delivered oracular utterances on individual topics, without giving
explanations of demonstrations, and without even properly enumerating the perceptual data. We, on the other hand, try to adopt initial pr inciples which are all evident
to anyone exper ienced in music, and to demonstrate what follows from them.11

This account of harmonics reveals Aristoxenus to be in conict not only with the
Pythagoreans (the unnamed antagonists who dismiss perception and explain pitch
relations through ratios) but also with earlier har monic theor ists whose empir ical
work he found decient because they did not explain their methods of proof or
properly descr ibe their observations.
Aristoxenuss alter native was to apply Aristotles intellectual method to music
more rigorously than did Aristotle himself .12 This entailed restricting the account
of music to ter ms and concepts that could properly be said to belong to the domain
of music. It excluded descr iptions that made recourse to ratios (which are in the
domain of number) or to theor ies about the propagation of physical sound (which
11. Aristoxenus, Elementa harmonica, in Harmonic and Acoustic Theory (vol. 2 of Greek Musical Writings),
ed. and trans. Andrew Barker, Cambridge Readings in the Literature of Music (Cambr idge: Cambridge
University Press, 1989), 149 50.
12. For commentar y on Aristoxenuss approach, see Barker, Greek Musical Writings, 2: 66 69, 119.

conc e p tual i z i ng mu s i c
are in the domain of physics). Once the denition of these basic musical concepts
was accomplished, an explanation of the entire domain of music could then follow.
Aristoxenuss demonstration proceeds in three steps. First, he identies two
forms of vocal motion, the continuous and the intervallic. In the continuous for m,
which is associated with speech, the voice appears to traverse space without stopping, until the point of silence. In the intervallic for m, which is associated with
singing, the voice appears to stand still at specic points, and then pass over some
interval of space before coming to rest at another point.13 The various pitches upon
which the voice pauses when singing constitute musical notes; the spaces between
these notes are musical intervals.
Aristoxenuss second step is to draw distinctions among the various musical
intervals. The rst distinction is made with regard to magnitude, which reects the
amount of space between the two notes of the interval. That such a space exists is
inferred from the difference between the two pitches that adjoin the interval; the
size of the space can be reckoned in ter ms of how many other notes could be put
inside it.14 The second distinction is made with regard to concord and discord. Aristoxenus identies the concordant intervals as the fourth, fth, and octave (and their
octave duplications). These are the only concordant intervals he accepts as determined by the intr insic nature of melody all other intervals are by denition
discordant.15
The nal step toward assembling the basic denitions and pr inciples of harmonics is the der ivation and division of the tone. Aristoxenus denes the tone as the
difference between the rst two concords (the fourth and the fth) and explains
that it can be divided in half (yielding the semitone), in thirds (yielding the least
chromatic diesis), or in fourths (yielding the least enhar monic diesis, which is the
smallest interval recognized as melodic).16 These distinctions allow him to locate
various notes within the tetrachords of the different genera and thereby to specify
the scalar structure of each.
With these steps, Aristoxenus lays out the basic mater ials for his account of musical organization. Singing involves a specic way of using the voice that creates
musical notes and musical intervals. Intervals can be distinguished according to size
and whether they are concordant or discordant. Concordant intervals the fourth,
fth, and octave are accepted as axiomatic to melody and thus representative of
the basic mater ials from which the various intervals of the Greater Perfect System
can be developed.

13. Aristoxenus, Elementa harmonica, 133.

14. Aristoxenus, Elementa harmonica, 136.
15. Aristoxenus, Elementa harmonica, 139. Aristoxenus further distinguishes between composite and
incomposite intervals (p. 137); however, this distinction is not necessary for a basic understanding of his
theory.
16. Aristoxenus, Elementa harmonica, 140. The different dieses apply to the three genera mentioned
above: the half-tone is used in the diatonic genus; the third-tone is used in the chromatic genus; and the
fourth-tone is used in the enhar monic genus. For further discussion of Aristoxenuss Elementa harmonica, see Mathiesen, Apollos Lyre, 319 34.

11

12

i nt roduc t i on

Greek Theory and Cognitive Structure

As mentioned, the music theor ies of Pythagoras and Aristoxenus belong to a world
remote from our own. Not only did these theor ists have to grapple with the most
basic of principles, but also the music they would descr ibe is a microtonal one that
is primarily concer ned with the successive notes of melody rather than the simultaneous notes of harmony. Despite this or perhaps because of it Pythagorean
and Aristoxenian accounts of musical organization give us a glimpse into how theories are formed and, more important, the cognitive processes that are basic to these
theories. In particular, three cognitive processes can be seen at work: categorization,
cross-domain mapping, and the use of conceptual models.

categorization

Our ability to categor ize things is a cognitive process so

basic and so pervasive that it can easily escape our notice.Were you to lift your eyes
from this book and survey your sur roundings, you might well see chairs, lamps,
tables, and other books; were you outside, you might see trees, birds, clouds, cars, and
bicycles. If you considered the other things that populate your day, you might think
of friends and family members, facial expressions and gestures, actions and activities.
Your recognition of these things reects the categor ies through which we structure
our thought: to recognize a book is to identify it as a member of the category book;
to recognize a tree is to identify it as a member of the category tree. Categorization
occurs in all sensor y modalities and throughout the range of mental activities: we
categor ize smells and sounds, thoughts and emotions, skin sensations and physical
movement.
Categor ies are not just basic to thought; they also give insight into our thought
processes. At one time it was thought that categor ies reected the structure of the
real world, but recent research has shown that the categor ies humans use are shaped
by their interactions with their environments. Our reasons for developing and
employing a given category are part and parcel of the category itself: categor ies are
not only not given by nature, but also they are subject to change and modication
as our thought unfolds.
Two categor ies basic to Pythagorean and Aristoxenian music theory are those for
consonant or dissonant intervals. Consonant intervals (such as the octave, fth, and
fourth) are fundamental to the conceptualization of Greek music: they mark the stable pitches of the Greater Perfect System and are the source of derivation for all further intervals, both consonant and dissonant. The process of categor ization is also
exhaustive: any interval that can be conceived belongs to one of these two categories.
This is not to say, however, that consonant and dissonant intervals are given by
nature in any simple way, Pythagoras and the blacksmiths notwithstanding.17 As one
17. I should note here that psychoacousticians distinguish between musical consonance, which is a
cultural construct framed relative to a particular set of musical practices, and sensory consonance, which
is a consequence of how sound waves are processed by the hear ing mechanism (which involves the
cochlea and the auditory cortex). Sensory consonance is thus a fairly straightforward product of nature.
Although musical consonance has its basis in sensory consonance, there is some freedom in how the sensory data are inter preted.

conc e p tual i z i ng mu s i c
example, consider the way Aristoxenians and Pythagoreans classied the interval of
an octave plus a fourth. Aristoxenians considered the interval a consonance, since it
was simply the combination of two smaller consonances. Pythagoreans, in contrast,
classied intervals according to the numer ical ratio for med by their constituent
pitches. As explained by the anonymous (and thoroughly Pythagorean) author of
the Sectio canonis (fourth century b.c.), consonant intervals are those whose ratios are
either multiple (of the for m [mn]:n) or epimor ic (of the for m [n + 1]:n). Dissonant
intervals are those whose ratios are epimer ic (of the for m [n + m]:n, where m is
greater than 1 and neither equal to nor a multiple of n).18 Because the octave plus
a fourth had the epimer ic ratio 8:3, it was regarded as a dissonance.
Another example of how categor ies shape our understanding of phenomena is
provided by Greek theor ists treatment of thirds and sixths. Although thirds and
sixths sound fairly consonant, they were nonetheless categor ized as discords. Two
factors bear on this classication. First, forming thirds and sixths requires using the
movable pitches of the Greater Perfect System at best, a third or a sixth will
involve only one of the stable pitches bounding the constituent tetrachords of the
system. Thirds and sixths were intervals that necessar ily varied in size, and so they
were placed among the dissonances. Second, in the classication of intervals Greek
theory followed a tradition of dichotomous categor ies: there was concord, discord,
and nothing else. By contrast, neither of these factors played a part in the music theory of early India. Indian music theor ists were consequently free to focus on the
qualitative aspect of intervals rather than on their cor respondence with the xed
notes of a tuning system and to construe intervallic relationships as concordant, discordant, or neutral.19
These two examples show that while the categor ies for consonant and dissonant
intervals may be basic to Pythagorean and Aristoxenian theor y, just how they are
dened reects the context and goals of categorization: consonance and dissonance
are not naturally occur ring properties, but ways of constructing an understanding of
musical organization.
Of course, there are numerous other categor ies important for Pythagorean and
Aristoxenian music theor y, including those for pitches, intervals, and numer ical
ratios. These categor ies and others are basic to the sort of systematic account of
musical phenomena provided by these theor ies indeed, it is simply not possible to
have a theory of music, or of anything else, without rst having categor ies.

cross-domain mapping

Cross-domain mapping is a process through

which we structure our understanding of one domain (which is typically unfamiliar or abstract) in ter ms of another (which is most often familiar and concrete). For
example, one way to think about the elusive concepts of electr ical conductance is
in terms of a hydraulic model: ipping the light switch tur ns on the juice, and elec18. Sectio canonis, in Harmonic and Acoustic Theory (vol. 2 of Greek Musical Writings), ed. Andrew
Barker, Cambridge Readings in the Literature of Music (Cambr idge: Cambridge University Press, 1989),
193.
19. Lewis Rowell, Music and Musical Thought in Early India, Chicago Studies in Ethnomusicology
(Chicago: University of Chicago Press, 1992), 157 60.

13

14

i nt roduc t i on

trical cur rent ows to the light bulb to light the room. By this means we take what
we know about a fairly concrete and familiar source domain the ow of water
and other liquids and map it onto a rather abstract and unfamiliar target domain:
that of electr icity. As a wealth of research on analogy and metaphor has shown, the
process of mapping structure from one domain to another is basic to human understanding.
One place cross-domain mapping is evident is in the Pythagorean and Aristoxenian construal of interval. Because musical pitches are ephemeral and virtually
intangible, relationships between pitches musical intervals represent something
of a challenge to understanding. One way to meet this challenge is to map structure
from the physical world onto music, a process evident in Nicomachuss story of
Pythagoras and the blacksmiths. Pythagoras hears harmonious sounds, traces their
origins to the blacksmiths hammers, and then proceeds to conduct various experiments using weights equivalent to those of the hammers. These exper iments lead,
among other things, to a highly pragmatic objectication of musical pitch, as pitches
are translated into physical objects that can be weighed, studied, and preserved. By
performing a mapping from the concrete physical domain proper to the blacksmiths hammers onto the domain of musical sound, Nicomachuss story allows us
to structure the latter domain in ter ms of the for mer. Of course, musical notes are
not physical objects that can be weighed, studied, and preserved they remain
ephemeral and virtually intangible. Nonetheless, we are so accustomed to the mapping between concrete physical objects and musical sound that we sometimes have
to be reminded that notes are not endur ing physical objects.
Aristoxenuss construal of musical interval involves a slightly different mapping.
As we have seen, according to Aristoxenus, when the voice moves intervallically, it
appears to stand still at a given place (a musical pitch) and then pass over an interval of space (a musical interval) before coming to rest at another place (another
musical pitch). Underlying this account is a mapping from the familiar domain of
two-dimensional space onto that of music. This mapping allows us to apply the
methodology of measur ing space to music. The difference between two linear measures yields a third measure; similarly, the difference between the intervals of a fth
and a fourth yields the interval of a tone. Since linear measures can be easily divided
into equal halves or thirds or fourths, the musical tone can be similarly divided,
something impossible from the Pythagorean perspective.
On closer inspection, the Pythagorean and Aristoxenian construals of interval are
indeed incommensurate. From the Pythagorean perspective, pitches are physical
objects, and an interval descr ibes the relationship between these objects. From the
Aristoxenian perspective, pitches are breadthless points that simply mark out an
expanse of two-dimensional space, and an interval is the expanse itself. Each mapping gives an account of interval, but each leads to a different conceptualization of
musical structure. This point can be generalized for music theory as a whole: mapping structure from a nonmusical domain onto music is a way of creating musical
structure, and different mappings will lead to different accounts of musical structure.

conce ptual mode ls

Both categor ization and cross-domain mapping

provide the basis for fundamental ontological assertions about musical mater ials: this

conc e p tual i z i ng mu s i c
interval is a consonance; the pitches of an octave are physical objects. They can also
lead to conditional statements: if the interval is an octave, then it is a consonance; if
a pitch is an object, then its properties are measurable. Propositions like this are basic
to conceptual models, which act as guides to reasoning and inference. In their simplest for m, conceptual models consist of concepts in specied relationships, which
pertain to a specic domain of knowledge.
For an example of a conceptual model, let us return to the classication of consonant and dissonant intervals presented in the Sectio canonis, according to which all
consonant intervals have either multiple or epimor ic ratios, and all dissonant intervals have epimer ic ratios. This classicator y system relies on a conceptual model
that organizes concepts related to interval, concord, discord, and the three classes of
ratios. The simple patter n of inference that follows from this model is that if an
interval has a multiple or an epimor ic ratio, it is a concord; if it has an epimer ic
ratio, it is a discord.
Integral to this model are the products of categor ization and cross-domain mapping. Two types of categories are involved in the model: those pertaining to music
(the categor ies of concord and discord) and those pertaining to number (the multiple, epimoric, and epimer ic ratios). Cross-domain mapping cor relates the two types
of categor ies by construing musical interval as a relationship between two objects
(namely, musical pitches) to which magnitudes (in the for m of numbers) can be
assigned. Specic classes of ratios can then be used to distinguish between the musical categor ies.
The robustness of this particular conceptual model is reected in the debate over
the status of the octave plus a fourth that continued into the Middle Ages. In the
second centur y a.d., Ptolemy showed the speciousness of the cor relation of concord with multiple or epimor ic ratios and argued for a classication of intervals
based on empir ical evaluation and the postulate that a concord added to a concord
produces a concord.20 Although Ptolemy still used ratios to descr ibe various intervals, they were no longer part of the conceptual model through which intervals
were classied into concords and discords. In the sixth century, Boethius presented
both the Pythagorean and Ptolemaic models but took no position on which he preferred.21 After Boethius, when an author wished to invoke the author ity of the
Pythagorean approach, the Pythagorean model of intervallic classication was cited;
when an author wished for a more empirically satisfying classication, the Ptolemaic model was used.22
Conceptual models provide the rst level of organization for concepts. They are
too limited and localized, however, to provide the comprehensiveness we expect
from theor ies of music. Theories achieve this comprehensiveness by integrating
20. Ptolemy, Harmonics, in Harmonic and Acoustic Theory (vol. 2 of Greek Musical Writings), ed. Andrew
Barker, Cambridge Readings in the Literature of Music (Cambr idge: Cambridge University Press, 1989),
286 90.
21. Boethius, Fundamentals of Music, 81 82, 169.
22. For a discussion of these modes of reasoning, see C. Andr Barbera, The Consonant Eleventh
and the Expansion of the Musical Tetractys: A Study of Ancient Pythagoreanism, Journal of Music Theory 28 (1984): 191 223. Barberas treatment of Aurelians Musica disciplina on p. 210 is especially illuminating.

15

16

i nt roduc t i on

clusters of conceptual models. And as we shall see in the following chapters, conceptual models also play a role in categor ization and cross-domain mapping. There,
as in theor ies, they provide guides for reasoning and inference about specic and
circumscr ibed domains of knowledge.

cognitive processes and music theory Much has been left out
of this discussion of theories of music from Greek antiquity, with respect to both the
theories themselves and the cognitive processes behind them. To be sure, these theories are of a different order than Swanns musings on Vinteuils sonata. Nonetheless, the cognitive processes we have seen at work in Pythagorean and Aristoxenian
theory are the same processes through which we organize our understanding of the
world as a whole. Just how this occurs how categor ization, cross-domain mapping, and the use of conceptual models shape our theor ies of music and guide our
analyses of musical works is the subject of the remainder of this book.

overview
I have divided the chapters that follow into two parts. In the rst, I present a
detailed overview of research on the three cognitive processes highlighted in this
introduction. This overview is itself framed around specic musical topics, such as
motivic transfor mation, text painting, and the ways in which we structure our
understanding of a specic musical domain. The research that has been done in
cognitive science over the past three decades has been extensive and far ranging, and
one of the jobs of this portion of the book is to br ing this work to bear on basic
issues of musical understanding. Another objective is to show in some detail how
these processes relate to one another and how they form the bedrock for our
thought about music. The second part of the book moves from this foundation to
analytical studies of specic musical issues. These issues include relationships
between categor ization and musical syntax, the problem of musical ontology, textmusic relations, and conceptions of musical for m and musical hierarchy.
Chapter 1 begins the overview of research in cognitive science with a close look
at processes of categorization. For centur ies, writers in the West regarded categor ies
as xed and immutable, and any variation in categor ization was taken as evidence of
the failure of the human intellect to deal with the structure of the real world. It took
the pioneer ing work of Eleanor Rosch and others in the 1970s to show that category structure was not as simple as rst believed. In particular, some levels of categorization are preferred over others, and some members of a category are regarded
as better representing the categor y than others (a phenomenon known as graded
membership).
The key to how this research can be applied to music is provided by the musical motive (or, as Proust would have it, the motif ). Motives are generally reckoned to
be one of the basic building blocks of musical works, but they are also a bit slippery:
the same motive typically assumes a number of diverse shapes over the course of
a work. Thinking of a motive as a cognitive categor y makes it possible to account
for its identity, as well as its diversity, and reveals how aspects of categorization are
embodied by musical mater ials. These preliminary applications of categorization to

conc e p tual i z i ng mu s i c
music also show ways musical mater ials can be organized over the course of a work
and offer an explanation of how it is possible to have musical concepts that are independent of language.
If categor ization can be said to be the source of musical concepts, cross-domain
mapping is the means by which these concepts are placed in cor relation with others. Chapter 2 examines the process of cross-domain mapping in some detail, beginning with the work of cognitive linguists who, in the 1980s, proposed that metaphor
was a basic structure of understanding. This proposal gained added weight when it
was shown that metaphor ical projection (which is one way to accomplish crossdomain mappings) was a general process not restricted to linguistic expressions but
grounded in embodied exper ience.
One example of cross-domain mapping that involves music in a rather immediate way is the technique of text painting, a compositional device that aims to represent in music specic images summoned by the text of a vocal work. Text painting provides a point of departure for the exploration of how cross-domain mapping
is manifested in our understanding of music, as it leads to an extension of crossdomain mapping called conceptual blending. In a conceptual blend, elements from
two cor related domains are projected into a third, giving rise to a r ich set of possibilities for the imag ination. As I show in the latter part of chapter 2, text painting
can lead to such blends, as can program music.
Chapter 3, which focuses on conceptual models and theor ies, gets to the heart of
the perspective on cognition developed in part I. My point of departure is research
by Jeanne Bamberger on childrens representations of musical structure. In my
analysis of Bambergers study of one specic eight-year-old boy, I show the part
played by categor ization and cross-domain mapping in the conceptual models used
by this boy to come to ter ms with a musical environment. I also show how these
models are combined to for m a theor y of music and how this theor y changes in
response to changes in the task at hand. This close-up glimpse of the structure and
role of conceptual models and theor ies leads, in the middle of the chapter, to a more
generalized character ization of these knowledge structures, which I connect with
work on similar structures in articial intelligence, cognitive anthropology, ethnomusicology, and developmental psychology. In the latter part of the chapter, I return
to music theory and explore the role of conceptual models and theor ies (that is,
theories framed relative to a cognitive perspective) in analyses by Jean-Philippe
Rameau and Heinr ich Schenker, two of the best-known music theor ists of the last
three hundred years.
Although the features of cognitive structure discussed in part I might seem to
be relatively detailed, in truth all are associated with relatively high-level cognitive
processes. My reason for focusing on this level is quite simple: it allows me to engage
in issues of immediate and occasionally central importance to music scholarship and
to do so in a way that connects with extensive research in cognitive psychology and
cognitive linguistics. Part II explores this possibility in g reater depth by considering various problems of musical understanding from the perspective on cognitive
structure and music theory developed in part I.
In chapter 4, I turn to the matter of how musical mater ials are organized within
a work more properly, the problem of musical syntax and, by extension, musical

17

18

i nt roduc t i on

semiotics. Although semioticians are usually quick to grant that music has a syntax,
they are more doubtful about whether its semantic level has any depth. By taking
a close look at how composers make use of categor ies of musical events in this
case, the way Mozart and Beethoven use motives in the opening movements of
three str ing quartets I am able to provide insight into how musical mater ials are
organized in the service of musical discourse, as well as how features of this organization contr ibute to meaning construction as a whole.
In chapter 5, I confront a somewhat larger problem one that may seem ir redeemably abstract: the problem of what counts as a work of music. I view this problem, usually called the problem of musical ontology, as one of cultural knowledge
and try to show that, as opposed to being hopelessly recondite, the problem is of
immediate importance for understanding music. By approaching the entire work of
music as a categor y a category that includes all the scores, performances, representations, and such that are said to be of the piece I develop a model for the
cultural knowledge upon which judgments about musical ontology are made.
Determinations of what counts as an instance of a particular musical work are thus
one of the ways members of a musical community construct and negotiate their
identity. My examples for this chapter are two songs taken from the traditions of
popular music and jazz: I Got Rhythm and Bye Bye Blackbird. The latter offers
an intr iguing case of how the cultural knowledge relative to which deter minations
of musical ontology are made can become complicated when implicated in the layered discourse structures Mikhail Bakhtin called double-voiced discourse, and
which were extended to African American culture through Henry Louis Gates Jr.s
notion of Signifyin(g).
Chapter 6 returns to the analysis of individual musical works by pursuing one
of the entailments of text painting noted in chapter 2: under certain circumstances,
combinations of words and music, through the process of conceptual blending, create worlds for the imagination well beyond those that spr ing from words or music
alone.Where only a few fairly circumscr ibed instances of text-music relations were
considered in chapter 2, here research on conceptual blending is applied to the
whole of ve Lieder from the nineteenth centur y. These analyses offer a way to
esh out the theory of conceptual blending as it applies to music and provide a further perspective on musical syntax. In these songs, we see words and music combining to create r ich domains in which the imagination can play, as well as discover
how musical syntax shapes our understandings of the words themselves.
The nal analytical chapter (chap. 7) turns to music theory itself, specically to
the theor ies of musical for m and hierarchy that go back to the eighteenth and early
nineteenth centur ies. Accounts of the for m a musical work takes, or of how its elements relate to one another, are basic to theor izing about music indeed, we can
see these emerging in the course of M. Swanns ruminations on Vinteuils sonata
but at times it seems that theorists are talking about quite different things. For instance,
there are two common ways to talk about musical for m: the rst approaches for m
as der iving from the assembly of relatively static building blocks that are combined
to create the nished work; the second approaches for m as an emergent property of
the work, which becomes manifest only as the music unfolds over time. The rst
approach yields a view of musical for m that is quite static, the second a view rather

conc e p tual i z i ng mu s i c
more dynamic. Using the analytical framework provided by cross-domain mapping,
I discuss the source of these two models for musical for m, as well as two models for
musical hierarchy, and explore some of the ways these models interacted over the
course of the history of music theory.
In the conclusion I return to M. Swann and to his nal encounter with Vinteuils
sonata after a year in which it became thoroughly intertwined with his love affair
with Odette, the courtesan with whom he had become acquainted around the same
time he rst encountered the andante. This will provide a frame for a review of the
points made in the preceding chapters and an instrumentality for drawing conclusions from the whole.

cognitive structure,
theory, and analysis
A central claim of this book is that through developing an appreciation of how
aspects of cognitive structure shape our understanding of music we can better appreciate the active role of theories of music in that understanding. A further claim is
that our analyses of musical phenomena from the most mundane and localized of
accounts to the most abstract and comprehensive similarly reect cognitive structure, in that every analysis is based on some sort of theory of music. Musical analyses are in truth dialogues, and not just dialogues between the analyst and an imagined audience: musical analyses are also dialogues between the analyst and some
body of theoretical knowledge. Analysis rarely, if ever, simply cor roborates a theory:
analysis pulls theor y and pushes it, extending and chang ing theory just as it also
extends and changes our understanding of musical phenomena.
The analyses I present throughout this book are no different, except that they
engage cognitive theory as well as music theor y. The intent of the analyses is to
show how our understanding of particular musical phenomena can be character ized
in terms of specic cognitive processes and structures and thereby connect that
understanding to research in cognitive science as a whole. The analyses are not intended
as denitive statements about how we can account for such understanding; they are
intended to be the initiation of a dialogue with cognitive theory, a dialogue whose
purpose is to expand our knowledge of both music and cognition. Analysis is thus
a central concer n of what follows, but, unlike cognitive structure or theor y, I have
not treated it as a central topic for investigation. Instead, analysis will be a fundamental tool to explore both and to provide new insight into the conceptual worlds
wrought by musical sound.

19

This page intentionally left blank

part i

aspects of cognitive

structure

This page intentionally left blank

chapter one

categorization

n the opening eleven measures of the prelude to Richard Wagners Tristan und
Isolde there are three successive statements of a melodic motive that foreshadows,
and then later accompanies, the tragic arc of the opera. Three times this motive
emerges from silence, swells through the cellos and English hor n, and then sinks
back into the stillness from which it came. After the third statement, the motive disappears, then reappears a few long minutes later near the end of the prelude. Here
it is stated rst in the cellos and violas (and completed by the oboes and bassoons),
then in the English hor n, and nally in the bass clar inet. After the last of these, the
prelude dissolves into the rst scene of the opera. All of these statements are shown
in example 1.1.
Although our rst inclination is to treat these various statements as somehow the
same that is, all can be regarded as statements of the opening motive from Wagners Tristan und Isolde there are actually notable differences among them. There
is, of course, the change in the instrumentation used for the statements at the end of
the prelude, which contrasts with the distinctive timbral stamp put on the motive in
the opening measures. There is a similar diversity in the pitches through which successive statements are given voice, since versions of the motive begin on either A3
(or Ab3), B3 (or B2), or D4. Although there are signicant similar ities in the rhythmic guration used for the various statements, there is no patter n of duration and
accentuation common to all. And there are also somewhat smaller variations: the
rst statement begins with a minor sixth instead of the major sixth used by the others, and the fourth is the only one shaped by both a diminuendo and a rallentando.
Despite their differences, these seven melodic fragments sound similar enough to
one another to be regarded as functionally equivalent. Such equivalence was important to Wagners compositional style, for it provided the mater ial foundation for
works uninter rupted by conventional for ms. By recalling specic motives at crucial points within his musical and dramatic discourse Wagner was able to create a
sense of shape and unity that seemed inevitable without being predictable. By the
time of Tristan, however, such motives were no longer simple building blocks to
be assembled and per muted but had become innitely malleable constructs. Carl
Dahlhaus, recognizing both the structural importance and mutability of these

23

24

aspe c t s of cog n i t ive st ruc ture

example 1.1
Isolde, Prelude

Statements of the opening motive from Richard Wagners Tristan und

Vnc. 1

Eng. Hn.

1.

Vnc. 5

Eng. Hn.

2.
cresc.

dim.

Vnc.

Eng. Hn.

3.
cresc.

85

Vnc., Vla.

Ob., Bn.

4.
dim.
88

Vnc., Vla.

Ob.

5.
Eng. Hn.
101

6.

Bs. Cl. 104

7.
pi

motives within the opera, likened them to woven threads that surface only to divide
and then disappear from sight.1
To understand this music to make sense of the sonic texture Wagner weaves
requires being able to assimilate these various musical phrases into a single cognitive
construct and then recall that construct, often after an hour or more of Wagnerian
effusion. Understanding Wagner or most music, for that matter requires being
able to think in ter ms of categor ies of musical events. In this chapter, I explore what
recent research into categor ization can tell us about this process. First, however, I
develop a slightly clearer picture of two basic aspects of understanding music, which
will provide the g roundwork for applying research on categor ization to situations
of the sort presented by the opening of Tristan. I do this by reviewing a theory of
musical motive developed by one of the many students of Wagner. Not that Wag1. Carl Dahlhaus, Richard Wagners Music Dramas, trans. Mary Whitall (Cambr idge: Cambridge University Press, 1979), 63.

cate g ori zat i on

ner ever actually taught composition. But the inuence of his music was enor mous,
and many lessons were taken away from his operas. Among these were lessons on
compositional technique absorbed, rened, and retransmitted in both treatise and
score by Arnold Schoenberg.

musical coherence and motive

Schoenberg thought of himself rst and foremost as a Brahmsian, but, under the
inuence of Alexander von Zemlinsky, he also found a passion for Wagner. Schoenberg noted that his early works showed the inuence of both masters: In my Verklrte Nacht the thematic construction is based on Wagner ian model and sequence
above a roving harmony on the one hand, and on Brahmss technique of developing variation as I call it on the other.2 Wagner was also known for his harmonic innovation, although Schoenberg took pains in one of his late essays to show
that Brahmss harmonies were every bit as progressive as Wagners.3 Nonetheless,
it was Wagners harmonic language that was perceived as posing the most for midable challenge to composers of the late nineteenth century, for the free treatment of
dissonance integral to this language loosened prevailing conceptions of tonal organization and called into question the laws of tonality that had for merly held sway.
From Schoenbergs perspective,Wagners music presented few dangers for tonality. As Schoenberg had argued in his Harmonielehre of 1911, tonality had its or igins
in acoustic phenomena.4 But shaping music in confor mance with acoustic phenomena was not all there was to musical composition: composition also required
insight into how the human mind worked. In an essay that attempted to sort out
some of the problems raised by the emancipation of the dissonance Schoenberg
wrote:
Tonalitys origin is found and rightly so in the laws of sound. But there are other
laws that music obeys, apart from these and the laws that result from the combination
of time and sound: namely, those governing the working of our minds. This latter
forces us to nd a particular kind of layout for those elements that make for cohesion and to make them come to the fore, often enough and with enough plasticity so that in the small amount of time granted us by the ow of the events, we can
recognize the [musical] gures, grasp the way they hang together, and comprehend
their meaning.5

In Schoenbergs view, the laws governing the workings of our minds require the
composer to wr ite in such a way that listeners can quickly recognize musical gures
and the way they cohere. The listener, upon grasping this coherence, will then be
able to comprehend the work.
Key to musical coherence and comprehension were musical motives. According
2. Arnold Schoenberg, My Evolution (1949), in Style and Idea: Selected Writings of Arnold Schoenberg, ed. Leonard Stein, trans. Leo Black (New York: St. Martins Press, 1975), 80.
3. Arnold Schoenberg, Brahms the Progressive (1947), in Style and Idea, 398 441.
4. Arnold Schoenberg, Harmonielehre (Leipzig:Verlagseigentum der Universal-Edition, 1911); trans.
as Theory of Harmony by Roy E. Carter (Berkeley: University of California Press, 1978), 23 25.
5. Arnold Schoenberg, Opinion or Insight? (1926), in Style and Idea, 259.

25

26

aspe c t s of cog n i t ive st ruc ture

to Schoenberg, comprehension starts with recognition, and recognition starts with

basic musical gures that is, motives: Motive is at any one time the smallest part
of a piece or section of a piece that, despite change and variation, is recognizable as
present throughout.6 More specically, motives consist of intervals and rhythmic
patterns combined to produce a shape or contour that, once recognized, can be easily remembered.7 A particularly str iking illustration of Schoenbergs concept of
motive is provided by his analysis of a subordinate theme from the rst movement
of Brahmss Sextet in Bb major, Op. 18 (1860), shown in example 1.2.8
As Patricia Car penter and Severine Neff have noted, rhythm is central to
Schoenbergs concept of motive.9 This centrality is clearly evident in his analysis of
the Brahms sextet. Each family of motives is distinguished by a specic rhythmic
patter n: all of the a motives use a qde pattern; and all of the b motives use a
qdeq pattern. (Both patter ns, of course, contribute to the marvelous waltz-like feel
of the theme.) Contour also serves to distinguish the two families, although here
some variation among the different motive forms can be noted. Motives a1, a2, and
a4 all preserve the basic contour of motive a: the second note of the motive is lower
than the rst note, and the last note of the motive is higher than the rst note.
Motive a3 preserves only a portion of the contour of a: although the second note of
the motive is lower than the rst, so is the last note. In a similar fashion, motives b1
and b3 preserve the basic contour of motive b (the second note is lower than the
rst note, and the third note is lower than the second), while b2 preserves only a
portion of the contour of b (the second note is lower than the rst note, but the
third note is higher than the second).
As this example shows, Schoenbergs concept of motive is broad and dynamic:
a family of motives can be distinguished simply by rhythmic gure or contour;
motive forms can be various and need have only some of their features in common.
This conception is rather different from, for example, Heinrich Schenkers or Rudolph
example 1.2
mm. 84 93

Schoenbergs analysis of the Brahms Sextet, Op. 18, rst movement,

85

89

a2

a1

b2

b3

a3

b1

a4

6. Arnold Schoenberg, The Musical Idea and the Logic, Technique, and Art of Its Presentation, ed. and
trans. Patricia Car penter and Severine Neff (New York: Columbia University Press, 1995), 169.
7. Arnold Schoenberg, Fundamentals of Musical Composition, ed. Gerald Strang and Leonard Stein
(London: Faber and Faber, 1967), 8.
8. Schoenberg, Brahms the Progressive, 417.
9. See the commentary in their edition of Schoenbergs The Musical Idea and the Logic, Technique, and
Art of Its Presentation, 27 29.

cate g ori zat i on

Rtis concept of motive. Both Schenker and Rti regard specic intervallic relationships as constituent of motive; rhythm and contour are regarded as secondar y
aspects of motivic organization.10 The difference between Schenkers or Rtis conception of motive and that of Schoenberg lies with Schoenbergs notion of coherence and the role it plays in making music comprehensible. Coherence comes about
when the various parts that make up a musical entity are connected in such a way
that those parts similar to other entities become prominent. The work is most comprehensible to the listener when the ar rangement of these parts is such that their relationship to each other and to the whole is manifest.11 Thus there is no need to
restrict the distinguishing features of a family of motives to specic intervallic relationships, nor is there a need to require that all motive forms share exactly the same
features. All that is required of motive is recognizability and a potential for connection to other motive forms. This potential for connection contr ibutes to coherence,
which is, in turn, the basis for comprehensibility.
According to Schoenberg, then, the process of comprehension starts with recognizable bits (motives) that are easily remembered. Motives hang together not simply because their constituent parts are connected to one another but because these
connections emphasize similar ities to other motives. Coherence thus reects properties shared by collections of motives; it is not, properly speaking, a property of any
one individual motive. Motive forms are of necessity variable, for differences between
forms reveal most clearly what is typical of the collection of motives as a whole.
And, although attention to coherence is important to the composer who wishes to
craft a convincing work, the apprehension of coherence is essential to the listener
who would make sense of that work.
Many composers, working in the long shadow cast by Ludwig van Beethoven
across the nineteenth century, relied on musical motives to make their works coherent. As Edward Cone has noted, the use of thematic transfor mation as a unifying
device became increasingly important for composers of the later nineteenth century, composers whose lengthy movements often embraced stylistic extremes of tempo,
meter, texture, and mood.12 However, the prevalence of the device also poses problems for the listener, for it is often difcult to sort out relationships among various
motivic entities. Consider the relationship of derivation, in which later motive
forms are derived from earlier motive forms; thus motive b3 in example 1.2 is said
to be der ived from motive b. As straightforward as this might appear, tracing der ivational relationships is rarely a simple task: motive b can itself be understood to be
derived from one of the motives prevalent in the preliminary subordinate theme
heard starting in m. 61 of Brahmss sextet (shown in ex. 1.3), which has the same
rhythmic patter n but the reverse contour. Could b3 then be regarded as der ived
directly from this earlier motive form, or is it still to be regarded as der ived from b?
Schuberts Wanderer Fantasie, Op. 15 of 1822 presents an even more complicated
10. Heinrich Schenker, Der freie Satz, Neue musikalische Theor ien und Phantasien, 3 (Vienna: Universal-Edition A.G., 1935); Rudolph Rti, The Thematic Process in Music (New York: Macmillan, 1951). I
discuss histor ical construals of motive below in the rst part of chap. 7.
11. Arnold Schoenberg, Coherence, Counterpoint, Instrumentation, Instruction in Form, ed. Severine
Neff, trans. Severine Neff and Charlotte M. Cross (Lincoln: University of Nebraska Press, 1994), 20 23.
12. Edward T. Cone, On Der ivation: Syntax and Rhetor ic, Music Analysis 6 (1987): 238.

27

28

aspe c t s of cog n i t ive st ruc ture

example 1.3
61

Johannes Brahms, Sextet, Op. 18, rst movement, mm. 61 64

in tempo

dolce

dolce

dolce
pizz.

dolce
pizz.

situation. The work is based on an ar ietta from Schuberts 1816 cantata Der Wanderer, Op. 4/1, D. 489, which is quoted in the works second movement Adagio;
the relevant passage from the cantata appears in example 1.4a. If we have the ar ietta in mind, we can hear the motivic mater ial of the rst movement as der ived
from the ar ietta and then presented instrumentally in the subsequent Adagio (the
motive from the rst movement is labeled c in ex. 1.4b; its use in the Adagio is
demonstrated in ex. 1.4c).13 The temporal order of the process of derivation is thus
reversed: c is heard as der ived from c1. If we are coming to the work fresh, we will
most likely hear the Adagio as der ived from the rst movement: c1 is thus der ived
from c. Finally, if we know the ar ietta but have temporar ily forgotten it, its reprise
in the Adagio, anticipated by the derived motives of the rst movement, can be a
moment of great rhetor ical impact: the pianistic musings of the rst movement have
summoned forth the ar ietta that now stands before us.
It would seem that, as Cone argued, hearing relationships among motive forms
is an ineluctably subjective affair. Following his argument still further, how we hear
these relationships is important to how we respond to a work, for, as Schuberts fantasy suggests, it infor ms our understanding of musical rhetor ic. The most dramatic
case occurs in a work in which there are multiple der ivations of motivic mater ial.
In the course of such a work we can lose our sense of how motive forms are connected, until the composer often unexpectedly reveals how the mater ials relate
to one another by bringing them into rapprochement. The rhetor ical aim of this
sort of compositional strategy is what Cone calls epiphany, and, when successful,
13. While the annotations in ex. 1.4 focus on the distinctive rhythmic gure common to each of the
excerpts, it is worth noting that each passage also features a move to the dominant.

cate g ori zat i on

example 1.4 Franz Schubert, (a) Der Wanderer, Op. 4 No. 1, D. 489, mm. 23 26; (b)
Wandererfantasie, D. 760, mm. 1 3; (c) Wandererfantasie, mm. 189 90
a

23

Die Son

ne dnkt mich hier so

kalt, die Bl

te welk, das Le

ben alt,

b
1 Allegro con fuoco ma non troppo

c
c
189

Adagio

c1

it can compel the listener to realize a previously unsuspected or at most unconrmed relationship among diverse motives.14 Put another way, the strategy that
leads to epiphany compels the listener to realize aspects of musical coherence that
become evident only over the course of time.
Motives and motivic relationships are important for both the coherence and the
rhetor ic of music, but understanding the part they play in musical understanding
poses some unique challenges. If the recognition of motives is central to the process
of comprehending music, it must occur very rapidly. This rapidity has led some to
argue for cognitive modules dedicated to the processing of such basic musical materials.15 The theory of cognitive modular ity, an adaptation of faculty psychology rst
proposed by the philosopher Jerry Fodor, bears some resemblance to the notion of
competence discussed br iey in my introduction.16 The coherence of musical
14. Cone, On Derivation, 246.
15. See, for example, David Temperley, Motivic Perception and Modular ity, Music Perception 13
(1995): 141 69; and Eugene Nar mour, The Analysis and Cognition of Basic Melodic Structures: The Implication-Realization Model (Chicago: University of Chicago Press, 1990).
16. Jerry A. Fodor, The Modularity of Mind: An Essay on Faculty Psychology (Cambridge, Mass.: MIT
Press, 1983). According to the theor y Fodor proposed in The Modularity of Mind, cognitive structure is
organized into modules dedicated to specic processing tasks. There are thus modules for such things

29

30

aspe c t s of cog n i t ive st ruc ture

materials, however, is an emergent property that is less well explained from a modular perspective since it involves the active compar ison and evaluation of a number
of motive forms.
One model that can r ise to these challenges is offered by recent research into
categor ization, which has demonstrated that categor ization is both extremely rapid
and evaluative. From the perspective provided by this research, the coherence of
music reects our ability to group musical events into categor ies: the a and b motives
of example 1.2 each constitutes a cognitive categor y, with ve members in one
category (a, a1, a2, a3, and a4) and four in the other (b, b1, b2, and b3). To descr ibe
the laws of the mind that Schoenberg noted laws that force the composer to
arrange musical elements so that they cohere is to descr ibe the process of categorization. In what follows, I shall review research on processes of categor ization
and show how it can be used to account for the role of motive as a starting place
for higher-level cognitive processes and for relationships among diverse motive
forms.

processes of categorization
For the most part, the categor ies through which we structure our understanding of
the world are not cause for a great deal of reection. To all appearances, these categories are simply givens. A book is a book, a tree is a tree, and any equivocal cases
are just a matter of insufcient knowledge. This commonsense view of categorization was the predominant view in philosophy and psychology at least since the time
of Plato: categories, because they were constitutive of knowledge, were regarded as
stable and universal. It was only in the 1950s that the adequacy of this view began
to be called into question, most notably by Ludwig Wittgenstein in his Philosophical Investigations. There, in the midst of a discussion of language games, Wittgenstein
paused to consider the category of game itself and found that category structure
was far more complicated than it rst appeared.17 Research carried out in the 1960s
and 1970s showed that numerous aspects of the way humans categor ize could not
be reconciled with the traditional view of categorization, and by the 1980s a new
view of categorization had begun to develop. The categor ies through which
humans organize their understanding of the world came to be seen as ways of having knowledge, rather than as reecting what could be known.
The research crucial to developing this new view of categor ization focused on
two topics. The rst concer ns the relationship between the process of categorization and the taxonomic structures to which it gives r ise. The second concer ns the

as vision, hearing, language processing, and, presumably, the identication of musical motives. Fodor has
since made arguments against modular ity; see his recent The Mind Doesnt Work That Way (Cambridge,
Mass.: MIT Press, 2000).
17. Ludwig Wittgenstein, Philosophical Investigations, trans. G. E. M. Anscombe (New York: Macmillan, 1953), 31 32. For summar ies of the recent history of research into categor ization, see Eleanor Rosch,
Categor ization, in Encyclopedia of Human Behavior, ed.V. S. Ramachandran (San Diego: Academic Press,
1994), 1: 513 23; and George Lakoff, Women, Fire, and Dangerous Things: What Categories Reveal about the
Mind (Chicago: University of Chicago Press, 1987), chap. 2.

cate g ori zat i on

internal structure of categor ies that is, how members of a category relate to one
another.
Most of the categor ies we use in daily life can be organized into hierarchically
arranged taxonomies. The cat prowling around my feet is a calico cat, a housecat, a
feline, a mammal, and a living organism. The levels in taxonomies of this sort are related
by inclusion: feline includes all of the different sorts of housecats, a category that
includes all of the different sorts of calico cats, a category that includes my particular
cat. In categor izing things, it seems logical to work from the top of the hierarchy
down, moving from the most abstract level to the most concrete. I note rst that the
thing before me is a living organism, and as I increase my level of discrimination
determine exactly what sort of organism it is. In actually categor izing things in the
world around us, however, we start neither at the most specic level (the cat
prowling around my feet) nor at the most general (living organism) but at a level in
between; this level has come to be called the basic level. Research on the basic level
tells us where, for practical pur poses, the process of categorization begins. The rst
section that follows reviews research on the basic level and shows how it supports
Schoenbergs contention that musical comprehension begins at the level of the
motive.
It is similarly a commonplace to regard each level of a taxonomy as having clear,
xed boundar ies: a creature either is or is not a housecat. Categories are thus unchanging anchors of understanding, their structure immutable and constant. The
categories we actually use in daily life, however, cannot be so neatly character ized.
These categor ies have a somewhat more complicated structure: membership is by
degrees; category boundar ies are relatively uid and often hazy; and category structure can change over time. In these cases, the process of categorization is a highly
dynamic one, requiring that we evaluate potential categor y members relative to
existing categor y members which, together, dene a standard of typicality. Categories with a variable membership are thus said to show typicality effects. The second section that follows reviews research on categor ies that illustrate typicality
effects and shows how it can be used to model the coherence and variability of
motive forms basic to Schoenbergs theory of motive.

Basic-Level Categor ization

the basic leve l

The psychologist Roger Brown was perhaps the rst to

note that the categor ies we most frequently employ are not at the lowest level of a
taxonomy (and concer ned with individuals), nor are they at the highest level (and
concer ned with broad classes); instead, they are in the middle of a taxonomy, at a
level of maximum utility.18 For instance, we typically call the thing in our purse or
pocket a dime, even though we might also call it money or a metal object (names at the
superordinate level of the taxonomy), or my favorite 1956 dime (which would be at
the subordinate level of the taxonomy).
18. Roger Brown, How Shall a Thing Be Called? Psychological Review 65 (1958): 14 21; idem,
Social Psychology (New York: Free Press, 1965), 319 21.

31

32

aspe c t s of cog n i t ive st ruc ture

table 1.1. Sample Taxonomies Showing the Basic Level
Superordinate

Basic Level

Subordinates

Musical instrument

Guitar

Folk guitar
Classical guitar
Grand piano
Upright piano
Delicious apple
Mackintosh apple
Freestone peach
Cling peach
Kitchen table
Dining room table
Floor lamp
Desk lamp

Piano
Fruit

Apple
Peach

Furniture

Table
Lamp

Adapted from Eleanor Rosch et al., Basic Objects in Natural Categor ies, Cognitive Psychology 8
(1976): 388.

Subsequent research into ethnobotanical systems of classication by Brent Berlin

and his associates supported the importance of mid-level classication and gave
empirical evidence that this level is psychologically basic: at this level, people name
things more readily, languages have simpler names, categories have greater cultural
signicance, things are remembered more readily, and things are perceived holistically.19 Research directed specically at mid-level classication was car ried out by
Eleanor Rosch and her associates in the early 1970s. In their report, this level was
named the basic level, a term which has been generally adopted in the literature on
categor ization.20 Examples of sample taxonomies showing the basic level are given
in Table 1.1.
Rosch and her associates suggested that two contrasting pr inciples inuence the
taxonomic level at which people prefer to categor ize.21 The rst is the efciency principle, according to which people prefer to minimize the number of categor ies they
19. Brent Berlin, Speculations on the Growth of Ethnobotanical Nomenclature, Language in Society 1 (1972): 51 86; idem, Ethnobotanical Classication, in Cognition and Categorization, ed. Eleanor
Rosch and Barbara B. Lloyd (Hillsdale, N.J.: Erlbaum Associates, 1978), 9 26; Brent Berlin, Dennis E.
Breedlove, and Peter H. Raven, Principles of Tzeltal Plant Classication: An Introduction to the Botanical
Ethnography of a Mayan-Speaking People of Chiapas (New York: Academic Press, 1974).
20. Eleanor Rosch, Carolyn B. Mervis,Wayne D. Gray, David M. Johnson, and Penny Boyes-Braem,
Basic Objects in Natural Categor ies, Cognitive Psychology 8 (1976): 382 439.
21. This inter pretation of Roschs work is drawn from Lawrence W. Barsalou, Cognitive Psychology:
An Overview for Cognitive Scientists (Hillsdale, N.J.: Erlbaum Associates, 1992), 181 84. See also Rosch et
al., Basic Objects in Natural Categor ies; Eleanor Rosch, Principles of Categorization, in Cognition
and Categorization, ed. Eleanor Rosch and Barbara B. Lloyd (Hillsdale, N.J.: Erlbaum Associates, 1978),
27 48; Carolyn B. Mervis and Eleanor Rosch, Categor ization of Natural Objects, Annual Review of
Psychology 32 (1981): 89 115; and Mary E. Lassaline, Edward J.Wisniewski, and Douglas L. Medin, Basic
Levels in Articial and Natural Categor ies: Are All Basic Levels Created Equal? in Percepts, Concepts, and
Categories: The Representation and Processing of Information, ed. Barbara Bur ns, Advances in Psychology, 93
(Amsterdam: North-Holland, 1992), 327 78.

cate g ori zat i on

must consider in making a categor ization. For instance, if you were concer ned with
things in my house, categor izing at the level of musical instrument would yield the most
efcient categor y, since the number of contrasting categor ies at this level (such as
furniture and fruit) is signicantly less than at the next level (guitar, piano, apple, peach,
table, lamp). The second pr inciple is the informativeness principle, according to which
people tend to maximize the infor mativeness of their categor izations. Since the
most infor mation about any entity is found at the most specic level of a taxonomy,
you would use grand piano to categor ize the thing sitting in the living room. Rosch
and her associates proposed that the inter mediate level of a taxonomy (in this case,
piano) optimizes both efciency and infor mativeness and is thus the preferred level
for basic categor izations. A number of empirical operations converge at the basic
level. The basic level is the highest level whose members have similar and recognizable shapes; it is also the most abstract level for which a single mental image can
be formed for the category. The basic level is also the highest level at which a person uses similar motor actions for interacting with categor y members. The basic
level is psychologically basic: it is the level at which subjects are fastest at identifying
category members, the level with the most commonly used labels for categor y
members, the rst level named and understood by children, the rst level to enter
the lexicon of a language, and the level with the shortest pr imary lexemes.22
In her initial work, Rosch assumed that the basic level depended on bundles of
perceived-world attr ibutes that for med natural discontinuities and that such attr ibutes were inherent in the real world. She later came to doubt this assumption, and
subsequent research has shown that the basic level is variable. It depends on cultural
context and level of expertise and on whether one is dealing with categor ies that
occur naturally, as humans interact with their environments, or with categor ies created for exper imental ends.23 Other research has indicated that the basic level is
strongly deter mined by the ways in which humans interact with their environment
and that it is highly dependent on the way humans perceive the spatial conguration of physical parts.24 In all cases, however, Roger Browns fundamental insight has
held up: the preferred categor ization for most of the things we encounter occupies
a maximally useful level in the middle of a taxonomy.
22. Rosch et al., Basic Objects in Natural Categor ies; Eleanor Rosch, Human Categor ization,
in Studies in Cross-Cultural Psychology, vol. 1, ed. Neil Warren (London: Academic Press, 1977), 1 49; Barbara Tversky and Kathleen Hemenway, Objects, Parts, and Categories, Journal of Experimental Psychology: General 113 (1984): 169 93.
23. On Roschs reservations about the objectivity of the basic level, see Rosch, Principles of Categorization; on the relationship between expertise and the basic level, see James W. Tanaka and Marjor ie
Taylor, Object Categor ies and Expertise: Is the Basic Level in the Eye of the Beholder? Cognitive Psychology 23 (1991): 457 82; on the contrast between natural and articial categor ies in this context, see
Lassaline et al., Basic Levels in Articial and Natural Categor ies.
24. On relationships between the basic level and the environment, see Barbara Tversky and Kathleen Hemenway, Categor ies of Environmental Scenes, Cognitive Psychology 15 (1983): 121 49; on the
basic level and humans interactions with physical objects, see Tversky and Hemenway, Objects, Parts,
and Categor ies; Irving Bieder man, Recognition-by-Components: A Theor y of Human Image
Understanding, Psychological Review 94 (1987): 115 47; and Lawrence W. Barsalou, Deriving Categories to Achieve Goals, Psychology of Learning and Motivation: Advances in Research and Theory 27
(1991): 1 64.

33

34

aspe c t s of cog n i t ive st ruc ture

the basic level and musical motive s

In a close analogue to
what are, perhaps, the most distinctive features of basic-level categor ies, Schoenbergs theory of musical coherence begins not with individual musical events or
with four- or eight-measure phrases, but at a level somewhere in between. Although the motive is the smallest recognizable part of a musical work, it is in fact
made up of still smaller parts namely, its constituent pitches, intervals, and durations.
The cognitive salience of the motive thus mir rors that of the basic level: in both
cases, the focus is on the whole rather than on the parts. Each of Schoenbergs families of motives has a distinctive shape or contour that allows us to distinguish it
from other motives. Similarly, the basic level is the highest taxonomic level at which
category members have similarly perceived overall shapes and the highest level at
which a single mental image can reect the entire categor y. Finally, motive recognition is not something given directly by perception and can change with expertise:
what counts as a motive, just as what counts as the basic level, is not immutable.
To illustrate cor respondences between a musical motive and a basic-level category, let us consider the opening of Beethovens Fifth Symphony, given in example
1.5. The passage is a familiar one, and it illustrates important matters of musical
organization that we will later explore in Wagners Tristan.25 To the extent that we
hear a motive in these opening gestures, we hear a gure that compr ises four
notes. The motive has a distinctive shape or contour that allows us to distinguish
it from other motives (although such are in rather short supply in this particular
example). Lastly, there is a certain minimal level of expertise involved in picking out
mm. 1 2 and its ilk as the motive. One can easily imag ine an alter nate hear ing
that would take mm. 1 2 and mm. 3 4 as the main thematic mater ial for this section; the ensuing music would then represent a sort of fragmentation of this thematic mater ial, rather than successive statements of various for ms of the motive.26

summary Research on basic-level categor ies provides support for Schoenbergs proposal that musical comprehension beg ins at the motivic level, for such a
level maximizes both efciency and infor mativeness. In addition, there appear to be
a number of correlations between musical motives and basic-level categor ies. Both
are concer ned with wholes, rather than the parts of these wholes; shape (whether as
visual property or auditory analogue) is important to both; and both are subject to
the inuence of expertise. Finally, while entities such as motives may be basic to
the process of comprehending music, they are actually relatively high-level cognitive constructs (as are basic-level categor ies). Although such constructs are fundamental to reason (and may even represent an important starting point for conscious
25. That a passage from Beethoven might be used to shed light on Wagners compositional strategies
is not accidental, for Beethovens compositional style had a profound inuence on Wagner. For discussion, see Klaus Kropnger, Wagner and Beethoven: Richard Wagners Reception of Beethoven, trans. Peter
Palmer (Cambr idge: Cambridge University Press, 1991); and Thomas S. Grey, Wagners Musical Prose:
Texts and Contexts (Cambridge: Cambridge University Press, 1995).
26. It is worth noting, however, the role of the fermatas in suggesting that mm. 1 2 are one entity
and mm. 3 4 another, since these add time for the listener to process the musical infor mation presented
by Beethoven. This additional time may also play a role in what is considered the most typical form
of the motive, a topic explored in greater depth later in this chapter.

example 1.5 Ludwig van Beethoven, Symphony No. 5 in C Minor, Op. 67, rst
movement, mm. 1 38
1

Allegro con brio ( = 108)

Flutes

Oboes

a2
Clarinets in B

Bassoons

Horns in E

Trumpets in C

Timpani in C, G

Allegro con brio ( = 108)

Violin I

Violin II

Viola

Cello

Bass

(continued)

36

aspe c t s of cog n i t ive st ruc ture

example 1.5

(continued)

a2

reasoning processes), they are the product of myriad other still more-basic processes
that operate below the level of consciousness.

Typicality Effects

graded structure and a typology of categories

As noted,

the variability of motive forms was important to Schoenbergs notion of musical

coherence, for this variability tended to emphasize how motive forms were similar to

cate g ori zat i on

example 1.5

(continued)

14

cresc.

cresc.

cresc.

cresc.

cresc.

cresc.

cresc.

cresc.

cresc.

cresc.

cresc.

cresc.

(continued)

one another. Research on categor ization has demonstrated an analogous structure for
numerous categor ies. Membership in these categor ies is not xed but is instead
graded through a dynamic process in which the attr ibutes of potential category
members are compared with the attr ibutes most typically found within the category.
As an example of such a graded structure, consider the category bird. Experimental
rankings show that subjects view robins and sparrows as the best examples of birds, with

37

38

aspe c t s of cog n i t ive st ruc ture

example 1.5

(continued)

20

owls and eagles lower down in the rankings, and ostriches, emus, and penguins among
the worst examples.27 All are considered members of the category bird, but some better represent the category than others. Consequently, category structure is graded
according to typicality: category members range from the most typical to the least typ27. Eleanor Rosch, On the Inter nal Structure of Perceptual and Semantic Categor ies, in Cognitive Development and the Acquisition of Language, ed. Timothy E. Moore (New York: Academic Press, 1973),
111 44; idem, Cognitive Representations of Semantic Categor ies, Journal of Experimental Psychology:
General 104 (1975): 192 233.

cate g ori zat i on

example 1.5

(continued)

27

(continued)

ical, with the former securely inside the bounds of the category (robins and sparrows)
and the latter in danger of being excluded from the category (emus and penguins).
Graded membership has been shown to be pervasive among what are conventionally called natural categories, so named based on their emergence from the
interaction of humans with their natural environments.28 For the sake of clarity, and
28. On the pervasiveness of natural categor ies, see Lawrence W. Barsalou, The Instability of Graded
Structure: Implications for the Nature of Concepts, in Concepts and Conceptual Development: Ecological
and Intellectual Factors in Categorization, ed. Ulric Neisser (Cambr idge: Cambridge University Press, 1987),
102. On the ter m natural applied to certain categor ies, see Barsalou, Cognitive Psychology, chap. 2.

39

40

aspe c t s of cog n i t ive st ruc ture

example 1.5

(continued)

33

cresc.

cresc.

cresc.

cresc.

cresc.

cresc.

cresc.

to avoid certain of the associations summoned by the ter m natural, I call this sort
of category a Type 1 category.
Type 1 categor ies stand in contrast to categor ies in which membership is based
on individually necessar y and jointly sufcient conditions, such that the members
of category x must have features y and z, and if something has features y and z it is
a member of category x. Because the conception of categories of this latter sort is
often traced backed to the work of Aristotle, such categor ies have been called clas-

cate g ori zat i on

sical categories; because the structure of these categor ies is not typically an analog
representation of the natural world, such categor ies have also been called articial.29
I call this sort of category a Type 2 category. Most evidence indicates that Type 2
categories simply represent a specialized for m of Type 1 categor ies take a Type 1
category, specify limits for the categor y through the imposition of necessar y and
sufcient conditions for category membership, and you create a Type 2 category.30

prototype effects and frame structure

Eleanor Rosch and

Carolyn Mervis, in their pioneer ing work on Type 1 categor ies, suggested that such
categories were organized around a stable cognitive construct called a prototype,
which encapsulated the statistically most-prevalent features of members of the category and against which potential category members were compared.31 Type 1 categories were thus said to exhibit prototype effects. It is clear, however, that the
prototype offers more structure than simply a record of the attr ibutes cor related
with a particular category: although wings and beak are both highly cor related with
bird, we know that wings are used for locomotion and beaks are used for eating. Relationships between attr ibutes, as well as the values assigned these attr ibutes, are thus
important to the function of the prototype.
As a way of captur ing the relational structure inherent in most categor ies,
Lawrence Barsalou proposed using the notion of a frame, as developed in articial
intelligence research, to represent category structure.32 As Barsalou has shown, the
29. On classical or articial categor ies, see Edward E. Smith and Douglas L. Medin, Categories and
Concepts (Cambridge, Mass.: Harvard University Press, 1981), chap. 1; Sharon Lee Armstrong, Lila R.
Gleitman, and Henry Gleitman, What Some Concepts Might Not Be, Cognition 13 (1983): 263 308;
Lila R. Gleitman, Sharon Lee Armstrong, and Henry Gleitman, On Doubting the Concept Concept,
in New Trends in Conceptual Representation: Challenges to Piagets Theory?, ed. Ellin Kofsky Scholnick
(Hillsdale, N.J.: Erlbaum Associates, 1983), 87 110; Lakoff, Women, Fire, and Dangerous Things, chap. 2;
Barsalou, Cognitive Psychology, chap. 2.
30. Although the typology of categories I introduce here places a burden on those familiar with the
traditional labels, it has two advantages. First, it avoids the complications caused if natural is inter preted
as given by nature; while Type 1 categor ies result from humans interactions with their environment,
they are not given by that environment in any simple way. Second, the typology classies categor ies on the
basis of their internal structure. This is especially useful when consider ing research on categor ization, since
articial categor ies (such as things to pack in a small suitcase, as discussed by Barsalou in his Deriving Categories to Achieve Goals) often have the same basic structure as do natural categories (such as bird).
31. Eleanor Rosch and Carolyn B. Mervis, Family Resemblances, Cognitive Psychology 7 (1975):
573 605. See also James Hampton, Prototype Models of Concept Representation, in Categories and
Concepts: Theoretical Views and Inductive Analysis, ed. Iven van Mechelen, James Hampton, Ryszard S.
Michalski, and Peter Theuns, Cognitive Science Ser ies (London: Academic Press, 1993), 67 95.
32. Barsalou, Cognitive Psychology, chap. 7; idem, Frames, Concepts, and Conceptual Fields, in
Frames, Fields, and Contrasts: New Essays in Semantic and Lexical Organization, ed. Adrienne Lehrer and
Eva Feder Kittay (Hillsdale, N.J.: Erlbaum Associates, 1992), 21 74; idem, Flexibility, Structure, and Linguistic Vagary in Concepts: Manifestations of a Compositional System of Perceptual Symbols, in Theories of Memory, ed. A. F. Collins, S. E. Gathercole, M. A. Conway, and P. E. Morris (Hillsdale, N.J.: Erlbaum
Associates, 1993), 29 101.
The exploration and implementation of frame structures in work on articial intelligence is demonstrated in Marvin Minsky, A Framework for Representing Knowledge, in The Psychology of Computer
Vision, ed. Patrick Henry Winston (New York: McGraw-Hill, 1975), 211 77; Marvin Minsky, The Society
of Mind (New York: Simon and Schuster, 1985); and Douglas B. Lenat and R. V. Guha, Building Large

41

42

aspe c t s of cog n i t ive st ruc ture

attributes
size

values
large
medium

color

individuals

small

brown

white

wren

cardinal

red

Bird

green
chirps

sound

clucks

chicken

sings
squawks

locomotion

flies *

parrot

runs

figure 1.1 A partial frame for the category bird (asterisks indicate the values of the
prototype that are stored in memory)

basic structure of a Type 1 category can be conceived of as a relational network

made up of relatively abstract attr ibutes, to which concrete values are assigned by
each individual. For the categor y bird, for example, attributes might include size,
color, sound, and locomotion, as shown in gure 1.1. For the attr ibute size, values might
include large, medium, and small. Each member of the categor y exemplies specic
values for each of the attr ibutes: a wren is small, brown, chirps, and ies; a chicken is
large, white, clucks, and (for the most part) runs.
Of course, individuals with none of the attr ibutes of the category would be
excluded from it. Prototype effects ar ise from the compar ison of the values represented by individuals with the values of the prototype for the categor y that are
stored in memory (indicated with aster isks in g. 1.1). If the values of the prototype
for bird were small, brown, chirps, and ies, then a wren would be most typical of the
category, a male cardinal (small, red, sings, ies) would be somewhat less typical, and
a chicken (large, white, clucks, runs) would be least typical.

applications to music Returning to the opening of the Fifth Symphony, we can see that, as suggested by Schoenbergs account of motive, there are a
Knowledge-Based Systems: Representation and Inference in the Cyc Project (Reading, Mass.: Addison Wesley,
1989).

cate g ori zat i on

number of different for ms that Beethovens motive takes in the rst thirty-seven
measures of the movement, and these vary with respect to contour, dynamics, and
orchestration. Approached from the perspective of categorization, each motive form
can be viewed as a member of the categor y motive forms from the opening of Beethovens Fifth Symphony.33 Attributes of this category might include things such as
orchestration, dynamic, and melodic prole. In the following discussion, I show how
these three attr ibutes could play a part in deter mining membership in the category
motive forms from the opening of Beethovens Fifth Symphony; however, I should emphasize that, for present pur poses, what is of concer n is less the specic attr ibutes and
more the model of categor ization they illustrate.
To character ize melodic prole, it is useful to account for intervallic relationships
among motive forms in ter ms of diatonic intervals. That is, within the diatonic context of the movement, the descending third from G4 to Eb4 can be reckoned to be
equivalent to the descending third from F4 to D4, even though the for mer is major
and the latter is minor. Here I represent a descending third with 3, an ascending
third with +3, and a unison with a u.With this in mind, the attr ibute-values for the
rst two statements of the motive would be tutti, fortissimo, and u, u, 3.
Although rhythmic prole is an equally important attr ibute of the category, it is
not as useful for distinguishing among the various motive forms found in the opening measures shown in example 1.5. All the motive forms found there would assign
the same value to this attr ibute: short, short, short, long (where short is equivalent to
an eighth note and long varies between a quarter note and three tied half notes). In
consequence, the attr ibute rhythmic prole is relatively transparent in the opening of
the symphony, as is the attr ibute played by instruments; both will be omitted from the
following account of category structure.
Figure 1.2 presents a frame diag ram for the categor y motive forms from the opening of Beethovens Fifth Symphony; it shows the values that thirteen instances of the
motive assign to the attr ibutes orchestration, dynamic, and diatonic melodic prole. The
result is a rather dense diag ram (owing in part to the number of individuals categorized), which gives an intr iguing and rather counter intuitive picture of what is
most typical of the various motive forms gathered there. Based on the evidence provided by these individuals, the most typical for m of the motive would assign the
values solo, piano, and u, u, 3 to the relevant attr ibutes. Although the last value is
what we might expect, the rst two are not: it is difcult to think of Beethovens
theme as anything other than tutti and fortissimo.
Were we to take into account a further fteen instances of the motive, the prole
33. What follows is by no means the rst attempt to use research on categor ization to model variation among musical elements; notable are the studies of Robert L.Welker,Abstraction of Themes from
Melodic Variations, Journal of Experimental Psychology: Human Perception and Performance 8 (1982): 435 47;
and Lucy Pollard-Gott, Emergence of Thematic Concepts in Repeated Listening to Music, Cognitive
Psychology 15 (1983): 66 94. The following extends this work by taking advantage of research done since
the early 1980s and by endeavoring to give a more thorough account of the process of categor izing
musical events. Mention should also be made of Robert O. Gjerdingens discussion of typicality in A
Classic Turn of Phrase: Music and the Psychology of Convention, Studies in the Cr iticism and Theor y of
Music (Philadelphia: University of Pennsylvania Press, 1988), 93 94, which anticipates many of the
points made here.

43

44

aspe c t s of cog n i t ive st ruc ture

attributes

values

individuals

tutti

orchestration

solo
ensemble

mm. 1-2

mm. 3-5

mm. 6-7

mm. 7-8

f, ff

dynamic

cresc.

mm. 8-9

mm. 10-11

piano
mm. 11-12

mm. 12-13

u, -2, -2
mm. 14-15

diatonic
melodic
profile

u, u, -2
mm. 15-16

u, u, -3
mm. 16-17

u, u, -4
mm. 17-18

u, +2, +2
mm. 18-19

figure 1.2 Diagram of the category structure for the category motive forms from the
opening of Beethovens Fifth Symphony, mm. 1 21

of typicality would be skewed even further away from what we might expect (this
is shown in g. 1.3, where the frame diagram has been rearranged somewhat in the
interests of legibility). The difference between the evidence provided by gures 1.2
and 1.3 and our usual recollection of how Beethovens theme sounds is due in part
to a limitation of frame diagrams: they are essentially statistical representations of the
most common values assigned attr ibutes by individuals. They may not represent the
most important attribute values for a given category, however.

beyond the prototype: conceptual models

Recent empir ical work on categor ization has shown that category structure is even more variable
than had been assumed by the researchers who rst proposed the prototype theory
of category structure.34 In most Type 1 categor ies, some attr ibutes are more important to categor y structure than others. As a result, some of the statistically most
prevalent features of members of a category may be disregarded in for mulating the

34. Armstrong et al., What Some Concepts Might Not Be; Gleitman et al., On Doubting the
Concept Concept ; Barsalou, The Instability of Graded Structure; Lawrence W. Barsalou, Intraconcept Similar ity and Its Implications for Interconcept Similar ity, in Similarity and Analogical Reasoning, ed.
S.Vosniadou and A. Ortony (Cambr idge: Cambridge University Press, 1989), 76 121.

cate g ori zat i on

mm. 1-2

mm. 3-5

tutti

mm. 25-26

solo

mm. 26-27

ensemble
mm. 6-7

mm. 7-8

mm. 8-9

f, ff
cresc.

mm. 27-28

mm. 28-29

mm. 29-30

piano
mm. 30-31

mm. 10-11

mm. 11-12

u, -2, -2

mm. 31-32

u, u, -2
mm. 32-33

mm. 12-13

u, u, -3
mm. 14-15

mm. 15-16

mm. 16-17

u, u, -4

mm. 33-34
(vln. I)

u, u, u,

mm. 33-34
(vln. II, vla.)

u, +2, +2
u, u, +4

mm. 17-18

u, -3, +2
mm. 18-19

mm. 22-24

mm. 34-35

mm. 35-36
(vln. I)

u, +3, u

mm. 35-36
(vln. II, vla.)

+4, u, u

mm. 36-37

figure 1.3 Diagram of the category structure for the category motive forms from the
opening of Beethovens Fifth Symphony, mm. 1 37

prototype. For example, attributes such as natural environment and means of locomotion have a strong inuence on how we categor ize whales, leading us to disregard
such taxonomic attr ibutes as mode of respiration and means of reproduction. In consequence, a whale seems closer to a prototypical sh than to a prototypical mammal,
despite its proper Linnean categor ization.35 Research has also shown that judgments
of typicality change over time, even when no new individuals are categor ized. It
appears that, rather than constituting a stable cognitive entity, the prototype for
a category is actually uid and variable and it changes with the circumstances of
categorization.
An alter native view of category structure suggests that categor ies are organized
around conceptual models. In brief, conceptual models consist of concepts in specied
relationships; each model pertains to a specic domain of knowledge. Such models
are central to reasoning and inference, and they will be discussed in g reater detail
in chapter 3; in what follows, the emphasis is on the role conceptual models play in
categorization, rather than on the structure of such models. With regard to catego35. For a discussion of some of the factors involved in this particular categor ization, see John Dupr,
Are Whales Fish? in Folkbiology, ed. Douglas L. Medin and Scott Atran (Cambr idge, Mass.: MIT Press,
1999), 461 76.

45

46

aspe c t s of cog n i t ive st ruc ture

rization, the primary function of the conceptual model is to supply a guide for reasoning about accepted and potential members of the category. This is accomplished
through a simplied representation of category structure that incor porates knowledge about what values are most typical for a select group of attributes for the given
category. These attr ibutes are selected according to the goals of categor ization,
which are themselves infor med by more global conceptual models applicable to a
broad range of categorization tasks.36 The conceptual model for a g iven category
thus reects infor mation of the sort summar ized in frame diagrams but conditioned
by knowledge about the overall goals of categor ization.
As a means of developing a preliminary understanding of the concepts and relationships constituent of a conceptual model for a category, let us return once more
to the opening of Beethovens symphony and consider a conceptual model for the
category motive forms from the opening of Beethovens Fifth Symphony. The model, diagrammed in gure 1.4, consists of four correlated conceptual elements, character ized
as the things necessary for a melodic fragment to count as an instance of the motive:
the rhythmic patter n h
a statement by the full orchestra
a fortissimo dynamic
an intervallic patter n consisting of three repeated notes followed by a
descending diatonic third37
These elements g ive priority to select attr ibutes of the motive (rhythmic patter n,
orchestration, dynamics, and intervallic prole) and assign default values to these
attributes that is, they dene what counts as a typical member of the category.
This network of attributes and values then provides a guide for categor izing the
various versions of the motive that occur in the opening measures. Accordingly, versions of the motive such as those of mm. 1 2 and mm. 22 24 (which instantiate all
aspects of the conceptual model) would be judged most typical of the category;
motive forms such as those of mm. 6 7 and mm. 29 30 (which instantiate only
two aspects of the conceptual model) would be less typical; and motive forms such
as those of mm. 7 8 and mm. 35 36 (which instantiate only the rhythmic aspect of
the conceptual model) would be least typical of the category.
What, then, of the evidence provided by gure 1.3? That is, what can one make
of the lack of correlation between the conceptual model and the individual category members, since most of these would be reckoned to be less typical members
according to the conceptual model for the categor y? The answer to this question
requires delving a bit deeper into the way conceptual models both reect and shape
our understanding of music.
36. Barsalou, Deriving Categor ies to Achieve Goals; Lawrence W. Barsalou, Wenchi Yeh, Barbara
J. Luka, Karen L. Olseth, Kelly S. Mix, and Ling-Ling Wu, Concepts and Meaning, in Chicago Linguistics Society 29: Papers from the Parasession on the Correspondence of Conceptual, Semantic, and Grammatical
Representations, ed. Kathar ine Beals, Gina Cooke, David Kathman, Sotaro Kita, Karl-Er ik McCullough,
and David Testen (Chicago: University of Chicago, Chicago Linguistics Society, 1993), 23 61.
37. I should emphasize that the conceptual models I descr ibe here are by no means the simplest elements of conceptual structure. Each conceptual cluster of the present model, for instance, could be represented by a still smaller conceptual model cor relating the concepts that make up the cluster.

cate g ori zat i on

qqq h

rhythmic pattern

motive forms from the

opening of Beethovens
Fifth Symphony

stated with full

orchestra

stated at fortissimo
dynamic

u, u, -3
intervallic
pattern

figure 1.4 Conceptual model for the category motive forms from the opening of
Beethovens Fifth Symphony

The conceptual model shown in gure 1.4 gives undeniable priority to the
statements of the motive that open the movement. The preference given the rst
statements of musical mater ials we encounter is not absolute there are, of course,
works in which the pr incipal mater ials emerge only gradually but it does reect
the fact that we must process musical mater ials in time and that it is often the case
that signicance accrues to the rst events in any psychological process. Note that
Beethoven departs from his opening mater ials only gradually; in particular, divergences from the intervallic patter n of the motive (the aspect of the typicality shown
by g. 1.3 that best matches the proposed conceptual model) occur somewhat later
in the opening measures and cor relate with the compositional strateg ies of development and transition.38 The initial statements of the motive thus represent anchor
points for the process of elaboration that Beethoven undertakes: the variations evident in gure 1.3 are variations on the musical mater ials summar ized in the conceptual model of gure 1.4.
The conceptual model of gure 1.4 is also infor med by our ideas about musical
themes: that is, such themes should be strongly and clearly stated at pr ivileged points
within a movement (such as the beg inning of the movement or after a change of
key or tempo), after which they may be modied or embellished in confor mance
with compositional strategies. Our local model for the motive of Beethovens symphony is thus shaped by a global model of what constitutes an appropriate musical
theme. Such global models are developed through abstraction from any number of
local models, and, perhaps more important, they are taken from the broader base of
knowledge that constitutes culture.39 For instance, in the case of Beethovens symphony, the relevant global model for a musical theme reects the inuence of nineteenth-century German and Austrian musical culture and, in particular, the efforts
38. Relationships between processes of categor ization and compositional strategy are developed
more fully in chap. 4.
39. This perspective on culture is explored more fully in chaps. 3 and 5.

47

48

aspe c t s of cog n i t ive st ruc ture

of the nineteenth-centur y critic and theor ist A. B. Marx and others to come to
terms with Beethovens compositional style.40
Our global model of what counts as a theme (which applies to any number of
works by any number of composers) is thus in rapport with the local models through
which we organize our understanding of specic musical works. Without a global
model, it can be difcult to decide what aspects of a phenomenon are relevant for
categorization. Without local models, there is no way to anchor global models to
specic phenomena.
It should be emphasized that the conceptual model sketched here and diagrammed in gure 1.4 represents limited features of a network of interrelated
propositions that might be used to guide reasoning about how motive forms in the
opening of Beethovens symphony relate to one another. The model does not represent, in a simple way, what someone has in mind when confronting this music.
The simplicity of the model (which reects in part the global models relative to
which it is framed) contr ibutes to its efciency. It is specic enough to allow for
clear distinctions between the motive of the opening measures and other thematic
mater ials in the symphony (such as the new theme that enters in m. 63). It is general enough, however, to accommodate the wide range of different for ms the opening motive takes. The model also reects our intuition that there is just one main
form of the opening motive, despite evidence to the contrary. Intuitions such as this
contribute to the plausibility of Type 2 categor ies, with their cr isp boundar ies and
clear denitions. Features of the conceptual model for a g iven categor y will often
be transfor med into the necessar y and sufcient conditions character istic of Type
2 categor ies: take a Type 1 category and specify limits for it by regarding the constituent elements of the conceptual model relative to which it is structured as necessary and sufcient conditions for category membership, and you create a Type 2
category. In descr iptions of music, this process yields textbook denitions of the
sort that often lead to considerable debate among music theor ists. Although such
debates appear to be about music, they are in fact about how to dene the categories through which we organize our understanding of music.

summary: typicality effects and musical motive Categories that show typicality effects offer a way to model both the coherence and the
variability of motive forms basic to Schoenbergs theory of motive. The property of
coherence is analogous to the shared attr ibutes of category members that is, our
40. Marxs is perhaps the most important nineteenth-century voice within the tradition of analyzing the thematic aspects of Beethovens music. Although Marx did not restrict himself to a denition of
theme as such, preferring a broader and more organic account of musical organization, he did offer the
notion of the Satz. Formed from motives, the Satz is a complete self-sufcient musical idea and represents the rst coherent unit of musical organization. See Adolph Ber nhard Marx, Musical Form in the Age
of Beethoven: Selected Writings on Theory and Method, ed. and trans. Scott G. Burnham, Cambridge Studies in Music Theor y and Analysis, 12 (Cambridge: Cambridge University Press, 1997), 68. The Satz is
roughly comparable to Schoenbergs notion of a Grundgestalt, as discussed in Schoenberg, The Musical
Idea, 169. It is no accident, of course, that Marxs formal theor ies reect his efforts to make explicit
Beethovens compositional strategies. That is, to the extent that our notion of what constitutes a theme
resonates with Marxs denition, that notion is infor med by works such as the Fifth Symphony. Further
discussion of Marxs notion of a motive is taken up in chap. 7.

cate g ori zat i on

sense that a collection of things coheres as a categor y reects the attr ibutes shared
by those things. The variability Schoenberg noted is analogous to the variation
among members of a Type 1 category. Just as certain motive forms are taken as the
source for the der ivation of further motive forms, so certain category members will
be regarded as more typical of the category than others. While typicality effects
reect, to a certain extent, statistical invariances among the attr ibute-values demonstrated by category members, they are also a function of the goals of categor ization.
These goals shape the categorys conceptual model, which represents what is regarded
as most typical of the category and which also provides a guide for reasoning and
inference about potential category members.
An important difference between collections of musical motives and cognitive
categories is that the for mer must be processed under the temporal constraints of
musical perfor mance. Under most circumstances, these constraints will infor m the
conceptual model for a categor y of motive forms, with priority given to those
forms that occur early in a work. As the example of Schuberts Wanderer Fantasie
suggests, however, other factors may cause our concept of what constitutes the central motive of a work to change as the piece unfolds.

categorization and
wagnerian leitmotiven
Wagners Leidensmotiv
Approaching the seven statements of the opening motive from Tristan und Isolde
given in example 1.1 as a cognitive categor y offers a way to explain why it is that
we consider them the same even when they are manifestly different, and it gives
some insight into their role as a point of entry into Wagners sonic fantasy. Knowing
something about the structure of cognitive categor ies also gives us a way to glimpse
features of Wagners compositional strategy, which both exploits and challenges our
ability to categor ize.
Before we explore the motive as a category, let us rst situate the motive in its
original context. As shown in example 1.6, the motive is accompanied by harmonies
(provided by the clar inets and bassoons) and a r ising chromatic line in the oboes.
Hans von Wolzogen, in his classication of the motives from Tristan und Isolde (which
he referred to as Leitmotiven), called the main motive the Leidensmotiv (the suffer ing
motive) and the r ising chromatic line the Sehnsuchtsmotiv, or the yearning motive.41
The har mony on the downbeat of m. 2 is traditionally known as the Tristan chord.
As Carl Dahlhaus has noted, all of these elements are intertwined to such an extent
that taking any one of them out of context distorts the whole.42 Nonetheless, it is the

41. Hans von Wolzogen, Thematischer Leitfaden durch die Musik zu Richard Wagners Tristan und
Isolde: Nebste einem Vorworte ber den Sagenstoff des Wagnerschen Dramas (Leipzig: Feodor Reinboth,
1890), 15. This work was also published in an English edition as Guide through the Musical Motives of
Richard Wagners Tristan and Isolde: With a Preface on the Legend and the Poem of Wagners Drama (Leipzig:
Feodor Reinboth, 1889).
42. Dahlhaus, Richard Wagners Music Dramas, 63.

49

50

aspe c t s of cog n i t ive st ruc ture

example 1.6

Opening motivic group from Richard Wagners Tristan und Isolde

1

changes to the Leidensmotiv that are the most noticeable (not the least because it is
the rst sounding element of the ensemble), and so in the following I will focus on
it, with the understanding that it is but one part if a rather prominent one of a
larger whole.43
With these thoughts in mind, let me propose that the categor y statements of the
Leidensmotiv is organized around a conceptual model consisting in two cor related
elements, character ized as the things necessar y for a melodic fragment to count as
an instance of the motive:
an ascending sixth, followed by three descending half-steps
the rhythmic patter n Z eqd q eqd e44

, ,

Again, this model is greatly simplied, for in addition to ignor ing the mater ials that
invariably accompany the Leidensmotiv, nothing is said of dynamics or orchestration, and the notion of a motive or theme is simply assumed (as part of the
global model that frames this local model). The model is sufcient, however, to distinguish the Leidensmotiv from other motivic mater ials in Wagners music drama
and to evaluate the typicality of different versions that appear over the course of the
work.
Turning to the versions of the motive given in example 1.1 we can note the following range of typicality: statements 1, 2, 6, and 7 are in complete confor mance
with the model (they are typical of the category), but statements 3, 4, and 5 confor m
rather less well (they are less typical of the category).Viewed from the perspective of
compositional strategy,Wagner begins with two thoroughly typical members of the
category and then follows them with a markedly less-typical member that adds a
note and changes the rhythmic prole. After a considerable span of music, he returns to the motive, but again in a less than typical for m indeed, the two temporally compressed versions of mm. 85 90 enact the sense of reminiscence they are
meant to evoke by presenting to our ears a distillation of the most salient aspects of
43. Robert P. Morgan has recently offered an analysis of the Tristan Prelude which focuses more on
harmonic structure than on the appearance of motive forms and which offers an intr iguing alter native
to what I present in the following. See Morgan, Circular Form in the Tristan Prelude, Journal of the
American Musicological Society 53 (2000): 69 103. A selection of analyses of the Prelude spanning the
twentieth century can be found in Robert Bailey, ed., Prelude and Transguration from Tristan and Isolde
(New York:W.W. Norton, 1985).
44. The actual duration of the concluding note of the motive is variable; in all cases, however, its
duration is of moderate length relative to that of the other notes that make up the motive.

cate g ori zat i on

the motive. Ten measures further (beginning in m. 100) there are two nal appearances of the motive before the curtain goes up, each a thoroughly typical for m.
Wagners compositional strategy thus falls into three stages. In the rst stage (associated with statements 1 and 2), he establishes the motive. In the second stage (associated with statements 3, 4, and 5), he presents less typical for ms of the motive; these
are cor related with a larger set of developmental strategies, with which the central
portion of the Prelude is concer ned. Finally, in the third stage (associated with statements 6 and 7), he returns to more typical for ms of the motive, preparatory to concluding the Prelude and beginning the music drama proper.45
Two things emerge from this initial sketch. First, in the different versions of the
Leidensmotiv that appear in the Prelude, we have a category that develops over
time. Because categor ization is an active response to the environment, it always has
a temporal dimension: a comparative categor ization of birds, for instance, demands
that we evaluate each in turn.With music the temporal aspect is slightly more prominent, since musical entities are thoroughly ephemeral: not only must we evaluate
each musical entity in tur n, but also this evaluation is practically the only evidence
that the entities ever existed at all. Wagner is clearly conscious of the way the Leidensmotiv functions through time and uses it to g ive shape to the Prelude as a
whole. Second, the view of Wagners compositional strategy der ived from comparing the versions of the motive in the Prelude, fascinating as it might be, is also misleading: the sort of circularity that it suggests is fundamentally at odds with his goal
of perpetually progressive music. By looking at further instances of the Leidensmotiv over the course of Tristan und Isolde we can develop our understanding of both
of these things and rene our inter pretation of Wagners treatment of the motive in
the Prelude.

The Leidensmotiv and Musical Syntax

The Leidensmotiv is prominent at ve moments in Tristan und Isolde: in act 1, scene
2, mm. 26 33 (p. 36; see ex. 1.7a) and scene 5, mm. 439 41, 456 58, and 468 72
(pp. 175 78; ex. 1.7b); in act 2, scene 3, mm. 260 78 (p. 448; ex. 1.7c); and in act 3,
scene 2, mm. 93 107 (pp. 594 96; ex. 1.7d) and scene 3 (near the conclusion of the
opera), mm. 153 57 (p. 630; ex. 1.7e).46 In each case, there are three entr ies of the
motive, beginning (as did those in the opening of the Prelude) on A (or, in the case
of the rst entry in act 3, scene 2, Ab), B, and D.
Each group departs from the conceptual model proposed above in various ways.
The versions of act 1, scene 2 have one less pitch (and, of course, one less interval)
than the model and a different rhythmic prole. Those of act 1, scene 5 have a different rhythmic prole. In act 2, scene 3, the versions of the motive are quite close

45. This syntactic patter n is analogous to the Beg inningMiddleEnd paradigm discussed by Ko
Agawu, but here it is expressed purely in ter ms of categor ies, rather than in ter ms of tonal organization.
See Victor Ko Agawu, Playing with Signs: A Semiotic Interpretation of Classical Music (Princeton, N.J.:
Princeton University Press, 1991), chap. 2.
46. Measure numbers are given by scene; page numbers refer to the Dover reprint of the C. F. Peters
1911 edition of the score: Richard Wagner, Tristan und Isolde (New York: Dover Publications, 1973).

51

52

aspe c t s of cog n i t ive st ruc ture

to the model, but the rst two have extended endings, while those of act 3, scene
2 have an altered rhythmic prole, so that their overall duration is greatly expanded.
Finally, the versions of the motive in act 3, scene 3 are greatly compressed (as are
those of act 1, scene 2). Within each group, however, the versions of the motive are
relatively consistent (with the exception of those of act 1, scene 5, about which
more in a moment), although the last statement (beginning on D) is always a bit different from the preceding two. This consistency suggests that each g roup of three
statements of the Leidensmotiv forms a syntactic unit, which represents its own
local standard of typicality. These units are themselves members of a superordinate
category (to which we could give some sort of cumbersome name like successions
of the Leidensmotiv beginning on A/Ab, B, and D) whose members span almost the
entirety of Wagners music drama.47
For the most par t, the statements that make up these syntactic units are conjunct, and, although they may stretch across a number of measures, the connection
between the successive statements is quite obvious. The one exception comes
with the three statements of act 1, scene 5, which occur at a climactic moment in
the drama. Tristan has just dr unk what he believes is a death potion. Isolde wrests
the cup from him, sings Verrther, ich tr ink sie dir! (Betrayer! I dr ink to
you!), and then empties what remains in the cup. Unbeknownst to either, the
draught they have consumed will lead not to death but will instead cause them to
fall even more deeply in love. The melody for Isoldes unintentionally ironical
Ich tr ink sie dir! is the version of the Leidensmotiv that begins example 1.7b.
For the fteen measures of music that follow, the stage directions read: She
drinks, then throws the cup away. Both [Tristan and Isolde], seized with awe, in
the greatest excitement but motionless, look xedly into each others eyes, in
which the expression of deance of death soon gives way to the glow of love.
The cellos then enter with the second statement of the Leidensmotiv, beginning
on B. Further stage directions follow, to accompany another ten measures of
music: Trembling seizes them: they clutch convulsively at their breasts and pass
their hands over their foreheads. Unison cellos and violas then br ing in the last
statement of the Leidensmotiv, beginning on D. In this instance, what draws the
three versions of the motive together into a single categor y is not so much their
similar ity to one another (although this is still quite audible) or their temporal
proximity, but the way Wagner pr ivileges them within his musical discourse. Not
only is there nothing else to compete with them at the moment of their appearance, but also their par ticipation in this pivotal moment suggests that they are
autonomous agents within the drama itself .
We should, if but briey, consider the notion of syntax developed in the preceding paragraphs. First, the basic idea of syntax that I employ here is what Websters
calls a connected or orderly system. I esh out this idea further in chapter 4; for
now, it is enough to say that this appears to be how at least some music theor ists
47. Robert P. Morgan has also pointed out that each successive phrase within these units takes up the
same notes outlined in the top voice of the preceding one, making it seem that each phrase emerges from
the one just before; see Morgan, Circular Form in the Tristan Prelude, 95, 102.

example 1.7 Statements of the Leidensmotiv from Richard Wagners Tristan und Isolde:
(a) act 1, scene 2, mm. 26 33 (Richard Wagner: Smtliche Werke [WSM], vol. 8, part 1,
mm. 313 20); (b) act 1, scene 5, mm. 439 41, 456 58, 468 72 (WSM; vol. 8, part 1,
mm. 1754 56); (c) act 2, scene 3, mm. 260 78 (WSM, vol. 8, part 2, mm. 1890 1908);
(d) act 3, scene 2, mm. 93 107 (WSM, vol. 8, part 3, mm. 1301 15; (e) act 3, scene 3,
mm. 153 57 (WSM, vol. 8, part 3, mm. 1581 85)
a
26

Eng. Hn.

cresc.

meno
b
439

Isolde:
468

456

Ich trink

sie

Vnc.

dir!
Eng. Hn.

Vnc., Vla.
sehr ausdrucksvoll

c
260

Eng. Hn.

pi
266

273

espress.
d
93

Hns., Bns.

Hns., Bs. Cl.

ausdrucksvoll
101

Eng. Hn., Bs. Cl., Hns.

ausdrucksvoll
e
153

Vnc.

dim.

Vla.

Vn. II

poco cresc.

dim.

54

aspe c t s of cog n i t ive st ruc ture

construe the notion.48 Second, the structural integrity of the syntax associated with
the Leidensmotiv emerges only over time. Although all of its elements are present
from the opening moments of Wagners music drama, conr mation of the syntax
must wait until well into the opera. Third, this syntax is specic to Tristan und Isolde.
Not only would we expect a different motive to be used for another work, but this
motive would almost certainly be deployed in different ways. Musical syntax, at least
with regard to this example, is thus created over time and is idiosyncratic to one
particular composition. Finally, semantic value may become associated with syntactic structure, something evident in Wagners manipulation of the three statements of
the motive in the climactic moment that occurs in act 1, scene 5. By disrupting or
modifying his musical syntax, Wagner can create a type of meaning that is thoroughly and exclusively musical.

Syntactic Structure and Compositional Strategy

As a general type, the category statements of the Leidensmotiv has a temporal span
coextensive with that of Tristan und Isolde. More specically,Wagner carefully articulates the category over the course of the drama, with the motive appear ing in syntactic units of three successive statements so that, with each entry of the motive on
A (or Ab), we expect to hear two further entr ies, replicating the patter n set in the
opening measures. Wagner relies on this expectation in act 1, scene 5 to underline
both the tragedy and passion of Tristans and Isoldes love. From this perspective, the
further appearances of the Leidensmotiv in the Prelude after the initial three seem
somewhat anomalous, for here they are only in pairs: statements on B and D in
mm. 85 90, and on Ab and B in mm. 100 106. But Wagner is playing a deep, even
subliminal, game: buried beneath the overheated music in mm. 80 83 are three
rhythmically displaced versions of the Leidensmotiv, all beginning on Ab (as shown
in ex. 1.8). Although little more than wraiths appear ing in the midst of a fury of
motives, these statements have their continuation indeed, their consequence in
the entrances on B and D in the measures that follow. This hidden syntactic unit
implies that the next round of statements, which start with the entrances on Ab and
B in mm. 100 106, should be followed by one additional entrance of the Leidensmotiv on D.Wagner withholds the third statement, however, and instead calls for the
curtain to go up: in response to our anticipation of a third statement of the motive,
Wagner gives us an entire opera.49
The shadowy appearances of the Leidensmotiv in mm. 80 83 of the Prelude
and the somewhat less-typical versions that follow in mm. 85 90 suggest that Wag48. Cone, in On Der ivation, pairs the organizational system of syntax with the expressive tools
of rhetor ic. As another example of this sort of approach, Robert P. Morgan, in a recent essay, describes
tonality as compr ising not only the abstract collections of diatonic pitches (plus their chromatic alterations), which have signicant symmetr ical properties of their own, but various asymmetr ical syntactical
functions, harmonic and linear, that transfor m these collections into a full-edged compositional language (Morgan, Symmetr ical Form and Common-Practice Tonality, Music Theory Spectrum 20 [1998]:
2, emphasis added).
49. Morgan points out that the har monic syntax for these nal statements is also incomplete; see
Circular Form in the Tristan Prelude, 83.

example 1.8 Statements of the Leidensmotiv from Richard Wagners Tristan und Isolde,
Prelude, mm. 80 83
zu 3.

80
Fl.

Hb.
3

Kl.A.
3

E.H.
3

1.
F.
2.

espress.

pi

espress.

pi

F.
Hr.

3.

E.
4.
E.
1. 2.

Fg.

3.

pi
Bk1.A.
espress.

1.
Tr.F.
1. 2.

Ps.

3.

Btb.

Pk.
cresc.

1.V.
pi

3
3

2.V.
pi
Br.
pi
espress.
Vc.
pi
Kb.
pi

(continued)

example 1.8 (continued)

82

zu 8.

Fl.

Hb.
3

Kl.A.
3

E.H.
3

F.

F.
Hr.
E.

E.
1. 2.

Fg.

3.

Bk1.A.
1.

1. 2. zus.

Tr.F.
pi
1. 2.

Ps.

pi
3.
pi

Btb.

Pk.

1.V.
3

3
3

2.V.

Br.

Vc.
trem.
Kb.

cate g ori zat i on

ner is both playing with and relying on our capacity to remember the motive (a
capacity directly related to our ability to categor ize it). But what Wagner wants is
not the sort of recognition that calls attention to itself , distracting us from the ow
of his drama as we become aware of its artice, but something much more subtle.
In consequence, the motive undergoes continual change, as the different versions
collected in example 1.7 show, the only constant being the opening sixth and
descending half steps. (Contrary to Schoenbergs concept of a motive admittedly
developed with reference to Brahms rhythmic guration is only weakly linked
to the identity of the Leidensmotiv within the opera.) At the same time, the
repetitions of the motive proper to each syntactic unit which set up local, if
ephemeral, standards for typicality give each unit the illusion of stability and thus
of properly representing the Leidensmotiv.What results is a syntax that is in constant transfor mation.
Further proof of Wagners attention to musical memory can be seen in the relative atypicality of two of the rare isolated statements of the Leidensmotiv within
Tristan und Isolde. One such statement occurs in the third scene of act 1. Isolde,
recounting to Brangne how her wrath against Tristan was tur ned to pity in her
rst encounter with him, sings Er sah mir in die Augen. Seines Elendes jammerte
mich! Das Schwert ich lie es fallen! (He looked into my eyes. His anguish
touched my heart! The sword I let it fall!). As shown in example 1.9a, the middle phrase is sung to a modied version of the Leidensmotiv. This statement begins,
unlike any other, on G4, and the chromatic passage that leads away from the E5
extends to four descending half-steps. And, relatively early in the long dialogue with
Tristan that makes up scene 2 of act 2, Isolde sings Wie schmerzte tief die Wunde!
(How deeply the wound smarted), again using a less-typical version of the Leidensmotiv (as shown in ex. 1.9b). The statement is doubled and or namented by the
strings, and its chromatic descent continues unhindered through six half steps. Both
of these atypical appearances of the Leidensmotiv atypical in their for m relative
to the conceptual model of the motive, and atypical in their isolation function as
remembrances of the motive rather than statements of it.
Carl Dahlhaus, in compar ing the treatment of themes in the rst movement of
Anton Br uckners Sixth Symphony to Brahmss method of developing variation,
noted Bruckners use of a similar strategy.When Bruckner returns to a theme, the
example 1.9 Isolated statements of the Leidensmotiv from Richard Wagners Tristan und
Isolde: (a) act 1, scene 3, mm. 135 38 (Richard Wagner: Smtliche Werke [WSM], vol. 8, part
1, mm. 672 75); (b) act 2, scene 2, mm. 341 43 (WSM, vol. 8, part 2, mm. 886 88)
a
135

Isolde:
b

Isolde: wie

Sei nes E

len

des

jam

mer te

mich

341

schmerz

te

tief

die

Wun

de!

57

58

aspe c t s of cog n i t ive st ruc ture

correspondence with the or iginal version is often only approximate. Dahlhaus

comments,
That the one version is able to substitute for the other means, aesthetically, that instead
of developing variation, where each variant represents . . . a consequence of the preceding one and a prerequisite for the next one, Bruckner makes use of an analytically
elusive but clearly perceivable similar ity by association, which makes the later version
seem like a wr itten-out memor y of the earlier one. The logic of discourse, as conceived by Brahms, gives way to a system of approximate cor respondences.50

The preceding analysis of transfor mations to the Leidensmotiv suggests that a more
apt compar ison would have juxtaposed Bruckner to Wagner rather than to Brahms,
for it is written-out memor ies such as those shown in example 1.9 that contribute much to the power of Wagners music dramas: we respond to such memories not with our intellect but in a way that seems far more intuitive and thus far
more deeply felt.
Wagner, of course, could have had no explicit knowledge of how humans structure their understanding through categor ization, since empir ical evidence for this
cognitive process did not begin to accumulate until a hundred years after the 1865
premiere of Tristan und Isolde. And yet Wagner almost assuredly had an implicit
understanding of what Schoenberg called the laws that govern the working of our
minds. Wagners music shows that he knew how to lay out his musical mater ials
to make them come to the fore, often enough and with enough plasticity so that
his listeners could comprehend their meaning. His music also shows that he knew
how to challenge this ability, to modify the structure of categor ies of musical materials so that these same mater ials seemed new, or to have an entirely different ontological status (changing from a theme into a remembrance of a theme). An account
of the way Wagner transfor med the Leidensmotiv cannot exhaust his inexhaustible
music, but it can give a glimpse into the processes through which we come to understand his music, as well as the way Wagners compositional strateg ies both exploit
and challenge that understanding.

categorization and
conceptualizing music
Schoenbergs theory of motive, shaped by his understanding of the constraints on
musical composition imposed by human cognitive processing (his laws of the
mind), offers insight into how we organize our understanding of music in ter ms of
categories of musical events. Motives were, for Schoenberg, where musical recognition started, and the coherence of musical motives was what made musical comprehension possible. As we have seen, these two functions are replicated in cognitive
categor ies. Our preferred level of categor ization is at a level in the middle of a taxonomy the basic level which, like the motive, strikes a compromise between
efciency and infor mativeness. The categor ies we use most often in everyday life
50. Carl Dahlhaus, Nineteenth-Century Music, trans. J. Bradford Robinson (Berkeley: University of
California Press, 1989), 273.

cate g ori zat i on

are not dened in ter ms of individually necessary and jointly sufcient conditions.
Instead, they have variable membership: some members are more typical of the category than others, and membership in the category is graded rather than an all-ornothing affair. Musical motives show a similar amount of variation, and it is through
developing a notion of what they have in common a conceptual model for what
is typical of the category that we begin to comprehend the musical organization
of works like those that fascinated Schoenberg.
As the analysis of Wagners use of the Leidensmotiv shows, Schoenbergs conception of a motive is not adequate for all circumstances (especially in its reliance on
rhythmic guration), nor are composers always content to establish a single model
for a motive and stick with it. Indeed, they may vary it constantly (as did Wagner in
the case of the Leidensmotiv), or they may combine the models for different motive
forms in any number of ways (as Cones work on the problem of motivic der ivation
shows). Although focusing on motives offers a place to start, it is on the more general notion of a musical category free as it is from associations with Beethovens
compositional process, fascinating as that is that I want to focus. While a collection of assorted motive forms is a good example of a musical categor y, categor ies
can be much more various and structured around whatever set of musical relationships seems best to account for what is salient about a particular repertoire. The relevant units may be harmonic (as in a chaconne), involve a repeated bass patter n (as
in a passacaglia), or combine the two (as in the linear cadential patter ns that Susan
McClary has shown provide a structural framework for music of the sixteenth century);51 music of the twentieth centur y suggests an even greater number of possibilities. And music (construed as patter ned sound) may be just one of a number of
contributing factors: one could also envision musical categor ies infor med by the
text being set (of particular importance for liturgical music) or the affect that is to
be summoned (a concer n for some types of dramatic music). In all cases, the choice
and structure of the relevant musical categor ies will be guided by the global conceptual models relative to which the local models for each category are framed.
My proposal, then, is that categories are where our conceptualization of music starts.
If to think is to think in terms of categories, then to think of music is to think in terms
of musical categories.Viewed this way, categories constitute an analogue or perhaps
replacement for the notion of a group as it has been developed by Fred Lerdahl and
Ray Jackendoff. For Lerdahl and Jackendoff, musical understanding begins with the
process of segmenting a musical work into contiguous sequences of pitch-events, drum
beats, or the like.52 Each such segment constitutes a group, and groups are arranged into
a strict hierarchy that ultimately yields the entire work of music.53 Musical categories
51. Susan McClary, Conventional Wisdom: The Content of Musical Form (Berkeley: University of California Press, 2000), 14 16.
52. Fred Lerdahl and Ray Jackendoff, A Generative Theory of Tonal Music (Cambridge, Mass.: MIT
Press, 1983); the process of grouping is descr ibed on p. 13; the denition of a group is given on p. 37.
53. By strict hierarchy, I mean a hierarchy in which the boundar y points at each superordinate
level confor m to boundar y points at the level immediately subordinate. Lerdahl and Jackendoff dene
this sort of organization as hierarchical; see Lerdahl and Jackendoff, A Generative Theory of Tonal Music,
13, 37. Conceptions of hierarchy as they are applied to music are discussed in greater detail in the latter
part of chap. 7.

59

60

aspe c t s of cog n i t ive st ruc ture

can absorb many of the functions of groups specically, the role of the group as the
rst level of organization for musical phenomena but do not require absolute contiguity or str ict hierarchical organization. As a result, a theory of musical categor ies
can easily account for the multiple strands of polyphonic music in a way that grouping theory cannot. Each of the overlapping entr ies of a work in canon, for instance,
could be regarded as a member of a category; this sort of analysis is simply impossible with grouping theory, since overlaps between groups are not per mitted.
If categories are where our conceptualization of music starts, they must be
closely linked to concepts themselves. At the r isk of some confusion, I would like to
go further and suggest that musical categor ies are concepts. Now, to attempt any
denition of a concept is to (at least potentially) embroil myself in a long-standing
debate within philosophy and cognitive science.54 To skirt this debate, I concentrate
on a denition of concept that reects recent work in the mind sciences and the
brain sciences, and that can be applied to music. Although I think this denition can
inform the larger debate on concept, I shall leave the substance of the debate for
another time.
The denition of concept I develop here is inuenced by the work of Gerald
Edelman, who has been interested in developing a biological approach to consciousness. Edelmans denition of a concept is developed as a descr iption of the
capacities necessar y for the control of complex interactions between an organism
and its environment:
An animal capable of concepts is able to identify a particular thing or action and control its future behavior on the basis of that identication in a more or less general way.
It must act as if it could make judgments based on recognition of category membership or integrate particulars into universals. This recognition rests not just on perceptual categor ization (although a concept may have a highly sensory content) but, to
some degree, must also be relational. It can connect one perceptual categor ization to another even in the absence of the stimuli that tr iggered these categor izations.55

Thus to have concepts involves not only the process of categor ization but also recognizing relationships between categor ies.56 What is also important is that having
concepts is a capacity that is not limited to humans, a point also made by Donald
Grifn.57 Concepts are thus not necessar ily tied to language.
54. For a recent and stimulating sally in this debate, see Jerry A. Fodor, Concepts: Where Cognitive Science Went Wrong (Oxford: Clarendon Press, 1998).
55. Gerald M. Edelman, The Remembered Present: A Biological Theory of Consciousness (New York:
Basic Books, 1989), 141.
56. Other works that either make a strong connection between categor ies and concepts or assert
their equivalence include James Hampton and Danile Dubois, Psychological Models of Concepts, in
Categories and Concepts: Theoretical Views and Inductive Analysis, ed. Iven van Mechelen, James Hampton,
Ryszard S. Michalski, and Peter Theuns, Cognitive Science Ser ies (London: Academic Press, 1993),
11 33; Barsalou, Flexibility, Structure, and Linguistic Vagary in Concepts; Barsalou et al., Concepts
and Meaning; Smith and Medin, Categories and Concepts; and Gregory L. Murphy and Douglas L.
Medin, The Role of Theories in Conceptual Coherence, Psychological Review 92 (1985): 289 316.
57. Donald R. Grifn, Animal Minds (Chicago: University of Chicago Press, 1992), chap. 6. A similar approach can be seen in the work of Douglas Hofstadter and his associates, which has resulted in a
number of computer programs that deal with relatively compact bundles of information they call concepts; see Hofstadter and the Fluid Analogies Research Group, Fluid Concepts and Creative Analogies:
Computer Models of the Fundamental Mechanisms of Thought (New York: Basic Books, 1995).

cate g ori zat i on

With this perspective in mind, let me propose that a musical concept has three
character istics. First, it is a product of the process of categor ization. A musical category, then, is quite literally where our conceptualization of music begins. Second,
a musical concept is an essential part of the means through which we guide present
and future actions. These actions thus constitute a sort of indirect evidence for a
cognitive structure almost as ephemeral as music itself. Third, a musical concept can
be related to other concepts, including concepts associated with bodily states (both
physical and emotional), perceptual categor ies (including sound, which, after all, is
not necessar ily music), and linguistic constructs.
From the perspective provided by this denition, language in the sense of natural language is not required in order to have musical concepts. I have, of course,
used language to character ize musical categor ies, as well as the conceptual models
around which musical categor ies are organized. But we could imagine a listener
perhaps a particularly astute listener, but perhaps one no more skilled than Prousts
M. Swann who develops a musical category without recourse to language.While
language is still nearly indispensable for communicating features of musical concepts
and for developing such concepts in r icher contexts, it is not required.58 The notion
of concepts independent of language is more than a little provocative, and with consequences that extend beyond music and the general debate on concept noted
above. Where some authors have excluded music from the conceptual realm
(because, by their denition, concepts require language),59 music can join dance, the
visual arts, and any of the other nonlinguistic modes of human expression as a properly conceptual activity.
This perspective on musical understanding also offers a new way to think about
musical syntax. Music theory has traditionally (if not explicitly) character ized syntax in ter ms of relationships between categor ies of musical objects or events. For
instance, a perfect authentic cadence, as it is usually dened, consists (in part) of a
dominant chord followed by a tonic chord, with bass motion from 5 to 1, and with
the soprano concluding an octave above the bass. The denition involves categor ies
like scale step (5 and 1), harmonic function (dominant and tonic), voice (bass and
soprano), and interval (the octave). Members of these categor ies, when placed in
specic relationships (succession and above), produce a component of musical
syntax (the perfect authentic cadence).60 Musical syntax can also be conceived of
58. The idea that language was not required for musical analysis was one of the ideas behind Hans
Kellers wordless analyses. See Keller, Wordless Functional Analysis no. 1: Mozart, K. 421, The Score and
I.M.A. Magazine 22 (1958): 56 64; idem, The Musical Analysis of Music, in Essays on Music, ed.
Christopher Wintle (Cambr idge: Cambridge University Press, 1994), 126 28; idem, Functional Analysis no. 9a: Mozarts Piano Sonata in A Minor, K. 310, in Essays on Music, ed. Christopher Wintle (Cambridge: Cambridge University Press, 1994), 129 38.
59. Ray Jackendoff, Consciousness and the Computational Mind (Cambridge, Mass.: MIT Press, 1987);
Mark DeBellis, Music and Conceptualization (Cambridge: Cambridge University Press, 1995).
60. It is worth noting that Type 2 categor ies are chiey involved in this denition: within a given
context, something either is or is not an example of 5; something either is or is not a dominant chord;
and so on. Such categor ies are typical of the sorts of taxonomies established by formal music theor y.
However, note that at least one categor y (for voice) is less explicitly dened and functions more like a
Type 1 category. For instance, the soprano is, by convention, the highest sounding voice within a multivoice texture, but in instrumental works it may actually sound well below the range of human sopranos (and thus be a somewhat atypical example of a soprano voice).

61

62

aspe c t s of cog n i t ive st ruc ture

as an emergent property of sequences of musical events. Wagners treatment of the

Leidensmotiv offers one example of this, but the same can be said of most of the
musical works that interest us: it is through the development of local levels of syntax that musical works become distinctive, and it is the multivalence of this syntax
that makes repeated listening to the same work a rewarding exper ience. Of course,
syntactic structure that is shared between musical works is also important, and in
most music there is a dialogue between the two analogous to the dialogue between
the local and global models relative to which musical categor ies are framed.61 Elaborating this approach to musical syntax is the main goal of chapter 4. There we will
see how composers rely on syntactic conventions to provide a framework for the
development of more local patter ns of syntax specic to individual works.
Before we can more fully investigate the role musical concepts play in understanding and the way musical syntax comes about, we need to consider two important issues. First, there is the issue of how relationships between concepts are established. This will be taken up in chapter 2, which explores how conceptual domains
such as music and language are cor related through the process of cross-domain
mapping. This process allows structure from one conceptual domain to be brought
to bear on another and for relationships among concepts to be established. Second,
there is the issue of where conceptual models come from and how they are used to
guide reasoning. Chapter 3 shows how concepts and conceptual relationships are
combined into conceptual models and how conceptual models are connected
together to create theor ies of music.

61. My perspective on musical syntax owes a debt to Eugene Nar mours construal of musical style.
Narmour distinguishes between intraopus style (which is specic to a g iven work) and extraopus style
(which can be seen across a number of works). See Narmour, The Analysis and Cognition of Basic Melodic
Structures; and idem, The Analysis and Cognition of Melodic Complexity: The Implication-Realization Model
(Chicago: University of Chicago Press, 1992).

chapter two

cross-domain mapping

bout a quarter of the way through the Credo of his Pope Marcellus Mass
(printed 1567), Giovanni Pierluigi da Palestrina indulges in a marvelous bit of
text painting with telling effect. The text Palestr ina sets here is Qui propter nos
homines et propter nostram salutem descendit de clis (Who for us men, and for
our salvation, came down from heaven). As shown in example 2.1, with the rst
statement of the word descendit, each voice begins a scalar descent. Christs
descent from heaven is thus represented with a cascading fall through musical space,
a ser ies of overlapping movements down the musical scale.
Palestrinas text painting is a str iking embodiment of the conventional construal
of pitch as high and low. This way of thinking about pitch relations has a venerable tradition. Its basic elements can be seen in Aristoxenuss descr iption of pitch
relations in ter ms of two-dimensional space, which was discussed in the introduction
to this volume. Pitches are understood as points in space, and musical intervals are
reckoned in ter ms of distances between these points. However, the spatial or ientation
crucial to Palestr inas text painting the cor relation of the up and down of
physical space with specic pitch relations, such that a musical scale can descend
was seldom used in Greek music theory. As Andrew Barker has noted, the standard
Greek for what would now be called high-pitched is oxys, which meant sharp,
pointed, or keen-edged; its musical opposite was barys, which meant heavy
(but not, in opposition to oxys, blunt).1 Just when up and down consistently
came to be cor related with musical pitch is unclear, but the linkage was in place at
least by the beginning of the tenth century, a good six hundred years before Palestrina wrote his Mass. From around the tenth century, thus, musicians in the West
began writing about and depicting pitch in ter ms of high and low, mapping
structural relations from the domain of vertically or iented, two-dimensional space
onto the domain of music.
Perhaps more remarkable than the long tradition of construing pitch relations in
terms of up and down are the ready reminders of how arbitrary a construal it is.

1. Andrew Barker, ed., Harmonic and Acoustic Theory (vol. 2 of Greek Musical Writings), Cambridge
Readings in the Literature of Music (Cambr idge: Cambridge University Press, 1989), 69, n. 2.

63

64

aspe c t s of cog n i t ive st ruc ture

example 2.1
mm. 53 58

Giovanni Pierluigi da Palestrina, Credo of the Pope Marcellus Mass,

53

Cantus
tem

de

scn

dit

de c

lis.

Altus
de scn dit de c

lis, de scn

dit

de c

lis.

Tenor I
tem

de

scn dit de

lis.

Tenor II
de

scn

dit de c

lis.

de scn

dit de c

lis.

de scn dit de

lis.

dit de c

lis.

Bassus I

Bassus II
tem

de scn

Consider, for instance, up and down on the piano: how can D4 be above C4
on the piano when they are both on the same hor izontal plane? Think of playing the
two notes on the cello to play the higher D4, we have to move our left hand
down, so that it is closer to the ground. In fact, there are countless reminders of the
articiality of describing pitch in ter ms of high and low. And yet Palestr inas text
painting seems anything but arbitrary there seems to be an aptness to his portrayal
of the descent from heaven that goes beyond mere traditions of depiction.
In this chapter, I explore the function of cross-domain mappings of the sort that
underlie this small bit of text painting in the Credo of the Pope Marcellus Mass.
Cross-domain mapping plays two important roles in musical understanding: rst, it
provides a way to connect musical concepts with concepts from other domains,
including those associated with language; second, it provides a way to ground our
descriptions of elusive musical phenomena in concepts der ived from everyday
experience. Both of these contr ibute to the establishment of relationships between
concepts, relationships that are fundamental to the prospect of theorizing about
music. In the rst section that follows, I provide an introduction to the theor y of
cross-domain mapping as it has been developed in recent work by cognitive linguists and others. This introduction provides a framework for explaining what
makes mappings possible and why some mappings are more effective than others.
Cross-domain mapping also makes it possible to cor relate the musical domain
with others, such as the domain of physical space or of gesture. Under certain circumstances, such cor relations provide the basis for r ich worlds of the imagination.

c ro s s - doma i n map p i ng
Correlating musical pitches with vertically or iented, two-dimensional space, for
instance, leads quite naturally to an imag inary world in which pitches become
things that move through space: the successive notes of a scale g radually descend
and ascend; in other passages, some notes leap, while still others fall. Within this
imaginary world, each traversal of space has a specic and unmistakable sound
that is, descent sounds one way, ascent another. And this is not something limited to
text painting of the sort demonstrated by Palestrina, as any number of cartoon
soundtracks conr m. Never mind that actual traversals of space sound nothing like
those of the hyperkinetic Roadrunner or the hapless Wile E. Coyote; if the cor relation between the domains is properly established, elements from each will blend
together to create novel relationships and elements. In the second section that follows, I descr ibe the process that leads to this sort of blending and show the role conceptual blending plays in text painting and program music.

an introduction to
cross-domain mapping
Cross-Domain Mapping and Metaphor
The theory of cross-domain mapping is a product of a generalized approach to linguistic metaphor rst taken by George Lakoff and Mark Johnson in 1980. Perhaps
the most common conception of metaphor is of a literary device, a manifestation of
the gural use of language to create colorful if imprecise images. Lakoff and Johnson accumulated a substantial body of evidence demonstrating that metaphor was
not simply a manifestation of literary creativity but was, in fact, pervasive in everyday discourse.2 As an example, consider the way up and down are used to characterize emotions, consciousness, and health:
emotions
Im feeling up. My spir its rose. Im feeling down. I fell into a depression. My spirits sank.
consciousness
Get up. Im up already. He rises early in the mor ning. He fell asleep.
health
Hes at the peak of health. Shes in top shape. He came down with the u.
Each character ization suggests not a literal representation of the spatial domain
implied by the or ientation updown but, instead, uses our knowledge of physical
space to structure our understanding of emotions, consciousness, or health.
Based on evidence provided by a large number of similar examples of the
2. George Lakoff and Mark Johnson, Metaphors We Live By (Chicago: University of Chicago Press,
1980). Their work has recently been extended in Lakoff and Johnson, Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought (New York: Basic Books, 1999).

65

66

aspe c t s of cog n i t ive st ruc ture

appearance of metaphor ical constructions in everyday discourse, Lakoff and Johnson proposed that metaphor was a basic structure of understanding through which
we conceptualize one domain (the target domain, which is typically unfamiliar or
abstract) in ter ms of another (the source domain, which is most often familiar and
concrete). Further study has provided a wealth of empirical evidence for this proposal and contr ibuted to the development of the eld of cognitive linguistics.3
Fundamental to the theor y of metaphor that sprang from Lakoff and Johnsons
work is a distinction between conceptual metaphors and linguistic metaphors. A
conceptual metaphor is a cognitive mapping between two different domains; a linguistic metaphor is an expression of such a mapping through language. For instance,
the conceptual metaphor state of being is orientation in vertical space maps
relationships in physical space onto mental and physical states.4 The cross-domain
mapping wrought by this conceptual metaphor then gives rise to innumerable linguistic expressions. Some of these expressions are commonplace, such as Maxwell
seems a bit down today. Others summon a r ich imagistic world, such as that of
John Keatss Ode to a Nightingale:
My heart aches, and a drowsy numbness pains
My sense, as though of hemlock I had drunk,

Or emptied some dull opiate to the drains

One minute past, and Lethe-wards had sunk.

Here the descent to the mythical r iver gives a physical cor relate to the narcotic state
of the nar rator: the act of sinking is mapped onto a melancholy emotional state.
Thus the same conceptual metaphor (state of being is orientation in vertical
space) is behind both linguistic metaphors, one commonplace (Maxwell seems a
bit down today), the other poetic.
With respect to music, the high and low used to descr ibe pitches reect the
conceptual metaphor pitch relationships are relationships in vertical space.
This metaphor maps spatial or ientations such as updown onto the pitch continuum. The mapping yields a system of metaphors replete with possibilities for
describing musical pitch. We can speak in ter ms of pitch contour (meaning successions of pitches, which are located at different places in pitch-space), gesture (mean3. For a review of the empir ical evidence supporting metaphor as a basic cognitive process, see Raymond W. Gibbs, Jr., The Poetics of Mind: Figurative Thought, Language, and Understanding (Cambridge:
Cambridge University Press, 1994). For discussion of the link between the study of metaphor as a cognitive process and the central concer ns of cognitive linguistics, see George Lakoff, The Invariance
Hypothesis: Is Abstract Reason Based on Image-Schemas? Cognitive Linguistics 1 (1990): 39 51.
4. By convention, conceptual metaphors are represented in capital letters. Thus love is a journey
designates a conceptual metaphor of general application, and Love is a journey a specic linguistic
expression based on that metaphor.
The conceptual metaphor state of being is orientation in vertical space is a variant of the states
are locations conceptual metaphor discussed by George Lakoff and Mark Turner in More Than Cool
Reason: A Field Guide to Poetic Metaphor (Chicago: University of Chicago Press, 1989). On cross-domain
mapping as a general phenomenon, see Lakoff and Turner, More Than Cool Reason, 4; George Lakoff,
The Contemporar y Theory of Metaphor, in Metaphor and Thought, 2nd ed., ed. Andrew Ortony
(Cambridge: Cambridge University Press, 1993), 202 51; Gibbs, The Poetics of Mind; and Gilles Fauconnier, Mappings in Thought and Language (Cambridge: Cambridge University Press, 1997).

c ro s s - doma i n map p i ng
ing successions of pitches with a specic directionality and contour), and musical
space (meaning a three- or four-dimensional extension of the basic two-dimensional mapping).5 This system is g iven graphic representation in traditional musical notation: notes that are the result of more rapid vibrations of the sounding
medium are placed higher on the page than notes that result from less rapid vibrations (with the exception of sharps and ats). The two-dimensional space of the
musical page thus cor relates with the spatial or ientation ascr ibed to pitch.6 The
systematic quality that results from mapping spatial or ientations onto the pitch
continuum thus leads to an entire vocabulary for descr ibing relationships among
pitches that provides a r ich set of possibilities for furthering our conceptualization
of music.
As common as conceiving of pitches as high or low seems, not all cultures
describe pitch relationships in purely spatial ter ms. As I noted here, Greek theor ists
of antiquity used oxys (sharp or pointed) and barys (heavy) to character ize
pitches. And traversing histor ical distance is not the only way to discover alter native
conceptualizations of pitch relations. Consider three examples in which it is culture,
rather than time, that creates distance:
1. Steven Felds research has shown that the Kaluli of Papua New Guinea
describe melodic intervals with the same ter ms they use to character ize
features of waterfalls. For instance, in Kaluli sa means waterfall, and a
mogan is a still or lightly swirling waterpool; sa-mogan is the ow of a waterfall into a level waterpool beneath it. Sa-mogan is also used to descr ibe a
melodic line that descends to a repeated note, the contour of which replicates that of a waterfall owing into a pool. In contrast, there are no
specic names for ascending intervals, which nonetheless do occur in Kaluli
song.7 Behind this account of musical intervals is the conceptual metaphor
pitch relationships are waterfall characteristics, which provides the
basis for a r ich set of descriptive terms that capture some aspects of melody
but not others.
2. In Bali and Java pitches are conceived not as high and low but as
small and large.8 Here the conceptual metaphor is pitch relationships are relationships of physical size, a mapping that accurately
5. For further discussion of mappings between the spatial and pitch domains, see Arnie Walter Cox,
The Metaphor ic Logic of Musical Motion and Space (Ph.D. diss., University of Oregon, 1999).
6. It is worth noting that the spatial domain of the musical page is itself metaphor ical, since the way
the page is or iented in physical space that is, whether it is propped on a music stand or lying at on a
table, turned right side up or upside down does not change what we regard as the top and bottom of
the page. Two factors, each independent of the actual or ientation of the page in physical space, help
establish the spatial domain of the page: the conventional or ientation of the symbols on the page (either
rightside up or upside down) and the relative distance of the symbols on the page from the reader.
Rightside up symbols that are farthest from the reader are at the top of the page; rightside up symbols that are nearest the reader are at the bottom of the page.
7. Steven Feld, Flow Like a Waterfall: The Metaphors of Kaluli Musical Theory, Yearbook for Traditional Music 13 (1981): 30 31.
8. Personal communication from Benjamin Br inner, 8 July 1997. See also Wim van Zanten, The
Tone Mater ial of the Kacapi in Tembang Sunda in West Java, Ethnomusicology 30 (1986): 85.

67

68

aspe c t s of cog n i t ive st ruc ture

reects the nor ms of acoustic production: small things typically vibrate

more rapidly than large things. This acoustic fact is represented throughout the numerous parts of the gamelan, the collection of instruments central to the musical practice of Bali and Java.
3. The Suy of the Amazon basin do not have an extensive vocabulary for
describing pitch relationships. When they are descr ibed, however, it is in
terms of age: pitches are conceived not as high and low but as young
and old. The conceptual metaphor that guides this mapping is pitch
relationships are age relationships, which accurately reects the Suys
beliefs that the pitch of the voice becomes deeper with age.9
With each of these conceptions of pitch relationships it is easy to focus on what
they lack rather than what they offer. That is, it is natural (if not quite excusable) to
reckon difference in somewhat chauvinistic terms: that the Kaluli do not have terms to
describe ascending intervals, where we do; that the way pitch relationships are characterized in Bali and Java does not transfer into graphic representations with the same
ease as do high and low. It has to be remembered, however, that mapping high
and low onto music has its own limitations: high and low cannot reect the subtle play of owing water, nor do they provide much of an explanation for how acoustic
features correlate with pitch relationships. Such differences show that each mapping
between domains makes some conceptualizations possible, while it disables others.

Image Schema Theory

The variety of conceptual metaphors used to character ize pitch relations leads to
the question of the ultimate g rounding of the process of cross-domain mapping.
Even if we grant that we understand a target domain (such as pitch relationships)
in ter ms of a source domain (such as or ientation in vertical space), how is it that we
understand the source domain in the rst place? Mark Johnson answered this question by proposing that meaning was grounded in repeated patter ns of bodily experience.10 These patter ns give rise to what Johnson called image schemata, which provide the basis for the concepts and relationships essential to metaphor. An image
schema is a dynamic cognitive construct that functions somewhat like the abstract
structure of an image and thereby connects together a vast range of different experiences that manifest this same recurring structure.11 Image schemata are by no
means exclusively visual the idea of an image is simply a way of captur ing the
organization infer red from patter ns in behavior and concept for mation.
As one example of an image schema, consider the verticality schema, which
might be summar ized by a diagram of the sort given in gure 2.1.We grasp this structure repeatedly in thousands of perceptions and activities that we exper ience every
day. Typical of these are the exper iences of perceiving a tree, our felt sense of standing upr ight, the activity of climbing stairs, forming a mental image of a agpole, and
9. Anthony Seeger, Why Suy Sing: A Musical Anthropology of an Amazonian People (Cambridge:
Cambridge University Press, 1987), 100 02.
10. Mark Johnson, The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason
(Chicago: University of Chicago Press, 1987).
11. Johnson, The Body in the Mind, 2.

c ro s s - doma i n map p i ng

figure 2.1 Diagram of the

verticality schema

watching the level of water rise in the bathtub. The verticality schema is the abstract
structure of the verticality experiences, images, and perceptions. Our concept of
verticality is based on this schema, and this concept is in turn invoked by the various
conceptual metaphors that use vertical space as a source domain through which to
structure such target domains as emotions, consciousness, health, and musical pitch.
By denition, image schemata are preconceptual: they are not concepts, but they
provide the fundamental structure upon which concepts are based. In consequence,
it is important to emphasize that any diagram used to illustrate an image schema is
intended to represent the key structural features and inter nal relations of the
schema; it is not meant to summon a r ich image or mental picture that we somehow have in mind and use actively to structure our thought. More directly, whatever actually occupies our thoughts is not, by denition, an image schema. We can
conceive of image schemata, just as we can conceive of any of a number of nonconceptual or preconceptual cognitive processes. We can also note general patter ns
in the way concepts are structured, which can be attr ibuted to image schemata.
However, there are, by denition, no image-schema concepts.
The relationship between the verticality schema and our character ization of
musical pitch with reference to the spatial or ientation updown is fairly immediate:
when we make low sounds, our chest resonates; when we make high sounds, our
chest no longer resonates in the same way, and the source of the sound seems
located nearer our head. The up and down of musical pitch thus cor relate with
the spatial up and down the vertical or ientation of our bodies. The verticality schema offers a straightforward way to explain why we character ize musical pitch in ter ms of high and low even when the actual spatial or ientation of the
means through which we produce pitches either does not reinforce the characterization or runs directly counter to it.
At present, the image schema remains largely a theoretical construct.Work across
a variety of elds, however, has made strong arguments for the importance of such
a construct, including that by Leonard Talmy in linguistics, Gerald Edelman in neuroscience, David McNeill in psychology, and Raymond W. Gibbs Jr. and Herbert L.
Colston in psycholinguistics.12 Recently, Lawrence Barsalou and his associates have
12. Leonard Talmy, Concept Structuring Systems (vol. 1 of Toward a Cognitive Semantics) (Cambr idge,
Mass.: MIT Press, 2000), 409 70; Gerald M. Edelman, The Remembered Present: A Biological Theory of Consciousness (New York: Basic Books, 1989), chap. 8; David McNeill, Hand and Mind: What Gestures Reveal
about Thought (Chicago: University of Chicago Press, 1992); Raymond W. Gibbs, Jr., and Herbert L. Colston, The Cognitive Psychological Reality of Image Schemas and Their Transformations, Cognitive
Linguistics 6 (1995): 347 78.

69

70

aspe c t s of cog n i t ive st ruc ture

developed a comprehensive theory of mental representation based on perceptual

symbols (which are analogical structures similar to image schemata) and have begun
to provide empir ical work to support the theory.13 Antonio Damasio, working from
a neurophysiological perspective, has made compelling arguments for the importance of the body to consciousness and thought.14 Although research on image
schemata and similar structures is still preliminary, it is also highly promising and
offers some of the best prospects for solving some of the problems of the relationship between mind and body that have dogged cognitive research throughout this
century.

The Invariance Pr inciple

The theory of image schemata provides a way to explain how conceptual metaphors are grounded. It does not, by itself, explain why some conceptual metaphors
seem intuitively better than others. For instance, the conceptual metaphor pitches
are fruit could provide the grounding for such expressions as You must play the
rst note more like an apple, the second more like a banana. Although such mappings are possible, they are certainly not common. Pitches and fruits just do not
seem to be a good match.
To account for why some metaphor ical mappings are more effective than others, George Lakoff and Mark Turner proposed that such mappings are not about the
imposition of the structure of the source domain on the target domain, but are
instead about the establishment of correspondences between the two domains.
These cor respondences are not haphazard, but instead preserve the image-schematic
structure latent in each domain. Lakoff and Turner for malized this perspective with
the Invariance Pr inciple, which Turner states as follows: In metaphor ic mapping,
for those components of the source and target domains deter mined to be involved
in the mapping, preserve the image-schematic structure of the target, and import
as much image-schematic structure from the source as is consistent with that preservation.15 According to the Invariance Pr inciple, mapping the spatial or ientation
13. Lawrence W. Barsalou,Wenchi Yeh, Barbara J. Luka, Karen L. Olseth, Kelly S. Mix, and Ling-Ling
Wu, Concepts and Meaning, in Chicago Linguistics Society 29: Papers from the Parasession on the Correspondence of Conceptual, Semantic, and Grammatical Representations, ed. Kathar ine Beals, Gina Cooke, David
Kathman, Sotaro Kita, Karl-Er ik McCullough, and David Testen (Chicago: University of Chicago,
Chicago Linguistics Society, 1993), 23 61; Lawrence W. Barsalou, Karen Olseth Solomon, and Ling-Ling
Wu,Perceptual Simulation in Conceptual Tasks, in Cultural, Psychological, and Typological Issues in Cognitive Linguistics: Selected Papers of the Bi-Annual ICLA Meeting in Albuquerque, July 1995, ed. Masako K.
Hiraga, Chris Sinha, and Sher man Wilcox (Amsterdam: John Benjamins, 1999), 209 28; Lawrence W.
Barsalou, Perceptual Symbol Systems, Behavioral and Brain Sciences 22 (1999): 577 609.
14. Antonio R. Damasio, The Feeling of What Happens: Body and Emotion in the Making of Consciousness (New York: Harcourt Brace, 1999), chaps. 5 and 6.
15. Mark Turner, Aspects of the Invariance Hypothesis, Cognitive Linguistics 1 (1990): 254. For
additional wr itings on the Invariance Pr inciple (which at rst was called the Invariance Hypothesis), see
Lakoff, The Invariance Hypothesis; Mark Turner, An Image-Schematic Constraint on Metaphor, in
Conceptualizations and Mental Processing in Language, ed. Richard A. Geiger and Brygida Rudzka-Ostyn,
Cognitive Linguistics Research, 3 (Berlin: Mouton de Gruyter, 1993), 291 306; and Mark Turner, The
Literary Mind (New York: Oxford University Press, 1996), chap. 3.
A preliminary discussion of a similar sort of topographical invariance, with applications to music, can

c ro s s - doma i n map p i ng
up down onto pitch works because of correspondences between the image-schematic
structure of components of the spatial and acoustical domains. Both space and the
frequency spectrum are continua that can be divided into discontinuous elements.
In the spatial domain, division of the continuum results in points; in the acoustic
domain, it results in pitches. Mapping updown onto pitch allows us to import the
concrete relationships through which we understand physical space into the domain
of music and thereby provide a coherent account of relationships between musical
pitches. Mapping various fruits onto musical pitches works less well because fruit
does not (in any ordinary way) constitute a continuum. To employ this mapping is
to highlight instead both the discontinuity among musical pitches and how they are
unlike one another (an emphasis on difference suggested by the idiomatic phrase
like apples and oranges).

Cross-Domain Mapping and Conceptual Models

According to cur rent theory, cross-domain mappings are grounded in repeated patterns of embodied exper ience called image schemata. These schemata provide the
basic structure employed in the mappings: the verticality schema is thus fundamental to our understanding of two- or three-dimensional spaces as having directionality and of musical pitch as high and low. Image schemata also constrain
the possibilities for mapping between two domains, a constraint reected in the
Invariance Pr inciple. Because the verticality schema can be applied to both the
spatial and the musical domains, we can use our understanding of the for mer to
structure our understanding of the latter. It remains to be explained why one mapping would be preferred over another why, for instance, we tend to descr ibe pitch
relations in ter ms of high and low rather than small and large.
On the f ace of it, both of these mappings are equally viable. Both draw on
aspects of our embodied exper ience: on the one hand, countless exper iences with
the seeming or igin of our own voices; on the other hand, countless exper iences
with the sounds g iven out by physical objects in the world around us. Both mappings allow us to descr ibe musical pitches as elements within a continuum. And
each mapping can be easily understood from the perspective provided by the
other. For instance, in Camille Saint-Sanss Le Carnaval des animaux, the music for
the elephant is played by the contrabass (large is low), that for the birds by the ute
(small is high).16 Although musicians educated in the West sense the novelty of the
mapping, they can nonetheless understand it perfectly well. In a similar f ashion,
musicians from Bali and Java, when confronted with Western conventions for notation, have few if any problems translating small and large into high and
low.17

be found in Peter Grdenfors, Semantics, Conceptual Spaces, and the Dimensions of Music, in Essays
on the Philosophy of Music, ed.Veikko Rantala, Lewis Rowell, and Eero Tarasti, Acta Philosophica Fennica,
43 (Helsinki: Philosophical Society of Finland, 1988), 9 27.
16. Camille Saint-Sans, Le Carnaval des animaux: Grande fantasie zoologique, ed. Felix Aprahamian
(Zurich: Eulenberg, 1974): Llphant (no. 5), p. 11, Volire (no. 10), pp. 23 27.
17. Personal communication from Sumarsam, 10 April 1998.

71

72

aspe c t s of cog n i t ive st ruc ture

The reason we prefer one mapping over another has to do with the global conceptual models we absorb from culture and that supply crucial support for the preferred mapping.18 In the West, the descr iption of pitch relations in ter ms of up
and down arose around the same time musicians began to develop ways of notating polyphonic compositions. These notational systems often relied, either directly
or indirectly, on the physical placement of symbols on the page.19 The attr ibution
of high and low to musical pitches is thus cor related with a system of notation
that permitted both the visualization and the preservation of musical works. In turn,
this notation relied on a global model that made three important assumptions: rst,
pitches could be regarded as objects that were independent of the sound source that
produced them; second, graphical symbols could be used to represent these pitchobjects; and third, the surface on which these symbols appeared was analogous to
physical space. In Bali and Java, the perfor mance of music was associated with the
rich palette of instruments through which the music was effected. Thus the attr ibution of small and large to pitches cor related with character istics of the musical instruments intr insic to musical perfor mance. The conception of musical pitches
as physical objects relies on a global model that does not, at some fundamental level,
disassociate a pitch from the sound source that produces it.20
As noted earlier, the character ization of pitch relations can be infor med by mappings other than high and low and small and large. In each case, the basic
mapping relies on embodied knowledge and on the cor relation of the musical
domain with a more concrete domain. The specic mapping chosen within a tradition of discourse about music reects not so much absolute musical structure as
it does the broader cultural practice within which music and its understanding are
embedded: mappings reect the conceptual models that are important to culture.
The cross-domain mappings employed by any theory of music are thus more than
simple cur iosities they are actually key to understanding music as a r ich cultural
product that both constructs and is constructed by cultural exper ience.

18. These models and the relationship between global and local models will be discussed in greater
detail in the next chapter.
19. With regard to an indirect reliance on physical placement, it was often a part of the pedagogy
of daseian notation (developed in the ninth century, apparently with an eye toward accommodating the
needs of organum at the fourth) to place high notes higher on the page, even though no such discr imination was required by the notation (since it relied str ictly on a limited repertoire of letter-like symbols to indicate different pitches). See Musica et Scolica enchiriadis una cum aliquibus tratatulis adiunctis, ed.
Hans Schmid, Verffentlichungen der Musikhistor ischen Kommission, Bayerische Akademie der Wissenschaften, 3 (Munich: Verlag der Bayerische Akademie der Wissenschaften, 1981); and the Musica
enchiriadis and Scolica enchiriadis, ed. Claude V. Palisca, trans. Raymond Er ickson, Music Theory Translation Series (New Haven, Conn.: Yale University Press, 1995). These treatises are cur rently thought to
have been wr itten no later than the tenth centur y and possibly as early as the middle of the ninth. For
further discussion of the histor y of this notational convention, see Mar ie-Elisabeth Duchez, La
Reprsentation spatio-verticale du caractre musical grave-aigu et llaboration de la notion de hauteur
de son dans la conscience musicale occidentale, Acta musicologica 51 (1979): 54 73.
20. It should be observed that a disassociation of pitch from sound does not obviate an imageschematic basis for understanding pitch relations. It only bears witness to a reication of pitch after its
initial conceptualization.

c ro s s - doma i n map p i ng
Cross-Domain Mapping, Music, and Music Theory

pale strinas text painting

The theory of cross-domain mapping

outlined in the preceding discussion provides a relatively simple way to account for
the effectiveness of Palestr inas text painting. According to this theor y, we use the
verticality schema, the product of countless bodily exper iences, to give physical
space an up and a down analogous to the up and down of our bodies. This
orientation is then mapped on to a metaphor ical musical space made up of pitches
and the relations (or distances) between them. Pitches that result from more rapid
vibrations of the sounding medium are regarded as higher than pitches that result
from less rapid vibrations of the sounding medium. As each of the voices in Palestrinas six-part texture takes up the word descendit, it begins to descend
through musical space. The notion of descent summoned by the text is thus given
sounding image by a specic ser ies of musical events.
Of course, we could also hear (through the ears of a Javanese musician) the
pitches growing larger, or hear (through the ears of the Suy) the pitches g rowing
older, or hear (through the ears of a Greek theor ist) the pitches g rowing heavier.
Distinct from these ways of hearing, Palestrinas text painting relies on a conceptual
model that character izes pitches as objects in two-dimensional, vertically or iented
space, in which up and down describe specic pitch relations. This model is
strongly associated with the conventions of musical notation as developed in the
West: as pitches get lower in sound, they are also wr itten lower on the musical
staff.
There is an additional reason why Palestrinas text painting is as convincing as it
is, however. The image of descent created by Palestrina relies on more than just the
general notion of descent, which could have been evoked in a variety of ways: by
a downward leap of a single interval; by a short sequence that alter nated descending
thirds with ascending seconds; or through having each voice enter in tur n, beginning with the highest voice and ending with the lowest. The scalar descent chosen
by Palestr ina, however, provides a str iking analogue for the descent of our bodies
through physical space (when this descent is unaided by articial means). Such a
descent involves a lessening of potential energy and a continuous action in one
direction, articulated by the regular transfer of weight from one leg to another.21

21. Most of us nd descent especially scalar descent well represented by the thought of walking
down a staircase, but I think it could be argued that walking down a hillside works as well. The reason
is that it is not so much the neat, two-dimensional image of stairs that is operative but the regular transfer of weight from one leg to another. Indirect evidence is provided by a striking anecdote related by
John Hockenberry, from a time when he was a reporter for National Public Radio. In order to get to a
group of Kurdish refugees on a remote edge of Iraqi Kurdistan dur ing the after math of the 1991 Gulf
War, Hockenberry, a paraplegic since 1976, had to abandon his wheelchair temporar ily and r ide on a
donkey. He comments on the effect of becoming reacquainted with a non-wheeled mode of transportation through the rhythm of the donkeys gait: It was walking, that feeling of groping and climbing and oating on stilts that I had not felt for fteen years. I had long ago grown to love my own wheels
and their special physical grace, and so this clumsy leg walk was not something I missed until the sensation came rushing back through my body from the shoulders of a donkey. Hockenberry, Moving Violations: War Zones, Wheelchairs, and Declarations of Independence (New York: Hyperion, 1995), 2 3.

73

74

aspe c t s of cog n i t ive st ruc ture

The act of singing a descending scale cor relates well with the basic structure of this
event: the relaxation many singers feel as they sing a descending scale matches the
lessening of potential energy; the temporary pauses on each note of the scale match
the regular transfer of weight, which articulates a physical descent. The text painting of descendit is thus supported by our embodied knowledge of descent, as well
as by the conventions of ascribing high and low to musical pitches.
In its exploitation and enr ichment of a mapping between the domains of physical space and music, Palestr inas text painting also gives us a glimpse into a process
of meaning construction. Conventionally, text painting is understood to operate
through a crude sort of mimesis: physical descents are represented by musical descents.
By contrast, I have argued that hear ing a succession of musical events as a descent
is an act that is thoroughly mediated.22 What mimesis there is is highly conditioned
by the choice of cross-domain mappings through which discourse about music is
structured; in turn, these mappings reect the global models of a given cultural perspective and histor ical moment. Palestr inas text painting is not just woven into this
web of meaning construction; it also spins its own threads. Some result from specic
features of physical descents summoned by the passage, some from the point within
the larger musical and dramatic discourse at which this str iking moment occurs, and
some from the sonic attr ibutes that get mapped back onto the notion of physical
descent. The meaning constructed is not, in the nal analysis, simple or direct but
multivalent and contingent, and it reects the r ich set of correspondences activated
by mapping between the two domains.

cross-domain mapping and music theory Because it provides

a way to br ing an integrated system of terms and structural relations to bear on
problems of musical understanding, cross-domain mapping plays an important part
in theor ies of music. Indeed, every theory of musical organization employs crossdomain mappings of one sort or another. Often, the appeal of such mappings is
strong, and the mappings seem intuitively correct (much as high and low seem
intuitively cor rect for the character ization of pitch relations). Further investigation,
however, reveals that these mappings are every bit as mediated as those that are less
systematically developed.
Rudolf Louis and Ludwig Thuilles character ization of tonality in their Harmonielehre provides a case in point. The relevant passage, which appears at the beginning of the rst chapter, runs as follows:
The unity within the diversity of all har mony is ensured through the law of tonality. This asserts that any succession of harmonies can become musically understandable only when each (independently appear ing) chord is perceived to be in a specic
relationship of dependence with respect to the principal chord that underlies the entire
harmonic context. Both melodic and har monic succession require a center, a point of
repose, around which everything twists and tur ns. For melodic relationships, this stationary middle point is the tonic note, the Tone, whose central place in the scheme of
the scale is made manifest by the melodic motions that we direct away from tonic and
22. This same argument, mutatis mutandis, applies to McClarys analysis of Robert Johnsons blues
mentioned above in the preface.

c ro s s - doma i n map p i ng
then back toward it again. The same role of a xed middle point and resting place that,
within melodic unity, falls to the tonic note, is played, with respect to har mony, by the
tonic chord: the consonant triad that has the tonic note as its root.23

The view set forth by this character ization is that musical organization is analogous
to that of the physical world. Music consequently obeys laws that are independent
of human beings tonality is, as it were, a natural force.24 Musical understanding relies on apprehending the har monic dependencies that reect these immutable
laws. The focal point of these dependencies is the center of the system either the
tonic note or the tonic chord which acts as both a center of gravity and an axis of
symmetry.25 To support this perspective, Louis and Thuille appeal to the empir ical
evidence provided by actual melodic and har monic progressions, which they believe show that every properly constructed melody tur ns around a single tonic note,
and that every properly constructed har monic progression tur ns around a single
tonic chord.
Louis and Thuilles character ization of tonality does a marvelous job of capturing the compelling coherence wrought by well-composed tonal music. Given the
evidence of ethnographic and histor ical studies of musical cultures, however, the
idea that Western European tonality is a naturally occur ring force seems doubtful,
as would be the notion of a science aimed at discovering the laws behind this force.
Also implicit in Louis and Thuilles theory is the concept that tonality exists apart
from musical syntax: successions of musical events do not g ive rise to the law of
tonality but serve only to provide evidence of its existence. According to this perspective, successions of musical events that do not provide evidence of tonal relations cannot be understood as music. Of course, such a view places rather profound
restrictions on what counts as music.
Theorizing about music requires that we bring order, even if of a tenuous sort,
to an ephemeral and often intang ible domain. Cross-domain mapping aids this
process by bringing systems of relationships to bear on the musical domain and by

23. Die Einheit in der Mannigfaltigkeit aller Har monie wird gewhrleistet durch das Gesetz der
Tonalitt. Dieses sagt aus, da irgendwelche Folgen von Har monien nur dann musikalisch verstndlich
wirken knnen, wenn jeder (selbstndig auftretende) Accord in einem bestimmten Verhltnis der Abhngigkeit von einem dem ganzen har monischen Zusammenhang zug runde liegenden Hauptaccord
aufgefat wird.Wie die melodische Folge, so bedarf auch die Accordfolge eines Centrums, eines ruhenden Pols, um den sich alles dreht und wendet. Fr die melodischen Beziehungen ist dieser feststehende
Mittlepunct die Tonica, der Ton, dessen centrale Stellung im Schema der Tonleiter dadurch zum Ausdruck gebracht wird, da wir deren melodische Bewegung von ihm ausgehen und zu ihm wieder
zurckkehren lassen. Dieselbe Rolle eines xen Mittle- und Ruhepuncts, die innerhalb der melodischen
Einheit der Tonica zufllt, spielt in harmonischer Hinsicht der Tonica-Accord: der consonierende Dreiklang, der
Tonica zum Grundton hat. Rudolf Louis and Ludwig Thuille, Harmonielehre, 7th ed. (Stuttgart: Carl
Grninger Nachf. Ernst Klett, 1920), 7.
24. I should point out that Louis and Thuille are not concer ned with tonality as it might be broadly
dened, but with tonality as it is represented in Wester n European music of the eighteenth and nineteenth centur ies. This is made all the more clear by their character ization of tonality in har monic ter ms,
something that has limited applicability to a fairly wide range of music cur rently recognized as tonal.
25. In the pages that follow this excer pt, Louis and Thuille explore in greater detail the idea that
tonic is at the center of a symmetr ical scheme of pitches and har monies. See Louis and Thuille, Harmonielehre, 8 10.

75

76

aspe c t s of cog n i t ive st ruc ture

correlating r ich, if at times complicated, images with essential musical concepts.

Such mappings are a way to both structure our understanding and extend it
indeed, much work in music theory is devoted to explor ing the ramications of
these mappings. Nonetheless, each mapping reects the cultural values and imperatives relative to which it is framed, not the verities of absolute musical structure.
Tonality is not, in a simple way, a centr ic and symmetr ical system, any more than it
is a reection of logic (as Hugo Riemann would have it), psychological energetics
(as Ernst Kurth would have it), or gravity (as Arnold Schoenberg would have it).26
Instead, tonality is a way of understanding musical organization. The cross-domain
mappings through which notions such as tonality are articulated are not simply
essential to theor ies of music; in fact, they constitute what counts as music in the
rst place.

Summary
Cross-domain mapping is a general cognitive process through which we structure
an unfamiliar or abstract domain in ter ms of one more familiar or concrete. Crossdomain mapping plays two important roles in musical understanding. First, it provides a way to connect musical concepts with concepts from other domains. As we
saw here, pitch relations within the domain of music have been connected with
concepts associated with vertical space, waterfalls, physical size, and human aging.
Each such mapping made possible systematic accounts of the ways pitches related to
one another. Second, cross-domain mapping allows us to g round our descr iptions
of elusive musical phenomena in concepts der ived from everyday exper ience, since
the structural relations basic to cross-domain mapping have their source in repeated
patterns of bodily exper ience that is, in image schemata.
As we have seen, the mappings we use to structure our discourse about music are
not accidental but reect two constraints. One constraint is the Invariance Pr inciple, which proposes that the best cross-domain mappings are those that preserve as
much of the image-schematic structure of both target and source domains as possible. The other constraint is provided by the global conceptual models relative to
which cross-domain mappings are framed. Taken together, these constraints suggest
that cross-domain mappings not only provide a way to structure our understanding of music but also shape our ideas about what we include under the rubric
music. For an element to count as musical, it must be able to serve as a target for
the cross-domain mappings that guide our discourse about music.
Because cross-domain mapping offers a way to connect what are often elusive
musical concepts with concepts from more concrete domains, and because these
connections give rise to integrated systems of terms and relations, cross-domain
mapping is essential to our theor izing about music. Indeed, as we shall see in the
26. Hugo Riemann, Musikalische Logik: Hauptzge der physiologischen und psychologischen Begrndung
unseres Musiksystems (Leipzig: C. F. Kahnt, 1874); Ernst Kurth, Die Voraussetzungen der theoretischen Harmonik und der tonalen Darstellungssysteme (Bern: Akademische Buchhandlungen von Max Dreschel, 1913);
Arnold Schoenberg, Harmonielehre (Leipzig:Verlagseigentum der Universal-Edition, 1911); trans. as Theory of Harmony, by Roy E. Carter (Berkeley: University of Califor nia Press, 1978).

c ro s s - doma i n map p i ng
next chapter, the concepts der ived from processes of categor ization and the relations
created by cross-domain mapping provide the basic mater ials for music theory.

conceptual blending
Palestrinas text painting relies on our understanding of pitches as high and low
and on our exper iences with physical descents down staircases, slopes, and hillsides
to create a vivid aural representation of the text. However, the imaginary world
summoned by Palestr ina also extends beyond the immediate bounds of text and
music. Using the basic cor relation between text and music as a point of departure,
we can enter an imag inary domain in which the pitches to which descendit is
sung become objects descending through musical space.Within this domain, every
physical descent is accompanied by sounds that result from a smooth transition from
very rapid to less rapid vibrations of the sounding medium. This extension of
Palestr inas imaginarium cannot be predicted simply from linking the domain of
text with the domain of music. It results instead from blending elements and events
from these two domains to create a new one with its own structures and relations,
a domain populated by such things as pitch-objects and the sound of descent.
It must be admitted that an imaginary domain populated with pitch-objects moving through musical space is a somewhat rareed one. Even so, the process of conceptual blending through which it comes about is itself exceedingly common. For
instance, concepts about humans and concepts about animals are often brought
together in childrens stor ies to produce talking animals. Such creatures are powerful
devices in storytelling, for they offer nar rative possibilities beyond those offered by
characters with only human or animal attr ibutes. In a like fashion, the combination of
musical concepts with those from other domains creates possibilities for meaning construction that reach far beyond those of music alone. As an introduction to a methodology for explor ing such conceptual blends in greater detail, let us turn to one of the
talking animals from recent literature and discover what he can tell us about the
process behind the imaginary world summoned by Palestr inas text painting.

Talking Animals and

Conceptual Integration Networks
The Old Grey Donkey, Eeyore, stood by himself in a thistly cor ner of the forest, his
front feet well apart, his head on one side, and thought about things. Sometimes he
thought sadly to himself , Why? and sometimes he thought, Wherefore? and
sometimes he thought, Inasmuch as which? and sometimes he didnt quite know
what he was thinking about. So when Winnie-the-Pooh came stumping along, Eeyore
was very glad to be able to stop thinking for a while, in order to say How do you
do? in a gloomy manner to him.27

Thus A. A. Milne introduces a new character into the stor ies he crafted for his son,
in this case a character built around a stuffed toy donkey Christopher Robin had
27. A. A. Milne, The World of Pooh: The Complete Winnie-the-Pooh and the House at Pooh Corner, illustrated by E. H. Shepard (New York: E. P. Dutton, 1957), 45 46.

77

78

aspe c t s of cog n i t ive st ruc ture

received as a Chr istmas present. Eeyore is, unquestionably, a donkey, fond of thistles
and content to stay outside in all sorts of weather. But Eeyore is also endowed with
human character istics: he is able to talk, to be perversely gloomy, to work his
own convoluted chains of logic, and, on occasion, to be capable of exquisite irony.
In fact, Eeyore is a blend of some of the concepts associated with donkeys and with
humans.
In order to study conceptual blends such as that represented by Eeyore, the
rhetor ician Mark Turner and the linguist Gilles Fauconnier developed the notion
of conceptual integration networks (CINs).28 Each CIN consists of at least four circumscr ibed and transitor y domains called mental spaces. Mental spaces temporar ily
recruit structure from more-gener ic conceptual domains in response to immediate
circumstances and are constantly modied as our thought unfolds.29 For instance,
Milnes sketch of Eeyore sets up two closely related mental spaces. The rst is that
of the Old Grey Donkey solitary, graceless, and phlegmatic. The second is that of
a somewhat morose and plodding intellect, tangled in its own thoughts and happy
enough to leave them behind at the rst opportunity for social interaction. Aspects
of these two spaces are combined in a third space, producing the character of
Eeyore. Turner and Fauconnier use CINs to for malize the relationships between
the mental spaces involved in a conceptual blend, to specify what aspects of the
input spaces are imported into the blend, and to descr ibe the emergent structure
that results from the process of conceptual blending.
The CIN for the conceptual blend used by Milne is diagrammed in gure 2.2.
Its network involves four interconnected mental spaces, which are shown as circles.
Central to the network are two cor related input spaces, the donkey space and the
human space. The solid double-headed ar row linking these two spaces indicates
28. Research on conceptual blending and conceptual integ ration networks is discussed in Gilles
Fauconnier and Mark Turner,Conceptual Projection and Middle Spaces, UCSD Department of Cognitive Science Technical Report 9401 (San Diego: University of Califor nia, San Diego, 1994); idem,
Blending as a Central Process of Grammar, in Conceptual Structure, Discourse, and Language [based on
papers presented at the First Conceptual Structure, Discourse and Language Conference, San Diego,
1995], ed. Adele E. Goldberg (Stanford, Calif.: Center for the Study of Language and Infor mation, 1996),
113 30; idem, Principles of Conceptual Integration, in Discourse and Cognition: Bridging the Gap: [Second] Conceptual Structure, Discourse, and Language Conference, Buffalo, 1996, ed. Jean-Pier re Koenig (Stanford, Calif.: Center for the Study of Language and Infor mation, 1998), 269 83; idem, Conceptual
Integration Networks, Cognitive Science 22 (1998): 133 87; Mark Turner and Gilles Fauconnier, Conceptual Integration and Formal Expression, Metaphor and Symbolic Activity 10 (1995): 183 204; idem,
Conceptual Integration in Counterfactuals, in Conceptual Structure, Discourse, and Language, II, ed. JeanPierre Koenig (Stanford: Center for the Study of Language and Infor mation, 1998), 285 96; Turner, The
Literary Mind; Fauconnier, Mappings in Thought and Language; Seana Coulson, Semantic Leaps: Frame-Shifting and Conceptual Blending in Meaning Construction (Cambridge: Cambridge University Press, 2001),
115 202. The most comprehensive study of conceptual blending as of this writing is Gilles Fauconnier
and Mark Turner, The Way We Think: Conceptual Blending and the Minds Hidden Complexities (New
York: Basic Books, 2002).
29. The theor y of mental spaces is developed in Gilles Fauconnier, Mental Spaces: Aspects of Meaning Construction in Natural Language, 2nd ed., with a foreword by George Lakoff and Eve Sweetser (Cambridge: Cambridge University Press, 1994); Fauconnier, Mappings in Thought and Language; Gilles Fauconnier and Eve Sweetser, eds., Spaces, Worlds, and Grammar (Chicago: University of Chicago Press,
1996); see also Turner, The Literary Mind.

c ro s s - doma i n map p i ng

Generic space
mapping between
beings and beings,
traits and traits

Input space

Input space
Donkey space
donkey
slow-moving
eats thistles, stays
outside in all sorts
of weather
four legs, hooves

Human space
person
gloomy disposition
embraces discomfort
as natural state
can think and talk

Eeyore space
talking donkey
gloomy, plodding
character
clumsy, ill-at-ease

Blended space

gure 2.2 Conceptual integration network (CIN) for the anthropomorphic blend used
for A. A. Milnes Eeyore

that elements within them serve as structural cor relates: person is correlated with
donkey, and gloomy disposition is correlated with slow-moving. Guiding the process of
mapping between these domains is the generic space, which denes the core crossspace mapping and basic topography for the CIN.30 Throughout this network,
beings are mapped onto beings, and character traits onto character traits. Guided by
the conceptual framework provided by the gener ic space, structure from each of the
input spaces is projected into the fourth space, called the blend, which results in the
anthropomorphic character of Eeyore. The mapping is only partial, however,
reecting the limitations imposed by the gener ic space. Since the gener ic space
does not map between physical character istics, we do not expect Eeyore to be of
human appearance or to be human-sized.31
The dashed ar rows linking the gener ic space to the input spaces, and the input
spaces to the blended space, indicate the directions in which structure is projected:
30. The topography of a domain consists of the elements that populate the domain and the relations
that hold between them. By denition, all of the domains in a CIN must have a unifor m topography,
which is represented in its abstract for m by the gener ic space. The requirement of uniform topography
within the cor relational structure of a CIN can be seen as another manifestation of the Invariance Pr inciple discussed earlier in this chapter.
31. Inasmuch as a number of the characters in Milnes stor ies or iginated as stuffed toys, they are easy
to imagine as rather diminutive in size; in fact, this is how they are rendered in E. H. Shepards illustrations. In his text, Milne is relatively silent on this matter, although various contextual clues suggest that
the animals with which he deals are of the stuffed variety.

79

80

aspe c t s of cog n i t ive st ruc ture

from the gener ic space to the input spaces, and from the input spaces to the blended
space. The arrows are double-headed because, under certain circumstances, structure may also be projected from the blended space back into the input spaces, and
from the input spaces back into the gener ic space. The idea of anthropomorphic
animals that emerges in the blended space may thus inuence the way we think
about regular animals, leading us to talk to them and act toward them as though
they had human character istics. Similarly, our exper iences with actual animals and
people give life to the abstractions of the gener ic space. The double-headed ar rows
also serve as a reminder of the limitation of all of the diagrams of CINs I shall use:
mental spaces are dynamic structures, as are the CINs that are built from them. Thus
gure 2.2 represents a sort of analytical snapshot of this particular network, framed
with the intent of capturing its essential features, but making no claim to exhausting the possibilities for descr iption. Hints about how the CIN and its spaces may
develop can be gleaned from the diagram, but a full account would require a ser ies
of such snapshots.
Although the diag ram given in gure 2.2 is standard in the literature on mental spaces, it can lead to two misunderstandings. First, since the blend is at the bottom of the diagram, it gives the impression that concepts precipitate down into
the blend. Second, the function of the gener ic space can be a bit confusing, since
it does not seem to be directly involved in the blend. In the interests of clarifying these points I offer gure 2.3, which represents the essential components of
a four-space CIN in a slightly different for mat. Here the gener ic space is properly
represented as both the backg round and the foundation for the entire network.
The two input spaces are concrete representations of the abstract str ucture represented by the gener ic space, and the conceptual blend is a fur ther projection
from these.
An important aspect of the topography of CINs is the basic logic established by
the gener ic space. For the CIN of gure 2.2, the assumption is that beings visibly
manifest their character traits. Old donkeys are slow-moving and balky; morose
humans are given to gloomy pronouncements on the state of the world; and Eeyore
talks and acts like the clumsy, gloomy character he is. The topography of the network also guides three operations composition, completion, and elaboration
that produce emergent structure unique to the blend.
Composition puts together elements from the input spaces to create new entities
in the blended space, yielding the character of Eeyore: although donkeys cannot
actually think and talk, and humans do not have four legs or hooves, Eeyore has all
these traits. Completion extends the image suggested by the initial mapping from the
input spaces, drawing on our backg round knowledge of the circumstances summoned by the CIN. For instance, we know that gloomy characters expect the worst
of situations. Eeyore can thus be relied on for solemn pronouncements of impending disaster on even the sunniest of days, to greet each calamity as conr mation of
his estimation of the world, and to eat his thistles with little sign of relish. Elaboration is a more extensive operation than completion is; it develops the structure of the
blended space by building on the pr inciples and logic evinced by the blend. In
effect, the input spaces decrease in importance and the focus is directed toward the
rich imaginary possibilities of the blended space. At this point, we can start to wr ite

c ro s s - doma i n map p i ng
Blended
space

Input
spaces

Generic
space

figure 2.3

Alternative representation of a four-space CIN

our own stor ies about the Old Grey Donkey.Were we to place Eeyore aloft in an
airplane, we could be sure he would profess no enjoyment of the view but would
instead focus on the various aeronautical catastrophes that, from his perspective,
were almost certainly imminent. Given Eeyores clumsiness and the awkwardness of
his hooves, we would be much less likely to imagine him ying the plane, leaving
this to the more dexterous animals or to Chr istopher Robin.
Although the conceptual blend that yields anthropomorphic animals is important to Milnes story, it is not the only blend in evidence. Because Eeyore originated
as a stuffed animal, he retains some of the character istics of the species. Accordingly,
in the story in which he is introduced, he loses his tail only to have it recovered by
Winnie-the-Pooh and nailed back on by Christopher Robin, a sequence of events
unlikely were he based solely on real donkeys. This points to yet another conceptual blend, in which one of the input spaces is structured around stuffed, rather than
real, animals (and one that I shall not pursue in the present discussion).
A. A. Milnes Eeyore and most of the other characters from the Winnie-the-Pooh
stories provide clear examples of conceptual blending. As indicated, however, the
process of blending is not restricted to childrens stor ies. On the one hand, conceptual blending is also common in everyday discourse. Witness statements such as
The car is being stubbor n today I just cant get her to start, which blends concepts related to the physical properties of inanimate objects and those related to the
behavior of humans. On the other hand, conceptual blending also has its place in
literature. Consider Prousts descr iption of one feature of the spr ingtime walks along
the Msglise way in Combray:
We would leave town by the road which ran along the white fence of M. Swanns
park. Before reaching it we would be met on our way by the scent of his lilac-trees,
come out to welcome strangers. From amid the fresh little green hearts of their foliage
they raised inquisitively over the fence of the park their plumes of white or mauve

81

82

aspe c t s of cog n i t ive st ruc ture

blossoms, which glowed, even in the shade, with the sunlight in which they had
bathed.32

Here Proust blends features of the lilac trees with actions of which only animals or
humans are capable, thus creating a memorable image that perfectly captures the
intense sensual engagement prompted by spring.
Conceptual blending is a pervasive and often transparent cognitive process. A
given situation or story may involve any number of conceptual blends, as the character of Eeyore constructed from the attr ibutes of both stuffed and real animals
shows. Each such blend can be descr ibed through a conceptual integ ration network, which per mits a systematic descr iption of specic features of the blend. Finally,
the CIN associated with the character of Eeyore can be used to explain why the
Old Grey Donkey has some, but not all, of the features of humans, and why we can
easily imagine him taking part in some activities but not in others.

Conceptual Blends and Text Painting

In the conceptual blends just discussed, the input spaces for each CIN were summoned by language. As Fauconnier has noted, however, mental spaces are very general and are constructed for many cognitive purposes; language is but one way
to prompt the constr uction of a mental space.33 Under cer tain circumstances,
music can also prompt space construction. In the case of Palestrinas text painting,
the mental space is relatively circumscr ibed and focuses on an orderly progression
of pitches that lead resolutely toward a cadence. This space is in cor respondence
with that set up by the text, which focuses on Chr ists movement from the heavenly
to the mundane and the lessening of potential energy associated with physical
descents.
Correlations between the musical and textual spaces involved in Palestr inas text
painting set up the CIN shown in gure 2.4. The gener ic space for the CIN is structured around the notion of elements that are in directed relations conceived with
respect to a teleological framework.Within the network, motions through physical
and musical space are directed motions. As with the conceptual blend associated with
the character of Eeyore, the emergent structure of the blended space can be descr ibed
in terms of the operations of composition, completion, and elaboration.
Composition puts musical pitches together with the act of descent to yield
pitch-objects that descend. Composition also associates descent with a specic
sound, which actual descents may or may not produce. Completing this image, we
might infer that the lower the pitch, the closer it is to the ground and to the mundane.We could also infer that, since descent can be given a sound, ascent can also be
given a sound. Palestr ina in fact conr ms the latter inference in mm. 92 93 of the
32. Marcel Proust, Swanns Way (vol. 1 of Remembrance of Things Past), trans. C. K. Scott Moncr ieff
and Terence Kilmartin (New York: Vintage Books, 1981), 147 48. Further blends in Proust are discussed
in Turner, The Literary Mind, 91 92, 120 23; see also my concluding chapter.
33. Personal communication from Gilles Fauconnier, 8 September 1998. This perspective is also represented, somewhat less explicitly, in Fauconnier and Turner, Conceptual Integration Networks; and
Fauconnier, Mappings in Thought and Language.

c ro s s - doma i n map p i ng

figure 2.4

CIN for Palestrinas text painting

Credo, when, as shown in example 2.2, he sets ascendit with steadily r ising
pitches. Elaborating the blend, we might imagine pitch-objects doing all sorts of
things, not just ascending and descending. Something like this is behind Louis and
Thuilles character ization of tonality as a system of forces operating on pitches distributed throughout the eld of musical space.
The extension of the conceptual blend suggested by mm. 92 93 of the Credo
example 2.2
mm. 92 94

Giovanni Pierluigi da Palestrina, Credo of the Pope Marcellus Mass,

92

Et

scn

dit

in

lum

Et

scn

dit

in

lum

Et

scn

dit

in

lum,

Et

scn

dit

in

lum

se

83

84

aspe c t s of cog n i t ive st ruc ture

notwithstanding, Palestr inas text painting is a relatively isolated instance of meaning construction involving music and language. Indeed, text painting is usually
regarded as a clearly circumscr ibed compositional technique, most often restricted
to a single word or image. Under certain circumstances, however, music and language can combine to create rather more extended possibilities for meaning construction, as is shown by an instance of text painting in Giaches de Werts seven-part
madrigal Tirsi morir volea, rst printed in 1581.
The text for Werts madrigal is from a poem by Giovanni Battista Guar ini that
is probably the most popular madr igal text of the late sixteenth century. It recounts
a sexual encounter between a shepherd and a nymph.34 The passage relevant to discussion here appears in example 2.3; the text for the entire poem, with the verses
that appear in example 2.3 underlined, is as follows:
tirsi morir volea

thyrsis wished to die

Tirsi mor ir volea,

Gli occhi mirando di colei ch adora,
Quandella, che di lui non men ardea,
Li disse: Ahim, ben mio,
Deh non mor ir ancora
Che teco bramo di mor ir anch io.

Thyrsis wished to die,

gazing at the eyes of his beloved,
when she, whose ardor equaled his,
said: Alas, my love,
do not die yet,
since I long to die with thee.

Fren Tirsi il desio,

Chebbe di pur sua vita allor nire,
E sentea morte, e non potea mor ire;
E mentre il guardo suo sso tenea
Ne begli occhi divini,
E l nettar amoroso indi bevea,
La bella Ninfa sua, che gi vicini

Thyrsis curbed the desire,

which by then had almost ended his life:
he felt death near, yet could not die;
and while he kept his gaze
xed upon those eyes divine,
and drank from thence the nectar of love,
his pretty Nymph, who felt

Sentea i messi dAmore,

Disse con occhi languid e tremanti:
Mori, cor mio, chio moro.
Cui r ispose il Pastore:
Ed io, mia vita, moro.
Cos mor iro i fortunati amanti,
Di morte s soave e s gradita,
Che per anco mor ir tornaro in vita.

Loves heralds near,

said with languishing and trembling looks:
Die my heart, for I die.
At which the Shepherd replied:
And I, my life, die.
Thus the happy lovers died
a death so sweet and pleasant,
that in order to die again, they returned to life.

The text painting beg ins in mm. 31 32 when the word tremanti is set to a
written-out or nament (a tr ill or gruppo) in all parts. Tremanti (trembling) sets
up a mental space focused on an intense and partially involuntary physical reaction
to stress that produces repeated oscillating motions. The written-out or nament sets
up a mental space focused on the rapid alter nation of a minor chord with Eb3 in
the lowest voice, and a major chord with FS4 in the highest (which creates an
34. The context for Guar inis poem and Werts setting of it have been discussed by Laura Macy in
Speaking of Sex: Metaphor and Performance in the Italian Madrigal, Journal of Musicology 14 (1996): 1 34.

example 2.3

Giaches de Wert, Tirsi morir volea, mm. 28 39

28
C

ci

ni

Sen tea i mes

si da mo

re,

Dis

se

con oc

chi

ci

ni

Sen tea i mes

si da mo

re,

Dis

se

con oc

chi

ci

ni

Sen tea i mes

si da mo

re,

Dis

se

con oc

chi

ci

ni

Sen tea i mes

si da mo

re,

Dis

se

con oc

chi

31

Mo

ri, cor mio,

chio mo ro,

Mo

ri, cor mio,

chio mo ro,

Mo

ri, cor mio,

chio mo ro,

lan

guid e

tre

man

ti:

Cui

lan

guid e

tre

man

ti:

lan

guid e

tre

man

ti:

Cui

lan

guid e

tre

man

ti:

Cui

Cui

(continued)

example 2.3

(continued)

34

Mo

ri, cor mio, chio mo ro,

Mo

ri, cor mio, chio mo ro,

Mo

ri, cor mio, chio mo ro,

ris pos il Pa sto

re:

Ed io, mia vi ta, mo ro,

Ed io, mia

ris pos il Pa sto

re:

Ed io, mia vi ta, mo ro,

Ed io, mia

ris pos il Pa sto

re:

Ed io, mia vi ta, mo ro,

Ed io, mia

ris pos il Pa sto

re:

Ed io, mia vi
ta, mo ro,

Ed io, mia

37

Mo

ri, cor mio, chio

Mo

ri, cor mio,

Mo

ri, cor mio,

mo

ro.

Co s mo rir

no i

Co s mo rir

no i

chio mo ro.

Co s mo rir

no i

chio mo ro.

vi ta mo ro,

Ed io,

mia vi ta, mo ro.

Co s mo rir

no i

vi ta mo ro,

Ed io,

mia vi ta, mo ro.

Co s mo rir

no i

vi ta mo ro,

Ed io, mia

ta, mo ro.

Co s mo rir

no i

vi
ta mo ro,

Ed io,

mia vi
ta, mo ro.

Co s mo rir

no i

vi

c ro s s - doma i n map p i ng

Generic space
changes in
disposition
represent motion

Music space
written out trill in
all parts (cross
relation, alternation
of major and minor
triads)

Text space
tremanti (trembling)
intense and partially
involuntary physical
reaction to stress

Blended space
pitches summon
motion
sound-of-trembling

figure 2.5

CIN for mm. 31 32 of Werts Tirsi morir volea

augmented-second cross-relation).35 Similar ities between these spaces suggest correlations between their elements: each chord in the musical space cor relates with an
end-point of the physical motion summoned by the textual space, and the intensity of the physical reaction cor relates with the rapid alter nation of major and minor
chords and the cross-relation between the outer voices.
It is the cor relations between these spaces that set up the CIN shown in gure
2.5. The gener ic space for the CIN is structured around the idea that changes of
disposition represent motion.36 Physical trembling of the sort summoned by tremanti is one instantiation of this notion, and the rapid alter nation of musical materials is another. Once again, the emergent structure of the blend can be descr ibed in
terms of the operations of composition, completion, and elaboration. Composition
combines musical pitches with specic features of trembling, so that the pitches
summon a sense of physical motion even though the sound-source producing them
remains relatively xed: the singers of Werts madrigal need not actually tremble
when they sing tremanti. Composition also suggests that trembling has a sound,
35. At the time Werts madrigal appeared, terms such as minor chord and major chord were not
conventionally used to descr ibe simultaneous sonor ities. Writers on music dur ing the per iod preferred
to descr ibe sonor ities in ter ms of their intervallic makeup. In the present context, this more accurate terminology is just a bit unwieldy, and I have consequently used somewhat more familiar and concise modern ter ms. I trust their anachronistic application here will not prove unduly confusing.
36. This is analogous to the phi phenomenon noted by Gestalt psychologists. See Kurt Koffka, Principles of Gestalt Psychology (London: Routledge and Kegan Paul, 1935), 179.

87

87

88

aspe c t s of cog n i t ive st ruc ture

even though most actual trembling is done in relative silence. If we complete the
pattern suggested by the basic mapping, we can infer that once the physical action
stops the sound will stop. Finally, through elaboration we can imagine that successions of restricted groups of pitch mater ials could depict intense physical actions
more complex than trembling, perhaps even extending to the interaction of physical bodies.
Although the elaboration of a blended space is often left to our imagination, at
times it may be explicitly developed. This is indeed what Wert does in the measures
immediately following the setting of tremanti. Here are some of the salient features of the music of mm.33 38:
Starting with the pickup to m.33, the womens and mens voices engage
in a rapid alter nation of entrances.
The beginnings and endings of the entrances are elided, so that the

singing is seamless.

The spacing between entrances is gradually compressed, culminating in

the joining of mens and womens voices in mm.37 38; in fact, this is the
rst time in the madr igal that the two vocal groups sing together.
The setting is almost entirely homophonic within each group of voices.
The har monic mater ial is highly restricted, consisting almost entirely of
triadic sonor ities built on G, D, and F.
In m. 38, the music breaks off, and there is a half measure of silence.
The alter nation of restricted groups of pitch mater ial, the charged interaction of
the vocal groups, the sudden breaking off of the voices in m.38, and the context
provided by the blend produced by the text painting of tremanti all summon an
image of intense physical activity followed by a sudden suspension of that activity. It
is an image that gives sounding representation to the sort of death sexual orgasm
about which the poem obsesses. This image also offers a way to ground the double entendre exploited by Guarini, built on the paradoxical connection between life
at its most intense and death. Both sexual climax and death are marked by sharp,
nearly immediate contrasts: orgasm followed by quiescence; the clamor of life followed by the silence of death. Werts music, building from the common g round
shared by sexual climax and death, breathes life into a play of words that, by this
point in the sixteenth century, had become a rather common commonplace.37
It is important to emphasize that not all combinations of text and music will
yield rewarding conceptual blends. For a blend to occur, there not only needs to be
some cor relation between two domains, but the domains must also have a unifor m
topography. In some cases certain strophic songs, for instance the cor relation
between text and music is simply too general to generate a compelling blend. In
37. My inter pretation of connections between sexual climax and death is somewhat different from
that offered by Laura Macy. Noting that early physiological accounts saw both sex and death as an emission of spirit, Macy writes that this link is surely at the bottom of the metaphor ic use of the word death
to mean sexual climax (Speaking of Sex, 5).While I would not want to deny the importance of this
connection, my argument is that commonalities between the larger physiological processes of both
orgasm and death could also serve to ground the metaphor and that Werts music provides a musical analogue for these.

c ro s s - doma i n map p i ng
other cases formulaic songs composed for opera or reecting the more commercial side of popular music the cor relation between text and music may be so tenuous as to be virtually nonexistent. Also important are the conceptual models that
inform the inter pretation of the cor relation between the input spaces for a blend. A
careful listener who is nonetheless unaware of the context of Palestr inas music, for
example, will almost certainly realize that something is going on in the passage cited
in example 2.1 but may interpret the music as sad or losing energy. If conceptual models associated with the act of descent, or with this particular portion of the
Mass, are not activated, the listener may get no farther than this. Similarly,Werts climactic moment, for a listener who does not understand the text (either for lack of
knowledge of Italian or out of innocence of the double entendre), may simply be
heard as exciting. In both cases, the conceptual blends will remain latent: potential
but not actualized opportunities for the construction of meaning.
As Werts madrigal suggests, conceptual blends involving text and music may be
relatively extended. Chapter 6 will consider such blends in greater depth, showing
how they develop over the course of entire songs, how the same text set to new
music can give rise to two very different songs, and how music alone can elaborate
a blend rst set up by text and music.

Conceptual Blending and Program Music

In theory, conceptual blends may involve mental spaces associated with any domain
of thought. Musical domains could thus be cor related not only with domains associated with language but also with physical gesture or color.38 In the following, I
would like to consider the conceptual blends that occur when instrumental music
is associated with an extra-musical program. Although such programs are almost
always invoked through language, the connection between the musical and linguistic domains is much looser, and the concepts that emerge from the blend are somewhat more variable. The generality of such blends also allows them to cover greater
expanses of music in the example with which I am concer ned here, an entire
symphony.
Ludwig van Beethovens Symphony No. 6 in F major, Op. 68 (Pastorale) is
among the most well-known of instrumental works with which extra-musical programs have been associated. In the case of the Sixth Symphony, this association was
aided by the descr iptive titles Beethoven gave individual movements (such as
38. The correlations I refer to here are between the conceptual domains associated with music and
with colors or gestures; this ts with my notion of concepts as not necessar ily linguistic. However, I am
not thinking of synaesthesia, where stimuli applied to one of the ve senses produces responses in one or
more of the other senses as well; I view this phenomenon as functionally automatic and thus not involving concepts. Synaesthesia between music and color is common enough to have warranted relatively
extensive research, and no end of commentar y. See, for instance, Lawrence E. Marks, The Unity of the
Senses: Interrelations among the Modalities (New York: Academic Press, 1978); Wilson Lyle, Colour and
Music: An Introduction, Music Review 43 (1982): 261 64; Michael Hurte, Musik, Bild, Bewegung: Theorie und Praxis auditiv-visueller Konvergenzen (Bonn:Verlag fr systematische Musikwissenschaft, 1982); and
Jonathan W. Bernard, Messiaens Synaesthesia: The Cor respondence between Color and Sound Structure in His Music, Music Perception 4 (1986): 41 68.

89

90

aspe c t s of cog n i t ive st ruc ture

Scene by the Brook). As Richard Will has pointed out, the programmatic aspects
of the Sixth Symphony created difculties for those who sought to give an account
of the work, for the simplistic representation of natural sounds seemed in conict
with the received wisdom that Beethovens music referred to nothing outside
itself.39
Symptomatic of this uneasiness with Beethovens program was Donald Francis
Toveys outr ight rejection of any link between such natural sounds and the music of
the Sixth Symphony. Tovey writes, In the whole symphony there is not a note of
which the musical value would be altered if cuckoos and nightingales, and country
folk, and thunder and lightning, and the howling and whistling of the wind, were
things that had never been named by any man, either in connection with music or
with anything else.40 With the cor relation between the domain of nature and the
domain of music sundered, Tovey paves the way to approach the work as a perfect
classical symphony and to claim for it a place in the pantheon of Beethovens works.
Nevertheless, Toveys analysis of the symphony seems to speak of something
other than absolute music. Describing an important moment in the slow movement, he writes,
The deep shadow of this remote key of G at becomes still deeper as C at, which,
changing enhar monically to B natural, swoops round to our or iginal key B at. At the
outset of this wonderful passage the theme was that of the rst subject with the murmur of the brook becoming articulately melodious in the clar inet and the bassoon.
At the moment when the melody gathers itself up into a sustained phrase and makes
its enhar monic modulation, there comes a phenomenon full of deep meaning. From
this point nothing is left of the melody but sustained notes and bird-song tr ills; the
whole of the rest of the return to the main key is harmonic and rhythmic. In this as
everywhere else the movement remains true to type, a perfect expression of the happiness in relaxation.41

Here we have key areas that swoop, clarinets and bassoons that are articulate, and
melodies capable of independent motion, all within a movement that expresses the
happiness there is to be found in relaxation.
At rst glance, Toveys prose seems to retain the very correlation between the
domains of music and nature that he had earlier rejected, and to develop a conceptual blend based on this cor relation. Still present are the murmur of the brook, the
bird-song tr ills of melody, and the general atmosphere of the pastoral, all wedded
to Beethovens music. Closer consideration, however, shows that the cor relation of
music and nature cannot explain the structural attr ibutes of the imaginative domain
summoned by Tovey. Although nature includes living beings, music does not; at
best, music compr ises sounds that occur within a temporal framework. The independent agency implied by Tovey, evident in the articulate and goal-directed entities he imagines, also seems somewhat out of step with the commonplace view of
39. Richard Will, Time, Morality, and Humanity in Beethovens Pastoral Symphony, Journal of the
American Musicological Society 50 (1997): 271 329.
40. Donald Francis Tovey, Symphonies (vol. 1 of Essays in Musical Analysis) (London: Oxford University Press, 1935), 45.
41. Tovey, Essays in Musical Analysis, 1: 51. The reference is to mm. 78 85 of the slow movement.

c ro s s - doma i n map p i ng
nature as a domain in which events are guided not so much by individual volition
as by larger forces working from without. In short, it is one thing to link elements
of nature and musical sounds to create a domain in which the cuckoo sings with the
timbre of the clar inet and the sound of a brook is summoned by undulating patterns in the str ings. It is rather another to animate musical sounds and to endow
them with their own volition.
As it tur ns out, Toveys inter pretation does not rely on a simple cor relation
between music and nature. Instead, Tovey activates a much larger network of mental spaces to produce his inter pretation. Central to this network is one of the most
common of conceptual blends, that of anthropomorphism (which played an important role in the blend that produced A. A. Milnes Eeyore). The basic cor relation that
underlies this blend can be seen in statements like The car is being stubbor n today
I just cant get her to start. In this mapping, the domain of inanimate objects is
structured by the human domain; more specically, human being is mapped onto car,
and a volitional state proper to humans (stubbornness) is mapped onto the mechanical state of the car (not-starting).42 A blend that exploits this mapping might then
attribute additional volitional or emotive states to the vehicle (I think the car wants
to stay home, Shes mad because I havent changed her oil), or extend into the
nonvehicular domain (I dont think the toaster likes me, it keeps burning the
bread). In its most abstract for m, this blend involves two input spaces, as shown in
the CIN of gure 2.6. One recruits structure from the human domain, the other
from some non-human domain. The cor relation of these two spaces gives r ise to
the blended space of anthropomorphism, whose emergent structure reects the
counter part correlations of the input spaces. Within this space, non-human entities
become endowed with the character istics and intellects of humans while retaining
many of the features that make them distinct from humans. An anthropomorphic
car may be stubbor n and sullen, but it will also be able to stay outside year round
(even though it may not like to do so).
The gener ic space for the CIN connects entities with entities and states with
states. According to the logic proper to the network, the events that occur within
each source domain have causes that follow from the rules that govern the behavior
of entities that populate the domain. Stubbornness in the human domain reects personal idiosyncrasies and may be abetted by exter nal circumstances that render the
current situation unsatisfactory. According to conventional wisdom, people are stubborn either because they were born that way or because they are unhappy with the
way things are and will not take any action (other than being stubbor n) until things
change. Not-starting in the domain of automobile mechanics is caused by electromechanical conditions that are insufcient to initiate or produce sustained combustion. The car will not start because there is something wrong with the engine. It
is important to note that, although the gener ic space may have a structure that is
strongly image-schematic, generic spaces are often r ich with detail. Generic spaces
42. Note that to ascr ibe a state associated with a living entity to a nonliving entity is not the same
thing as to endow the latter with life. The statement The car wont start I think it died does not
entail a belief in animistic vehicles, but only cor relates a particular state associated with living beings
(being dead, or being no-longer-alive) with the mechanical state of the car (never-to-start-again).

91

92

aspe c t s of cog n i t ive st ruc ture

Generic space
mapping between
entities and entities,
states and states

Non-human domain
Non-human
entities and events
related to them

Human domain
Human entities
and events related
to them

Anthropomorphic
domain
Non-human entities
endowed with human
characteristics

figure 2.6

CIN for anthropomorphic blend

are not necessar ily primitive or schematic they only dene the basic topography
common to all of the spaces in the blend.
Tovey activates the mental space of anthropomorphism early in his essay when
he writes, The Pastoral Symphony has the enor mous strength of some one who
knows how to relax.43 Within the blend that results from cor relating the human
and musical domains, musical events become actors endowed with their own characteristics and capable of their own acts of volition. Thus the transition that follows
the rst subject of the opening movement of the symphony leads in three indolent
strides to a second subject which slowly stretches itself out over tonic and dominant
as a sort of three-part round.44 Tovey develops an anthropomor phic perspective on
nature more gradually, bringing it to fruition only in his account of the end of the
slow movement. Suddenly for a moment all is silent; we have no ears even for
the untir ing brook, and through the silence comes the voice of the nightingale, the
quaint rhythmic pipe of the quail, and the syllabic yet impersonal signal of the
cuckoo.45 Here elements of nature (brooks and birds) become actors, each endowed
with individual character istics and capable of individual acts of volition. In both
cases, the cor relation of human and non-human domains gives rise to imaginative
conceptual blends: symphonies that know how to relax, musical passages that str ide,
brooks that do not tire, and cuckoos capable of impersonal signals.
43. Tovey, Essays in Musical Analysis, 1: 46.
44. Tovey, Essays in Musical Analysis, 1: 47.
45. Tovey, Essays in Musical Analysis, 1: 52.

c ro s s - doma i n map p i ng

Generic space
anthropomorphic
space: non-human
entities with human
characteristics

Anthropmorphized
music
musical entities
and events with
human characteristics

Anthropomorphized
nature
entities and events
from nature with
human characteristics

Toveys interpretive
blend
storms with
the voices of timpani
and double basses;
storm climax/trombones;
rainbow/oboes

figure 2.7

CIN for Toveys analysis of the Beethoven Pastoral Symphony

Construing both music and nature in anthropomorphic ter ms makes it possible

for Tovey to connect the two in a further blend, shown in gure 2.7. The gener ic
space for this network is the blended space of gure 2.6. One of the input spaces is
that of anthropomorphized nature, which includes natural agents who act within
and respond to their native environment (nature). The other input space is that of
anthropomorphized music, which includes musical agents who act within and
respond to their native environment (music). In the blend, aspects of the input
spaces are combined to create unique entities and relations. Here the murmur of a
brook becomes an articulate melody for clar inets and bassoons; here Beethovens
music can be conceived as natural without having been reduced to nature.
The three operations that produce emergent structure in the blended space
composition, completion, and elaboration are all evident in Toveys account of
the end of the fourth movement: The stor m moves in grand steps to its climax.
This is marked by the entry of the trombones. . . . Then the stor m dies away, until
with the last distant mutter ings of the thunder the oboes give a long slow fragment
of bright sustained melody on the dominant of F. This has been aptly compared
with a rainbow.46 Composition unites the lives and interests of natural agents and
musical agents. In the nal twenty measures of the fourth movement, the thunder
mutters with the voices of the timpani and double basses, who also prepare the
tonality of the closing fth movement precipitated by the retreat of the stor m.
46. Tovey, Essays in Musical Analysis, 1: 55.

93

94

aspe c t s of cog n i t ive st ruc ture

Completion supplies additional structure. Beethovens terse title for the fourth
movement is Gewitter, Sturm (Thunderstorm, Tempest), which says nothing of
the after math of a stor m. Nonetheless, we know that rainbows often follow stor ms,
and so the bright sustained melody of the oboes can summon the image of a rainbow, a completion suggested but not necessitated by the input spaces.47 Elaboration
develops the blend by pursuing its logic. The stor m, as a magnicent entity, moves
with grand steps. The climax of its ter ror and wrath (enacted through musical
agents) is marked by the entry of new musical agents (the trombones). As these passions abate, musical agents g radually disappear (a change only implicit in Toveys
prose) until only a bare few remain. Perhaps most str iking about this elaboration is
that a single natural agent (the stor m) appears to be brought fully to life only through
a multiplicity of musical agents. Such a genesis, while inexplicable in the natural
domain, is fully consonant with the idiosyncrasies of the musical domain. There,
individual musical entities (such as movements or entire symphonies) are often
understood to emerge from the actions of the subentities (such as themes or tonal
areas) that they compr ise.
Tovey thus mobilizes a number of mental spaces to provide an account of
Beethovens symphony. The most active spaces are those that contr ibute to a blend
of anthropomorphized natural and musical elements. This blend gives Tovey a way
to descr ibe music that seems naturalized but does not reduce to a simple cor relation
of natural and musical sounds. Less active, but still important, are spaces involved in
blends fundamental to this pr imary network. These include spaces built up from the
domain of human entities and events, along with the domains of nature and music.
Toveys descr iption makes only spar ing use of these subsidiar y blends, since they
cannot provide the naturalized account of music that is the goal of his essay. Far in
the background is the gener ic space common to anthropomorphic blends, which
denes the core cross-space mapping and logic that underlie the entire network.

Summary
Conceptual blending is a dynamic process of meaning construction that involves
small, interconnected conceptual packets called mental spaces, which temporar ily
recruit structure from conceptual domains in response to local conditions. When
blending occurs, a portion of the structure from two cor related input spaces is projected into a third, blended space. As part of this process, the operations of composition, completion, and elaboration produce structure within the blend that is not
found within either of the input spaces. This structure only becomes possible
through the concepts and relations produced by conceptual blending.
The str ucture common to the mental spaces within a conceptual integ ration
network is reected in the gener ic space. This space denes the core cross-space
mapping and is organized according to a basic logic that remains consistent
throughout the network. Generic spaces are not necessar ily pr imitive or image47. This suggestion is also followed by the animators of Walt Disneys Fantasia, who not only supply a rainbow to accompany this moment in Beethovens symphony but integrate it within the multileveled, quasi-mythical story they set to the music.

c ro s s - doma i n map p i ng
schematic they only dene the basic topography common to all of the spaces in
the blend.
In a basic blend, four spaces will be active (although not necessar ily to the same
degree): the gener ic space, the input spaces, and the blend. It often will be the case,
however, that additional spaces will be activated, to a greater or lesser extent, as the
construction of meaning proceeds. In certain passages within Toveys account of
Beethovens Sixth Symphony, as many as six spaces may be active. In addition to the
blend, these include mental spaces built up from the domains of music and nature,
from the domains of anthropomor phized music and nature, and from the gener ic
domain of anthropomorphism. This multiplicity of spaces explains the complexity
of Toveys account, as well as its inter pretive richness. It also serves to explain the
seeming contradiction entailed by Toveys desire to deny the programmatic aspect
of Beethovens symphony while responding to the powerful impressions to which
it gives rise.
More generally, evidence from conceptual blending lends credence to the idea
that music is an independent conceptual domain. As I have shown in my analyses of
text painting and program music, musical concepts combine with concepts from
other domains to create blended concepts that suggest str iking possibilities for the
imagination. More to the point, musical syntax (from the perspective developed in
chapter 1) can be seen contr ibuting to the blended concepts associated with Werts
text painting and Beethovens program music. In the for mer,Werts deployment of
musical mater ials sets the stage for an enactment both by the voices of the singers
and by our imaginations of the sort of death with which Guar inis poem is concerned. As but one example from the latter, Beethovens conclusion of the fourth
movement and preparation for the fth (the distant mutter ings of the timpani
notwithstanding) provide a moment of contrast and articulation that clearly supports our imagining one process ending (the stor m) and another beg inning (the
shepherds song of joyful thanksgiving). And musical spaces can infor m our understanding of gener ic spaces: to hear Werts musical climax is to get new insight into
climax as a general phenomenon, be that climax sexual, cerebral, or even visual; to
hear Beethovens transition is to get a glimpse of how natural events like stor ms can
be thought of not simply as personages but also as processes.
Each of these aspects of conceptualizing music will be taken up in more detail in
chapter 6, within the context of the analysis of the nineteenth-century Lied. There
it will be possible to see, to a greater extent, how processes of cross-domain mapping and conceptual blending construct and contr ibute to our understanding of
both music and the world as a whole.

95

chapter three

conceptual models

and theories

s part of her efforts to understand the everyday knowledge with which people
make sense of music, the music theor ist Jeanne Bamberger has often asked
young children to develop ways to represent familiar melodies or simple rhythmic
patter ns. One way Bamberger does this is to have a child ar range a specially constructed set of bells, called Montessor i bells, so that he or she can play a given tune.
Bamberger then asks the child to invent instructions so someone else could play the
tune on the bells as theyve been ar ranged.
For those who have even a passing acquaintance with music, an encounter with
Montessor i bells can be a truly disor ienting exper ience. Developed as part of the
mater ials for the Montessor i curriculum, the bells are designed to encourage childrens exploration of sound and music.With this in mind, they are carefully constructed to be identical in appearance: the lowest-sounding bell looks just like the
highest-sounding one, as do all those in between. Only by striking each bell with
a small mallet can the different pitches of the various bells be discovered. Each bell
is attached to a wooden stem mounted on a wooden base so that the bells are physically independent of one another. There is, in consequence, no set ar rangement of
the bells, just as there are no visual cues as to their pitch: how the bells might be
arranged, whether in casual play or in the perfor mance of some tune, has to be
determined by each person who uses them. As a further challenge (or an additional
resource), a complete set of Montessor i bells contains duplicates: there are eight bells
with white bases, which make up a diatonic scale from C4 to C5, and thirteen bells
with brown bases, which make up a chromatic scale from C4 to CS5. Thus for each
white-based bell there is one brown-based bell with a matching pitch but again,
just which bells match must be discovered by the user.
In the course of her research, Bamberger discovered that children use one of two
general strategies when building a tune with the bells. Children following the rst
strategy ar range the bells in the order in which the pitches occur in the tune. To
build the rst phrase of Happy Birthday, for instance, the bells might be placed
one after another to create a bell-path of the sort shown in gure 3.1, with the
arrow indicating the order in which the bells are set down. (Note that two G-bells
are used, one for each time the pitch G occurs anew in the tune. That is, a separate

96

conc e p tual mode l s and th e ori e s

figure 3.1 Bell-path no. 1 for the rst phrase of Happy Birthday (open bases, white
bells; solid bases, brown bells; arrows show direction of play)

bell is not provided to accommodate repeated notes in immediate succession. In this

illustration, the second G-bell is taken from the set of brown-based bells.) Playing
this much of the tune is a relatively simple matter of retracing the building process
and str iking each bell in tur n (with the rst bell getting two hits). The spatial
arrangement of the bells, then, is in direct cor respondence with the succession of
pitches in the rst phrase of the tune.
Children following the second strategy start by putting the bells in order, typically from low to high. Once this is done, they gure out the sequence in which to
strike the bells in order to play the tune. Figure 3.2 illustrates the results of this strategy for the rst phrase of Happy Birthday. The bells are arranged from low to
high (left to r ight); numbers indicate the order in which the bells should be struck,
and arrows indicate the direction of motion. (Note that only one G-bell is required
by this strategy; however, the path through the bells is considerably more complex.)
As gure 3.2 shows, the placement of the bells no longer cor responds to the succession of pitches that make up the tune. The second strategy does not rely on a
spatial analogue for the succession of pitches in the tune, as does the rst. Compensation for this loss in immediacy is provided by a gain in exibility: to play a different tune, the child needs simply to discover what new succession of hits will produce the cor rect results the bells themselves do not need to be reordered.
Bamberger calls the representation produced by the rst sort of strategy gural
and that produced by the second formal.1 Figural representations highlight the
unique gures (or groupings of musical mater ials) that make the tune what it is.
1. Jeanne Bamberger, The Mind behind the Musical Ear: How Children Develop Musical Intelligence
(Cambridge, Mass.: Harvard University Press, 1991), 24 26; idem, Cognitive Issues in the Development
of Musically Gifted Children, in Conceptions of Giftedness, ed. Robert J. Sternberg and Janet E. Davidson (Cambr idge: Cambridge University Press, 1986), 398 99. In personal cor respondence (20 May
2000), Bamberger notes that for mal strategies, as far as she has seen them among both children and
adults, only occur with people who have had some training in music or who play an instrument.
Bamberger also observes that representation (the ter m I favor here) tends to imply more fully
worked-out cognitive processes than a ter m like description. Although I respect this distinction (especially where it applies to the activities of young children who may not really be concer ned with representing anything), I shall retain representation in my discussion, but with the understanding that it
applies to a quite general set of cognitive structures. Representation, dened thus, embraces imageschematic constructs, perceptual symbols, concepts, and actions whatever is necessar y to make realworld phenomena available (to re-present them) to the higher-level cognitive processes associated with
theories. For further discussion of this approach to representation, see Mark Turner, Design for a Theory of Meaning, in The Nature and Ontogenesis of Meaning, ed.Willis F. Overton and David S. Palermo
(Hillsdale, N.J.: Erlbaum Associates, 1994), 91 108; and Antonio R. Damasio, The Feeling of What Happens: Body and Emotion in the Making of Consciousness (New York: Harcourt Brace, 1999), 320.

97

98

aspe c t s of cog n i t ive st ruc ture

G
1
3
figure 3.2

A
2

B
5

C
4

Bell-path no. 2 for the rst phrase of Happy Birthday

Formal representations, in contrast, tend to obscure the unique character istics of the
tune but situate musical elements within a more general context applicable to any
number of tunes. Bambergers research suggests that gural representations are
among the earliest for med by children. Formal representations develop somewhat
later as the childs sense of how to organize and represent mater ials matures.
Bambergers primary interest in the distinction between gural and for mal representations was with the developmental changes that led to the transfor mation of
the one into the other. Nonetheless, the contrast between these two strategies also
points to different ways to construe relationships among musical concepts. That is,
gural strategies offer one way to character ize these relationships, and for mal strategies another. Perhaps more important, both remain options even after the passing of
childhood, as is shown by the transfor mational networks employed by David Lewin
in his analyses of late-nineteenth- and early-twentieth-century music.While Lewins
networks certainly have their abstract side they owe much to theor ies of mapping
that are employed in for mal algebra they also offer a way to represent unique relationships among musical events: they are gural as well as for mal.2
These different ways to think about the essential relationships that obtain within
what is putatively the same sonic domain (here, the rst phrase of Happy Birthday)
demonstrate the role of two basic cognitive structures conceptual models and theories in our understanding of music. Conceptual models, which are the more basic
structures, are made up of relatively limited cor relations of concepts and are specic
to a given domain. (We caught glimpses of such models in the discussion of categorization in chapter 1 and saw the role they played in cross-domain mapping in chapter 2.) Theories, in contrast, are extended structures that coordinate a number of conceptual models to provide a more comprehensive account of the world.3 Figural
2. See David Lewin, Generalized Musical Intervals and Transformations (New Haven, Conn.:Yale University Press, 1987); and idem, Musical Form and Transformation: Four Analytic Essays (New Haven, Conn.:
Yale University Press, 1993). Lewin makes explicit reference to Bambergers work in his analysis of Karlheinz Stockhausens Klavierstck III; see Lewin, Musical Form and Transformation, 45 47. I should note that
Bambergers work is by no means limited to early childhood development, and that she, too, takes the
position that gural and for mal strategies are not restricted to developmental stages in childhood. For
instance, in dialogues built into The Mind behind the Musical Ear, she attributes gural strateg ies to one
college-age interlocutor and for mal strategies to another.
3. As is evident from this initial character ization, theory, as I dene it, embraces cognitive structures far more modest than those with which moder n scientic theor y is occupied. The connection
between such cognitive structures and scientic theory is explored in greater detail in the next section.

conc e p tual mode l s and th e ori e s

strategies, on the one hand, rely on one set of conceptual models and lead to one
theory of musical organization. Formal strategies, on the other hand, rely on another
set of conceptual models and lead to a somewhat different theory of musical organization.
Both conceptual models and theor ies are central to our understanding of the
world as a whole; however, both often remain implicit. It is only when they are
teased out in the course of work such as Bambergers, or as the result of deliberate
reection on our thought processes, that either conceptual models or theor ies come
to the fore. In this chapter, I explore the structure and function of conceptual models and theor ies and show the part they play in both infor mal and for mal accounts
of musical organization. In the rst section I examine the role conceptual models
and theor ies play in the musical representations of one of Bambergers subjects, an
eight-year-old boy named Jeff. Jeff s encounters with music engage him in a struggle to come to ter ms with basic musical concepts, and they reveal music theory
that is, a theor y specic to music in the making. In the second section, I show
how the cognitive structures demonstrated by Jeff s nascent theor ies of music connect with similar structures proposed by researchers in a wide variety of elds, from
articial intelligence and cognitive science to anthropology and ethnomusicology.
This will serve to generalize the construal of conceptual models and theor ies beyond
Jeff s specic situation and lay out some of their additional properties. In the third
section, I turn to two of the more inuential theor ies of music within the last three
centur ies of the Western tradition: the music theory of Jean-Philippe Rameau and
that of Heinrich Schenker. My objective is to show how they employ the same
basic cognitive structures that can be seen in childrens gural and for mal strategies
for tune building.

developing a theory of
musical organization
Learning to Play Twinkle on the Bells
Jeanne Bamberger worked with Jeff for about six months in the Childrens Learning Lab at MIT, starting in the fall of 1975.4 She set him the task of building the
tune Hot Cross Buns, rst with the sonic resources provided by a simple computer program and later with the Montessor i bells. After Jeff achieved some facility with these tasks, Bamberger had Jeff concentrate on the tune Twinkle, Twinkle,
Little Star, using only the Montessor i bells. In consequence, at the point we take up
Jeff s story he has already lear ned some things about tunes and the properties of the
bells, although Twinkle, with its comparatively more complex structure, presents
new challenges. In what follows, the focus is on the rst phrase of Twinkle
(Twinkle, twinkle, little star, / How I wonder what you are; the musical notation
is given in ex. 3.1) and on two of the bell-paths Jeff constructed in order to play it.

4. Jeff is a pseudonym adopted by Bamberger to preserve the boys anonymity. An extensive

account of Bambergers work with Jeff is given in chaps. 6 12 of The Mind behind the Musical Ear.

99

100

aspe c t s of cog n i t ive st ruc ture

example 3.1

The rst phrase of Twinkle, Twinkle, Little Star

Twin kle, twin kle,

lit

tle

star,

How

won der

what you are.

In preparation for Jeff s work with Twinkle, Bamberger had selected eleven
bells from the complete Montessor i set and put them out on a table.5 Included were
all of the white-based bells (providing a complete C major scale) and three brownbased bells: the C-, E-, and G-bells. Once Jeff had familiar ized himself with the
bells, Bamberger proposed that he tr y building Twinkle. After beginning with
the C-bell, Jeff searched among the available bells until he found one that matched
the next pitch of the tune. He placed this bell next to the C-bell, and then played
both to re-create the beginning of the tune (that is, the pitches that t the words
Twinkle, twinkle). To nd the bell for little, he once again searched among the
available bells, and when he found the r ight one he placed it next in line. He again
played what he had assembled of the tune (Twinkle, twinkle, little) before returning to the remaining bells to nd the one cor responding to the next pitch of the
melody.6
Jeff s preference was to use the white-based bells rst. He turned to the brownbased bells only when there was no white-based bell left that would provide the
pitch for which he was looking; this rst happens with the recurrence of the note
G on the word star. As a way of giving a visual representation of the division
between the two subphrases (or gures) of Twinkle, Jeff left a space between
the rst four bells and the second four bells (marking the rst four bells as one
group and the second four bells as another). This produced the ar rangement shown
in gure 3.3. The result is a gural representation of the tune, with each successive
bell cor responding to a successive pitch of Twinkle. Performing the tune is simply a matter of beginning at one end of the bell-path and proceeding to the other
(but with some bells getting two hits, and others only one).
After Jeff had completed his bell-path, Bamberger asked him to come up with
instructions so someone else could play Twinkle on the bells as he had set them
out. Jeff s initial solution was quite straightforward: he simply drew eight short vertical lines, with each line cor responding to a bell. The cor responding notation
looked something like this:


His intent was that the player should start at the beginning and play straight ahead
to the end (mimicking, incidentally, his process of setting up the bell-path), some-

5. Bambergers account of Jeff s rst work with Twinkle can be found in The Mind behind the
Musical Ear, 183 92.
6. The strategy of playing through as much of the tune as he had built before searching for the next
pitch is one Jeff used throughout the tune-building exercises I discuss here.

conc e p tual mode l s and th e ori e s

C
figure 3.3

Jeff s initial bell-path for the rst phrase of Twinkle, Twinkle

thing made manifest when, at Bambergers prompting, he put in numbers to show

the sequence of actions:

1


2


3


4


5


6


7


8

With a bit of additional prompting, Jeff concluded his notation by adding another
set of numbers to show how many times each bell was to be struck:
2

1

2

2

2

3

1

4

2

5

2

6

2

7

1

8

With this, Jeff s rst session with Twinkle was complete: he had constructed
a bell-path with which he could play the rst phrase of the tune, and he had come
up with a notation that would allow another child to do the same (once the meaning of the numbers and lines had been explained).7

Conceptual Models and Theor ies in Jeff s First

Version of Twinkle
Jeff s actions, as he constructs a bell-path for Twinkle and then invents a notation
for both the path and the actions to be perfor med on it, are based on a number of
suppositions:
The pitch-events of the tune are embodied by the bells that reproduce
them.8 In this way, each pitch-event (which is invisible and ephemeral) is
given visible manifestation by a bell.
Each of the pitch-events of the tune is unique. It follows that there will
be a different bell for each pitch-event; Jeff does not, at this point, recognize the pitch equivalence of the two G-bells or of the two C-bells.
The succession of temporal events that make up the tune can be represented through the spatial disposition of the bells. Since this succession is
7. It is worth noting that Jeff s notation is tied to his bell-path: were the additional space between
the fourth and fth bells eliminated, his notation would almost assuredly lead to a distortion of the temporal spacing of the pitches of Twinkle.
8. By pitch-event I mean a pitched sound framed relative to some temporal context (so that it is
an event). As will be seen, because building the tune involves ar ranging the bells in space, the temporal aspect of pitch-events is almost always conated with spatial location.

101

102

aspe c t s of cog n i t ive st ruc ture

conceived of as linear, the bell-path is likewise linear; similarly, the pause

between the two subphrases is represented by a greater amount of space
between the fourth and fth bells.
Playing the tune is a matter of performing actions along a path specially
constructed for this tune. These actions are unidirectional the player
starts at one end of the bell-path and proceeds straight through to the
other and mimic the process of building the path. Playing the tune
constitutes one large, continuous motion, articulated by the smaller
motions of striking each individual bell.
These suppositions, the relationships between them, and the way they help guide
Jeff s actions as he sets up his bell-path can be character ized in ter ms of four sor ts
of cognitive constructs: concepts, conceptual models, conceptual domains, and
theories.
Concepts. Musical concepts, as noted in chapter 1, have the following attr ibutes: they are a product of processes of categorization, they are part of the
means through which we guide present and future action, and they are not
necessar ily linguistic. The concepts basic to Jeff s perspective include (but
are not limited to) those associated with pitched sounds, with the individual
Montessor i bells, with the pitch-events that make up the rst phrase of
Twinkle, and with the notion of performing the tune (as distinct from,
say, listening to the tune).
Conceptual models. Conceptual models consist of concepts in specied relationships. Three conceptual models play a prominent role in Jeff s gural
strategy. The rst model cor relates pitch-events and objects. Accordingly,
the constituent pitch-events of Twinkle are construed as objects that
endure even after the acoustic signal has faded away. This perspective is
reinforced by the physical and visual evidence provided by the bells. In
turn, each bell can stand for a pitch-event, even when it has not been
struck.9 The second model translates temporal relationships into spatial relationships. Events separated by equal amounts of time are inter preted as
objects separated by equal amounts of space; events separated by unequal
amounts of time are inter preted as objects separated by unequal amounts
of space. This model infor ms Jeff s arrangement of the bells on the table, so
that there is a greater space between the two bells that mark the end and
the beginning of the two consecutive subphrases (the fourth and fth bells)
than between the bells within each subphrase. The third model links physi-

9. While I believe that the conceptual models I descr ibe here accurately reect the cognitive structures that guide Jeff s actions, they should be generalized only with great caution. Bamberger notes that
some children believe that a bell will sound differently if it is moved to another place. Indeed, there is
good evidence that Jeff is more concer ned with spatial location than with pitch. For discussion of a particularly revealing moment in his development, see Bamberger, The Mind behind the Musical Ear, 204 09.
To simplify matters I shall continue to focus on pitch-events (since these ultimately infor m the meaning given to any particular location along the bell-path or action-path), with the understanding that the
full descr iption of an event requires specication of both its temporal and spatial coordinates.

conc e p tual mode l s and th e ori e s

cal actions with a sequence of sounds. As a result, Twinkle acquires a corporeal dimension to go with its acoustic dimension. Each action produces
another of the pitch-events of Twinkle, and, just as the pitches of Twinkle form a coherent whole, so does the sequence of physical actions.10
Conceptual domains. Each of these conceptual models sets up a conceptual
domain. The rst model sets up a domain of sound-objects, populated by
physical embodiments of ephemeral sonic events. The second model sets up
the domain of pitch-space, quite apparent in the ar rangement of bells on
the table. As Jeff has constructed it, pitch-space consists of temporal and
structural relationships between sound-objects. The third model sets up a
domain associated with a practiced sequence of actions, what Bamberger
calls a felt path.11 It should be noted that these conceptual domains are
not completely exclusive of one another: in general, conceptual domains
will overlap to the extent that the conceptual models on which they are
based share concepts and relationships.
Theories. Theories coordinate a number of conceptual models in order to
guide inference and reason. In this case, the three conceptual models
described above are coordinated by a theory Jeff s theory of Twinkle.
According to this theory, a representation of the tune can be constructed by
arranging a ser ies of unique sound-producing objects in space in order of
their appearance in the tune, producing a unidirectional path that, if traversed with the appropriate actions, will yield a sounding version of the
tune. Jeff s notation for Twinkle is a symbolic manifestation of this theory: each sound-producing object is represented by a separate line.12 The
uniqueness of these objects is indicated by the ser ies of numbers that is
placed beneath these lines; the actions that will yield the tune are indicated
by the numbers that are placed above.
There will be more to say about each of these constructs, but rst let us see how
Jeff s theory of Twinkle changes as Bamberger alters one aspect of the task of
constructing a representation of the tune.

Jeff s Second Version of Twinkle

Jeff continued to work with Twinkle for a number of weeks and became quite
adept at building the tune according to his rst theory. One day, as an outgrowth
of a pitch-matching task that she had asked Jeff to do, Bamberger changed the set of
10. In some cases, these physical actions are constitutive of the tune in a way that the ar rangement
of the bells is not. For instance, the bell-path gives no clues as to how many times each bell should be hit,
which, of course, is an integral part of the action-path.
11. Bamberger, The Mind behind the Musical Ear, 213.
12. My use of symbol with respect to Jeff s notation is not very heavily freighted the symbols
are at best ad hoc, and under different circumstances Jeff might use quite different symbols to accomplish
the same notational goals. However, I do want to say that Jeff s notation cor relates with his concepts
that is, the marks he puts down stand as symbols for some of the concepts in ter ms of which he understands Twinkle.

103

104

aspe c t s of cog n i t ive st ruc ture

C
figure 3.4

Jeff s modied bell-path for the rst phrase of Twinkle, Twinkle

bells Jeff had to work with by removing the brown-based bells.13 This left Jeff with
only one bell for each pitch-type. Jeff built the tune as usual until he came to the
place where, in accordance with the strategy based on his rst theory, he would nd
and position the brown-based G-bell. When he could not nd the bell, he stopped
his work and said, I need another bell.
Bamberger did not press Jeff further at this point, but a few days later she
returned to the task. This time, Jeff decided to look for the needed bell among those
he had already arranged in the bell-path. After rst trying the C-bell, he moved to
the G-bell (second in the bell-path) and discovered that it provided the pitch he was
looking for. Jeff eventually was able to build up the entire tune using only the
white-based bells, creating the bell-path and action-path shown in gure 3.4.While
this ar rangement of the bells is more parsimonious only six bells are needed
instead of eight the action-path through the bells is more complex, requiring
both switchbacks and leaps over intervening bells. As a result, the bell-path and
action-path are no longer in simple cor respondence with one another.
As before, Bamberger asked Jeff to come up with a notation that would enable
someone else to play Twinkle on the bells as he had ar ranged them. This time,
however, the complexities of the action-path required to play the tune on the new
arrangement of bells posed some challenges, and it was only after a number of
attempts that Jeff was able to develop the notation shown in gure 3.5. The notation shows two important innovations. First, by means of arrows, Jeff is able to show
just how one negotiates the switchbacks and leaps necessary to play Twinkle. Second, the numbers now refer to only one thing: the number of hits each bell gets.
The ser ial notation Jeff used for his rst version of Twinkle has been abandoned,
since there is no longer a one-to-one mapping of bells to pitches.

Conceptual Models and Theor ies in Jeff s

Second Version of Twinkle
While some of the suppositions behind Jeff s construction and notation of the second version of Twinkle are the same as those for his rst version, some are different. Unchanged are the notions that the pitch-events of the tune can be embod-

13. Bamberger did this in part to see if she could move Jeff from a gural representation of the tune
toward a for mal representation. For a fuller account of this aspect of Bambergers work with Jeff, see The
Mind behind the Musical Ear, chap. 10.

conc e p tual mode l s and th e ori e s

figure 3.5
Twinkle

Jeff s notation for his modied bell-path for the rst phrase of Twinkle,

ied by the bells that reproduce them and that playing the tune is a matter of performing actions along a path in a specially constructed pitch-space. Changed are
Jeff s ideas about pitches and about the relationship between temporal events and
spatial disposition. In a conceptual shift that ultimately had profound implications
for Jeff s understanding of musical organization (as Bambergers further work with
Jeff showed), he now recognizes that pitch-events are not entirely unique: some
pitches will be repeated over the course of a tune. A given bell can thus serve as a
sound-source for more than one pitch-event. This realization has the consequence
of loosening the relationship between when pitches happen in the tune and how
representatives of these pitches can be disposed in space. As a result, playing Twinkle is no longer a matter of following a unidirectional path through pitch-space.
Even though time might still be conceived of as moving unifor mly in one direction, the motions needed to play Twinkle are no longer unidirectional and ser ial:
the player moves forward and backward, sometimes str iking the bells ser ially, sometimes leaping over one or more bells to nd the next pitch of the tune.
The conceptual shift associated with Jeff s changed ideas about pitches, pitchspace, and temporal order is an example of what Douglas Hofstadter calls conceptual
slippage. As Hofstadter denes it, conceptual slippage is the context-induced dislodging of one concept by a closely related one, within the mental representation of
some situation.14 Here, the context is provided by Bambergers removal of the
brown-based bells from those Jeff had to work with in constructing Twinkle.
What has slipped is Jeff s concept of a pitch-event specically, how pitch-events
are unique. Jeff still clearly realizes that the G that comes after the rst C of Twinkle is different from the G that comes after the rst A after all, they occur at different points in the tune. He now knows, however, that both can be played on the
same bell. This in tur n means that the temporal uniqueness of pitch-events will not
necessar ily translate into spatial uniqueness: the two distinctly different Gs of the
tune can be found in the same place on the bell-path. Jeff s concept of pitch-event
as a specic pitch in Twinkle has been dislodged by a concept of pitch-event as a
general type of sonic event that is manifested at a particular point in the temporal progress of
Twinkle.
Because all of the conceptual models through which Jeff structures his under14. Douglas R. Hofstadter and the Fluid Analogies Research Group, Fluid Concepts and Creative
Analogies: Computer Models of the Fundamental Mechanisms of Thought (New York: Basic Books, 1995), 198.
The idea of conceptual slippage has proven quite important in Hofstadters work in that it introduced a
level of exibility and uidity into the computer programs his group developed, features that were absent
from earlier articial intelligence programs.

105

106

aspe c t s of cog n i t ive st ruc ture

standing of Twinkle incorporate the concept of pitch, the slippage of this concept
cascades through the system. The cor relation between sounds and objects provided
by the rst conceptual model is now looser than before, and the domain of soundobjects it sets up is populated with gener ic, rather than specic, entities. In consequence, the functional distinction Jeff made between brown-based bells and whitebased bells (preferring to work with the latter rst) is no longer necessar y now
there are only bells. The second model, which maps temporal relationships onto
spatial relationships, is more profoundly affected. While temporal and spatial events
are still connected ( Jeff s bell-path still has a beg inning and an end), the independence of temporal succession and the spatial ar rangement of sound-sources is far
more evident. Reecting this, the domain of pitch-space has been reshaped: once
pitches are conceived as gener ic entities, they start to have more to do with each
other than they do with the specic sounds that make up Twinkle. These alterations in the notion of spatial disposition then inuence the third conceptual
model. While the physical actions that produce Twinkle can still be seen as a
coherent whole, it is a whole made of a larger number of disparate parts (including
switchbacks and leaps). This creates a physical manifestation of the growing separation of the tune from the medium through which it sounds: one can now get lost
along the action-path in a way that would have been unlikely with the rst bellpath Jeff constructed.
As a result of these changes, Jeff s theory of Twinkle is changed as well.
According to this modied theor y, Twinkle can be built using various gener ic
sound-producing objects in space in the approximate order of their occur rence in
the tune. A relatively complex action-path is then imposed on this ar rangement,
which, if properly executed, will yield a sounding version of the tune. Again, Jeff s
notation for Twinkle is a symbolic manifestation of this theory: there are no
longer unique entities that cor respond to each of the constituent pitches of the
tune, and perfor ming the tune involves a ser ies of complex actions. In ter ms of the
two basic strategies Bamberger observed, Jeff has begun to move from a gural representation to a for mal one.

The Function and Structure of Theories

At this point, it will be useful to pause and consider ve features of theories revealed
by Jeff s two approaches to constructing Twinkle on the bells.
First, theories serve as a guide for actions, both mental and physical. Each of Jeff s
theories of Twinkle allows him to create an ar rangement of the bells that will
enable him to play the tune. Using these theor ies, Jeff would be able to nd a way
to arrange similar objects glasses lled with varying amounts of water, marimba
bars, pots and pans so that he could play Twinkle on them. Should something
change in one of these ar rangements (for instance, if one of the bells was covertly
switched for a bell with a different pitch), Jeff would most likely be able to nd the
problem and suggest ways of remedying it. In order to avoid confusion later, I want
to broaden the character ization of this feature and say that theor ies serve as a guide
to understanding and reasoning. Although this seems a bit g rand when applied to
Jeff s situation after all, Jeff is someone who is much more predisposed to action

conc e p tual mode l s and th e ori e s

than to talking about action it makes clear that the actions with which I am concerned are not of an automatic or unconscious sort but are a manifestation of either
implicit or explicit patter ns of reasoning.
Second, theories provide answers to conceptual puzzles.When Jeff, using his rst
theory of Twinkle, could not nd a brown-based bell to complete the tune, his
theory told him that he could not continue building the tune, and he stopped. Of
course, this is not a very satisfying answer to the puzzle, not the least because it
points to the inadequacy of the theor y as a guide for further actions. Jeff s second
theory of Twinkle, in which the concept of a pitch-event has slipped toward a
more gener ic perspective, is able to provide a solution by reinterpreting the function of one of the bells already in the bell-path, thereby permitting him to continue
building the tune.
Third, theories simplify reality. Jeff does not attempt to make the proportional
spacing of the bells replicate the temporal intervals between the pitch-events of the
tune it is enough that there is a somewhat larger space between the bells that end
the rst subphrase and those that begin the second. Jeff s notation (and, presumably,
his theory) also do not specify the tempo at which Twinkle should be played,
how hard the bells should be struck, whether the player should be seated or standing, and so on.
Fourth, theories are relatively extended cognitive structures. As analyzed here,
both of Jeff s theories involve at least three conceptual models, which descr ibe relationships among a number of concepts and which set up conceptual domains. Each
conceptual model can operate independently of the others: construing pitches as
objects does not necessar ily require arranging the objects in any particular way, nor
does it necessar ily involve physical actions perfor med on these objects (to which
mental rehearsal by musicians bears testament); temporal successions (particularly
those that do not involve music) can be represented spatially without involving
pitch-objects, and so on. Theories, however, coordinate such models into systems
for inference. The systematic quality of theories (in the sense that they involve
coordinated and thus inter related conceptual models) is evident in the way the
slippage of Jeff s concept of pitch-event cascades through his theory of Twinkle:
it affects each of the conceptual models involved.
Fifth, theories are dynamic.When Jeff s concept of pitch-event changes, his theory of Twinkle changes as well. Theories thus respond to changes in circumstances and (as Bambergers further work with Jeff shows quite clearly) can g row
even more extensive as they are developed to deal with a broader range of circumstances.
Although Jeff s efforts to construct and notate bell-paths for playing Twinkle
give a glimpse into how we structure our understanding of music, the view is limited by the specicity of his situation, which is that of a young boy struggling with
basic musical concepts. In order to know what it means to have a theory, and to
connect Jeff s functional and implicit theor ies with the abstract and explicit theories of writers like Rameau and Schenker, we need to consider in greater depth the
nature and function of the cognitive structures conceptual models and theor ies
basic to our understanding of the world as a whole.

107

108

aspe c t s of cog n i t ive st ruc ture

conceptual models and theories

As chapter 1 demonstrated, the last quarter of the twentieth century bore witness to
a great deal of research on categor ization research that fundamentally changed
our understanding of the process of categor ization and its role in human cognition.
And as shown by chapter 2, at about the same time, work in metaphor theory and
cross-domain mapping blossomed, resulting in similar changes to our understanding of the role played by mapping structure from one domain onto another. It does
not appear, however, that we will come to regard this time as one in which work on
conceptual models and theor ies came to fruition. While much interesting work has
been done work to which I shall refer in the following it has not converged on
broad agreement as to what counts as a conceptual model (or even what such structures should be called), nor has it produced any unied view of what constitutes a
theory.
Two factors have contr ibuted to this situation. First, the functions we call the
mind and the organ we call the brain are both stagger ingly complex indeed,
Gerald Edelman has called the brain the most complicated mater ial object in the
known universe.15 Further, mind and brain are not coextensive in any simple way:
mind functions extend at the least to the central nervous system, and much of what
the brain does is not directly involved with what we conventionally accept as mind
functions.16 Mind and brain thus admit of any number of structural descr iptions,
and in consequence there is great diversity in how wr iters have character ized constructs like conceptual models and theor ies. The second factor, which is related to
the rst, is that researchers have tended to focus either on conceptual-model-like
structures or on theor y-like structures, but rarely on both. The advantage of this
strategy is that it limits the amount of cognitive structure that has to be explained.
The disadvantage is that it makes it difcult to explain how relatively low-level
structures like conceptual models coordinate with relatively high-level structures
like theor ies.
In what follows, I give an account of cognitive structure that begins with concepts and extends through theor ies. In line with my analysis of the cognitive structure behind Jeff s strategies for building Twinkle, I distinguish among concepts,
conceptual models, conceptual domains, and theor ies. I should make clear that I
assume humans rarely, if ever, make explicit reference to structures of this sort in the
process of reasoning about a g iven situation. Instead, as is evident from Jeff s responses to the challenges presented by building tunes with the Montessor i bells, we
most typically are completely unaware of things like concepts and conceptual models, no matter how evident they are in our patter ns of reasoning. Concepts, conceptual models, conceptual domains, and theor ies are, by and large, implicit cognitive structures: they can be teased out through exper iment or analysis, but they need
15. Gerald M. Edelman, Bright Air, Brilliant Fire: On the Matter of Mind (New York: Basic Books,
1992), 17.
16. Things get even more complicated when emotions and feelings are factored into our account of
cognition, for there is strong evidence that these are both of the body and of the mind. Antonio
Damasio has offered compelling arguments on this point; see in particular Damasio, Descartes Error: Emotion, Reason, and the Human Brain (New York: Avon, 1994); and idem, The Feeling of What Happens.

conc e p tual mode l s and th e ori e s

not be explicit to be useful to us in reasoning about the world. I also do not want to
suggest that these constructs have unitary representations, either cognitively or
physiologically. Indeed, recent work indicates that the lines between entities such
as concepts, models, domains, and theor ies are more often blurry than clear and that
there are no simple neurobiological explanations for any of them. My distinctions
are thus pragmatic ones, intended to reect the different sorts of cognitive work
done at different levels of structural complexity and the compass of the conceptual
structures that result.
In the rst subsection that follows, my focus is on conceptual models and their
relationships to concepts and conceptual domains. In the second subsection, I turn
to theor ies and explore what recent research in anthropology and developmental
psychology can tell us about what it means to have a theory.

Conceptual Models

basic features of conceptual models

As I dene it here, a conceptual model consists of concepts in specied relationships. To get a better idea
what this means, let us return to one of the models used by Jeff in his initial attempts
to construct a bell-path for Twinkle, the model that translates temporal events
into spatial events. The concepts required for the model are event and object. Both
of these are very broad Type 1 categor ies: the typical event is something that happens in time; the typical object is something that persists over time. And, as always
with this sort of category, there are boundary cases that challenge what is typical: a
rainbow is in some respects an event (since it happens only as a result of certain
atmospher ic conditions), but, in other respects, an object (since it persists if
briey over time). In the model, events and objects are related to one another by
a cross-domain mapping between pragmatic construals of time and space, which
interprets durations (that is, the amount of time between events) as spatial dispositions (that is, the amount of distance between objects).17 Large units of time are represented by distantly spaced points, small units of time are represented by closely
spaced points, and equivalent temporal units get equivalent spacing. This way of
interpreting time can be seen in the distr ibution of marks for minutes on the face
of an analog clock: marks for ve-minute intervals are relatively far apart, marks for
one-minute intervals are relatively close together, and equivalent units get equivalent spacing.18
Conceptual models are drawn from two sources. The rst, and perhaps most
important, is culture, especially as it is represented by the people around us.We learn
conceptual models by observing and imitating the actions of others and by abstracting models from the lessons we are taught. Conceptual models are thus part of the
shared body of knowledge that constitutes culture. As Naomi Quinn and Dorothy
17. Modern physics, of course, does not retain the clear separation between the categor ies time and
space that is necessary for this mapping.
18. The spatial representation of temporal units on the face of an analog clock reects in part the
mechanism that moves the hands at a steady rate across the clock face. Compare this to the face of a sundial, on which the marks for equal units of time are not spaced equally.

109

110

aspe c t s of cog n i t ive st ruc ture

Holland put it, culture is not a peoples customs and artifacts and oral traditions, but
what they must know in order to act as they do, make the things they make, and
interpret their exper ience in the distinctive way they do.19 Conceptual models may
also be created through cross-domain mapping. The propositions and structure of
a model framed relative to one domain are applied to another domain; if the process
is successful (if it allows us to structure effectively our knowledge of some aspect of
the target domain), a new conceptual model is bor n.
Models, once lear ned or created, are stored in memory and called up to organize
our understanding of a given set of circumstances. Each of the models Jeff used in
his theory of Twinkle was lear ned or developed before he attempted to build the
tune (even if, as in the case of the model that cor relates pitch-events with objects,
it was lear ned only when he worked with Hot Cross Buns) and was recalled from
memory when needed for the new task.
Although there has been wide interest in constructs like conceptual models,
there has been an almost equally wide variety of construals of mental models of this
sort.20 There is general agreement on two basic features, however. First, conceptual
models are of relatively restricted extent that is, they structure a quite limited portion of our knowledge of the world. Second, conceptual models are abstractions
from exper ience they do not replicate the outside world in all its detail but model

19. Naomi Quinn and Dorothy Holland, Culture and Cognition, in Cultural Models in Language
and Thought, ed. Dorothy Holland and Naomi Quinn, Cambridge Readings in the Literature of Music
(Cambridge: Cambridge University Press, 1987), 4.
20. Analogues for conceptual models, and the relevant literature, include the following.
mental models: Philip Nicholas Johnson-Laird, Mental Models: Towards a Cognitive Science of Language, Inference, and Consciousness (Cambridge, Mass.: Harvard University Press, 1983); Lawrence W. Barsalou, Cognitive Psychology: An Overview for Cognitive Scientists (Hillsdale, N.J.: Erlbaum Associates, 1992);
Dedre Gentner and Albert L. Stevens, eds., Mental Models (Hillsdale, N.J.: Erlbaum Associates, 1983); Jane
Oakhill and Alan Garnham, eds., Mental Models in Cognitive Science: Essays in Honour of Phil Johnson-Laird
(East Sussex: Psychology Press, 1996).
idealized cognitive models: George Lakoff, Women, Fire, and Dangerous Things: What Categories
Reveal about the Mind (Chicago: University of Chicago Press, 1987); Robert N. McCauley, The Role of
Theories in a Theory of Concepts, in Concepts and Conceptual Development: Ecological and Intellectual Factors in Categorization, ed. Ulric Neisser (Cambr idge: Cambridge University Press, 1987), 288 309.
cultural models: Quinn and Holland, Culture and Cognition; Roy G. DAndrade and Claudia
Strauss, eds., Human Motives and Cultural Models, Publications of the Society for Psychological Anthropology (Cambr idge: Cambridge University Press, 1992); Roy G. DAndrade, The Development of Cognitive Anthropology (Cambridge: Cambridge University Press, 1995); Bradd Shore, Culture in Mind: Cognition, Culture, and the Problem of Meaning (New York: Oxford University Press, 1996).
cognitive domains: Ronald W. Langacker, Theoretical Prerequisites and Descriptive Application, vols. 1
and 2 of Foundations of Cognitive Grammar (Stanford, Calif.: Stanford University Press, 1987, 1992).
frames: Marvin Minsky,A Framework for Representing Knowledge, in The Psychology of Computer
Vision, ed. Patrick Henry Winston (New York: McGraw-Hill, 1975), 211 77; idem, The Society of Mind
(New York: Simon and Schuster, 1985); Douglas B. Lenat and R.V. Guha, Building Large Knowledge-Based
Systems: Representation and Inference in the Cyc Project (Reading, Mass.: Addison Wesley, 1989);
knowledge structures: Robert P. Abelson and John B. Black, Introduction, in Knowledge Structures, ed. James A. Galambos, Robert P. Abelson, and John B. Black (Hillsdale, N.J.: Erlbaum Associates,
1986).
Needless to say, these constructs are by no means equivalent to each other. However, all involve the
basic features I outline for conceptual models.

conc e p tual mode l s and th e ori e s

aspects of it as a guide to understanding. Modifying a distinction drawn by Clifford
Geertz, these are models for understanding a particular set of circumstances rather
than models of the circumstances themselves.21
Conceptual models are thus a very basic and yet abstract construal of how the
world is organized they are examples of how concepts rst become useful through
being related to other concepts.22 Because we cannot think without thinking in
terms of relationships of concepts, conceptual models are part of the warp and woof
of our knowledge of the world: they represent the way the world is. However, this
is not to say that conceptual models reect, in a simple way, an objective and unchanging reality, even if they are crucial to the way we construct our reality.
In earlier work, I proposed, following George Lakoff, that models of the sort I
am concer ned with here were idealized.23 Idealization is a way to account for
instances when there is a poor t between a conceptual model and actual circumstances. For example, the conceptual model that sets up the notion of a soundobject runs into something of an obstacle when it comes to the human voice, since
it is difcult to nd the physical object that cor relates with vocal sound. Most of the
time this difculty is ignored: confronted with the sounds of the human voice, we
continue to think in ter ms of sound-objects. A conceptual model concer ned with
sound-objects could thus be regarded as an ideal to which actual circumstances may
or may not confor m.
I now think it better to regard idealization as belonging to a later stage in conceptualization. As they typically function, conceptual models are immediate and
transparent: they represent our working notion of reality. Accounting for why a
conceptual model does not t a given situation belongs to a less immediate level of
cognitive structure occupied with reasoning and causal explanation. This is the level
of the theory. One job of theories, then, is to explain discrepancies between circumstances and conceptual models, even if this explanation is little more than the
recognition that the model works in most, but not all, cases. The distinction reected
in this perspective is one I shall maintain throughout the following: conceptual
models are immediate and not consciously evaluative; theories are always evaluative they provide an inter pretation of circumstances and possible courses of action
based on this inter pretation.
Conceptual models are a response to, not a simple reection of , the outside
world. They deal with only a small portion of that world and represent its structure
21. Geertz draws his distinction between models without reference to cognitive processing, focusing only (but quite importantly) on how models are used. A model for is used to guide physical relationships; a model of is a symbolic representation of a particular situation. See Clifford Geertz, Religion as a Cultural System, in Anthropological Approaches to the Study of Religion, ed. Michael Banton (New
York: Frederick A. Praeger, 1966), 7 8.While I assume that conceptual models have a symbolic representation, I do not assume these symbols are necessar ily available for conscious manipulation (that is, they
may be what Lawrence Barsalou calls perceptual symbols, which were mentioned above in chap. 2).
22. In function and structure, conceptual models bear a resemblance to the very simple, imageschematic stor ies descr ibed by Mark Turner in The Literary Mind (New York: Oxford University Press,
1996), chaps. 2 and 3. Such stor ies may involve little more than a record of the action of throwing a ball
but nonetheless are essential to our understanding of the world.
23. Lawrence Zbikowski, Large-Scale Rhythm and Systems of Grouping (Ph.D. diss.,Yale University, 1991), chap. 4.

111

112

aspe c t s of cog n i t ive st ruc ture

in ways that have more to do with cognitive efciency than with accuracy. Nonetheless, changes in the structure of the outside world will at times require changes
to a conceptual model. As we saw in the preceding discussion, the absence of the
brown-based bells caused Jeff s concept of a pitch-event to slip, resulting in changes
to the conceptual models basic to his theor y of Twinkle. On a more local level,
biological processes will have a similar, if more incremental, effect on the structure
of our conceptual models. For instance, every time we remember something, we
strengthen some synaptic connections and weaken others,re-wiring our brain on
a micro-level and subtly transfor ming the memor y we would seek to recall. Such
changes will have an inuence, slight but cumulative, on our conceptual models.
Conceptual models are thus dynamic at both the global and the local level, although,
because of their immediacy, this uidity is rarely evident.

representing conceptual models

My thinking about conceptual

models has been shaped in important ways by research in articial intelligence from
the 1970s and 1980s in particular, the work of Marvin Minsky and of Roger
Schank and Robert Abelson24 and I nd it useful to character ize conceptual
models in ter ms of simple propositions linked together to create a very basic system
of inference. I thus represented the conceptual models in chapter 1 as a collection
of interlinked nodes and shall use similar representations in chapters 4 and 5. I
should note, however, that I propose these models, and this way of representing
them, much in the spir it of that early work in articial intelligence: as a place to start
a descr iption of the processes basic to reason and inference. The models are generalized, and they are presented in a way that connects with a wide body of research.
But they are not meant to represent a true, or even empirically veried, picture of
what humans have in their heads when they reason about various situations. This
is not to say that there is no empir ical evidence for conceptual models, only that
my purpose in this book is to show how conceptual models are used rather than
to make an argument for exactly what they compr ise or how they are structured.
As I discuss in the concluding chapter, I believe much work has to be done before
a detailed picture of the structure and function of conceptual models can be
developed.

conceptual models and conceptual domains

In the conceptual model for sound-objects, the concept of a sound is linked with the concept
of an object through a process of cross-domain mapping. Elements from the invisible and ephemeral domain of sound are cor related with those from the visible and
enduring domain of objects, and structure is mapped between the domains, giving
rise to sound-objects. The domains involved in this mapping the domain of
sound and the domain of objects are broad and relatively gener ic. In compar ison,
the domain of sound-objects is rather more specic. I shall call the more specic
domains set up by conceptual models conceptual domains.
The specicity of conceptual domains is manifested in two ways. First, because

24. Minsky, The Society of Mind; Roger C. Schank and Robert P. Abelson, Scripts, Plans, Goals and
Understanding: An Inquiry into Human Knowledge Structures (Hillsdale, N.J.: Erlbaum Associates, 1977).

conc e p tual mode l s and th e ori e s

some of the relationships proper to a conceptual domain may not obtain in the
more gener ic domain (or domains) from which it descends, the scope of the conceptual domain is more limited: for example, not everything that can be said of
sound-objects can be said either of sounds or of objects. Second, conceptual domains
are typically associated with specic phenomena or exper iences. The conceptual
domain of sound-objects set up by Jeff s model pertains to his exper ience with the
Montessor i bells and with the task of building a bell-path for Twinkle. This particular domain is instantiated by the various bells used to construct the bell-path for
Twinkle, but it can be expanded to include things whose relationship to sounds
or objects is more abstract, such as marks on a page or physical places on a musical
instrument.
I regard conceptual domains as distinct from mental spaces; again, this is a departure from my earlier work. As descr ibed in chapter 2, mental spaces are complex and
transitory sites for meaning construction and for reasoning. Furnishing a mental
space requires drawing on a number of coordinated conceptual domains. One way
to achieve this coordination is through the recursive application of conceptual
models. Through recursion, the structur ing principles that give rise to a conceptual
model are subsequently applied to the elements of a conceptual domain or iginally
created by the same, or another, conceptual model. For example, consider applying
a conceptual model for anthropomor phism to the conceptual model for soundobjects. (Anthropomor phism, it will be recalled from the discussion of anthropomorphic blends in the second part of chapter 2, maps the properties of living beings
onto objects.) This application maps the properties of living beings onto soundobjects, yielding a mental space populated with animistic pitches. Admittedly, such
a space is fairly fantastic even those with a steadfast belief in the viability of an
organicist inter pretation of pitch structure usually rely on conceptual blends rather
than on the rather cur ious idea of animistic pitches but not entirely dissimilar
from mental spaces populated by pitches that obey the laws of classical physics.25
Another way to achieve the coordination of conceptual domains is by linking
conceptual models into a system for reasoning about the world that is, by creating a theory.

Theories
When I outlined the structure and function of theories in the preceding discussion,
I did so with reference to an eight-year-old boys attempts to ar range some rather
unusual bells so as to play the rst phrase and only the rst phrase of Twinkle,
Twinkle, Little Star. But is this really a theory in the way Darwins theor y of
natural selection is a theory, or in the way Einsteins special theory of relativity is a
theory?
Scientic theor ies, viewed from a broad perspective, have the features of theories
25. Organicist construals of pitch relationships are discussed in chap. 7. Early in his career, JeanPhilippe Rameau construed pitch relations in ter ms of Newtonian physics; for a discussion, see Thomas
Christensen, Rameau and Musical Thought in the Enlightenment, Cambridge Studies in Music Theory and
Analysis, 4 (Cambr idge: Cambridge University Press, 1993), 7 11.

113

114

aspe c t s of cog n i t ive st ruc ture

I outlined above. They serve as a guide for understanding and reasoning, and they
provide answers to conceptual puzzles. Thus Darwins theory sought to explain
(among other things) why nches on various islands of the Galpagos g roup had
different-shaped beaks and why similar variations could be observed in other
species.26 Scientic theor ies also simplify reality. For instance, Darwins theory did not
take into account the effect of inheritance on variation: if natural selection were
the only process affecting variation, variation should become less with each successive generation until it nally disappears altogether. However, because there is a
recombination of parental traits in offspr ing, each successive generation has new
variations. Finally, scientic theor ies are extended cognitive structures that are subject to change: Darwinian evolutionary theory was eventually rened to take into
account the effect of inheritance, as well as natural selection, on the evolution of
species.
Despite these similar ities, there are nonetheless signicant differences between
scientic theor ies and theor ies of the sort that Jeff used in his encounters with
Twinkle. Jeff s theories are of quite limited extent, pertaining only to the way
bells or similar objects might be ar ranged to play a particular song, and they have
the limited function of guiding his actions within this specic context. Scientic
theories, by contrast, attempt to account for as broad a range of phenomena as possible, and they offer predictions that can be proved true or false by any of a number of researchers working independently. These features have contr ibuted to the
success of scientic theor ies as a type, a success demonstrated by the incredible
explosion of knowledge in the hard sciences over the past two centur ies. It is no
wonder that explorations of the structure and function of scientic theor ies generated an extensive literature in the twentieth centur y and that scientic theor ies
have become a standard to which all other theor ies are compared.
The distance, then, between Jeff s theories and scientic theor ies is considerable,
enough so that using theory for the for mer might seem ill advised. Nonetheless,
work by anthropologists and others over the past thirty years has provided evidence
that scientic theor ies are but one sort of general theory and that the process of theorizing is both various and omnipresent. More recently, research by developmental
psychologists has shown that even very young children make use of theory-like
structures to guide their actions. These two lines of research, which I review briey
in the following paragraphs, do not diminish the distance between Jeff s theories
and those of modern science, but they do suggest that these two types of theory
represent points along a continuum rather than discrete entities.

cultural theorie s The notion of a cultural theory (as distinct from a

scientic theory) came to prominence in the 1960s and reected a growing aware26. One can see the puzzle that challenged Darwin emerging in his earliest wr itings. After noting,
in The Voyage of the Beagle (rst published in 1839, three years after the Beagle returned home), the gradation in the sizes of the nchess beaks, Darwin makes this prescient comment: Seeing this gradation
and diversity of structure in one small, intimately related group of birds, one might really fancy that from
an or iginal paucity of birds in this archipelago, one species had been taken and modied for different
ends. Charles Darwin, The Voyage of the Beagle, with an introduction by Leonard Engel (Garden City,
N.Y: Doubleday, 1962), 381 (chap. 17).

conc e p tual mode l s and th e ori e s

ness among anthropologists that Wester n models for representing knowledge,
including those of anthropology, were not always adequate or appropriate for nonWestern cultures. Here are three examples of the kind of knowledge that caused
anthropologists and those working in related elds to rethink what was meant by a
theory.
1. Research by Brent Berlin and his associates on ethnobotanical classication (mentioned in chapter 1 in connection with the way processes of
categor ization were reconceptualized dur ing the 1970s) demonstrated that
the taxonomic systems of indigenous cultures could be every bit as complex as those of Wester n biologists. Berlins further work has shown that
ethnobotanical classication reects, in great part, the order that is inherent in nature and that it also achieves a measure of comprehensiveness
by classifying unknown species through making reference to known
species.27
2. For more than a thousand years, long-distance noninstrumental navigation
has been practiced in Polynesia and Micronesia. Experienced navigators
routinely sail their outr igger canoes up to one hundred fty miles between islands in an area where less than 0.2 percent of the surface is land.
Edwin Hutchins perhaps uniquely qualied as someone with expertise
in navigation, ethnography, and articial intelligence has shown that this
method of navigation is based on a computational system distinct from
(yet comparable to) that of Western navigation.28
3. The cure for snakebite among the Rama Indians of Nicaragua is a complicated affair involving ritual, diet, beliefs in the super natural, and herbal
medicines especially prepared by snakebite doctors. For instance, after
someone has been bitten, the victim must make sure the snake is killed; if
the snake is not killed, the Rama believe the person cannot be cured.
While under cure, the patient is placed on a special diet, with foods prepared in a special way; the Rama believe that this is necessar y to avoid
endanger ing the lives of other members of his or her household. Although
it is difcult to sort out just which aspects of the snakebite cure have medical efcacy and which do not, the cure appears to be reasonably effective,
saving the patient over three-fourths of the time.29
One way to account for the stability and efcacy of the knowledge represented
by each of these sets of cultural practices is to construe it as organized by a theory
(a theory of botanical classication, or navigation, or snakebite cure). In that this
knowledge is specic to a particular culture or ethnographic context, anthropolo-

27. Brent Berlin, Ethnobiological Classication: Principles of Categorization of Plants and Animals in Traditional Societies (Princeton, N.J.: Princeton University Press, 1992); idem, How a Folkbotanical System
Can Be Both Natural and Comprehensive: One Maya Indians View of the Plant World, in Folkbiology,
ed. Douglas L. Medin and Scott Atran (Cambr idge, Mass.: MIT Press, 1999), 71 89.
28. Edwin Hutchins, Cognition in the Wild (Cambridge, Mass.: MIT Press, 1995), 65 93.
29. Franklin O. Loveland, Snakebite Cure among the Rama Indians of Nicaragua, in Medical
Anthropology, ed. Francis X. Grollig and Harold B. Haley (The Hague: Mouton, 1976), 97.

115

116

aspe c t s of cog n i t ive st ruc ture

gists have designated the cor responding theor ies cultural theor ies, folk theor ies, or
ethnotheor ies.30
One thing that distinguishes cultural theor ies from scientic theor ies is that scientic theor ies have an explicit tradition of verication: it has to be possible to
replicate exper imental results that support the theor y, or the theor y is cast into
question.While cultural theor ies clearly have efcacy the classication of a wealth
of plant life, navigation between distant islands, and cure from snakebite all actually
occur they are not subjected to the open and constant scrutiny that marks the
modern scientic method. Cultural theor ies, in contrast to scientic theor ies, organize a relatively stable body of knowledge that changes only gradually and most
often imperceptibly.31 Cultural theor ies, as a further manifestation of cultural
knowledge, thus both reect and constitute culture.
As I construe them, cultural theor ies need not be explicit; indeed, theories are
often manifested as patter ns of behavior rather than as explicit statements descr ibing how a specic situation should be understood or what actions are appropriate
for a given set of circumstances.32 This is particularly important for music, for theories of music may be both implicit and nonverbal.33 A glimpse of an implicit and
at least partially nonverbal theor y of music was provided by Steven Felds research
on the Kaluli of Papua New Guinea, which was touched on in chapter 2. Feld
showed that Kaluli ter minology centers around waterfall imager y, which yields a
number of highly distinctive ways to descr ibe the contour of musical descents. By
contrast, there is only a single ter m in Kaluli to descr ibe musical ascents.34 This does
not mean that the Kaluli conception of musical ascent is undifferentiated were an

30. Of course, ethnotheor y, as a sort of Third World shadow of theory, is not a very comfortable concept: theory always remains the pr ivileged ter m. For discussion of the related topic of the relationship between ethnopsychology and psychology, see Cather ine Lutz, Ethnopsychology Compared
to What? Explaining Behavior and Consciousness among the Ifaluk, in Person, Self, and Experience:
Exploring Pacic Ethnopsychologies, ed. Geoffrey M.White and John Kirkpatr ick (Berkeley: University of
California Press, 1985), 35 79.
For an illuminating discussion that also bears on this matter, framed with respect to folk models
(which have the function I here ascr ibe to theor ies), see Ladislav Holy and Milan Stuchlik, The Structure of Folk Models, in The Structure of Folk Models, ed. Ladislav Holy and Milan Stuchlik (London: Academic Press, 1981), 1 34.
31. Although histor ical evidence suggests that changes in the theor ies (and the cor responding conceptual models) that are constitutive of culture have often been so gradual as to escape notice, the incursion of Wester n commercial and technological culture into traditional societies dur ing the past two
decades has accelerated the rate of change markedly all of the researchers I cite note the effect of this
incursion into the cultural practices they study. Further discussion of this point, framed relative to musical practice, is taken up in chap. 5.
32. My construal of cultural theor ies as at least potentially implicit differs from that of Roy DAndrade, who has proposed that cultural theor ies are always explicit. See DAndrade, The Development of
Cognitive Anthropology, 172 73.
33. My thinking here follows that of John Blacking, who proposed that implicit theor y embraces
both verbal and non-verbal discourse about music. See Blacking, Music, Culture, and Exper ience, in
Music, Culture, and Experience: Selected Papers of John Blacking, ed. Reginald Byron, with a foreword by
Bruno Nettl (Chicago: University of Chicago Press, 1995), 228 31.
34. Steven Feld, Flow Like a Waterfall: The Metaphors of Kaluli Musical Theory, Yearbook for Traditional Music 13 (1981): 31 32.

conc e p tual mode l s and th e ori e s

ascending melody played improperly, they would be quick to call attention to the
mistake but that, within the broader cultural context relative to which musical
concepts are framed, ascents are not pr ivileged and thus require only the barest of
descriptive terms.35
In sum, cultural theor ies have the same general features that Jeff s theories and
scientic theor ies have. They serve as a guide to reasoning and inference, provide
answers to conceptual puzzles, and simplify reality. They are extended cognitive
structures, and they are subject to change. But cultural theor ies suggest two extensions to the conception of theory that I have outlined thus far. First, theories should
be seen as manifestations of cultural knowledge. Jeff s theor ies must thus be understood as developed within the context of a larger set of cultural practices, and so
should those of modern science. Second, theories can be implicit and to some
extent nonverbal.Without a doubt, explicit theor ies are immeasurably important to
the advancement of knowledge, for their precepts are available for public debate,
discussion, and modication. Even so, implicit theor ies make up a great part of our
cultural and pr ivate knowledge, knowledge that becomes visible only when these
theories are subjected to challenge or analysis.

theory in very young children

The fact that theor ies are everywhere in science, in culture, and in our personal construals of the world suggests that they are basic to human understanding. This perspective has received support from recent research by Alison Gopnik and Andrew Meltzoff, who have shown
that children use patter ns of reasoning quite similar to those of scientists to organize
their understanding of the world. Focusing on exper imental evidence gathered
from studies with nine- and eighteen-month-old children, Gopnik and Meltzoff
have demonstrated that childrens understanding of the world replicates the structure and function of scientic theor ies.36 For instance, six-month-old infants do not
appear to realize that spatial contact is required to move an object: they assume that
if the object moved shortly after they perfor med some action (like kicking their
legs), it was this action that caused the object to move. At around eight to ten
months, however, infants begin to develop a theor y of action on objects that is,
they appreciate that their actions can affect objects only if they are spatially in contact with those objects. As they grow older, children come to realize that this theory does not work in all circumstances: if an object is out of reach and she is given
a toy rake, the one-year-old will use the rake to make contact with the object

35. For other work on implicit theor ies of music, see Simha Arom, African Polyphony and Polyrhythm:
Musical Structure and Methodology, trans. Martin Thom, Barbara Tuckett, and Raymond Bond (Cambridge: Cambridge University Press, 1991), 139 40; idem, New Perspectives for the Descr iption of
Orally Transmitted Music, World of Music 23 (1981): 43; James R. Cowdery, The Melodic Tradition of Ireland (Kent, Oh.: Kent State University Press, 1990); and Ruth M. Stone, Let the Inside Be Sweet: The Interpretation of Music Event among the Kpelle of Liberia (Bloomington: Indiana University Press, 1982).
36. Alison Gopnik and Andrew N. Meltzoff, Words, Thoughts, and Theories (Cambridge, Mass.: MIT
Press, 1997). Those who have worked with infants and young children will appreciate the cleverness with
which the exper iments which produced this data were designed and the r igor with which they were carried out. For a report intended for a more general audience, see Alison Gopnik, Andrew N. Meltzoff, and
Patricia K. Kuhl, The Scientist in the Crib: Minds, Brains, and How Children Learn (New York: William
Morrow, 1999).

117

118

aspe c t s of cog n i t ive st ruc ture

(where the six-month-old might simply wave the rake at random), although this
contact alone will not make the object come any closer. By eighteen months of age,
children rene the theory of action on objects further still, and, now realizing that
just touching the object is not enough, an eighteen-month-old will adjust the rake
so that it captures the object cor rectly.37
According to Gopnik and Meltzoff s view, infants and young children develop
their knowledge much as scientists do. They start with an initial theor y for a particular set of circumstances (the relationship between actions and objects). They
conduct a variety of exper iments (reaching out to g rab objects, observing how
objects move as a result of their actions). Then they rene and adapt the theor y on
the basis of the exper imental results (when touching the object alone does not
produce the desired results, for example, infants and young children will modify
the theory to acknowledge that spatial contact must be of a specic sort). Indeed,
Gopnik and Meltzoff go even further to argue that it is not so much that there is
a scientist in the child as it is that there is a child in the scientist: the cognitive role
of theories in earliest childhood is the same as the cognitive role of theories at
maturity.
The strong claim that Gopnik and Meltzoff make in Words, Thoughts, and Theories that theory (scientic or developmental) is a means of nding truth and that
theorizing is an evolutionary adaptation conducive to the survival of the species
gives occasion for pause, if only because each theor y frames truth a bit differently.38 Nonetheless, the weaker version of the claim that the theor ies of children
resemble those of scientists is somewhat more promising. Thomas Kuhn, as part
of the central conceptual frame for his essay on thought exper iments, noted that
both children and scientists sometimes hold paradoxical views of phenomena and
that both lear n when these paradoxes are made evident.39 According to Kuhns
argument, thought exper iments which introduce no new evidence and would
thus seem to offer no opportunity for lear ning are one way of making paradoxical views of phenomena evident and thus advancing scientic knowledge. Although
only implicit in his discussion, Kuhn appears to consider the thought processes of
children to be analogous to those of scientists inasmuch as both exper ience the
same sort of nonempir ical learning.
While we have much to lear n about the cognitive structures that guide the
behavior of very young children it is hardly a pun to say that developmental psychology is still in its infancy evidence has begun to accumulate in support of the
37. Gopnik and Meltzoff, Words, Thoughts, and Theories, chap. 5.
38. Paul L. Harris offers a somewhat different argument against what Gopnik and Meltzoff call the
theory theory (that is, the theory that children have theor ies of the world), based on two points: rst,
that the theor ies of children and lay adults do not count as true theor ies; second, that the pursuit of truth
is not what guides the scientists day-to-day activities. As is evident from the preceding, I differ from Harris on the rst point; however I am in broad agreement with his second point. See Harris, Thinking by
Children and Scientists: False Analogies and Neglected Similar ities, in Mapping the Mind: Domain
Specicity in Cognition and Culture, ed. Lawrence A. Hirscheld and Susan A. Gelman (Cambr idge: Cambridge University Press, 1994), 294 315.
39. Thomas S. Kuhn, A Function for Thought Exper iments, in The Essential Tension: Selected Studies in Scientic Tradition and Change (Chicago: University of Chicago Press, 1977), 242 46.

conc e p tual mode l s and th e ori e s

notion that theor ies are a basic structure of human understanding.40 Whether the
theories of children are of exactly the same sort as those of modern science remains
to be seen; however, it does seem that these two sorts of theories have a similar
structure and that they are responses to the same impulse to understand the world.

cognitive theorie s and music theories

Based on his work

building Twinkle with the Montessor i bells, I do not think that there is a strong
case to be made for character izing Jeff as a scientist. Jeff is not really involved in a
wide-ranging search for knowledge, aided by a protocol of experiment and analysis. Jeff is instead doing his best to meet a fairly specic and, for him, relatively challenging task. He is similar to a scientist, however, in that theory is an important cognitive tool, serving to guide his actions and make the elusive world of sound a bit
more comprehensible. On the basis of evidence from anthropology and developmental psychology, Jeff is not alone in using theory in this way, for theor ies are part
of the very fabric of human thought. Given this context, the fact that Jeff should
have a theory of music, primitive as it may be, is not sur prising. Indeed, we should
only be sur prised if we could nd no evidence for the sort of cognitive theor ies I
have described. Even so, if the distance between Jeff s theor ies and those of modern science is considerable, the same can be said of the distance between his theories and those of music theory as a discipline. In the next section, I shrink this distance just a bit by showing how the same cognitive structure evident in Jeff s
theories can be seen in the mature work of two authors who made music theory
their lifes work: Jean-Philippe Rameau and Heinr ich Schenker.

conceptual models, theories,

and music theory
During the eighteenth century, Jean-Philippe Rameau (1683 1764) was widely recognized as the foremost opera composer in France. Rameau had intellectual aspirations, as well as musical ones, however, and these he rst brought to fruition in
his Trait dharmonie of 1722. The Trait offered two ideas that changed thinking
about musical organization throughout Europe. First, it suggested that the r ich but
complex lexicon of chordal sonor ities used in early-eighteenth-centur y music
could be simplied and rationalized by granting certain ar rangements of notes pr iority (the theory of chord inversion). Second, it proposed that musical organization
was governed by privileged relationships between specic har monic entities (the
theory of harmonic progression). In part because of the novelty of these ideas, in
part because of Rameaus famously treacherous prose, these theor ies were subject to
a variety of reinterpretations and misunderstandings by those who adopted them,
and Rameau himself continued to revise and rene his theor ies through no fewer
40. For further work on the role of theories in reasoning by children, see Frank C. Keil, Concepts,
Kinds, and Cognitive Development (Cambridge, Mass.: MIT Press, 1989); and Frank C. Keil, Daniel T.
Levin, Bethany A. Richman, and Grant Gutheil, Mechanism and Explanation in the Development of
Biological Thought: The Case of Disease, in Folkbiology, ed. Douglas L. Medin and Scott Atran (Cambridge, Mass.: MIT Press, 1999), 285 319.

119

120

aspe c t s of cog n i t ive st ruc ture

than ve additional treatises and a host of other wr itings until almost the last years
of his life.41
Heinrich Schenker (1867 1935) was an Austrian pianist, critic, and music theorist who supported himself primarily through giving pr ivate lessons. Schenkers
writings include the jour nal-like ser ies Tonwille and Meisterwerk, analyses of Beethovens piano sonatas and Ninth Symphony, and three main treatises: Harmonielehre
(1906), Kontrapunkt (1910 1922), and Der freie Satz (1935). Through these publications, Schenker developed two ideas that had a profound inuence on AngloAmerican music theor y during the latter half of the twentieth centur y. First, he
maintained that highly complex musical compositions were the result of elaborations of more basic and abstract structures. Second (and related), he held that musical works consisted of a number of hierarchical levels, proceeding from the most
basic and abstract to the most complex and concrete, and that each successive level
involved a further elaboration of a preceding level.42
The theor ies of music developed by Rameau and Schenker are complex, highly
formalized systems for descr ibing and analyzing music. In what follows, I consider
each from the perspective on theor ies developed in this chapter, using short musical analyses by Rameau and Schenker to provide glimpses into their respective theories. This will allow me to show some of the cognitive substructure of their theories, and it will also reveal some important features of extended theor ies of music.

Rameau on a Recitative by Lully

As do most of the music treatises of the early eighteenth centur y, Rameaus contain few if any analyses of actual works of music. An exception is found in his Nouveau systme of 1726, where he provides an analysis of an entire recitative from JeanBaptiste Lullys opera Armide of 1686, written on a libretto by Philippe Quinault.43
An excer pt from the analysis is given in example 3.2. The top two staves are as they
would have appeared in Lullys score: a vocal line with text, below which is the basse

example 3.2 Rameaus analysis (1726) of mm. 18 22 of Armides recitative from act 2,
scene 5 of Lullys Armide
18

A che vons . . . je fr mis!

ven geons nous

je sou

pi

re!

Basse
continue

7
Basse
fondamentale

41. For a more complete and more elegant account of Rameaus music theor y, see Christensen,
Rameau and Musical Thought in the Enlightenment.
42. The notion of hierarchy employed by Schenker is somewhat complex and is discussed further in
chap. 7.
43. Jean-Philippe Rameau, Nouveau systme de musique thorique et pratique (Paris: Ballard, 1726), 80 90.

conc e p tual mode l s and th e ori e s

continue part, with Arabic-numeral gures to indicate how the har monization
should be completed. Added beneath these is a staff for Rameaus basse fondamentale,
which encapsulated his analysis of Lullys music.
Rameaus basse fondamentale had its or igins in the precept that each chord had
one fundamental sound (in moder n ter ms, its root), from which the remaining
notes were derived. In his analyses and descr iptions of various successions of chords,
Rameau would notate each of these fundamental notes on an additional staff, thus
giving an account of what he considered to be the basis of any given harmony. The
basse fondamentale was more than just a record of chord roots, however: it also
allowed Rameau to g ive an account of musical syntax. The way Rameau accomplished this is more than a little complicated, but the essential points are as follows.44
To explain why chord Y followed chord X rather than chord Z, Rameau exploited
the notion of cadence prevalent in eighteenth-centur y music theory: chord Y followed chord X because the two were part of a nor mative cadence for mula. As part
of this for mula, chord X would invariably be a major chord to which was added a
minor seventh (in moder n ter minology, a dominant-seventh chord). Rameau came
to regard the dissonance introduced by this seventh as the mechanism that led to
cadence: the dissonant notes str ike together, and the collision leads to the appearance of the chord that completes the cadence what Rameau called the tonique.
The chord that preceded the tonique, with its fundamental bass a fth above that
of the tonique, was called the dominante-tonique.45 Of course, music of even the simplest sort involves more than tonic chords preceded by their dominants, but the
basic patter n gave Rameau a place to begin his explanation of musical syntax. This
he did by using the basse fondamentale to show the cadential relationships he took
to be the motive force in music, even if this meant that the note indicated in the
basse fondamentale could be related through only the most circuitous means to the
chord that actually appeared in the music.46
Returning to the passage from Lullys recitative, there are two places where the
note of the basse fondamentale is different from that of the basse continue. In m. 19,
Rameau indicates that the fundamental of the chord B-D-G is G, an inter pretation
wholly in keeping with moder n thought about chordal inversion. In m. 20, however, Rameau takes E as the fundamental of the C-E-G chord and implies that a D
should be added to the chord; thus the 7 above the E of the basse fondamentale.
This inter pretation reects his belief that the chord with A as its fundamental
should be preceded by its dominante-tonique, which has E as its fundamental and
which should be accompanied by the dissonant seventh; thus the E and D of m. 20.
As an additional complication, Rameau added the gure 9 to the basse continue
part of the version of Lullys score that appeared in the Nouveau systme. In this way
he was able to accept Lullys bass-note C and accommodate his own added D by
44. My discussion draws on that of Christensen in Rameau and Musical Thought in the Enlightenment,
106 23.
45. As mentioned in n. 25, Rameau employs a cross-domain mapping between the physical and
musical domains to character ize musical progression. Consequent to this mapping, individual pitches are
construed as physical objects that obey the laws of Newtonian physics.
46. For more on the function of the basse fondamentale as an extra-musical commentary on music,
see Allan Keiler, Music as Metalanguage: Rameaus Fundamental Bass, in Music Theory: Special Topics,
ed. Richmond Browne (New York: Academic Press, 1981), 83 100.

121

122

aspe c t s of cog n i t ive st ruc ture

suggesting that the latter should appear as a dissonant ninth above the bass. The fact
that the E of the basse fondamentale appears above the C of the basse continue is not
a problem for Rameau, since the C is explained by his notion of suppositionit
functions as an or namental note below the true bass note E.47 From the perspective
of modern music theory, Rameaus inter pretation of m. 20 is viewed as completely
wrongheaded, whereas his inter pretation of m. 19 is viewed as essentially r ight. It is
important to note, however, that both interpretations follow from Rameaus theory
of musical organization and its account of musical syntax.
Rameaus treatment of Lullys recitative in the Nouveau systme was not to be his
last. In 1753, Jean-Jacques Rousseau offered a competing analysis of Lullys recitative, taking Rameaus discussion as a point of departure.48 Rousseaus intent was to
show that French musicians even the most celebrated had no idea of how to set
text (in contrast to Italian musicians who, in Rousseaus view, did). Singling out the
passage cited in example 3.2, Rousseau wr ites
Certainly here is the most violent moment in the whole scene. It is here that the
greatest struggle is taking place in Armides heart. Who would believe that the
Musician has left all this ag itation in the same key, without the slightest intellectual
transition, without the slightest har monic distinction, in a manner so insipid, with a
melody so little distinguished and so inconceivably clumsy, that instead of the last verse
spoken by the Poet,
Achevons; je frmis. Vengeons-nous; je soupire
[End it; I tremble. Avenge myself; I sigh]
the Musician says precisely this:
Achevons, achevons. Vengeons-nous, vengeons-nous.

[End it; end it. Avenge myself; avenge myself.]49

Rameau responded to Rousseaus polemic with the 1754 Observations sur notre
instinct pour la musique, et sur son principe, which set out to prove two points: that harmony is the most important aspect of music (prior to both the melody or the text
that is sung) and that French music excels in the way it uses har mony.
The idea that har mony was the most important aspect of music was not new for
Rameau, for he had argued the point in his 1722 Trait. In the Observations, however, the argument takes a slightly different for m, based on the premise that our
understanding of music begins at an instinctual level that reects a natural pr inciple.
This pr inciple exists, Rameau argues, in the har mony that results from the reso47. For further discussion, see Chr istensen, Rameau and Musical Thought in the Enlightenment, 120 29;
and E. Cynthia Verba, The Development of Rameaus Thoughts on Modulation and Chromatics, Journal of the American Musicological Society 26 (1973): 69 91.
48. Jean-Jacques Rousseau, Letter on French Music, in Essay on the Origin of Languages and Writings Related to Music, trans. and ed. John T. Scott, Collected Writings of Rousseau, 7 (Hanover, N.H.: University Press of New England, 1998), 168 73.
49. Rousseau, Letter on French Music, 171 72. At this most violent moment, the princess and
sorceress, Armide, is about to take her revenge on the hero Renaud, who, enchanted, sleeps before her.
However, confusion freezes Armides vengeful hand as love casts a spell more powerful than that which
grips the sleeping hero. At the point singled out by Rousseau, Armide swings back and forth between a
steely deter mination to kill her foe and a desire to melt into his ar ms.

conc e p tual mode l s and th e ori e s

nance of every vibrating system, or corps sonore, including musical instruments and
the human voice.50 The essential elements of this har mony the principle it reveals
are the intervals of the octave, twelfth, double octave, and seventeenth, or, in a
more compact representation, root, third, and fth.
According to Rameau, these intervals are generated not only above the root (as
overtones) but below the root as well.51 These undertones gave Rameau a way to
explain the or igins of the minor mode. In the Observations, he also used the intervals below the root to explain the affectual qualities of the subdominant in contrast
with those of the dominant. The dominant is a product of the overtones; in consequence, it arouses feelings of vigor and joy. The subdominant literally, the underdominant is a product of the undertones; thus, it gives rise to feelings of weakness, gentleness, tenderness, and sadness.52
Having established this perspective on har mony and har monic function,
Rameau argued that Quinaults text was served perfectly by Lullys music, most particularly so in the case of the passage quoted in example 3.2. Rameaus argument,
illustrated by the example given in example 3.3, runs as follows: shortly before this
point, the music has been in E minor. With the D dominante-tonique of m. 18
(which sets the ang ry, ascending Achevons), the music establishes G major (Ton
de Sol in ex. 3.3). At je frmis, this G is converted into the dominante-tonique
of C. This represents a move to the subdominant of G (Ton dUt), which is accompanied by a melodic descent. With vengeons-nous, the dominante-tonique of D
is introduced. This represents a move to the dominant of G (Ton de Re), which is
accompanied by a melodic ascent. Finally, with je soupire, D is converted into the
dominante-tonique of G. Although this signals a return to G (Ton de Sol), it is also
a move to the subdominant of the immediately preceding key of D, and it is accompanied by a nal melodic descent. According to Rameau, Lully relies on our
example 3.3 Rameaus analysis (1754) of mm. 18 22 of Armides recitative from act 2,
scene 5 of Lullys Armide
[18]

[22]

A che vons . . . je fr mis!

ven geons nous . . . je sou

pi

re!

Basse
continue
Ton de Sol

Ton dUt Ton de Re

Ton de Sol

50. Jean-Philippe Rameau, Observations sur notre instinct pour la musique, et sur son principe (Paris: Prault
ls, 1754; facs. ed., Monuments of Music and Music Literature in Facsimile. Second Ser ies, Music Literature, 54 [New York: Broude Brothers, 1967]), 2.
51. The claim is rst made in Rameaus Gnration harmonique ou trait de musique thorique et pratique
(Paris: Prault ls, 1737; facs. ed. Monuments of Music and Music Literature in Facsimile. Second Ser ies,
Music Literature, 6 [New York: Broude Brothers, 1966]), 4 5.
52. Rameau, Observations, xiixiii.

123

124

aspe c t s of cog n i t ive st ruc ture

instinctual responses to these oscillations between the dominant and the subdominant to summon Armides swings between revenge and love: each move toward the
dominant arouses in us feelings appropriate to her revenge; each move toward the
subdominant arouses feelings appropriate to the dissolution of her resolve. Rameaus
analysis of this passage serves as the capstone to his argument in the Observations: the
excellence of French music, the appreciation of which had eluded Rousseau, is
embodied in the way it exploits our instinct for music in the service of musical
expression.
Rameaus revised analysis necessitated some important changes in his basse fondamentale for this passage. In the Observations, he reads the second half of each of
mm. 18 21 as including a dissonant seventh (which is in inversion in m. 19), reecting the dominante-tonique chords that establish the keys of G, C, D, and nally G.
Perhaps more important, the chord in the rst half of m. 20 no longer serves as the
dominante-tonique of the chord in the second half of the measure, for this would
undercut Rameaus assertion that C major is in effect at the beginning of the measure: the note in the basse fondamentale is now C instead of E, and there is no 7
above it. As Rameaus theory changed, so did his analysis.
Rameaus intent with this analysis, as with the theor y of music on which it is
based, is to explicate the pr inciples of musical organization and musical syntax. He
does this by focusing on har monies and har monic relationships, reading out the
larger implications of the basse continue that is, not only the notes that will complete the musical texture but also the musical sense made by the bass line itself
and making them explicit through the basse fondamentale. The project is a large
and complex one, and Rameaus analysis of Lullys recitative lets us glimpse only a
portion of all that he would try to say about musical organization. Nonetheless, this
vignette is enough to show how Rameaus music theor y is a theor y in the more
general sense developed in this chapter, for all of the ve features of theories are
clearly in evidence:

theories guide unde r standing and reasoning

Rameaus
theory gives an account of harmonic progression that is, it provides an explanation of why one chord follows another. Given a bit of context connected with a
particular chord, we could make a reasoned judgment about what the following
chord would most likely be. Through the notion of chordal inversion, his theory
also explains relationships between chords that compr ise the same pitch classes. In
the analysis shown in example 3.3, the har mony in the rst half of m. 19 is the
same as that in m. 22, even though their bass notes are different. Finally, by locating
the origin of music in the naturally occur ring corps sonore, Rameaus theory offers
a way to account for the emotions music evokes in humans. One set of emotions
(broadly, the happy ones) is associated with notes generated above the cor ps sonore;
another set (the sad ones) is associated with notes generated below the cor ps sonore.

theories provide answers to conceptual puzzles

One of
the profound puzzles that Rameau sought to solve in the Observations is why music
moves us. This he did by proposing that the cor ps sonore was not simply an acoustic
phenomenon but a natural one, too: we are moved by music in confor mance with

conc e p tual mode l s and th e ori e s

natural laws. A more parochial puzzle concer ns successions of bass notes that do not
confor m to his notions of harmonic progression (such as the C3 to A3 in m. 20 of
ex. 3.2, from Rameaus rst analysis in the Nouveau systme). Rameau suggests that
the given bass note has, essentially, an or namental function and the real bass note
(indicated in the basse fondamentale) is the one that confor ms to the proper progression of harmonies. Of course, the puzzle here could be seen as simply a
manifestation of the limitations of Rameaus theory it is one that disappears when
the theory changes, as it has by the time of the Observations. Nonetheless, the issue
raised by this puzzle/non-puzzle points up something important about theor ies and
conceptual puzzles: theories, through the way they descr ibe situations, tend to create puzzles, for which the theor y then offers solutions. When the theor y changes,
these puzzles often disappear, to be replaced by new puzzles specic to the new
theory.53

theories simplify reality

As a composer of many successful operas,

Rameau knew well that music consisted of more than just successions of dominant
and tonic chords. Focusing on this succession to the exclusion of others, however,
allowed Rameau to give an explanation of why one har mony followed another, and
thus to give an explanation of harmonic progression. And Rameaus account of the
way humans respond to music greatly simplies the emotional response to music,
reducing it to an essential binar ism at a variance with the range of emotion evoked
by effective musical works.

theories involve a number of conceptual models

Rameaus
complete theor y coordinates a large number of conceptual models, ranging from
models for pitch-events and intervallic relationships to models for or nament and
expressivity. Three models stand out in both of his analyses of Lullys recitative: the
conceptual model associated with the cor ps sonore, which uses basic acoustic properties to explain relationships among musical events; the conceptual model for
musical mechanics, which maps Newtonian physics onto musical events to explain
musical syntax; and the conceptual model for chordal inversion, which descr ibes
relationships between the notes that make up a har mony in terms of a set of dependencies created by chord generation. Although the notion of the cor ps sonore was
ultimately absorbed into general thinking about the acoustic or igins of music, the
latter two models had a profound effect on the music theory of succeeding generations: the success of Rameaus theory was such that later theor ists either incor porated these models into their own theor ies or found substitutes for them in alternative conceptions of chordal progression and inversion.

theories are dynamic As is evident from the two analyses of Lullys

recitative, Rameaus theor y of music changes over time: in the 1726 version, the
53. This situation is somewhat more common in the humanistic disciplines than in the hard sciences. In the latter, the puzzles are typically independent of humans they obtain whether humans are
observing the phenomena or not whereas in the for mer the puzzles (and the disciplines, for that matter) would not exist without humans.

125

126

aspe c t s of cog n i t ive st ruc ture

chord at the beginning of m. 20 acts as a dominante-tonique for the chord that follows; in the 1754 version (expanded to account for the drama evoked by the passage), the same chord represents a temporar y, but affectually important, harmonic
goal. It should be said that when Rameau changed his theor ies, it was not always for
the better: even though these changes allowed him to account for a wider range of
phenomena, they often introduced complications that made his theor y more
unwieldy.
Although Rameaus is a complex and highly elaborate theory, embedded in history and pertaining to a relatively limited musical repertoire, it nonetheless shares
signicant features with the cognitive theor ies that are the basic tools of human
understanding. To fully comprehend Rameaus theor y requires our reclaiming its
historical context and br inging our own conceptual models into cor respondence
with those on which his theor y is based. Certain models, such as that for chordal
inversion, pose few problems. Others, such as the conceptual model for supposition,
have little in common with our own models and require much greater effort to
understand. Nonetheless, the distance between his way of understanding and ours,
or between either of these and Jeff s, is only a matter of the concepts and conceptual models employed. The basic cognitive structures that make this understanding
possible are the same.

Schenker on a Waltz by Schubert

Although Heinr ich Schenker adopted some of Rameaus theoretical precepts, his
attitude toward them changed over the course of his wr itings.54 In his Harmonielehre
of 1906, Schenker speaks approvingly of Rameaus work, but after World War I,
Schenker became highly cr itical of all things French, Rameau included. His late
essay Rameau or Beethoven? is a sustained diatr ibe in which he portrays Rameau
as at best ignorant, at worst a malevolent force who contr ibuted to the downfall of
music.55 Rameau was not the only theor ist Schenker cr iticized, however. Even
before the war, Schenker had become convinced that Ger man music was being
pulled into mediocr ity and that music theor ists had aided and abetted the process.56
A central and explicit goal of Schenkers writings on music was to retard or ar rest
this process by showing the basis for the super iority of a select repertoire of music
associated with the her itage of German-speaking countr ies. Schenker believed that
only when the work of composers from Johann Sebastian Bach to Johannes Brahms
was once again fully appreciated could humanity be saved from its own excesses.

54. For a fuller account of how Schenkers attitude toward Rameau changed over the course of his
writings, see Harald Krebs, Schenkers Changing View of Rameau: A Compar ison of Remarks in Harmony, Counterpoint, and Rameau or Beethoven?, Theoria 3 (1989): 59 72.
55. Heinrich Schenker,Rameau or Beethoven? Creeping Paralysis or Spir itual Potency in Music?,
in The Masterwork in Music: A Yearbook, trans. Ian Bent, ed.William Drabkin, Cambridge Studies in Music
Theory and Analysis 5 (Cambr idge: Cambridge University Press, 1994 97), 3: 1 9.
56. Heinrich Schenker, Counterpoint: Book I: Cantus Firmus and Two-Voice Counterpoint, ed. John
Rothgeb, trans. John Rothgeb and Jrgen Thym, New Musical Theor ies and Fantasies, 2 (New York:
Schirmer Books, 1987), xxviixxx.

conc e p tual mode l s and th e ori e s

Basic to Schenkers music theory were two ideas: rst, that the forces that shape
music are independent of humans; second, that these forces deter mine fundamental aspects of musical syntax.57 These forces are made manifest in what Schenker
called the Naturklang, or chord of nature, consisting of the overtones generated from
a single fundamental, with priority granted to those closest to the fundamental.58
Couching his argument in strongly organicist ter ms, Schenker proposed that these
overtones were rst witness to the desire of the tone to propagate itself .59 This
desire also gave rise to various Urstze, or fundamental structures, each of which
began with either the third, fth, or octave of the fundamental. As shown in example 3.4, an Ursatz consists of a simple two-part counter point. The notes with upward stems make up the Urlinie (or fundamental line), which descends stepwise
from the Kopfton (or head note here, the third of the fundamental).60 Below this,
in one-to-one counter point, is the Brechung (or arpeggiation); in example 3.4, this
consists of an alter nation between the fundamental and fth of the Naturklang. The
Ursatz represents an idealized succession of musical mater ials, all of which are in
clearly specied relationships with each other. It thus provides an account abstract,
minimal, but nonetheless complete of musical syntax: that is, how musical materials are organized to produce the elements of musical discourse.
As Schenker would have it, the skilled composer, in intuitive response to the
powerful urges of the Naturklang, elaborates an Ursatz further and further until,
nally, a complete work is bor n. The possible elaborations of Urstze are manifold
example 3.4

A Schenkerian Ursatz beginning on the third of the fundamental

57. These two aspects of Schenkers music theory are discussed in a bit more depth in chap. 7.
58. Although there are obvious afnities between Rameaus cor ps sonore and Schenkers Naturklang, the latter is more complex in both the overtones proper to the system and their application to
musical structure. In general, the idea that there was an acoustical basis to musical organization was a
commonplace of nineteenth-century musical thought, and there is little evidence that Schenkers thinking on this matter was directly linked to that of Rameau.
59. It is self-evident that the urge to produce unending generations of overtones belongs to every
tone in equal measure. One might also compare this urge to that of animals, for it appears in fact to be
in no way infer ior to the procreative urge of a living being. [Es ist selbstverstndlich, da den Trieb,
Generationen von Obertnen ins Unendliche zu zeugen, jeder Ton in gleichem Mae besitzt. Man darf,
wenn man will, auch diesen Trieb einem animalischen vergleichen, denn er scheint in der Tat dem
Fortpanzungstr ieb eines Lebewesens durchaus nicht nachzustehen.] Heinr ich Schenker, Harmonielehre,
Neue musikalische Theor ien und Phantasien, 1 (Stuttgart: J. G. Cottasche, 1906), 13, p. 42.
60. Translating Schenkers terminology is a notor iously fraught issue. For a discussion, see Robert
Snarrenberg, Competing Myths: The American Abandonment of Schenkers Organicism, in Theory,
Analysis and Meaning in Music, ed. Anthony Pople (Cambr idge: Cambridge University Press, 1994),
29 56. In the following, I shall rely on Schenkers original ter ms (followed, at their introduction, by
common translations) as presented in Schenker, Der freie Satz, Neue musikalische Theor ien und Phantasien, 3 (Vienna: Universal-Edition, 1935).

127

128

aspe c t s of cog n i t ive st ruc ture

and, for all intents and pur poses, are coextensive with the repertoire Schenker
sought to explain. Within these possibilities, two main types of elaboration can be
distinguished. The rst is a manifestation of the desire of individual components of
the Ursatz to assert control over longer spans of music. This occurs through the
process of Verwandlung (transfor mation or prolongation), realized through a variety
of compositional stratagems.61 The second type of elaboration replicates whole sections of the Ursatz on subsidiar y levels of musical structure, thus presenting the
Ursatz in a sort of structural diminution. In both cases, the Ursatz is a rst level of
musical syntax, whose elements are subsequently transfor med or reiterated to create more extensive musical utterances. This notion of syntax is different from
Rameaus in two important respects. First, rather than descr ibing local, event-toevent relationships, Schenkers theory attempts to account for relationships over the
course of an entire work. Second, where harmony was the starting point for
Rameau, for Schenker, musical syntax starts with the contrapuntal voice-leading of
the Ursatz.
Of course, composers rarely if ever committed the successive elaborations of the
Ursatz to paper. According to Schenker, creating a musical composition that both
confor med to and properly expanded on an Ursatz was the work of genius. The
best a nongenius could do was to come to an appreciation of the nished work by
retracing the successive elaborations of the Ursatz through a ser ies of hierarchical
levels, which proceed from the backg round (the most minimal elaboration of the
Ursatz) to the foreground (which often resembles a simplied version of the
nished work).62 The result is a str iking image of musical structure that shows
musical syntax operating on a number of levels and within a number of different
time frames within each composition.
To represent the elaborations of an Ursatz and the hierarchical levels that
resulted, Schenker developed a unique set of graphical symbols that he used in his
analyses. These allowed him to show relationships among structural elements and
also made manifest the replication of elements across the various levels of an elaborative hierarchy. As an example, let us tur n to Schenkers analysis of the rst of
Franz Schuberts Valses nobles, Op. 77. The score for the work is given in example
3.5; Schenkers graphic analysis appears in example 3.6 (the Ursatz for the analysis is
the same as that g iven in ex. 3.4). Although not all of the features of Schenkers
graphic notation need concer n us here, a few will be useful for or ienting ourselves
and for understanding his analysis. Open note heads in the graph (with the exception of the F5 neighbor note after the double bars) represent components of the
Ursatz; generally speaking, slurs and beams mark off larger syntactic units. The
graph omits the rhythmic guration, lled-out chords, and much of the registral
detail of Schuberts waltz, as well as various subsidiar y passing and neighbor notes

61. The transfor mation (or prolongation) of elements of the Ursatz is more than a little complicated,
and Schenker used a number of terms in order to capture the variety of ways it is accomplished. See
Schenker, Der freie Satz, 45, p. 49.
62. Actual musical analysis most often proceeds in the reverse direction, going from completed
musical work, to foreground, to background.

conc e p tual mode l s and th e ori e s

example 3.5

Franz Schubert, Valse noble, Op. 77 No. 1, D. 969

8va

11

example 3.6

Schenkers analysis of Schuberts Op. 77 No. 1

and br ief arpeggiations. Nonetheless, it does not take too much imagination to see
(and hear) the outlines of the opening of the waltz in the rst portion of Schenkers
graph.
After the double bars, the connection between Schuberts music and Schenkers
graph is somewhat less immediate, and the graph becomes somewhat more complicated. Schenkers basic argument, however, is clear: mm. 9 11 serve as contrasting mater ial that prolongs the Kopfton, and after m. 12, the music returns to a
modied version of the opening mater ial that culminates in the completion of the
Ursatz. The completion of the Ursatz is accomplished in spite of two complications. First, scale degree 2 is not actually present in the music Schenker shows this
by placing the 2 of the Ursatz in parentheses. Second, the Ursatz is brought to com-

129

130

aspe c t s of cog n i t ive st ruc ture

pletion in a higher octave than that established by the Kopfton in m. 8, which runs
counter to Schenkers principle of obligate Lage (obligatory register).63
Schenkers theory, then, offers a view of Schuberts waltz as a sort of abstract discourse structure. The rst eight measures (the a1 section) serve as introduction, providing an Anstieg (initial ascent) from C5 through D5 to the Kopfton E5. The next
four measures (the b section) offer a contrast, prolonging the Kopfton through a
neighbor-note gure (indicated with Nbn in ex. 3.6). The nal four measures
(the a2 section) return to the opening mater ial, modied in order to complete the
Urlinie and br ing about the conclusion of the piece. According to Schenker, to the
extent that we understand this discourse structure, we understand how the Ursatz is
elaborated to create this unique composition. The force of this understanding is
such that we supply the D6 of the Ursatz at the conclusion of the waltz, even
though it is not literally present in the music.
In the hundreds of analyses that accompanied the exposition of his theor y,
Schenker showed evidence of similar discourse structures in the works of all of the
principal composers from Germany and Austria dur ing the eighteenth and nineteenth centur ies, and a few others besides. The theory of music that emerges, not
unlike the works whose structure it seeks to explicate, is fascinating in its extent and
complexity. Nonetheless, it is a theory with all of the features of a cognitive theory
outlined in this chapter.

theories guide unde r standing and reasoning

Schenkers
theory provides a broad account of musical syntax by explaining how musical materials relate to one another in confor mance with the rules of counter point and harmonic progression. His intent is to descr ibe the entirety of musical works. Given
this objective, and the organicist imperative relative to which the theory is framed,
Schenkerian analysis provides an explanation for the function and meaning of even
the smallest part of a musical composition: one can proceed, in an orderly fashion,
from the graph of example 3.6 back to the music of example 3.5. From a larger perspective, Schenkers theory gives an account of form, understood as the various roles
that different sections of music play within an overall discourse structure. The indication of a return to the opening mater ial of Schuberts waltz in m. 12 (indicated by
Schenkers a2 below the staff ) marks not only a reprise of the opening gure but
also a return to the syntactic process initiated by the opening of the waltz. Finally,
Schenker also offers an aesthetic theory: he believed that the works most worthy of
his theoretical accounts were also the most valuable aesthetically.

theories provide answers to conceptual puzzles

One of
the modest puzzles presented by Schuberts waltz concer ns the attention we should
pay to it. It is one of dozens if not hundreds of waltzes Schubert tossed off to gen-

63. The principle of obligate Lage is a reection of the urge of the notes of the Ursatz to return to
their source, which in all cases should involve a descent from the Kopfton to the octave above the fundamental (as in ex. 3.4). However, Schenker himself did not seem to regard obligate Lage as obligatory in
all cases he notes the Schubert waltz discussed here as an instance in which the pr inciple is violated.
See Schenker, Der freie Satz, 268, p. 175.

conc e p tual mode l s and th e ori e s

erate income. Does it bear the mark of his compositional genius, or is it simply
another bit of doggerel chur ned out to t the needs of the moment? Schenkers
analysis decidedly afr ms the for mer: whether it was wr itten in a few minutes or
over the course of several days, the waltz confor ms to and properly expands on one
of the Urstze basic to music. As with Rameaus account of Lullys recitative, we also
see the theory providing answers to puzzles specic to the theory. In the analysis of
the Schubert waltz, these are the complications presented by the apparent departures from the Urlinie in the concluding measures.

theories simplify reality

One of the supercial cr iticisms made of

Schenkers theory is that his graphic analyses leave out all of the surface detail that
makes a musical work distinctive and rewarding, and that register, dynamics, and
rhythmic guration are often obscured or ignored. Indeed, the graph given in
example 3.6 bears this out. This is not a deciency, however, but simply something
that marks Schenkers theory as a theor y: one that simplies reality as do all theories.64 The process of theorizing, then, is one of selecting data from a given situation
and using them to build a more comprehensive understanding of that situation.

theories involve a number of conceptual models

Schenkers
propensity for neologisms, his attempts to vivify the realm of tones through comparisons to human and animal domains, and his adaptations of the theoretical work
of earlier theor ists give rise to an astounding number of conceptual models. Two
in particular give his analysis of the Schubert waltz a distinctive avor: the conceptual model associated with the Naturklang, which supplies the organicist basis for
his account of musical syntax; and a conceptual model of hierarchy (also organicist,
and discussed in further detail in chap. 7) that cor relates numerous layers of musical syntax as a way of explaining how basic syntactic structures account for the syntax of an entire work.

theories are dynamic

In as br ief a summary as this, it is not possible to

show in any compelling way the dynamic aspect of Schenkers theory over time.
Indeed, about the best that can be done is to note Schenkers exibility with regard
to obligate Lage, which suggests that his approach to musical structure is not as r igid
as he would sometimes have us believe. Nonetheless, there is a wealth of research that
shows that Schenkers theory was a dynamic one, changing in response to both the
repertoire and problems he engaged.
Many claims have been made for Schenkers theory of music, a theory that, in
its ambition and extent, has few competitors in the Western tradition. No one to my
knowledge, however, has claimed that it is cut from the same cloth as the music theory of Rameau or that of an eight-year-old boy from the Boston area. Indeed, I
would not want to make such a claim, or at least not without considerable quali-

64. Ronald Langacker makes a similar point about the analyses produced by theor ies of language,
noting that there is certainly some truth in the view that analysis and descr iption inevitably distort subject matter since they cannot be their subject matter. Langacker, Foundations of Cognitive Grammar, 1: 14.

131

132

aspe c t s of cog n i t ive st ruc ture

cation. In the things that matter most to music theor ists, Schenkers theory is simply
of another order. Nonetheless, in its general features in the cognitive work that
it does, and in the way that it does it it shares much not only with Rameaus theory of music and with Jeff s theor ies of Twinkle but also with cognitive theor ies
as a whole.

Features of Extended Theor ies

The theor ies of Rameau and Schenker, unlike those of Jeff, are sophisticated intellectual constructs that can account for a wide range of complex musical phenomena. Such theor ies, which I shall call extended theories (in order to differentiate them
from more ad hoc theor ies like Jeff s), have three prominent features. First, extended
theories usually engage a problem or problems of long standing, like those of musical expression or musical coherence. Ancillary to this, extended theor ies often reveal
or create problems through the constructs they apply to music.
Second, extended theor ies, as the product of traditions of discourse that often
stretch across generations, are utterly embedded in culture. This was quite apparent
in the nationalism that infor med indeed, fueled both Rameaus and Schenkers
work. Although it might be easy to dismiss such cultural trappings as extraneous to
the business of music theory, judging from the wr itings of Rameau and Schenker,
it is hardly so: they see music as thoroughly implicated in the business of history and
nation.
Third, theories like those of Rameau and Schenker are invested in developing
systematic accounts of musical organization. This systematic quality can take the
form of developing an exhaustive account of musical structure. In some measure,
this is what Schenker aspired to, as did Allen Forte in his efforts to provide a theory of pitch organization for atonal music.65 The systematic aspect of a theory, however, may be somewhat more circumscr ibed, especially when its focus is musical
syntax. Here the goal is not so much to explain everything about musical organization but, rather, what is s in musical organization. This is closer to the goals of
Rameaus theory, and it tempered both the repertoire Schenker chose to deal with
and the problems he engaged.
The topic of musical syntax, especially as a manifestation of the systematic quality of extended theor ies, brings us back to the paradox that theor ies simultaneously
treat and create problems. The very notion of musical syntax the idea that musical discourse is in some way organized is in part a creation of theories that seek
to explain musical organization. Depending on the models music theory brings to
bear on the problem, musical syntax can be seen to be quite different things: a matter of how one har mony leads to another; an account of how basic contrapuntal and
harmonic frameworks are elaborated into full-edged musical works; or a different
explanation altogether. As shall be seen in the next chapter, my own approach to
musical syntax is different yet, involved as it is with the way musical mater ials are
arranged to take advantage of and to exploit our ability to categor ize musical
events.
65. Allen Forte, The Structure of Atonal Music (New Haven, Conn.:Yale University Press, 1973).

conc e p tual mode l s and th e ori e s

conclusions

Theories, as general cognitive constructs manifested in a wide range of human

activity, accomplish their work by correlating a number of conceptual models, each
of which sets up a conceptual domain. In some cases, these conceptual domains will
be quite similar to one another, but in others, the domains may have very little in
common. This means that the topography of any theory, reckoned in ter ms of the
conceptual domains it compr ises, will not necessar ily be unifor m. A theor y may
thus appear to make contradictor y claims when in fact it has simply shifted focus
from one conceptual domain to a second, and incommensurate, conceptual domain.
Thinking about a theory in terms of its constituent conceptual models can thus be
a tool for coming to understand it better, especially where these models are novel or
lead the theory in surprising directions.
Theories are also part of culture. The culture specic to music theory, it must be
admitted, is a quite rareed one. Nonetheless, the broader inuence of culture can
be seen here, too, as was demonstrated by the nationalistic claims made by both
Rameau and Schenker. These claims are not extraneous to Rameaus and Schenkers
theories but are in fact integral to them manifestations of the broader culture of
which these theor ies are a part and relative to which they are framed. This is not
to deny the uneasiness we may feel when reading the str ident claims of the music
theory of another era but only to say that they are not so very different from the
notions about pitch relationships and chordal inversion common within the same
cultural milieu.
Finally, theories of music can tell us something about cognitive structure as a
whole. Because the focal domain for music theory is a nonlinguistic one, theories of
music demonstrate how we have and cor relate nonlinguistic concepts. The abstraction of Rameaus and Schenkers theories notwithstanding, theories of music are
often action or iented. These actions can be of a quite visible kind, as demonstrated
by Jeff s attempts to set up different bell-paths for Twinkle, but they can also
involve actions less apparent to the eye: theories often suggest ways to attend to or
respond to music. Given their involvement with nonlinguistic entities and their
concer n with action, theories of music can be seen to be deeply involved with
embodied understanding in ways that few other highly developed cognitive theories are.
While the preceding chapters have sketched how work in cognitive science can
be applied to problems of musical understanding or, put another way, how our
accounts of musical understanding can be brought into line with what is cur rently
known about the mind and brain much work remains to be done to ll out this
picture. The chapters that follow make a start by showing how the framework
developed thus far can be used to give accounts of musical syntax, cultural knowledge, musical ontology, text-music relations, and the structure of theories. As I
noted in the introduction to this volume, I regard every analysis as based on some
theory or collection of theories. The theor ies I highlight in part II, however, are
theories about how basic cognitive processes are manifested in and exploited by
musical understanding. In consequence, the analyses that follow are of a somewhat
different character than musical analysis as it is usually practiced, for my intent is

133

134

aspe c t s of cog n i t ive st ruc ture

not only to shed light on specic issues of concer n to music scholars but also to
rene a theor y of music cognition. The examples around which I build my argument are not simply case studies, but are also tools that will help us probe further
into the issues discussed thus far.66
As should be clear, the view of cognitive structure, theory, and analysis that I have
presented thus far is of necessity preliminary, and it deals with only a small part of
the potential issues raised even when the domain of inquiry is restricted to the conceptual level of cognition. Nonetheless, as the following chapters will make clear,
even this partial view offers str iking new perspectives on musical understanding and
on what it means to theor ize about music.

66. This approach to the role examples play in explorations of the structure of thought is similar to
that noted by Mark Turner in his discussion of cognition and rhetor ic. See in particular Turner,Figure,
in Figurative Language and Thought, ed. Albert N. Katz, Cristina Cacciar i, Raymond W. Gibbs, Jr., and
Mark Turner (New York: Oxford University Press, 1998), 80.

part ii

analysis and theory

This page intentionally left blank

chapter four

categorization, compositional

strategy, and musical syntax

arly in his Theory of Semiotics, Umberto Eco surveyed the communicative

processes and cultural systems that might be included in the eld of semiotics,
moving through a list worthy of Jorge Luis Borges: zoosemiotics, olfactory signs,
tactile communication, codes of taste, paralinguistics, medical semiotics, kinesics and
proxemics, formalized languages, written languages, unknown alphabets, secret
codes, natural languages, visual communication, systems of objects, plot structure,
text theory, cultural codes, aesthetic texts, mass communication, and rhetor ic. Midway through this list he considered what he called musical codes. He began by
proposing that the whole of musical science since the Pythagoreans has been an
effort to descr ibe musical communication as a r igorously structured system.1 However, Eco also observed that, in spite of this tradition of structuralism from which
musical semiotics can draw, music poses ser ious difculties for semiotic theor y:
Music presents, on the one hand, the problem of a semiotic system without a
semantic level (or a content plane): on the other hand, however, there are musical
signs (or syntagms) with an explicit denotative value (trumpet signals in the ar my)
and there are syntagms or entire texts possessing pre-culturalized connotative value
(pastoral or thr illing music, etc.).2
Eco developed a somewhat fuller account of what he meant when he characterized music as a semiotic system without a semantic level a bit further on, just
prior to setting out the combinator ial rules for semiotic codes. The problem was
not that musical syntax cannot be inter preted indeed, Eco argued that any syntactic system can be inter preted on some level but that the inter pretation of
music is resolutely indeter minate. There is thus no depth to the semantic levels produced by musical syntax.3

1. Umberto Eco, A Theory of Semiotics, Advances in Semiotics (Bloomington: Indiana University

Press, 1976), 10. The idea that musical communication is a r igorously structured system is, of course,
another way to theor ize about music. Interest in such an approach has waxed and waned along with
musicians interest in numerology, mathematics, and science as possible models for musical organization.
2. Eco, A Theory of Semiotics, 11.
3. Eco, A Theory of Semiotics, 88 90. In advancing this point, Eco engages in a dialogue with Louis
Hjelmslevs distinction between semiotic and non-semiotic systems; see Louis Hjelmslev, Prolegomena to

137

138

analysis and theory

I nd Ecos comments interesting, not for the view of musical semiotics they
provide (since the passages to which I have referred are pretty much all he has to say
about music in A Theory of Semiotics), but for their clear focus on the importance of
musical syntax for musical understanding. Musical syntax is where any semiotics of
music must begin. In what follows, I build on this idea, using the perspective on
musical syntax introduced in chapter 1. There, I proposed thinking of syntax as a
connected or orderly system. In the case of music, organized sequences of musical
events properly recognized as syntactic are typically created through a set of musical practices shared among a number of musicians. Within most traditions of music
making, however, there is also some latitude in how musical mater ials can be
arranged, and the syntax specic to a particular musical work may emerge only over
time. This was the case with the Leidensmotiv from Wagners Tristan und Isolde, for
the importance of three successive statements of the motive was something that
became evident only once we were well into the opera. In situations such as this,
certain aspects of the systematic quality associated with musical syntax are determined by the expressive goals of the individual ultimately responsible for ar ranging the musical mater ials. My term for the way such expressive goals are realized is
compositional strategy.4
Compositional strategy, conceived in these ter ms, assumes an alliance between
syntax and processes of meaning construction: composers arrange musical mater ials with relatively specic expressive or communicative goals in mind. This approach
takes inspiration from recent work in cognitive linguistics, which does away with a
hard-and-fast distinction between syntax and semantics and instead construes language as made up of symbolic structures that consist of form-meaning pairs.5 A
similar perspective can be seen in Mark Turners work on the iconicity of rhetor ical gures. Turner argues that, in some cases, the for m of a rhetor ical gure is
matched to the meaning the speaker wishes to convey, connecting the imageschematic structure of the for m with the image-schematic structure of the meaning. Thus a rhetor ical gure based on repetition, such as anaphora (which involves
the repetition of the same word or group of words at the beginning of successive
clauses, sentences, or lines), can be used to summon the image of repeated blows,
as in an example attr ibuted to Longinus: By his manner, his looks, his voice, when

a Theory of Language, rev. ed., trans. Francis J.Whiteld (Madison: University of Wisconsin Press, 1963),
111 13. Hjelmslev, for his part, leaves open the question of whether music is a semiotic system in the
strict sense he develops.
4. It should be clear from the general way I have formulated the notion of compositional strategy
that these strateg ies are not the exclusive property of those traditionally recognized as composers. In
certain situations, performers and even listeners may use such strategies to organize musical mater ials.
Throughout this chapter, I shall simplify things by assuming that composers are the architects of compositional strategies, with the understanding that (mutatis mutandis) these strategies can also be implemented by those not conventionally recognized as composers.
5. Ronald W. Langacker, Conceptualization, Symbolization, and Grammar, in The New Psychology
of Language: Cognitive and Functional Approaches to Language Structure, ed. Michael Tomasello (Mahwah,
N.J.: Erlbaum Associates, 1998), 2. Langackers notion of a symbol is spelled out in g reater detail in his
Theoretical Prerequisites (vol. 1 of Foundations of Cognitive Grammar)(Stanford, Calif.: Stanford University
Press, 1987), 11 12, 56 58.

cate g ori zat i on, st rate g y, and sy ntax

he strikes you with insult, when he str ikes you like an enemy, when he str ikes you
with his knuckles, when he str ikes you like a slave. The efcacy of such a connection is straightforward enough as Turner notes, Involving members of the audience in the image schema of the iconic for m automatically involves them in the
basic structure of the meaning, thus moving them part way toward accepting the
whole.6 In what follows, I show how compositional strateg ies in music make use
of similar connections between expressive goals and the for ms expressions take.
Music being music, the meaning that is conveyed is not as direct as that which concerned Longinus, nor as quotidian as that which concer ns cognitive linguists.
Nonetheless, I argue for a direct connection between compositional strategy and
musical syntax: the organization of musical elements into systematic structures is
part and parcel of meaning construction in music.7
The particular focus of what follows is on the part processes of categorization
play in compositional strateg ies. As discussed in chapter 1, categor ies of musical
events have immediate salience for listeners and are thus important guides to any
understanding of how musical mater ials are organized. Building on the perspective
developed in chapter 3, we can say that they are basic to any theory of music. Inasmuch as composers try to shape their music to confor m to the capacities of listeners (trying, in Schoenbergs ter ms, to create musical coherence so that the work can
be comprehensible), processes of categorization will also play an important part in
compositional strategies. Musical categor ies are thus a place where the concer ns of
listeners and the concer ns of composers meet.
In the rst section of this chapter, I explore the role categor ies of musical events
play in the compositional strategies basic to a musical game attr ibuted to Wolfgang
Amadeus Mozart. The game in question is one of a genre, which seems to represent
the very antithesis of compositional strategy, since such games produce nished or
nearly nished musical works simply through tosses of the dice. Embedded within
their structure, however, is a set of rather traditional strategies for wr iting two-part

6. Mark Turner, Figure, in Figurative Language and Thought, ed. Albert N. Katz, Cristina Cacciar i,
Raymond W. Gibbs, Jr., and Mark Turner (New York: Oxford University Press, 1998), 50 51. Turners
discussion of the iconicity of rhetor ical for m appears on pp. 49 51. The quotation attr ibuted to Longinus is from Aristotle, On the Sublime, trans. W. Hamilton Fyfe, Loeb Classical Library, 199 (Cambr idge,
Mass.: Harvard University Press, 1995), 20, p. 190. Both a poetic and musical use of anaphora can be
found in Bob Dylans Knockin on Heavens Door, which rst appeared in his 1973 album Pat Garrett
and Billy the Kid. The chorus of the song consisting of reiterations of the line Knock, knock, knockin
on heavens door makes use of a repeated note in the melody, as well as a repeated word. Both repetitions cor relate with the meaning of the text. Note as well that the meaning is at least potentially gurative, rather than literal, since nowhere in the song is it assumed that heaven actually has a physical door
on which one can knock.
7. As I have character ized it here, compositional strategy bears a strong resemblance to the tactical
logistics of rhetoric, and it might thus be seen as another instance of the use of ideas drawn from rhetor ic
to conceptualize musical design (something discussed in greater detail in the rst part of chap. 7). As will
become apparent, I construe the idea of compositional strategy as rather more general than its rhetor ical counter part, and more involved with issues of immediate importance to musical expression (such as
how the temporal succession of musical mater ials should be ar ranged). This construal of compositional
strategy seems largely consonant with the approach Edward Cone took in his article On Der ivation:
Syntax and Rhetor ic, Music Analysis 6 (1987): 237 55, which was discussed in chap. 1.

139

140

analysis and theory

musical for ms, realized through the manipulation of categor ies of musical events.
During the eighteenth century, these same strategies were applied, on a much larger
scale, to sonata for m, guiding the way harmonic and thematic mater ials were laid out.
They could also be applied to melodic mater ial alone, however. In the second section
of this chapter, I show how this was accomplished through an analysis of the syntax
of the principal motive from the rst movement of Mozarts Dissonance Quartet
(K. 465). Taking the different versions of the motive as members of a Type 1 category,
the patter n of typicality descr ibed by these members as they appear over the course
of the movement shows strong similar ities to the basic strategies for sonata for m. The
syntax for these motive forms nonetheless occupies a different layer than the syntax
that guides the for mal organization of the movement as a whole.
As with the grouping of statements of Wagners Leidensmotiv, the syntax of the
principal motive from the rst movement of the Dissonance Quartet emerges
over the course of the movement. Both the emergence of musical syntax and the
independence of syntactic levels is further demonstrated by the rst movement of
Ludwig van Beethovens String Quartet, Op. 18 No. 6, which is taken up in the
third section of the chapter. Although there is little variation of the principal
motivic mater ial for the movement, Beethoven is nonetheless able to imbue two
versions of the motive with distinctly different syntactic functions. The categor ies
of motive forms that compr ise these versions then become elements at a higher
level of syntax. The nesting of syntactic functions can also be seen, in a slightly more
complicated way, in the rst movement of Beethovens String Quartet, Op. 18 No.
1, which is discussed in the fourth section of the chapter. There, different versions
of the main motive are distinguished by melodic structure and by compositional
strategy, and the categor ies for each become syntactic elements within the movement as a whole. In a concluding section I review the perspective on musical syntax developed over the course of the chapter and g ive a bit further consideration
to how syntax contr ibutes to the construction of meaning in music.

musical games and

compositional strategy
A Musical Game Attributed to Mozart
Europe in the latter half of the eighteenth centur y saw the blossoming of a number of musical games that allowed any person, without the least knowledge of
Musick, to compose minuets, marches, waltzes, and contredanses simply by tossing dice or spinning a top.8 As if by magic, a ser ies of throws or spins produced a
workable piece of music that was different every time the game was played.
The game descr ibed here the Musikalisches Wrfelspiel allowed its players to
8. One example is Pier re Hoegis A Tabular System Whereby the Art of Composing Minuets Is Made So
Easy That Any Person, without the Least Knowledge of Musick, May Compose Ten Thousand, All Different, and
in the Most Pleasing and Correct Manner (London: Welcker, 1770). For further infor mation about such
games and comprehensive lists of them, see Leonard G. Ratner, Ars combinatoria: Chance and Choice in
Eighteenth-Century Music, in Studies in Eighteenth-Century Music: A Tribute to Karl Geringer on His Seventieth Birthday, ed. H. C. Robbins Landon (New York: Da Capo Press, 1979), 343 63; Stephen A.
Hedges, Dice Music in the Eighteenth Century, Music and Letters 59 (1978): 180 87; and Sebastian Klotz,

cate g ori zat i on, st rate g y, and sy ntax

figure 4.1

Number table for mm. 1 8 of the Musikalisches Wrfelspiel

write a sixteen-measure waltz. It is one of several such games attr ibuted to Mozart
and is catalogued as K. Anh. C 30.01, but the ascr iption is hardly secure.9 Mozarts
authorship, in any event, is not my central concer n. Instead, I want to focus on the
compositional strategies that can be teased out of this particular game.
The game requires the use of a set of dice, and of two tables a number table
(the one used for the rst eight-measure per iod of the waltz is shown in g. 4.1)
Ars combinatoria oder Musik ohne Kopfzerbrechen: Kalkle des Musikalischen von Kircher bis Kir nberger, Musiktheorie 14 (1999): 231 45.
9. See Wolfgang Amadeus Mozart, Musikalisches Wrfelspiel: Eine Anleitung Walzer oder Schleifer mit
zwei Wrfeln zu componieren ohne Musikalisch zu seyn, noch von der Composition etwas zu verstehen, ed. Karl
Heinz Taubert (Mainz: B. Schotts Shne, 1956). On the matter of the authenticity of the attr ibution of
this particular game to Mozart, see Paul Lwenstein, Mozart-Kuriosa, Zeitschrift fr Musikwissenschaft 12
(1930): 342 46; Otto Er ich Deutsch, Mit Wrfeln komponieren, Zeitschrift fr Musikwissenschaft 12
(1930): 595; Herbert Gerigk, Wrfelmusik, Zeitschrift fr Musikwissenschaft 16 (1934): 359 63; Ludwig
Ritter von Kchel, Chronologisch-thematisches Verzeichnis smtlicher Tonwerke Wolfgang Amad Mozarts, 8th
ed., ed. Franz Giegling, Alexander Weinmann, and Gerd Sievers (Wiesbaden: Breitkopf and Hrtel, 1983),
581, 910; and Hedges, Dice Music in the Eighteenth Century, 183.
It is known from sketch studies (unrelated to the particular game under consideration) that Mozart
had an interest in musical diversions of this sort; see Kchel, Chronologisch-Thematisches Verzeichnis, 581.
And a musical dice game survives that is perhaps more reliably attributed to Mozart; it appears, in his hand,
on the verso of a manuscr ipt that contains a fragment also in Mozarts hand of the G minor Str ing
Quintet K. 516. See Hideo Noguchi, Mozart Musical Game in C Major, K.516f, Mitteilungen der Internationalen Stiftung Mozarteum 38 (1990): 89 101. K. 516f is signicantly different from the game to be discussed below, however, and the dispar ities that distinguish the two could be seen as arguing against the
authenticity of the latter one. The game from K. Anh C 30.01 that produces the waltz, and that I will consider, was published under Mozarts name soon after the composers death by Johann Julius Hummel
(Amsterdam and Berlin, 1793). This is hardly a guarantee of Mozarts authorship, since his name could
readily have been used simply to boost the sales of a work by another (and unknown) author.

141

142

analysis and theory

figure 4.2

Portion of the notation table for the Musikalisches Wrfelspiel

and a notation table (a portion of which is shown in g. 4.2). For the rst measure
of music, the player throws the dice and uses the result to nd a number in the rst
column of the table. Thus, the throw 11 gives the result 3 in the rst column. The
player then tur ns to the notation table, nds the music designated with a 3 there,
and uses it for the rst measure of the waltz. For the second measure of music, the
player throws the dice again and uses the result to nd a number in the second column of the number table. The throw 10 gives the result 142 in the second column;
the music that cor responds to 142 on the notation table is then used for the second
measure of the waltz. A complete waltz produced by the game is shown in example 4.1. Although the dance is hardly distinctive, it is nonetheless completely serviceable, equal to if not exceeding the quality of the Tafelmusik produced by knowledgeable amateurs of the period.
Of course, the point of such games was not the production of works of music,
but diversion. There were a number of ways the game could be played: a group of
players might sit in a circle and take turns throwing the dice, so that each successive measure was deter mined by a different player; players might each create their
own waltzes, which could then be compared with one another; the entire piece
might be wr itten down and then perfor med; or players might take turns perfor ming the piece as it progressed. In any event, the game offered an amusing and pleasant way to spend a portion of an evening, with the added benet that music produced by what appeared to be total chance was orderly and coherent, no matter
what the skill of the player.
In truth, chance played little part in the success of the music produced by such
games. Instead, what was required of the compilers of games such as this one were

cate g ori zat i on, st rate g y, and sy ntax

example 4.1
Dice throw
Result in
notation table

11

3
134

7
150

Waltz produced by the Musikalisches Wrfelspiel

11

10

142

114

85

99

7
91
1

6
125

3
117

5
176

9
51

8
168

8
89

5
170

a little knowledge about how to put the game together and an understanding of the
formal design of waltzes, contredanses, and the like. One rst needs to construct a
number table.10 A table for a sixteen-measure waltz using two dice, for instance,
starts with an 11 by 16 grid (for the eleven possible throws of the dice and sixteen
measures of the piece; the table shown in g. 4.1 provides only one-half of this
grid). The numbers 1 through 176 are then randomly distr ibuted on the 176 squares
of the grid to complete the table. To construct a notation table, it is necessar y to
write eleven variations on a sixteen-measure waltz, preferably arranging these on a
single sheet with the rst variation at the top and the eleventh at the bottom. The
various numbers of the number table are then mapped onto the individual measures
of the eleven variations. For instance, the number table used in our Musikalisches
Wrfelspiel would map the number 96 (found at the intersection of the rst row and
rst column) onto the rst measure of the rst variation, the number 22 onto the
second measure of the rst variation, and so on for all 176 measures of the eleven
variations. An example of how this might look, derived by working backward from
the number and notation tables, can be seen in example 4.2.11 Once each measure
10. Another approach would be for a number table to be adopted from a previous publication. In
fact, the number table for our waltz game is exactly the same as that for the the one published in the
Weimar Journal des Luxus und der Moden of February 1787. See Karl Heinz Taubert (ed.) Das Menuett:
Musikalisches Wurfelspiel, Paris 1786 (Zurich: Musikhaus Pan, 1988).
11. There is no way to know whether the eleven variations shown in ex. 4.2 are actually the source
variations for the game attr ibuted to Mozart: the number found in the rst square of the number table

143

144

analysis and theory

example 4.2 Eleven variations with numbers derived from the number table for the
Musikalisches Wrfelspiel
m.
96

2
22

3
141

4
41

5
25

6
122

7
14

1.

30

**2.

*1.

32

128

63

146

46

34

2.

31

**2.

*1.

69

95

158

13

153

55

110

3.

24

**2.

*1.

40

17

43

85

161

159

4.

100

**2.

*1.

48

163

45

5.

80

97

36

107

**2.

*1.
* The notes with stems down are used for the first ending.
** The notes with stems up are used for the second ending.

could, in theory, be mapped onto any of the variants of m. 1, and the game would still function properly. However, my focus in the following is not on establishing the or iginal for m of each of the eleven
variations but on the way the mater ial of each measure is varied. The principal function of ex. 4.2, which
puts each measure into its proper column, is to aid discussion of this aspect of the musical game.
Readers who would like to try their luck at composing a piece with the Musikalisches Wrfelspiel can
use ex. 4.2 (supplying their own dice). The number cor responding to each variation can be read as the
result of the throw of the dice; the columns denote the measures of the dance, as in the or iginal number table. Of course, playing the game with the number table and the notation table collapsed together
(as in ex. 4.2) takes most of the mystery and pretty much all of the fun out of the game.What remains
is the somewhat thin pleasure of nding out what combinations of measures chance will dictate.

cate g ori zat i on, st rate g y, and sy ntax

example 4.2

(continued)

10

11

12

13

14

15

16

70

21

26

112

44

109

14

117

39

126

56

174

18

116

83

66

139

15

132

73

58

145

79

90

176

34

67

160

52

170

25

143

64

125

76

136

93

(continued)

has been assigned a number, all of the measures are rewritten, in ascending numerical order, to create the notation table.
The permutational combination required by the game each of the variants of
m. 1 had to make sense with each of the variants of m. 2; each of the variants of m. 2

had to make sense with each of the variants of m. 3; and so on placed signicant

constraints on what could be varied and what sort of music was produced by the

game. In general, all of the variants of a given measure had to have the same har
-

145

example 4.2
m.
104

(continued)

2
157

3
27

4
167

5
154

6
68

7
118

6.

8
91

**2.

*1.

152

60

171

53

99

133

21

7.

127

**2.

*1.

119

84

114

50

140

86

169

8.

94

**2.

*1.

98

142

43

156

75

29

62

9.

123

**2.

*1.

87

165

61

135

47

147

10.

13

**2.

*1.

54

130

10

103

28

11.

37

106

**2.

*1.
* The notes with stems down are used for the first ending.
** The notes with stems up are used for the second ending.

146

example 4.2

(continued)

10

11

12

13

14

15

16

138

71

150

29

101

162

23

151

16

155

57

175

43

168

99

172

120

88

48

166

51

15

72

111

65

77

19

82

137

38

149

102

31

164

144

59

173

78

35

20

108

92

12

124

44

131

147

148

analysis and theory

mony, and they typically had the same bass note. The melody for each measure was
restricted to gener ic gures arpeggios, an occasional scalar run, or sometimes a
single note that could easily be concatenated with other such gures. In truth, the
compiler of the game wrote not so much variations on a waltz as variations on the
harmonic plan for a waltz, using melodic fragments that served the minimal contrapuntal plan of the whole.12 Rhythmic guration offered the chief means of
deriving the variants of a given measure, although changes to the beg inning
melodic note, overall melodic contour, or number of nonhar monic tones were also
used to render each variant unique. The result of a game constructed according to
this plan was a set of variations that was, for all practical pur poses, inexhaustible:
there are theoretically 11 that is, 45,949,729,863,572,161 different waltzes that
could be wr itten from the mater ials provided in example 4.2. The waltzes would
not be great music, but their for mulaic melodies and predictable har monic plan
were perfectly in keeping with dance for ms of the late eighteenth century.
The success of the compositions produced by such games is often attr ibuted to
the rationalization of musical mater ials effected dur ing the eighteenth centur y,
which allowed specic compositional functions to be attr ibuted to musical materials.13 Although such rationalization is quite evident in the work of music theor ists
like Joseph Riepel and Heinr ich Chr istoph Koch,14 and while it is supported by the
interchangeability of the components of example 4.2, it is not the only thing that
makes the waltz of example 4.1 believable as music.What is also at work is a compositional strategy that shapes the whole of the waltz. The strategy is evident in the
overall har monic plan for the dance, an outline of which is given in gure 4.3. The
waltz is divided into four-measure sections; the rst section (mm. 1 4) establishes
the tonic, the second (mm. 5 8) modulates to the dominant, and then the two sections repeat. The third section (mm. 9 12) remains in the dominant, the fourth section (mm. 13 16) returns to the tonic, and these, too, are repeated. Every one of the
waltzes generated by the game will have this basic har monic syntax. However, a succession of harmonies alone does not create convincing music Rameaus harmonic theor y notwithstanding, every succession of dominant to tonic does not
make for a cadence. What is crucially important for the compositional strategy
embodied by the Musikalisches Wrfelspiel is the way the mater ial for each measure
is varied. For instance, note that only one version of m. 8 appears in example 4.2
12. Mostly, the counter point above the bass is relatively free (within the str ictures of style). In a few
cases, however mm. 9 10, for example, where the dominant in position leads to a tonic , and at the
cadences that conclude each per iod the melodic possibilities are constrained by the overall counterpoint in the outer voices.
13. Ratner, Ars combinatoria, 344 45.
14. See Joseph Riepels Anfangsgrnde zur musicalischen Setzkunst (1752 68), included in his Smtliche
Schriften zur Musiktheorie,Wiener musikwissenschaftliche Beitrge (Vienna: Bhlau, 1996); and Heinr ich
Christoph Koch, Versuch einer Anleitung zur Composition (Leipzig: Adam Fr iedrich Bhme; Rudolstadt:
gedruckt mit Schr iften der Lwischen Erben, und Schirach, 1782 [vol. 1]; Leipzig: Adam Fr iedrich
Bhme, 1787 93 [vols. 2 3]; repr. Hildesheim: Georg Olms, 1969), esp. vol. 2. For a translation of portions of the latter, see Heinr ich Chr istoph Koch, Introductory Essay on Composition: The Mechanical Rules
of Melody, Sections 3 and 4, trans. Nancy Kovaleff Baker, Music Theory Translation Ser ies (New Haven,
Conn.:Yale University Press, 1983). The work of both Riepel and Koch is discussed more extensively
in the rst half of chap. 7 in this volume.

cate g ori zat i on, st rate g y, and sy ntax

Section 1

Section 2

C: I

Section 3
9

G: V

figure 4.3

G: V

IV

65
4#

II

65
4#

Section 4
10

11

IV

12

13

C: I

14

15

16

IV

II

Basic har monic plan for the Musikalisches Wrfelspiel

and, save for the ending of the tenth variation, only one version of m. 16.15 By contrast, none of the versions of m. 1 are identical. Shaping each of the trillions of
waltzes that might be generated by the game is a template that regulates the degree
and kind of variation that can occur within the sixteen-measure framework of the
dance.

A Template for Compositional Strategy

We can discover the plan of the template if we approach each column of measures
in example 4.2 as a Type 1 category of the sort discussed in chapter 1. Each measurecategory can be dened in ter ms of attributes (such as har mony, rhythmic guration, beginning melodic note, and so on), to which each of the members of the category (that is, the different variants included in the category) supplies values. Each
measure-category also has a distinctive prole of typicality, determined by the range
of values its members assign to the various attr ibutes and the weight given any particular set of values. For example, the members of measure-category 16 assign two
different values for the attr ibute rhythmic gure, but assign one of these ten out of
eleven times (as shown in g. 4.4). With respect to the attr ibute rhythmic gure, we
can say that the prole of typicality for measure-category 16 is clearly dened. By
comparison, the members of measure-category 15 assign seven different values for
the attr ibute rhythmic gure; one value is assigned ve times, the other six are
assigned once each. The prole of typicality for rhythmic guration is thus somewhat less clearly dened for measure-category 15.
Two patterns can be glimpsed from gure 4.4. The rst concerns which attributevalues change from measure-category to measure-category and which stay the
same for instance, there is no rhythmic gure in common between measurecategor ies 15 and 16. The second patter n concer ns changes in the prole of typi15. The lack of variation in mm. 8 and 16 means that only (1114)  2 merely 759,499,667,166,482
different waltzes can be wr itten using this particular game.

149

analysis and theory

measurecategory 16

measurecategory 15
rhythmic
figure

figure 4.4

number of
times used

rhythmic
figure

number of
times used

qqqqqq
Q
E

q
Q Q Q

10

q q q
Q Q Q

q q
Q
E

qqqqqq

qq q q
QQ Q Q

q qqqq
Q
E

qq q qq
Q E

qq q q
Q
E

qqqqqq

150

Values assigned the attr ibute rhythmic gure in measure-categor ies 15 and 16

cality between measure-categor ies as can be seen, there is a relatively sharp contrast between the proles of measure-categor ies 15 and 16. These patter ns, taken
together, create the template that shapes each of the waltzes generated by the game.
A somewhat more complete perspective is provided by looking at the values
assigned to just two attr ibutes harmony and rhythmic gure in the rst section of
the waltz, as illustrated by gure 4.5. As can be seen, there is little change in the values that members of measure-category 1 and measure-category 2 assign to these
attributes: one new rhythmic gure is introduced and one is dropped, but otherwise
the categor ies are (at least with respect to these two attr ibutes) identical. While
measure-category 3 retains most of the rhythmic gures of measure-category 2, the
values assigned the har mony are completely different and markedly more various.
Finally, measure-categor y 4 has signicant changes in the values assigned both
attributes: two new rhythmic gures come to prominence, and the attr ibute-values
for har mony change once more.

cate g ori zat i on, st rate g y, and sy ntax

q q q
Q 

qqqqqq
Q

q q q
Q Q Q

q q q
QQQQQQ

q qqqq
Q

m. 1 11

m. 2 11

m. 3
m. 4 4

5
7

V5

q q q
Q E

qqqqqq
Q
E

qqqqq
Q

q qq q
Q E

1
2

figure 4.5 Attribute-values for harmony and rhythmic gure in measure-categor ies 1
through 4 (left side: the values assigned to the attr ibute harmony by the members of
measure-categor ies 1 through 4 and the number of members within each measurecategory that assign that value; right side: the values assigned to the attr ibute rhythmic gure
and the number of members within each measure-category that assign that value)

There are also changes to the prole of typicality for the measure-categor ies in
this section. While the prole for measure-categor ies 1 and 2 is identical a typical variant will have a root-position tonic har mony and use either the rst or second rhythmic gure it is less clearly dened for measure-categor y 3. Although
the number and range of rhythmic gures do not change dramatically, there are
now four possible values for the attr ibute har mony, and only a bit of additional
weight is put on one of these attr ibute-values (the root-position dominant is
assigned by ve of the eleven members of the category). Things come back into
focus with measure-category 4 the most typical variant is a rst-inversion harmony using either of the last two rhythmic gures but the prole of typicality is
still not quite as clearly dened as it was in the rst two measure-categor ies.
The syntax of this section (reckoned in ter ms of harmony and rhythmic gure)
is one in which new mater ials are gradually introduced and displace previous material. At the same time, measure-categor ies with a fairly clear prole of typicality give
way to ones that have a less clearly dened prole (a trend reversed somewhat by
measure-category 4). Similar syntactic patter ns structure the remaining three sections of the waltz. The most important features of the syntactic template for the
waltzes produced by the game, taken as a whole, are as follows:16
New attr ibute-values especially for rhythmic gure are introduced
throughout the rst three sections. At the beginning of section 4, this process
is temporar ily halted and, coincident with the return C major in m. 13, rhythmic gures are for the rst time brought back.
Cadences (such as those that occur in measure-categor ies 7 8 and 15 16)
are marked by a distinctive treatment of attribute-values: values that occur in

16. In developing this character ization of the changes in attr ibute-values and typicality that can be
found in the measure-categor ies collected in ex. 4.2, I considered the attr ibutes of contour, number of
non-chord tones, bass motion, and starting pitch of the melody, as well as har mony and rhythmic gure.
I found, however, that focusing on the latter two attr ibutes provides the clearest picture of typicality
within and across the measure-categor ies. In consequence, my account of the syntactic template
tends to emphasize these attr ibutes.

151

152

analysis and theory

no other measure-category occur here, and the number of different attr ibutevalues within each measure-categor y is shar ply attenuated. In particular, the
measure-categor ies that cor respond with har monic ar rivals (that is, 8 and 16)
have the most clearly dened prole of typicality of any of the sixteen.
The syntax of sections 1 and 2 involves complementary processes: section
1 begins with measure-categor ies that have clearly dened proles of typicality; the proles of the measure-categor ies at the end of the section are less
clearly dened. Section 2 begins with measure-categor ies that have less clearly
dened proles of typicality; at the end of this section, the proles of the
measure-categor ies are clearly dened.
The syntax of sections 3 and 4 involves contrasting processes: section 3 has
the broadest range of attribute-values of any of the sections (there are, for
instance, over two times as many different rhythmic gures in section 3 as
there are in section 1) and the most poorly dened proles of typicality. By
compar ison, the range of attribute-values is much more limited in section 4,
and the clar ity of typicality is perhaps second only to section 1.
The compositional strategy embodied in this syntactic template is one that
would be familiar to aspir ing composers of the eighteenth century, for it is in complete accord with the compositional strateg ies for sixteen-measure dance for ms
proposed in treatises of the per iod. Theorists like Riepel and Koch taught that the
tonic should be plainly stated in the rst section (the measure-categor ies with tonic
harmony in this section have the most clearly dened prole of typicality); there
should be clear cadences at the end of the second and fourth sections (in the musical game, this is accomplished by the distinctive treatment accorded the end of sections 2 and 4); and the third section should be planned so that it does not become
independent from the whole in Kochs words, it must be for med so that it loses
its completeness and is bound more closely with the following fourth melodic section17 something achieved in the template by the large amount of variation
among the members of measure-categor ies 9 through 12.
At a more abstract level, the compositional strategy incor porated in the template
is, of course, similar to late-eighteenth-century conceptions of what is now conventionally called sonata for m.18 These conceptions divided sonata for m into two main
17. Koch, Introductory Essay on Composition, 100. Kochs discussion of sixteen-measure dance for ms is
on pp. 78 117.
18. The classic study of eighteenth-century conceptions of sonata for m is Leonard Ratners Harmonic Aspects of Classic Form, Journal of the American Musicological Society 2 (1949): 159 68. The viability of eighteenth-century conceptions of sonata for m, and their relevance for recent literature on the
topic of form, have been given a thorough discussion recently in Mark Evan Bonds, Wordless Rhetoric:
Musical Form and the Metaphor of Oration, Studies in the History of Music, 4 (Cambridge, Mass.: Harvard
University Press, 1991). A thoughtful cr itique of Bondss work can be found in Peter Hoyt, Review of
Mark Evan Bonds, Wordless Rhetoric, Journal of Music Theory 38 (1994): 123 43.
It should be noted that recent scholarship has argued for a substantially more complicated view of
sonata for m than that presented by either Ratner or Bonds. For an insightful discussion, see Scott Bur nham, The Second Nature of Sonata Form, in Music Theory and Natural Order from the Renaissance to the
Early Twentieth Century, ed. Suzannah Clark and Alexander Rehding (Cambr idge: Cambridge University Press, 2001), 111 41.

cate g ori zat i on, st rate g y, and sy ntax

parts, whose functions were determined by harmonic syntax: the Exposition (which
set out the tonic and, in major keys, closed on the dominant); and the Development
and Recapitulation (which closed on the tonic).19 The Exposition presented the pr incipal mater ials of the movement, much as sections 1 and 2 of the musical game present tonic and dominant; the Development departed from these mater ials, as does
section 3 of the game; and the Recapitulation returned to the mater ial of the Exposition, just as there is a reprise of previously used mater ial in section 4 of the game.
Given these similar ities, we can see that the conception of compositional strategy behind the template provides a link between modest sixteen-measure dance
forms and sonata for m: both rely on the same syntax for organizing musical materials.20 This points out something rather important about the teaching of composition in the eighteenth centur y: the highly detailed accounts of dance for ms by
Koch and others did indeed set out the means by which modest utilitar ian works
could be generated, but they also schooled students in compositional strategies that
could be applied to larger works. These larger works, then, were not simply generated by repeated applications of the compositional techniques descr ibed for minuets
and the like, but by the reinterpretation of these techniques as strateg ies for composition that could be applied to categor ies of musical events. When these categories were realized over longer spans of time, extended works (such as those by
Mozart and Beethoven discussed in the following sections) were born.

Typicality Effects and Musical Syntax

What emerges from this analysis of the Musikalisches Wrfelspiel is the signicance of
typicality effects for compositional strategy. Important structural points in the syntactic template the establishment of tonic in the opening measures, the cadence on
the dominant at the end of the second section, and the cadence on the tonic at the
end of the fourth section were all marked by measure-categor ies with clearly
dened proles of typicality. Indeed, a functional distinction between tonic and
dominant can be made simply on the basis of the amount of variation present
among the members of the measure-categor ies associated with each. Taking the
template as a whole, there is less variation in (and thus a clearer prole of typicality for) measure-categor ies associated with the tonic (that is, measure-categor ies
1 2, 4, 13 14, and 16); there is comparatively more variation in (and consequently
a less clear prole of typicality for) measure-categor ies associated with the dominant
(that is, measure categor ies 3, 6, 8, 10, 12, and 15).21
19. In the following, I capitalize Exposition, Development, and Recapitulation when they
refer to components of sonata for m, reserving the uncapitalized versions of the ter ms for the compositional techniques through which musical mater ials are laid out, developed, and reprised. There is, of
course, a certain measure of anachronism in applying these ter ms to eighteenth-century works, especially
when sonata for m is conceived as a two-part structure; nonetheless, the ter ms do single out compositional strategies recognized, if not so named, by eighteenth-century theorists.
20. The absence of an explicit link between the discussion of the appropriate compositional strategy
for dance for ms and that for sonata for m has been noted by Bonds; see Wordless Rhetoric, 26 27.
21. I interpret the C major har mony of measure-category 11 as the subdominant of G major; note
the FS3 in the bass of the sixth variation.

153

154

analysis and theory

What also emerged was the possibility of describing the syntactic structure for
a two-part form purely in ter ms of a general set of compositional strategies realized
through the manipulation of categories of musical events. There is no doubt that, in
the case of the waltz created by the game, harmonic syntax is of central importance;
nor is there any doubt that those who have descr ibed for m with an emphasis on
harmonic syntax do it with the assumption that this syntax also embodies compositional strategies applied to other aspects of music. Nonetheless, the account I have
provided shows how certain compositional strategies viewed as a generalized set
of processes on musical mater ials operate independently of the strategies associated with har monic syntax.
With these thoughts in mind, let us tur n to some actual musical compositions
as opposed to recreational games that make use of some of the compositional
strategies seen in the Musikalisches Wrfelspiel, as well as more novel strategies that
are appropriate for extended and adventuresome instrumental works.

mozart, string quartet in c major

(k. 465), first movement. alleg ro
The Allegro of Mozarts String Quartet in C major (K. 465) was one of Arnold
Schoenbergs favorites for illustrating motivic development of the sort discussed in
chapter 1 of this volume. He found the work appropriate for these pur poses in part
because of the prominence of the motive among the various musical mater ials and
in part because it is easy to nd traces of the motive even after it has left the scene
(thus supporting the organic conception of musical composition of which Schoenberg was so fond).22 Because my concer n here is not with developing an account of
the way a musical whole is der ived from a single theme, but with the motive as an
element in an unfolding musical discourse, my account approaches the syntax of the
principal melodic motive as independent from syntactic levels occupied with harmony and stylistic for mulae, and independent from musical signs or syntagms of the
sort noted by Eco. I shall, however, consider how these are cor related with the syntax of the principal motive toward the end of my analysis.
Mozart completed the quartet in C major dur ing January of 1785, and in September of that year it was published by Artaria in Vienna as one of a set of six quartets dedicated to Joseph Haydn.23 The rst movement of K. 465 begins with an
Adagio (mm. 1 22) remarkable for its har monic peregrinations and rather odd
pitch juxtapositions, which together ear ned the work its nickname as the Dissonance Quartet. The ensuing Allegro, by contrast, has hardly anything dissonant
22. For a discussion of Schoenbergs analyses of this movement, see the Commentar y by Patricia
Carpenter and Severine Neff in their translation of Schoenbergs The Musical Idea and the Logic, Technique, and Art of Its Presentation (New York: Columbia University Press, 1995), 53. Schoenberg rst developed his idea of developing variation which he later, and famously, applied to the music of
Brahms in an analysis of this movement made in 1917; see Arnold Schoenberg, Coherence, Counterpoint,
Instrumentation, Instruction in Form, ed. Severine Neff, trans. Severine Neff and Charlotte M. Cross (Lincoln: University of Nebraska Press, 1994), 41.
23. For a fuller account of the quartet and its companions, see John Irving, Mozart: The Haydn
Quartets, Cambridge Music Handbooks (Cambr idge: Cambridge University Press, 1998).

cate g ori zat i on, st rate g y, and sy ntax

about it. The Allegro is cast in sonata for m, and the pr incipal melodic motive appears
prominently in ve places: the opening and end of the Exposition (mm. 23 51 and
91 106, respectively); the Development section (mm. 107 54); and the opening
and end of the Recapitulation, extending into the Coda (mm. 155 70 and 211 46).
My discussion concentrates on the rst three of these, with br iefer comments on
the beginning of the Recapitulation and the Coda.

Measures 23 51 of the Allegro

Example 4.3 g ives the rst twenty-nine measures of the Allegro. The principal
motive of the movement is presented in the rst violin in mm. 23 24, and then
again in mm. 25 26 (symbolized with  in ex. 4.3). Using an analytical perspective
that will infor m each of the remaining analyses, I shall approach the various for ms
of this motive as a cognitive category, structured around a conceptual model. I shall
leave open the question of the status of these conceptual models that is, whether
they are empir ically veriable constructs or simply convenient analytical tools
and proceed on the general understanding developed in chapter 1 and br iey
reviewed at the beginning of this chapter: categor ies are basic structures in ter ms
of which we understand the world, they are organized around conceptual models,
and they represent a common g round where the concer ns of listeners and composers meet.
The model for the categor y consists of three cor related conceptual elements,
character ized as the things necessary for a melodic fragment to count as an instance
of the motive:
a rhythmic patter n that spans two measures of common time:
ch, h q
a rather specic diatonic contour, consisting of four elements (each is
shown in ex. 4.4); in order of succession, they are
(i) a sustained rst note
(ii) an ascending stepwise motion through a third
(iii) an upward skip of a third
(iv) a downward step
an implied har monic change from the rst to the second measure of the
motive
The contour of the motive gives rise to two distinctive intervals: the fth spanned
by ii and iii; and the fourth between i and the last note of iv. Although the implied
change in har mony is masked in mm. 23 24 (owing to the violas pedal point), in
most other cases it is quite apparent.
After mm. 23 26, there are three further instances of the motive that cor respond
to this model: two in mm. 31 34 (with the motive accompanied by the second violin), and one in the cello in mm. 44 45 (with the cellos concluding half note taking the place of the quarter-note and quarter-rest of the original). Somewhat lesstypical for ms appear in the rst violin in mm. 27 28 and in the ascending cascade
of mm. 45 48 ( in ex. 4.3). In each case, the very end of the motive has been

155

example 4.3

W. A. Mozart, String Quartet K. 465, rst movement, mm. 23 51

Allegro

23

28

33

156

example 4.3

(continued)

38

cresc.

cresc.

cresc.

cresc.

43

cresc.

cresc.

cresc.

cresc.

48

157

158

analysis and theory

example 4.4 Schoenbergs analysis of the main motive from W. A. Mozart, String Quartet K. 465, rst movement
i

iii

ii

iv

modied: in mm. 45 48, the most signicant change is to the end of the rhythmic
gure; in mm. 27 28, both the end of the rhythmic gure and contour patter n have
been changed. One additional version of the motive, heard in the rst violin in
mm. 35 36 ( in ex. 4.3), is still less typical, retaining most of the rhythmic patter n
of the original, but only element ii of the contour.
One other bit of musical mater ial (shown with  in ex. 4.3) seems to demand
afliation with the motivic category. As but a subcomponent of the motive the
ascending gesture that ends on a strong beat it is hardly typical of the category.
Indeed, hearing the fragment as a member of the category requires that it appear in
close proximity to more typical members of the categor y.When a similar gesture
occurs in mm. 62 63 and 64 65 in the midst of the rst subsidiar y theme (as
shown in ex. 4.5), it is decidedly subservient to the alter nation of ascending and
example 4.5

W. A. Mozart, String Quartet K. 465, rst movement, mm. 61 66

?

61

64

cresc.

cresc.

cresc.

cresc.

cate g ori zat i on, st rate g y, and sy ntax

descending skips that dominates the musical scene, having lost its connection to the
principal motive. The change in the way we hear this fragment in mm. 48 51,
as belong ing to the categor y that includes the pr incipal motive; in mm. 62 65, as
a rather gener ic bit of ller is an example of the phenomenon of conceptual slippage noted in chapter 3, but one applicable to purely musical concepts. The context
that induces this slippage is the absence of the principal motive and the presence of
new mater ial associated with the rst subsidiary theme.
The melodic syntax for these opening measures, then, involves setting out the
most typical members of the category rst indeed, these provide the basis for the
conceptual model for the categor y.24 Less typical members of the categor y follow
(in mm. 27 28 and 35 36), and, in mm. 37 43, the motive temporar ily disappears
altogether (save for the  fragments). It then reappears in its most typical for m in
m. 44, immediately followed by three slightly less typical versions. After this, there
are further  fragments that lead to the gener ic cadential mater ial that prepares for
the entrance of the rst subsidiary theme in m. 56.
Taken as a whole, the motive forms presented in mm. 23 51 are a relatively normative Type 1 category.What is str iking, however, is the dynamic shape of the category as it is presented over the course of time. Although typical members dominate the musical texture in the rst half of the section, they momentar ily disappear
from the texture in the second half , to be brought back at an important moment
preceding the transition to the rst subsidiary theme. The importance of the principal motive for the melodic syntax of mm. 23 51, then, is less a function of its constant presence and more a function of its memorability and cognitive utility. These
make possible a musical syntax of both exposition and development: exposition as
the motive is rst established in its most typical for m; and development as lesstypical variants are then introduced.

Measures 90 106 of the Allegro

After its gradual departure in mm. 44 51, the motive disappears completely for
forty measures, during which time two subsidiary themes are presented in the dominant. When the motive does return, it is at a moment of opportunity, following a
nal emphatic cadence of the sort that usually signals the conclusion of a section
or that prepares the way for additional thematic mater ial. By bringing back the
main motive at this point, Mozart ties together the beginning and end of the Exposition. The bonds are relatively loose ones, however, for it is a variant of the main
motive rather than the or iginal that is introduced, and the compositional strategy
applied to the statements that follow is closer to development than it is to exposition.
As shown in example 4.6, the motive enters in the second violin in m. 91, just
as the rst violin achieves its cadential G5.25 The rst violin responds by taking up
24. As noted in chap. 1, it is also possible for the conceptual model for a category of motive forms
to emerge over time, although the longer the process is delayed, the more the conceptual model comes
into conict with the global model for a musical theme.
25. The sustained note at the beginning of the motive allows Mozart to introduce it under the cover
of cadence, something he exploits at various developmental moments throughout the movement.

159

160

analysis and theory

example 4.6

W. A. Mozart, String Quartet K. 465, rst movement, mm. 90 106

90

94

98

cresc.

cresc.

cresc.

cresc.

102

cresc.

cresc.

cate g ori zat i on, st rate g y, and sy ntax

the motive as well, beginning a short game of motivic tag in which some for m of
the motive enters in each of the succeeding measures. Here, the initial motive forms
are of the  variety, highly similar to the or iginal for m but with a slightly modied
ending. Although har monic change is still evident, the constant overlap of motives
limits the har monic possibilities to the tonic and dominant of G major. The sense
of harmonic progress that accompanied the or iginal statement of the motive is
markedly attenuated: contr ibuting factors are the constant alter nation between
tonic and dominant, the cellos pedal point in mm. 91 95, and the halving of harmonic rhythm that follows in m. 96.
This reprise of the motive bears witness to new, and somewhat more extensively varied, members of the motivic category. In mm. 93 94 and 94 95, the second and rst violins present embellished versions of the motive forms they have
just stated (indicated with a  in ex. 4.6); were it not for the temporal proximity
of these to the preceding motive forms, it might be difcult to make a case for
their der ivation from the or iginal. More obvious is the inverted for m of the motive
that appears in the rst violin in mm. 96 99, and which is then used in the simultaneous presentation of motive forms in all four instr uments in mm. 103 104.
Finally, the  fragment that provided a measure of unity at the opening of the
exposition reappears in mm. 100 103 to serve as a link between the motive forms
that precede and follow it.
Given the absence of the or iginal for m of the motive in these measures, a case
could be made for modifying the conceptual model for the category so that variants
of the sort shown with a  in example 4.6 would be regarded as typical (indeed,
such a move would be the rst step along a slippery slope of motivic development
of the sort that Wagner leads us down with his transfor mations of the Leidensmotiv). Nonetheless, retaining the sense that we have moved away from what is typical of the category is important for what happens next. If the repeat sign of m. 106
is observed and the Exposition reprised, the typicality of the original version of the
motive will be reafrmed: we will have moved from the edges of the category (as
represented by mm. 91 106) back to the center. If we instead proceed from m. 106
into the Development, we will move yet farther toward the fr inges of the category.

Measures 107 35 of the Allegro

As shown in example 4.7, Mozart begins the Development with a slightly modied
version of the or iginal motive, but with changes now made to the middle of the
motive rather than to the end: he starts the ascending-third gure a step higher than
the sustained note, instead of repeating the sustained note at the beg inning of the
gure. As a result, the motive, as a whole, now spans a fth and implies no change of
harmony. The chromatic alteration in the rst violin at m. 111 moves the motive a
bit further away from its typical for m, but it is the version in the viola at mm. 112 13
that accelerates the process of transfor mation. In this version (indicated with a  in
ex. 4.7), the ascending-third gure is replaced with an ar peggio, which becomes a
feature of the following motive forms. At m. 117, the arpeggio gure takes over, and
the subsequent dissolution of the motive leads to a momentar y cadence on the
dominant of A minor in m. 121.

161

example 4.7
107

W. A. Mozart, String Quartet K. 465, rst movement, mm. 107 35

112

cresc.

cresc.

cresc.

cresc.

116

162

example 4.7

(continued)

120

124

128

132

163

164

analysis and theory

At m. 121, the cello returns to the for m of the motive used at the beginning of
the Development (but now within the context of A minor). After two statements,
the rising arpeggio gure once again replaces the ascending third within the
motive, after which point it displaces the main motive or its variants altogether (as
indicated by the preponderance of  symbols after m. 125 in ex. 4.7). In m. 147
(which is not g iven in ex. 4.7), the for m of the motive used at mm. 107 and 121
makes a nal appearance (beg inning with the cellos G2, but pointing toward
C minor). Given the events that followed the previous two appearances of this
form of the motive, this third appearance seems to suggest yet another episode of
development. Nonetheless, in m. 151, the motive yields to an ascending cascade of
arpeggio-gures (all outlining a G dominant-seventh), which br ings the Development to a close.
As was the case in the Exposition, in the Development more typical for ms of the
motive are followed by less typical for ms.What is distinctive is the repetition of this
pattern in close order (which occurs three times within the forty-odd measures of
the Development), and the emergence of the ascending  gure as the means for
Mozart to advance his musical argument. While the slightly varied for m of the
motive that marks the beginning of each patter n has prominence as a point of reference for this musical argument (and thus represents what is typical of thematic
mater ial), it lacks the sense of balanced har monic progression summoned by the
original for m of the motive. Given the variation of the motive that takes place over
the course of mm. 107 35, however, just what is typical of the or iginal motive is
something of an open question by the time we arrive at the dominant-seventh of
mm. 151 55.

Measures 155 60 and 211 46 of the Allegro

The question of motivic identity posed by the Development is answered by the
understated but nonetheless condent return of the original motive in mm. 155 58,
shown in example 4.8. Although the accompaniment is somewhat different than
that used in mm. 23 26, the motive itself is unchanged from its or iginal for m: this
is the motive at its most typical. Mozart is not content to leave it at that, however.
The accompaniment for the second statement of the paired motives, in mm. 163 66,
is not only fuller, but it br ings back the overlapped motive forms used at the end of
the Exposition and at the beg inning of the Development. These slightly varied
forms of the original (indicated with  in ex. 4.8) recall the developmental cast of
the earlier passages, as do the ascending  gures in the cello. Mozart nonetheless
quickly tur ns away from any further dissolution of the main motive and instead
moves with dispatch toward the rst subsidiary theme, which commences at m. 176.
The end of the Recapitulation restates the end of the Exposition almost exactly:
mm. 91 106 are transposed into C major for mm. 211 26. From this point, the
music proceeds either back to the beg inning of the Development (if the repeat is
taken) or forward to the Coda (which occupies mm. 227 46). As shown in example 4.9a, the latter grows out of the opening of the Development but uses yet
another modication of the original motive. This version owes much to that used
in mm. 211 13 (the transposed quotation of mm. 91 93), but applies a chromatic

example 4.8

W. A. Mozart, String Quartet K. 465, rst movement, mm. 155 70

155

159

163

167

166

analysis and theory

inection to the note at the beginning of the ascending-third gure. A further variant then concludes the movement, which is shown in example 4.9b: the end of the
motive, shortened from the or iginal half-notequarter-note gure to the successive
quarters of mm. 227 29, is shortened yet again to two successive eighth-notes. Although a similar gure occurs in mm. 229 30, that variant refers back to mm.
27 28, for the successive eighths lead immediately to additional mater ial (in the
case of mm. 229 30, a repetition of the ascending-third gure). In mm. 242 45, the
eighths simply represent a compression of the original motive, and it is with this
abbreviated for m of the motive that the movement ends.
Although the concluding section recalls the opening of the Development,
Mozart allows the prospect of further divergence from the main motive to linger
for only a moment before banishing it with a vigorous and tr iumphant cadence
in C major in mm. 234 35. Nonetheless, it cannot be said that the movement
ends with an unequivocally typical for m of the motive: if the or iginal for m of the

example 4.9 W. A. Mozart, String Quartet K. 465, rst movement: (a) mm. 227 30;
(b) mm. 242 46
a
227

cresc.

cresc.

cresc.

cresc.
b
242

cate g ori zat i on, st rate g y, and sy ntax

motive represents the stable center of the cognitive categor y of motive forms, the
conclusion does not return us to that center. Two factors infor m the ending; both
are related to global models, which provide a context for the conceptual model
that denes what is typical of this motivic categor y. First, the ending is har monically quite stable: the nal version of the motive introduced in mm. 242 45 reinforces this stability by outlining scale deg rees 5 and 1 and by placing the ar rival
on 1 as close as practicable to the r st beat of the metr ic cycle while retaining
the appogiatura gure character istic of the motive. Second, the most typical
version of the motive, in accordance with the global model for music themes relative to which it is framed, represents musical mater ial to be acted upon it is a
theme in the sense of a topic for discourse. At this point in the movement, however, discourse must be concluded: what is required is not a highly typical version
of the main motive but a reminiscence of the motive suited to the pur poses of
ending.

Compositional Strategy, Categor ization,

and the Allegro
Although I chose not to emphasize it in the preceding discussion, the overall compositional strategy that shapes the Allegro (as opposed to its pr incipal motive alone)
has, not sur prisingly, a high deg ree of conformance both with models for sonata
form proposed by music theor ists dur ing the late eighteenth century and with the
syntactic template for the musical game discussed above. The rst main section of
the work (mm. 23 106, which I have called the Exposition) moves from tonic to
dominant, with themes or thematic groups for each key area. The second main section (mm. 107 204) begins by moving through a number of different key areas (the
Development) and then tur ns to music that emphatically conr ms the tonic (the
Recapitulation). The syntax of this for m is articulated by important cadential
arrivals (such as the ar rival on the tonic in m. 44 or the ar rival on the dominant in
m. 91), and although melodic themes are important for projecting a key and rendering a particular key area distinctive, they are not the dr iving force behind this
portion of the musical discourse.
Running parallel to this for mal syntax, but independent from it, are the compositional strateg ies that shape the categor y compr ising the various versions of the
principal melodic motive of the Allegro. While the most typical members of this
category are associated with tonic (and, in their typicality, serve to emphasize the
importance of tonic within the for mal syntax), members with a high degree of conformance to the conceptual model for the category are also used to initiate presentations of the motive at points not associated with the tonic (such as the end of the
Exposition or beg inning of the Development). Each of these presentations of the
motive (whether on or off the tonic) leads to an increasing number of less typical
members of the category, a process that ends either with the introduction of a more
typical for m (as in m. 44 or m. 155), or with a move toward new thematic mater ial
(as happens following m. 49). Through the accumulating evidence of these processes
we come to understand the motive as a basis for Mozarts musical discourse

167

168

analysis and theory

something to be introduced, varied, and ultimately reprised. A similar understanding does not develop in connection with the subsidiary themes rst introduced in
the dominant dur ing the Exposition; instead, these are melodic mater ials with the
highly localized syntactic function typically attr ibuted to such themes by eighteenthcentury theor ists.
Understanding the pr incipal motive as the basis for Mozarts musical discourse,
however, is not the same thing as regarding it as the generator of discourse throughout the movement. As I pointed out in connection with example 4.5, snippets of
material reminiscent of some other aspect of the motive can be found elsewhere in
the movement.26 Nonetheless, without the context provided by more typical members of the motivic category, we tend not to hear such fragments as related to the
principal motive and thus do not hear them contr ibuting to the discourse of which
the motive is a part.
It is important to emphasize that the syntax associated with the pr incipal motive
of the Allegro is not entirely unique to this movement. As Robert Gjerdingen has
shown, composers of the late eighteenth century relied on a variety of stylistic formulae for both the substance and shape of their melodic and har monic mater ial,
and these for mulae infor m the compositional strateg ies applied to the motive.27
However, the way this particular motivic category unfolds how many typical
members of the category are presented before variation ensues, the relative typicality of subsequent members, and the amount of time before more typical members
return is something specic to the Allegro.
The principal motive of the Allegro should also be understood as an instance of
the sort of musical topics to which eighteenth-centur y composers regularly made
recourse (and to which Eco referred in his acknowledgement of musical signs).
Mozarts motive is a good example of the singing style as character ized by Leonard
Ratner, and the lyr ical presuppositions of this topic do much to shape just what
components of the motive are called on for the pur poses of variation.28 Nonetheless, I think Mozarts motive is also more than simply a musical topic, for in its status as a remarkable musical categor y remarkable in the variety of members the
category compr ises, and remarkable in the way the manipulation of these members
contributes to Mozarts overall musical argument it transcends topicality to become an essential element within the musical discourse peculiar to the Allegro of
the C major quartet.

26. A thorough analysis that makes an argument for the omnipresence of the principal motive can
be found in Bonds, Wordless Rhetoric, 102 10.
27. See in particular Robert O. Gjerdingen, A Classic Turn of Phrase: Music and the Psychology of Convention, Studies in the Cr iticism and Theor y of Music (Philadelphia: University of Pennsylvania Press,
1988); idem, Revisiting Meyers Grammatical Simplicity and Relational Richness, in Cognitive Bases
of Musical Communication, ed. Mari Riess Jones and Susan Holleran (Washington, D.C.: American Psychological Association, 1992), 225 43; and idem, Courtly Behaviors, Music Perception 13 (1996):
365 82.
28. For examples and discussion of musical topics, see Leonard G. Ratner, Classic Music: Expression,
Form, and Style (New York: Schirmer Books, 1980), 9 29. For a discussion of musical topics of this sort
within the specic context of musical semiotics, see Victor Ko Agawu, Playing with Signs: A Semiotic
Interpretation of Classical Music (Princeton, N.J.: Princeton University Press, 1991), 17 34.

cate g ori zat i on, st rate g y, and sy ntax

beethoven, string quartet in bb major,

op. 18 no. 6. first movement

Joseph Kerman devoted the rst quarter of his classic study of the Beethoven str ing
quartets to the Op. 18 set (which date from 1798 to 1800), rightly believing that
these early works held important keys to understanding how Beethoven transformed what had been a relatively modest genre into the ne plus ultra of chamber
music.29 But Kerman had little patience with the rst movement of the sixth quartet, for he found scant evidence of the sort of motivic development that would
become the hallmark of Beethovens mature style.30 In truth, the opening motive of
the work undergoes hardly any transfor mation over the course of the movement: its
syntax is not easily explained in ter ms of relative degrees of typicality among category members. What is involved instead is a categor y that includes two distinctly
different for ms of the motive, forms distinguished not only by the mater ials from
which they are fashioned but also by the way they are used. The different uses of
these two forms suggests a higher level of syntax that includes both, and this in tur n
gives a glimpse into the process by means of which musical topics, of the sort discussed by Ratner and alluded to by Eco, are formed.
Ratner, for his part, character ized this movement as pure buffa, a dancing cartoon.31 As can be seen in example 4.10, the two forms of the principal motive contribute much to the effect, providing a musical Tweedledum and Tweedledee to the
listeners Alice.32 The rst of these ( in ex. 4.10) sketches a rapid ascent through a
broken ar peggio on the tonic and, aided in the rst case by a slightly overemphatic
fortepiano, sets out the rst beats of mm. 3 and 5 as points of arrival. The second for m
of the motive ( in ex. 4.10), suave and somewhat understated, introduces the dominant and is metr ically more open. In mm. 7 and 9, it ends on the second main beat
of the measure; in its third appearance, it ends on the rst beat of the measure, but
only after its ending is extended by mm. 12 13. The overall har monic contour of
the  form is also relatively at: in mm. 5 7, for instance, it begins and ends on Bb2,
while m. 6 is preoccupied with a simple alter nation of the fth and root of the
dominant chord.
29. With this character ization I mean to take nothing away from the accomplishments in this genre
by Haydn or Mozart, but only to suggest that it is with Beethoven that the quartet eventually becomes
a site for the active contestation of musical values and that this represents a new possibility for chamber
music. See, for instance, Daniel K. L. Chua, The Galitzin Quartets of Beethoven: Opp. 127, 132, 130
(Princeton, N.J.: Princeton University Press, 1995). The issue for musicologists, then, is the extent to
which the early quartets prepare the way for this transfor mation of the genre.
30. Joseph Kerman, The Beethoven Quartets (New York:W.W. Norton, 1966), 71 74.
31. Leonard G. Ratner, The Beethoven String Quartets: Compositional Strategies and Rhetoric (Stanford,
Calif.: Stanford Bookstore, 1995), 89.
32. In Through the Looking Glass, Alice quotes a nursery rhyme about Tweedledum and Tweedledee,
and, after hear ing this, the two brothers reply:
I know what youre thinking about, said Tweedledum; but it isnt so, nohow.
Contrar iwise, continued Tweedledee, if it was so, it might be; and if it were so, it would be; but
as it isnt, it aint. Thats logic.
Lewis Carroll, The Annotated Alice: Alices Adventures in Wonderland and Through the Looking Glass,
introduction and notes by Martin Gardner, illustrated by John Tenniel (New York:W.W. Norton, 2000),
181. In a note to this passage, Gardner observes that the nursery rhyme may descend from a bit of doggerel by John Byrom (an eighteenth-century hymn writer and teacher of shorthand), which recounted
the rivalry between George Fr ideric Handel and Giovanni Battista Bononcini.

169

example 4.10
mm. 1 30

Ludwig van Beethoven, String Quartet, Op. 18 No. 6, rst movement,

Allegro con brio

Violino I

Violino II

Viola

Violoncello

11

16

cresc.

cresc.

cresc.

cate g ori zat i on, st rate g y, and sy ntax

example 4.10
21

(continued)

26

cresc.

cresc.

cresc.

cresc.

Commenting on the deliberate and somewhat stagy entrances of  and  is a

fragmentar y form of the motive ( in ex. 4.10). This consists of only a tur n and a
leap (rst ascending, then descending); its entrances reverse the expansion of the
distance between strong metr ic accents occasioned by the extension of mm. 12 13,
shrinking it rst to one-measure and then (in mm. 17 18) half-measure intervals.
This process of acceleration culminates in an ar rival on the dominant in m. 18,
embellished with a run in the cello that returns the music to the tonic and to the
 form of the motive.
To better understand the distinctive features of and relationships between these
three forms of the motive, as well as their progress through the movement, let us
consider conceptual models for each; diagrams for the models are given in gure
4.6. The conceptual model for the categor y of  forms consists of ve cor related
conceptual elements, character ized as the things necessar y for a melodic fragment
to count as an instance of this version of the motive:
the rhythmic patter n C qqqq hd
the motive begins with a tur n gure (here, around 5)
the continuation of the motive involves a broken ar peggio with ascending
contour (here beginning on 3 and ascending through a sixth to 1)

171

172

analysis and theory

b.

a.
C q q q q q q q q h.

C qqqq q q q q h q

begins with turn

figure

begins with turn

figure

broken arpeggio
with ascending
contour

C qqqq q

rhythmic pattern

rhythmic pattern

rhythmic pattern

c.

overall contour flat,

with internal skip
of a fifth

begins with turn

figure

ends with ascending

or descending skip

no change of
harmony

harmony alternates
I-V-I

no change of
harmony

melody with
accompaniment

stated in dialogue

stated in dialogue

figure 4.6 Conceptual models for the three forms of the principal motive from
Beethoven, String Quartet, Op. 18 No. 6, rst movement, mm. 1 30: (a) the  form;
(b) the  form; (c) the  form

there is no change of harmony implied by the motive

the motive is presented as a melody supported by accompaniment in the
remaining str ings33
Each of these elements could, of course, be broken down into smaller parts. This is
particularly true of the rhythmic patter n, which is extended by the  form of the
motive and truncated by the  form. As indicated by gure 4.6b, the conceptual
model for the  form retains most of the rst two components but alters the last
three. The  form continues this process, retaining only the tur n gure that is the
motives most character istic feature.
Slightly varied restatements of the  and  forms of the motive in mm. 19 29
the texture thickened somewhat, and with the two violins taking part in the dialogue proper to the  form of the motive lead to a long pedal on the dominant
of F major preparatory to the statement of the second theme, which commences
in m. 45.What is remarkable, from the perspective on musical syntax I have developed thus far, is the nearly complete lack of typicality effects among these mater ials.While it would be quite possible to include them all in a single category for the
motive, framed relative to a conceptual model that combines the essential elements
of those included in gure 4.6, there does not seem to be a compelling argument
for regarding either the  and  form as more typical of the category. A better case
could be made for conceiving of the  form as a less typical member of the category (if only on account of its fragmentar y nature), but it is a bit hard to say just
33. There are, of course, other ways the motive could be character ized. For instance, the rhythmic
gure I identify does not include the very rst note of the piece. My chief reason for seizing upon this
gure is for compar ison with other for ms of the motive. Should one wish to include the rst note, the
remaining for ms of the motive would represent even more radical departures from the standard established by the  form.

cate g ori zat i on, st rate g y, and sy ntax

what it is less typical of . It could be der ived from either the  or the  form, and
without further variants it remains somewhat difcult to settle the question.
Subsequent to a cadence on F major in m. 80 that wraps up the exposition of the
second theme, the  form of the motive reappears, as shown in example 4.11.
Although it has been transposed to F, and its third statement in the cello is accompanied by parallel tenths in the viola and by a reprise of the passagework from the
bridge in the rst violin, it is otherwise exactly the same as when it appeared in the
context of the rst theme g roup. This restatement has two consequences. First, it
conr ms the character of the  form of the motive that was established in the
opening measures that is, it reinforces the conceptual model outlined in gure
4.6b. Second, it draws out a function of  implicit in the initial exposition of motive
forms, which was to close off musical discourse.Where the  form of the motive
sets out the conditions for musical discourse establishing a two-measure hypermeter and opening the melodic range to cover more than two octaves the 
motive, with its pedestr ian alter nation of tonic-dominant-tonic and restricted
range, provided closure. Following m. 80, the motive has a similar function, but now
it serves to close off the discourse of the entire Exposition.
The syntactic functions of the  and  forms of the motive are due in part to
example 4.11
mm. 79 87

Ludwig van Beethoven, String Quartet, Op. 18 No. 6, rst movement,

79

84

cresc.

cresc.

cresc.

cresc.

173

174

analysis and theory

their resemblance to similar gures in other compositions:  has all the features of
a prototypical opening gesture within eighteenth-century style, and  has much in
common with prototypical closing gestures. It should be emphasized, however, that
how such motives are used is also important. As Janet Levy has pointed out, Joseph
Haydn, at least on occasion, used closing gestures as his opening mater ial.34 The use
of a closing gesture in this way is in part comic it creates a piece that seems to end
before it has begun but it also makes possible some marvelously ambiguous
seams: one is not sure whether a particular passage is a beginning or an ending. The
possibility of this ambiguity testies to the importance of the way musical mater ials are used within a particular composition, as distinct from how similar mater ials
are used in other compositions. For a gesture to signify opening, it has to be used
in openings; for a gesture to signify closing, it has to be used for closings.
The  form of the motives function as an opening gesture is conr med when
it initiates the Development, which begins with a tutti statement of the motive,
immediately echoed by a solo statement in the rst violin (as shown in ex. 4.12).
These slightly less than typical versions of the motive (which has heretofore always
appeared as melody with accompaniment) are followed immediately by the  form,
which tosses a repetition of the head of the  form just heard back and forth
between the upper str ings. In mm. 103 110, two further variants of the  form
appear, which extend the motive both metr ically (to four measures) and intervallically (to a compass of an eleventh, and then a twelfth). These expansions lead to G
minor, which initiates a rapid cycle of keys explored through the medium of the
thematic mater ial that rst served as a br idge between Bb and F back in mm.
33 43.
In m. 139, there is a weak cadence on F, followed immediately by a reappearance
of the  form of the motive (shown in ex. 4.13). Stated, as it is, in F major, it recalls
the end of the Exposition; however, this particular version does not confor m completely to the conceptual model diag rammed in gure 4.6b. The rhythmic prole
has been changed and is now the same as that of the  form; the dialogue presentation typical of  forms is absent, replaced here by one unaccompanied and one
accompanied statement. Nonetheless, although close to thirty measures of the
Development remain, this entry of the  form of the motive signals the end of the
harmonic exploration initiated by the variant  forms in mm. 103 10: F major having been regained, the remainder of the Development simply serves to postpone the
inevitable return to the opening mater ial. Just as it has each time before, the  form
of the motive conr ms a closing of harmonic structure.
With the return to Bb in m. 175, there is an exact reprise of the rst eighteen
measures of the movement, after which the  form of the motive is presented in a
series of close-order repetitions that initiate the br idge to the second theme. At the
conclusion of the movement, the  form returns once more, as it did in
mm. 80 86, only this time conr ming a nal close on Bb.

34. Janet Levy, Gesture, Form, and Syntax in Haydns Music, in Haydn Studies: Proceedings of the
International Haydn Conference, Washington, D.C., 1975, ed. Jens Peter Larsen, Howard Serwer, and James
Webster (New York:W.W. Norton, 1981), 355 62.

example 4.12
mm. 91 110
91

Ludwig van Beethoven, String Quartet, Op. 18 No. 6, rst movement,

96

cresc.

cresc.

cresc.

101

cresc.

cresc.

cresc.

cresc.

106

cresc.

cresc.

cresc.

cresc.

cresc.

176

analysis and theory

example 4.13
mm. 138 43
138

Ludwig van Beethoven, String Quartet, Op. 18 No. 6, rst movement,

As Kerman noted, there is little development of the principal motive in the rst
movement of Beethovens Quartet, Op. 18 No. 6. This is not to say that the motive is
uninvolved in the syntax of the movement, however, only that the musical discourse
is not, in any signicant way, about this motive. Thus the  form of the motive signals an opening for musical discourse, but it is not discourse about . Even when the
 form makes an attempt to grab center stage at the beginning of the Development,
the effort ends up being little more than comedic bluster: knowing only how to
begin, it cannot quite gure out how to continue, and after two attempts (in
mm. 102 10), it gives up. Similarly, the  form of the motive closes off discourse, but
it is not always a discourse in which  has been involved (as in its appearances at the
end of the Exposition and Recapitulation, and in the Development). Thus the two
motive forms, by virtue of the way they are used, contr ibute to a larger syntax concerned not with them but with the progress of the movement as a whole.
Despite their similar ities, I believe these two motive forms are distinct enough in
their features to be regarded as separate, though related, categories (something illustrated by the conceptual models of g. 4.6). The syntactic function that accrues to
each over the course of the movement with the  form used only for openings,
and the  form used only for closings gives further support to this distinction.
Nonetheless, I regard distinguishing the motive forms on the basis of syntactic function to belong to a later and more theoretical stage of understanding. Accordingly,
categor izing the motives is relatively immediate, while understanding the way they
function within the work emerges only over time.
As I have noted, these two forms of the motive bear resemblance to stylized
musical gestures that signal openings and closings. However, I would argue that, at
least on the abstract level concer ning processes that pertain to musical categor ies,
their function within the movement does not der ive from these gestures. To be
blunt, I do not believe knowledge of such stylized gures is necessary to understand
Beethovens musical argument. That it would enr ich this understanding I have no
doubt, but this is not the same as asserting such knowledge as a precondition for
making sense of the discourse of the movement. Instead, the function of the  and

cate g ori zat i on, st rate g y, and sy ntax

 forms of the motive is an emergent property consequent to the way they are constituted (as contrasting yet similar motivic structures) and the way they are used
within this movement.
The notion that Beethovens use of these motives bestows particular functions
on them further suggests how musical topics come to signify the things they do.
Granted, the musical mater ial for any topic needs to meet certain prerequisites of
harmony and counter point and confor m to general expectations about affect. But
beyond this, one must acknowledge the signicance of the way composers have
used particular musical gestures that is, their prevalence within a certain histor ical and cultural milieu in establishing the things these gestures come to signify.
Put another way, signication is not an inherent property of musical mater ials but
derives from how they are deployed in particular works of music that is, from the
syntactic function with which they are invested.

beethoven, string quartet in f major,

op. 18 no. 1, first movement
Where the discourse of the rst movement of Beethovens Bb major quartet tends
not to be about its opening motive, the discourse of the rst movement of the F
major quartet has difculty being about anything else: as A. B. Marx famously
observed, the opening motive occurs no fewer than 131 times within the 427 measures of the movement.35 Within this multiplicity, two categor ies of motive forms
can be distinguished, differentiated both by their constituent features and by the
compositional strategy that is applied to them. The distinction between these two
categories becomes especially plain in the Development, when a new and highly
dynamic variant of the motive appears that combines features of both. The importance of this variant for developing Beethovens musical argument is afr med in the
Coda, which revisits all three forms of the motive.
The compositional strategy associated with these various motive forms is relatively complicated, and it was not completely successful in Beethovens rst attempt
at the movement, a version descr ibed in great detail in a thorough study by Janet
Levy.36 In what follows, I discuss how this strategy is manifested in the Exposition,
Development, and Coda of the second (and nal) version of the F major quartet, and
then turn to Beethovens rst version to explore how it was brought to its nal for m.

The Exposition
In a manner similar to the rst movement of the Bb major quartet, the F major
quartet opens with a pr incipal motive that takes two distinctly different for ms. The
rst ( in ex. 4.14) is assertive, direct, and ends on the rst beat of the measure. The
second ( in ex. 4.14) is voiced in a lighter manner, is har monically uid, and ends
35. Adolph Ber nhard Marx, Ludwig van Beethoven: Leben und Schaffen, 6th ed., ed. Gustav Behncke
(Berlin: Otto Janke, 1908), 203.
36. Janet M. Levy, Beethovens Compositional Choices: The Two Versions of Opus 18, no. 1, First Movement, Studies in the Cr iticism and Theory of Music (Philadelphia: University of Pennsylvania Press, 1982).

177

example 4.14
mm. 129

Ludwig van Beethoven, String Quartet, Op. 18 No. 1, rst movement,

Allegro con brio

Violino I

Violino II

Viola

Violoncello
8

16

cresc.

cresc.

cresc.

cresc.

23

cate g ori zat i on, st rate g y, and sy ntax

on the third beat of the measure. Joining these is a fragmentary form ( in ex. 4.14)
whose character is less distinctive, and which plays a much more subsidiary role in
the Exposition. This fragment consists of little more than the stylized tur n gure
common to all for ms of the motive (and which was almost certainly the basis for
Marxs impressive tally). It is this feature of commonality that, as Ratner has noted,
allows the motive to t seamlessly into a variety of different contexts.37
As with the preceding analyses, we can character ize the distinctive features of
and relationships between various motive forms by consider ing conceptual models
for each; diagrams for the models are given in gure 4.7. The conceptual model for
the category of  forms consists of ve cor related conceptual elements, characterized as the things necessar y for a melodic fragment to count as an instance of this
version of the motive:

the rhythmic patter n T qd q

the rst part of the motive outlines a stylized tur n gure
the motive ends with a small downward skip
the motive is stated tutti
the motive has a relatively static har monic context

Save for the change in dynamics from the rst to the second pair of statements
piano replaced by forte all four of the statements of this for m of the motive in the
opening measures confor m to this model. Put another way, the members of the category presented in mm. 1 4 and 9 12 assign a relatively restricted range of values
to the attr ibutes picked out by the conceptual model.
Where the members of the  motivic category are all quite similar to one
another, the members of the  category are rather more diverse. In mm. 13 18, the
b.

a.
3
4 q. q q q q

3
4 q. q q q q

rhythmic pattern

motive ends with

small downward skip

stated tutti
relatively static
harmonic context

rhythmic pattern

3
4 q. q q q q

first part of
motive outlines stylized
turn figure

first part of
motive outlines stylized
turn figure

c.

rhythmic pattern

g
first part of
motive outlines stylized
turn figure

motive ends with

upward skip, followed
by step down
stated solo
with accompaniment
relatively fluid
harmonic contex

figure 4.7 Conceptual models for the three forms of the principal motive from
Beethoven, String Quartet, Op. 18 No. 1, rst movement, mm. 1 29: (a) the  form; (b)
the  form; (c) the  form
37. Ratner, The Beethoven String Quartets, 12 13.

179

180

analysis and theory

rst measure of each statement of the motive is unaccompanied, an appogiatura

over a fully-diminished seventh chord is introduced (in mm. 14 and 16), and the end
of the last statement (in mm. 17 18) has a slightly modied rhythmic prole.38 Versions of the motive that appear in mm. 21 26 are similar to those of mm. 13 18,
but include echoes provided by the  fragment as part of their accompaniment. The
category of  motives is, as a consequence, much more diverse: category members
represent a relatively wide range of attribute-values.
In the remainder of the Exposition, not only are these two categor ies of motives
maintained but also Beethoven continues to treat them as he did in laying out the
rst key area. The  motives (for the most part) continue to stay close to the conceptual model around which their motivic category is structured, and the  motives
(for the most part) continue to diverge in various ways from the conceptual model
around which their motivic category is structured. The Exposition thus reinforces
both the structure of the two categor ies and the compositional strateg ies through
which they were rst presented.
After the opening thematic group, the  form of the motive is prominent in four
places, putting in appearances twice in the transition to C major: once in the middle of the second thematic group, and preparatory to the cadences that conclude the
exposition. In each case, one of two relatively minor variants is used.
The rst variant, which initiates the transition that begins in m. 30 (as shown
in ex. 4.15a), differs from the model in that it is stated solo against accompaniment and in alter nation with a competing gure in the rst violin and then
concludes with a skip down an octave. From a har monic perspective, this
variant is more static than the or iginal  form in that it prolongs but a single
pitch-class (here, C) through the combination of the tur n gure and an
octave skip; Ratner has noted that, in this for m, the motive evokes the sound
of a musette.39
The second variant of the  form of the motive, introduced by the viola
in mm. 41 42, is similar to the rst, but it concludes with a repeated note
instead of an octave skip (as shown in ex. 4.15b): rather than a single pitchclass, now only a single pitch (Eb3) is prolonged. Appearing against the 
form is a  form (in the rst violin), whose features are discussed below.
This same variant of  is used by the cello in mm. 72 77 in the midst of
the second theme group (see ex. 4.15c).40 An interesting complication here is
that the gure with which this appearance of the variant competes (stated in
the rst violin) can easily be heard as a further variant of the  form of the
motive: admittedly, only the rhythmic prole and the beg inning of the tur n
gure are retained (and, unlike other  variants, this one is sequenced upward), but placed against the second  variant the similar ities are striking.
38. In the rst version of the quartet, the initial statement of the  form of the motive (in mm. 5 6)
was also unaccompanied. The difference in texture, however, is less str iking than the clar ication of harmonic context provided by the accompaniment Beethoven added, which helped emphasize the contrast
between the  and  forms.
39. Ratner, The Beethoven String Quartets, 14.
40. Note that the cello appears here an octave lower than in most published scores. This reading of
the passage reects that of the new Beethoven Werke.

example 4.15 Ludwig van Beethoven, String Quartet, Op. 18 No. 1, rst movement,
variants of the principal motive: (a) mm. 30 35; (b) mm. 41 48; (c) mm. 72 77;
(d) mm. 101 8
a
30

b
41

45

cresc.

cresc.

cresc.

cresc.

(continued)

182

analysis and theory

example 4.15

(continued)

c
72

cresc.

cresc.

cresc.
d
101

105

Heard as a variant of the  form of the motive, it offers the rst real challenge
to the standard of typicality for the motivic category.
The rst variant of  returns in the cello in mm. 101 104 (shown in ex.
4.15d) against a gure in the rst violin that can be related to the  form of
the motive; this passage is discussed in a bit more detail below.

cate g ori zat i on, st rate g y, and sy ntax

In contrast to the ubiquity of the  form of the motive, the  form makes only
two appearances after m. 30. The rst, already mentioned above, is in m. 42 (see ex.
4.15b). Here the motive appears with an altered rhythmic prole (lacking its rst
note and using the variant ending rst introduced in m. 18), and, in its role as counterstatement to the modied  form in the viola, with less independence than
before. The third entrance of the variant (in m. 46) departs further still, with its ending a durational augmentation of that used in the preceding measures. A similar
augmentation of the tail of the  form is heard in mm. 89 91 (as shown in ex. 4.16)
after a general pause that inter rupts the mater ial from the second thematic g roup.
Save for this augmentation, this variant of the  form of the motive is quite close to
the very rst version of , heard in mm. 5 6.
The variant heard in mm. 89 91 is the last explicit reference to the  form of
the motive before the conclusion of the Exposition. Nonetheless, the appogiatura
gure that came to be a feature of  beginning in m. 14 gures quite prominently
in the counterstatement to the nal appearance of the  form in mm. 101 104
(shown in ex. 4.15d). When this counterstatement is embellished with a repeated
turn gure in m. 106 (answering the embellishment of the  variant in m. 105), the
connection becomes somewhat more plausible: the counterstatement could be
heard as a further variant of the  form of the motive. This inter pretation is supported by two things: rst, there is the close relationship between the  and  forms
of the motive throughout the Exposition (exemplied by their juxtaposition in
mm. 41 48). Second, the tur n gure typical of all for ms of the motive is replicated,
in compressed for m, by the combination of the trill and ter mination on the third
beat of m. 102 and the appogiatura of m. 103 (and further varied in mm. 104 105).
The possibility of extending the category of  motives to include the counterstatement of mm. 102 105, together with the connection between the category of
 motives and various counterstatements (in mm. 41 48, 72 77, and 101 108),
suggests a way our concept of the  form of the motive might slip further still. If we
return to the transition that begins in m. 30 (ex. 4.15a), we note a similar ity between

example 4.16
mm. 88 91

Ludwig van Beethoven, String Quartet, Op. 18 No. 1, rst movement,

88

cresc.

cresc.

cresc.

cresc.

183

184

analysis and theory

the accompanimental gure in the rst violin and the  variant from mm. 17 18:
the rhythmic patter n and contour of mm. 31, 33, and 35 is the same as that of m. 18.
Our return to these measures becomes literal rather than imaginary if the Exposition is repeated: hearing mm. 30 35 after the several pair ings of the  form of the
motive with counterstatements and the extensions of the  form that occur during the rst pass through the Exposition could provide enough pressure to cause
signicant slippage within our concept of the  form. This, in turn, might lead us
to embrace the accompanimental gure of mm. 30 35 within that for m when we
encounter it for the second time.
Perhaps more remarkable than the conceptual uidity of the  form of the
motive is the resistance of the  motive to change: despite numerous juxtapositions
with rival mater ials, it remains constant. Indeed, these two categor ies of motive forms
can be further distinguished by the compositional strategies associated with each.
Over the course of the Exposition,  forms are subjected to relatively little variation,
and most members of the category are close to the conceptual model outlined in
gure 4.7a. In contrast,  forms are varied almost from the outset. While the conceptual model for the category of  forms still makes sense it captures the salient
features of members of the category as presented in the opening measures and serves
to distinguish them from members of the category of  motives the range of values that members assign to the attr ibutes picked out by the conceptual model is
much broader, and a given variant is rarely repeated. The difference between the
compositional strategies applied to each category of motive forms lends a distinctive
character to them:  as stable in the face of various challenges;  as an agent of
change and transition. Although qualities such as these are part of the elemental
materials of sonata for m, as they are manifested in the rst movement of the F major
quartet they also pose a kind of problem. The sort of uidity with which the  form
of the motive is associated operates pr incipally in contrast to the stability of the 
form. By itself, it is not quite sufcient to create the dynamic energy the sparks of
real transfor mation that the Development section will require. For this sort of
drama, a new form of the motive is required, but one that draws on the musical
materials and syntax developed over the course of the Exposition.

The Development
After a br ief A major our ish, the Development revisits  and , recalling the opening of the Exposition (as shown in ex. 4.17; the example does not include the A major
ourish).While each motive departs slightly from its respective conceptual model
the  motive includes light accompaniment in the upper str ings, and the  motive
acquires an echo for the most part the conceptual model for each category is
reafrmed. Beethoven then introduces a new form of the motive that recongures the
properties of  and  and transfor ms the possibilities for musical discourse.
The new motive form (shown with a  in ex. 4.17) is launched in m. 129
through a ser ies of imitative entrances. Although motive forms have appeared at
one-measure intervals before, here they differ only in the octave in which they are
stated. This manner of presentation makes reference to a topic to which Beethoven
often made recourse in the Development sections of his instrumental works: the

example 4.17
mm. 119 46

Ludwig van Beethoven, String Quartet, Op. 18 No. 1, rst movement,

119

125

132

(continued)

186

analysis and theory

example 4.17

(continued)

137

142

learned style that includes fugue and canon.41 This new motive form resembles a
member of the category of  motives. It is new in its isolation, harmonic stasis, and
clear assertion of the downbeat of each measure as a point of initiation (rather than
a point of articulation within a larger gesture, as it is in the second measure of a typical member of the  category of motives). However, the dissonant interval for med
by the downward leap consequent to the tur n gure (a diminished seventh Bb3
to CS3 in mm. 129 30) demands continuation beyond the downbeat, a uidity
reminiscent of the  form of the motive.42 This amalgam of features and the way
they dominate the measures that follow suggest that  constitutes a new category of
41. For a discussion of the learned style, see Ratner, Classic Music, 23 24.
42. Given both the emphasis of the large downward leap and the fact that the leap is dissonant, the
 form of the motive is more ambiguous at the hypermetric level than either the  or  forms: where
the previous for ms of the motive can be heard in ter ms of pairs of strong and weak measures, such a
grouping in the case of the  motive is much more arbitrary. This is not to say that deter mining which
measure of the  and  forms of the motive is strong or weak at any given point is a simple matter, however. Janet Levy and Roger Sessions, for instance, give interpretations of the accentuation of mm. 1 2
that are the exact opposite of one another. See Levy, Beethovens Compositional Choices, p. 10, n. 28; and
Roger Sessions, The Musical Experience of Composer, Performer, Listener (Princeton, N.J.: Princeton University Press, 1950), 13 14.

cate g ori zat i on, st rate g y, and sy ntax

figure 4.8 Conceptual model for the  form of the principal motive from Beethoven,
String Quartet, Op. 18 No. 1, rst movement, mm. 129 44

motive forms. The conceptual model for this categor y (diagrammed in g. 4.8)
consists of ve cor related conceptual elements, character ized as the things necessary
for a melodic fragment to count as an instance of this version of the motive.While
this model has much in common with those diag rammed in gure 4.7, it also
demonstrates enough differences to warrant its being used to distinguish the  form
from any of the other for ms of the motive important in the movement.
Beethoven follows mm. 129 34 with two similar sections, the rst beginning
in m. 135, the second in m. 141. Each section begins with a slightly different patter n
of entrances among the instruments, and each section prolongs a single fully diminished seventh chord: CS (mm. 129 34), FS (mm. 135 40), and B (mm.
141 46).43 Each section also includes a variant of the  fragment, based on the tur n
common to all for ms of the motive but with a sforzando on the second beat: it rst
appears in the second violin in mm. 133 34, and subsequently in the viola and cello
(mm. 139 40) and rst violin (mm. 145 46). In truth, the variant of  is difcult to
hear, for within these imitative sections it must compete with the qd gure that
comes to dominate the rst half of each measure in which the fragment appears.
Nonetheless, as can be seen in example 4.18, the fragment is gradually transfor med,
43. This is in fact the same sequence of fully diminished seventh chords initiated in mm. 14 16;
however, that sequence led not to a B chord but to a Bb pedal preparatory to the rst full cadence on
F in mm. 19 20.
In his analysis of mm. 129ff., Ratner inter prets each fully diminished seventh chord as a leading-tone
harmony with a nor mal resolution; see Ratner, The Beethoven String Quartets, 15. Indeed, on the very last
eighth-note of mm. 128 and 129 there is a D minor chord, as there is a G minor chord in mm. 139 and
140, a C minor chord in mm. 145 and 146, and F and Bb minor chords in mm. 147 50. However, I hear
these har monies as simply embellishing the fully diminished seventh chords that dominate these measures. This hear ing is not without its theoretical problems it suggests the kind of dissonant prolongation that has vexed a number of music theor ists over the past two decades but it has the advantage of
capturing the intensity and uncompromising nature of this passage.

187

188

analysis and theory

example 4.18
mm. 145 54
145

Ludwig van Beethoven, String Quartet, Op. 18 No. 1, rst movement,

150

over the course of mm. 147 51, into the  fragment as it or iginally appeared. So
transfor med, it leads in a cascading sequence to the ar rival on the dominant of F
major in m. 167. In one sense, however, this is not the  fragment as it was rst
heard in mm. 22 27, for that fragment was one-half of a two-measure motive. By
contrast, the fragment of m. 151 and following is heard as the extension, in one
voice, of the imitative entrances initiated in m. 129.
With its abrupt introduction of the lear ned style, combination of features from
both the  and  forms of the motive, and prolongation of dissonant har monies, the
 form of the principal motive provides elements of contrast and continuity with
the preceding music. And compared with the various changes visited upon  and 
in the course of the Exposition, the repetitions of mm. 129 46 both on the level
of the individual motives that make up the imitative passages and on the level of the
separate sections that prolong three different fully diminished seventh chords
make for an island of stability within the changing fortunes of the principal motive.
Nonetheless, this moment of stability provides its own disruptions, and not only
because it is based around a sequence of fully diminished seventh chords. The balance that had existed between the  and  forms of the motive one associated

cate g ori zat i on, st rate g y, and sy ntax

with a syntax of stability, the other with a syntax of uidity, and each necessary for
the other has now disappeared: what is left after  has done its work is only the
chatter ing of the  form of the motive. This, and passagework scales over a dominant pedal, is all that remains until the  form of the motive reasserts itself at the
beginning of the Recapitulation.

Recapitulation and Coda

The appearance of motive forms in the Recapitulation, which begins with a literal
return to the opening of the movement in m. 179, follows the same general outline
as that of the Exposition. Missing, however, are almost all of the members of the 
category of motives. After the initial response to the  motives (mm. 183 86, an
exact repetition of mm. 5 8),  disappears, along with other vestiges of the opening twenty-odd measures, including the sequence of fully diminished seventh
chords from mm. 14 16 that had their extension in mm. 129 46. Beethoven
replaces mm. 9 29 with a br ief episode that veers off toward Gb major, which in
its latter half makes reference to the transition that began in m. 30, and which also
absorbs the function of the statement and counterstatement of motive forms in
mm. 41 48.44 The consequence of these changes is that the  form of the motive
goes virtually unchallenged in the Recapitulation, for both the  form and the syntax of uidity with which it was associated are absent. What we are left with is a
clear impression of the stability even obstinateness of the  form of the principal motive.45
Beethoven does offer a nal challenge to this stability in the Coda, which begins
in m. 274. After a pair of short ascending passages in staccato quarter notes
(mm. 274 81) that recall the quarter notes that followed  in the Development and
a brief descending gure built from the paired sixteenth-note portion of the stylized tur n gure (mm. 282 83), the main motive reenters in the rst violin in m. 284
(see ex. 4.19). With the violas entrance at the octave below in m. 285, and continuation with the same sort of staccato quarter-notes that began the Coda, the reference is clear: Beethoven means to summon the  form of the motive, although
without its fortissimo entrances or dissonant har monies. On the one hand, the subsequent entr ies (beginning in m. 288) hint at the possibility of further development
as the motive is tossed back and forth between different voices; on the other hand,
they all serve to prolong the dominant. The possibility of development becomes a
bit more real with the move to the submediant, in m. 294, and the reprise of a variant of the  form of the motive.While the rhythmic gure is different from a typical member of the  category, the syntactic references are clear: the motive is
sequenced upward (as it was in mm. 13 18, 21 26, and 123 27), and its har monic

44. Levy notes that these changes tend to draw together various features of the original transition;
see Levy, Beethovens Compositional Choices, 79 83.
45. The impression of the solidity of the  form of the motive is given further support by
Beethovens decision to eliminate the repeat of the second section of the movement (that is, the Development and Recapitulation), one of the signicant changes he made between the rst and second versions of the quartet.

189

example 4.19
mm. 284 302

Ludwig van Beethoven, String Quartet, Op. 18 No. 1, rst movement,

284

290

cresc.

cresc.

cresc.

cresc.

297

cate g ori zat i on, st rate g y, and sy ntax

context is thoroughly uid, including four fully diminished seventh chords (in
mm. 295, 296, 298, and 299) and a cadential dominant (in mm. 300 301). Against
the resolution of the over this dominant, the  form of the motive nally enters
(in the rst violin in m. 301), and for the rst and only time in the movement, its
concluding downward leap outlines a falling 5-1, signaling an end to the peregrinations suggested in the preceding measures.
After this closing gesture, there is one more bit of dialogue between  and  (not
shown in ex. 4.19). The cello, beginning and ending on F2, repeats the variant of 
that ends with a repeated note (in mm. 302 303 and 304 305, a variant rst heard
in mm. 41 42). It is answered by a variant of  reminiscent of that just heard in
mm. 294 299 (played in thirds in the rst and second violins in mm. 306 307 and
308 309). True to for m, these  variants pull away from F major by introducing its
b6 (Db), but they are answered by a cadential gesture that returns the movement, for
once and for all, to F major.
With regard to the pr incipal motive of Beethovens F major quartet, the Coda
has a summational function that is twofold. First, the Coda revisits each of the three
categories of motive forms that appeared over the course of the movement, doing
so in reverse order of their initial appearance. Second, the Coda revisits the syntax
with which each of these categor ies was associated. The Coda begins (much as did
that of the rst movement of Mozarts Dissonance Quartet) with a reference to
the Development; the subsequent allusions to  suggest a challenge to the integrity
of the principal motive (through fragmentation) but also take place within a context of overall stability (with the bulk of the entrances linked to the dominant of F).
The  motives that follow these  forms exemplify the syntax with which  is most
often associated: although the broad range of features exhibited by members of the
category of  motives is not in evidence, the har monic uidity typical of the category is well represented by the fully diminished seventh chords prominent in the
passage. And the nal  motive, with its summoning of cadence, is the very embodiment of general, processive stability, yielded by the compositional strateg ies most
often used for the category of which it is a member.

Changes between the First and Second Versions

of the F Major Quartet
Janet Levy observed that one of the principal differences between the rst and second versions of the F major quartet is the greater concision of the second version.46
Beethoven accomplished this in a number of ways, which together resulted in the
elimination of a signicant number of statements of the  form of the principal
motive. Basic elements of the plan Beethoven followed in his revisions are evident
in the changes made to the episode involving the  motive initiated in the midst
of the second key area. The passage as or iginally wr itten is given in example 4.20.
In the fourteen measures spanned by the episode, there are seven clear statements of
the  motive. In the second version of the movement, this episode was cut back to

46. See Levy, Beethovens Compositional Choices, 24.

191

example 4.20 Ludwig van Beethoven, String Quartet, Op. 18 No. 1 (rst version), rst
movement, mm. 72 85
72

cresc.

cresc.

cresc.

cresc.

76

80

cresc.

cresc.

cresc.

cresc.
83

cate g ori zat i on, st rate g y, and sy ntax

six measures (which are given in ex. 4.15c), with only three statements of the 
form of the motive. But perhaps more important than simply reducing the appearances of the  motive, Beethoven appears to have rethought how it ts into the
overall musical fabric, as a closer inspection of his rst attempt at this episode
reveals.
The or iginal version of the passage divides into two parts, the rst in C minor
(mm. 72 79), the second in C major (mm. 80 86). In both parts, there are three
main statements of the  motive: in the C minor section these are in the second
violin, and each ends with a large ascending leap; in the C major section the statements are given in the cello and end with a repeated note. (The latter for m of the 
motive is heard in a pr ior episode, given in ex. 4.15b, which is substantially the same
in both versions of the movement.) Each of the second violins statements is
answered by a counterstatement in the rst violin. The counterstatements in
mm. 73 and 75 beg in with a tur n gure and share their rhythmic prole with the
 form of the motive. On the other hand, the counterstatement of m. 77 answers
with the same variant of the  motive that will be heard momentar ily in the cello.
The cellos three statements of this  variant are not given similar counterstatements
but are answered instead by passagework in the rst violin, which ultimately leads
out of the episode.
When revising this passage, Beethoven cut the C minor section, salvaging the
rst violins counterstatements of mm. 73 and 75. These he placed an octave lower,
but still in the rst violin (see ex. 4.15c). He omitted the counterstatement of m. 77,
changing the rst violins music so that it starts with what appears to be another
counterstatement, but which then dissolves into passagework that leads out of the
episode. The effect of these alterations is not just concision but a view of the 
motive that is changed in three ways. First, by dropping the three statements of the
motive made by the second violin in the C minor section, Beethoven eliminated a
further variant namely, one that ends with a large leap from the category of 
forms. Second, the counterstatements made by the rst violin, now closer in register to the cellos statements of the motive, make a direct if limited challenge to the
structure of the  motivic category. Third, changing the last counterstatement so
that it leads out of the episode (rather than echoing the cellos third statement of the
 motive, as the or iginal version of m. 77 would) lets the challenge stand, albeit only
for a moment or two. Thus, in the revised version of the passage Beethoven has not
only achieved great concision, but he has also accomplished two strategic goals: he
has tightened the boundar ies of the  motivic category, and he has intimated how
these boundar ies might be challenged.
A similar elimination of outright statements of the  motive and modication of
the structure of motivic categor ies occurs with the nal statement of the motive at
the end of the Exposition. In the rst version of the passage, shown in example
4.21, the musette variant of the  motive (in mm. 109 and 111) is answered by a
counterstatement with a tr ill and appogiatura (in mm. 110 and 112). When the
phrase is repeated (in mm. 113 16), the counterstatement is embellished by
repeated tur n gures, but the  motive is not. In Beethovens revisions, both the
counterstatement and the musette variant are embellished when the phrase is
repeated (mm. 105 108 of ex. 4.15d). Although this removes two explicit state-

193

194

analysis and theory

example 4.21 Ludwig van Beethoven, String Quartet, Op. 18 No. 1 (rst version), rst
movement, mm. 109 16
109

113

ments of the  motive (and, consequently, could be seen to weaken the structure
of that category), the change makes it possible to hear the rst violins counterstatements as further variants of the  motivic categor y (a possibility raised in the
earlier discussion of ex. 4.15d).
A further result of Beethovens decision to reduce the number of appearances
of  in the revised version of the movement was a clearer distinction and more
even balance between the  and  forms of the motive. In the Development of the
rst version of the movement, for instance, the two forms of the principal motive
blend into one another. As shown in example 4.22, the entrances of the  form
subsequent to the A major our ish that opens the Development are rst answered
by an  form of the motive and then by a variant of  (mm. 127 32). Appearing
above the extended tail of this answering  is another  (a quite typical member
of the categor y), which leads, after a pause, to yet one more entrance of  (itself
answered by a  fragment, as it was in m. 21 and following). Thus an episode
beginning with the  form of the motive ends with the  form of the motive,
blurring the distinction between the two. In Beethovens revised version (ex. 4.17),

example 4.22 Ludwig van Beethoven, String Quartet, Op. 18 No. 1 (rst version), rst

movement, mm. 127 43

127

131

138

196

analysis and theory

the two for ms are successive but do not overlap:  and  stand out more clearly
as separate entities.
Beethovens revisions to the Development also resulted in a shar per distinction
between the  form of the motive and the  and  forms, especially where the syntax of the motivic categor ies is concer ned. In the or iginal version, the answering
statements by the second violin (rst  in m. 128 of ex. 4.22, then  in m. 130) both
enter at the octave above the viola. They thus anticipate the imitative entrances of
 in mm. 137 and following.When these answers are removed in the second version
(cf. ex. 4.17, m. 119 and following), the entrances of  seem fresher the introduction of the lear ned style seems more of an innovation. The entr y of  in the rst
version of the movement, moreover, leads to a consonant har mony (the dominant
of G minor), and the downward leap at this point is a minor sixth (ex. 4.22,
mm. 137 38). In the revised version,  leads to a dissonant har mony (ex. 4.17,
mm. 129 30), and the downward leap is a diminished seventh. In sum, in the or iginal version of the Development  seems to grow out of a generalized idea of the
principal motive ( and  blurring together); after Beethovens revisions, the introduction of  is more striking, and it contrasts shar ply with the statements of  and
 that precede it.
Revisions to the Coda offer a nal bit of evidence that the changes Beethoven
made were with the intent of more clearly differentiating between different versions
of his principal motive. The original version of the Coda, starting at the point that
references to the pr incipal motive become explicit, is given in example 4.23. In the
rst half of this excer pt, Beethoven relies heavily on the  form of the motive, providing statements of it that beg in variously on G and C in the rst violin, and Bb
and A in the second violin. Because the latter have an ascending skip after the tur n
gure, they could also be heard as a variant of the  form of the motive, but the
quality of the  form of the motive still pervades. In m. 312, with the shift away
from the dominant to the submediant, the  form of the motive enters in ear nest,
with a clear reference to the successive entr ies between  and  over the fully
diminished seventh chords found in the Exposition. This ser ies of  motives leads
ultimately to a conventional, if deliberately under played, cadence.
In the second version of the Coda (ex. 4.19), the  motive is withheld, and the
rst references are to the  form. As noted in the discussion of example 4.19, the
syntax of the passage owes much to the uidity associated with the process of developing musical mater ials and extends to the  motive that enters in m. 294. When
the  form of the motive nally returns in m. 301, it is coincident with a perfect
authentic cadence and signals a denitive end to the possibility of further development raised by the entrances of the  motive.
The changes Beethoven made to the rst movement of the F major quartet,
then, had two important results. First, they created a more balanced relationship
between the  and  forms of the motive in the Exposition. Second, the distinctions between the three main categor ies of motive forms are more sharply drawn.
These changes make the syntax associated with each categor y somewhat more
apparent that is, we can more easily hear the role of each motive form in the discourse of the movement as a whole.

example 4.23 Ludwig van Beethoven, String Quartet, Op. 18 No. 1 (rst version), rst

movement, mm. 302 19

302

307

cresc.

cresc.

cresc.

cresc.

312

198

analysis and theory

Compositional Strategy and Multiple Levels

of Musical Syntax

As A. B. Marx noted, the principal motive of the rst movement of Beethovens F

major str ing quartet is a nearly constant presence, never far from the listeners ear.
The function of this dense network of motives, however, is not so much to lend the
movement coherence after all, Beethoven actually reduced the number of times
the  form of the motive appeared when he revised the movement but to set up
a musical discourse separate from but coordinated with the har monic and thematic
syntax of sonata for m. This discourse involves not simply a syntax specic to the
principal for m of the motive (as in the case of the rst movement of the Dissonance Quartet), but higher levels of syntax of the sort seen in the sixth quartet
from the Op. 18 set.
Beethovens compositional strategy depends on clear distinctions between different for ms of the motive. These distinctions are made not only on the basis of
audible features of the motive (of the sort captured in the conceptual models of g.
4.7) but also on the compositional strategies that lend a distinctive character to each
category of motive forms:  as stable in the face of various challenges and  as an
agent of change and transition. By the end of the Exposition, a delicate balance has
been established between the two, a balance not in evidence in the rst version of
the quartet, where the category of  motive-forms dominated.
This balance is maintained through the rst part of the Development, but it is
disrupted by the introduction of the  form of the motive and the topic of the
learned style (ex. 4.17, m. 129). It is only with the Recapitulation that the focus
comes back to the  form of the motive, where it stays (with little or no competition from ) until the Coda. The disruption created by  and the recuperation facilitated by  suggest a function for the topic of the lear ned style within Beethovens
compositional practice: the fugato on the  form of the motive creates the impression of development by temporar ily suspending the nor mative syntax for the projection of tonal areas associated with  and , a syntax restored only with the
reprise of  at the beginning of the Recapitulation. This function of  is in evidence again at the end of the movement when, after hints at further development
and har monic uidity in the Coda,  returns as the melodic voice of the nal
cadence in F major.
The distinctions among these three forms of the principal motive distinctions
that make possible a musical discourse that is focused on the changing fortunes of
the motive over the course of the movement, and that were made more apparent
in Beethovens revisions to the movement suggest that motivic analysis could be
rened by attending to how motives are used, rather than to how successfully they
can be traced back to some Ur-type. Such a renement makes possible an appreciation of multiple levels of syntax levels that interact with the overall har monic
and thematic syntax of the musical work, and which collectively engage more and
less immediate levels of cognitive processing. Ultimately, an appreciation of this
renement leads to a better understanding of how compositional strateg ies shape
our conceptions of musical organization.

cate g ori zat i on, st rate g y, and sy ntax

conclusions

For Umberto Eco, one of the mysteries of semiotics was that music could have a
fairly well developed syntax and yet no apparent semantic depth. Music meant
something, but outside of conventionalized signs like those associated with military
or pastoral music, it was quite difcult to say what this was. In this chapter, I have
not tr ied to address this problem head on; indeed, describing the or igins and scope
of musical meaning is an endeavor far more ambitious than what I undertake here.
I have, however, tried to argue that musical syntax and semantics are better viewed
as points along a continuum than as discrete domains. More directly, composers (as
well as others who br ing music to life) make use of strategies that disrupt and
redene musical syntax as part of a process of meaning construction.
My path into these compositional strategies has been through categor ization, on
the assumption that categor ies of musical events represent a common ground where
the concer ns of composers and listeners meet. The viability of this assumption was
demonstrated by the structure of the waltzes produced by the musical dice game.
Where formal structure needed to be clearly dened (for instance, at the end of
each half of the waltz), the range of category members within each measurecategory was quite nar row, and what was typical of the relevant measure-category
was quite apparent.Where formal structure could be more open (for example, at the
beginning of the second half of the waltz), the range of category members was quite
wide and typicality harder to deter mine. Typicality could thus be used to dene formal syntax, a syntax that not incidentally coordinated with and thus conr med
harmonic syntax.
A somewhat different example of a compositional strategy organized around
typicality was seen in the category for the pr incipal motive of the rst movement of
Mozarts Dissonance Quartet. At important moments in the movement at the
beginning and close of the Exposition and Recapitulation, and in the Coda either
thoroughly or highly typical for ms of the motive made their appearance. These
were followed by less typical members of the motivic categor y, leading either to a
reprise of a typical for m or to new musical mater ial. This basic process a syntax of
typicality yielding to atypicality was repeated three times in the course of the
Development, ending only with the appearance of a thoroughly typical for m that
signaled the start of the Recapitulation. While the compositional strateg ies that
shaped the different versions of the principal motive ran parallel to the for mal syntax of the movement, they were nonetheless independent. The motive thus emerged
as a topic for musical discourse, whose progress could be followed along with the
other strands of musical discourse compr ised by the movement.
By contrast, the principal motive of the rst movement of the Bb major quartet
from Beethovens Op. 18 never functioned as a topic for musical discourse. Although the motive forms fell into two distinct categor ies, individual members never
strayed far from their respective conceptual models. In consequence, the immediate level of syntax for the categor ies of motive forms was not particularly interesting, since members were almost always quite typical of their categor y. However,
over the course of the Exposition each category acquired distinct functions: one ()

199

200

analysis and theory

signaled the opening of musical discourse; the other () signaled the close of discourse. This emergent syntax was then exploited in the Development and more
generally served to show how musical topics of the sort descr ibed by Ratner could
come into being.
It was in Beethovens F major quartet from Op. 18 that we saw both of these levels of syntax come together. As with the Bb major quartet, Beethoven introduced
two categor ies of motive forms ( and ) in the Exposition, with a third () playing a decidedly subsidiary role. A compositional strategy focused on typicality was
applied to both categor ies, but where the range of variation for  was relatively narrow, for  it was quite wide. Contextual pressures on  even suggested that the concept of what counted as a member of the category could slip to include a number
of rather remote variants, a kind of slippage that never seemed to threaten . By the
end of the Exposition, and even into the beg inning of the Development,  and 
appeared to maintain a delicate balance, the for mer stable in the face of various
challenges, the latter an agent of change and transition. Beethoven then introduced
a new form of the motive () to disrupt this balance, creating a species of development of high drama:  and  were swept from the stage, vanquished by a volatile
combination of properties drawn from each. Balance within the movement as a
whole (but not between  and ) was restored with the return of  at the start of
the Recapitulation. Even the developmental urges of the Coda could not threaten
this equilibr ium. The compositional strategies applied to the pr incipal motive, then,
operate on both a local level (within a g iven category) and global level (among
groups of categor ies). Again, these syntactic levels are independent of but coordinated with those for the overall har monic and thematic structure of the movement the syntactic strands associated with the pr incipal motive are multiple, but
they do not exhaust the syntax of the whole.
In sum, these analyses suggest that the construction of meaning in music can be
achieved through the way composers choose to deploy the elements of musical syntax. Of course, compositional strategy is not the only source of meaning construction in music I have also pointed out the contr ibutions of musical topics and
cross-domain mapping to musical semantics but it is one to which our knowledge of categorization can be protably applied. As the meeting place for the concerns of composers and listeners, categor ies of musical events are important to both
compositional strategy and musical syntax, for they represent a means through
which uniquely musical meaning can be created.

chapter five

cultural knowledge and

musical ontology

s part of a search for an inter pretive theory of culture and consequent to a cr itique of early work in cognitive anthropology, Clifford Geertz momentar ily
and somewhat famously tur ned his thoughts to music as an example of culture.
Taking a Beethoven str ing quartet as his case in point, he suggested that no one
would identify the quartet with its score, with the skills and knowledge needed to
play it, with the understanding of it possessed by its perfor mers or audience, with a
particular perfor mance of it, or with some mysterious entity transcending mater ial
existence. There was one perspective, however, that Geertz thought would meet
with general agreement: That a Beethoven quartet is a temporally developed tonal
structure, a coherent sequence of modeled sound in a word, music and not
anybodys knowledge of or belief about anything, including how to play it, is a
proposition to which most people are, upon reection, likely to assent.1
By this example, Geertz intended to put to rest, for once and for all, the notion
that culture exists as some sort of program inside peoples heads. Nonetheless, if
music is not knowledge or belief, then what is it? Geertzs denition of music as a
coherent sequence of modeled sound is simply too broad, for it would have to
include str ings of Morse code, speech in an unknown tongue, and the cr ies of various animals. Indeed, from the perspective of this denition, there is, in principle,
nothing to distinguish music from noise or noise from music.2 Music is a category
constructed by humans, not a substance or set of relations oating free in the world.

1. Clifford Geertz, Thick Descr iption: Toward an Inter pretive Theory of Culture, in The Interpretation of Cultures: Selected Essays (New York: Basic Books, 1973), 11 12.
2. Much can and has been said regarding the relationship between music and noise. Three sources
from the last ninety years represent the range of issues engaged by the topic: Luigi Russolo, The Art of
Noises (1916), trans. Barclay Brown (New York: Pendragon Press, 1986); Jacques Attali, Noise: The Political Economy of Music, trans. Brian Massumi, with a foreword by Fredric Jameson and an afterword by
Susan McClar y, Theory and Histor y of Literature, 16 (Minneapolis: University of Minnesota Press,
1985); and the improvised performances by Derek Bailey, Pat Metheny, Gregg Bendian, and Paul Wertico
on The Sign of 4 (Knitting Factory Works, [CD] KFW 197, 1997).
I indicate the media of the recordings I cite with the following abbreviations: CD = compact disk;
LP = 33 RPM record; 78 = 78 RPM record.

201

202

analysis and theory

As such, the notion of music is dependent upon and constrained by human knowledge, which presumably resides with individual humans. Having said this, it must
also be said that the continued existence of a Beethoven str ing quartet does not
depend on the personal knowledge of any one individual. Music is indeed a cultural
fact, and culture is largely public and inter personal.
Cultural knowledge, as represented by Geertzs example, is thus a bit of a paradox. On the one hand, it is not something that is deter mined by any one individual:
culture is, as Geertz maintained, public. On the other hand, individuals need such
knowledge if they are going to participate in culture, and it is through the participation of individual humans that culture and cultural knowledge is maintained.
The problem, as framed by Geertzs example of a Beethoven quartet, is more
specically one of musical ontology that is, the ontological status of a work of
music. In the West, it is common to regard pieces of music as having an objective
status, speaking of a work of music as one might of a painting or sculpture. This
habit of thought persists in spite of the transience of musical phenomena: unlike a
painting or sculpture, a musical work has no endur ing mater ial existence. The
enigma of works that are of profound cultural importance and yet are totally
ephemeral has led to a number of attempts by philosophers and others to account
for the objective status of the work of music and, by extension, music itself; thus the
problem of musical ontology.3 A contrasting perspective has developed among ethnomusicologists who, working chiey with non-Western music, have noted the uidity of musical practice and a concomitant absence of any ascription of permanence
to the products of musical activity. These observations render doubtful any universal solution to the problem of the ontology of the musical work, or even music,
since many cultures do not show linguistic or practical evidence of equivalent concepts.4 And yet, musical practice within these cultures, while uid, is certainly not
3. Thoughtful discussions of the problem of musical ontology can be found in Lydia Goehr, The
Imaginary Museum of Musical Works: An Essay in the Philosophy of Music (Oxford: Clarendon Press, 1992);
Philip V. Bohlman, Ontologies of Music, in Rethinking Music, ed. Nicholas Cook and Mark Everist
(Oxford: Oxford University Press, 1999), 17 34; and Nicholas Cook, At the Borders of Musical Identity: Schenker, Corelli and the Graces, Music Analysis 18 (1999): 179 233. See also the essays collected
in Michael Talbot, ed., The Musical Work: Reality or Invention? (Liverpool: University of Liverpool, 2000).
4. Examples of cultures the languages of which lack a word equivalent to music or concepts
equivalent to that of a musical work include the Hausa of Nigeria, the Macuma Shuar of Ecuador, and
the Mapuche of Argentina; Charles Keil reports such a lack in over twenty languages of the African continent and gives detailed discussion of the methodological problems involved in studying such cultures.
See David W. Ames and Anthony V. King, Glossary of Hausa Music and Its Social Contexts (Evanston, Ill.:
Northwestern University Press, 1971), ix;William Belzner, Music, Modernization, and Wester nization
among the Macuma Shuar, in Cultural Transformations and Ethnicity in Modern Ecuador, ed. Norman E.
Whitten, Jr. (Urbana: University of Illinois Press, 1981), 735 36; Carol E. Robertson-DeCarbo, Tayil
as Category and Communication among the Argentine Mapuche: A Methodological Suggestion, Yearbook of the International Folk Music Council 8 (1976): 39; and Charles Keil, Tiv Song (Chicago: University
of Chicago Press, 1979), 27 52.
I should make clear that what I nd dubious is a universal solution to the problem of the ontology
of a musical work. I do not think it is impossible, however, to speak about music in the case of a culture that does not show linguistic or practical evidence of such a concept, as long as one does not attempt
to colonize that culture by insisting that its members, when they try to make sense of their own experiences and cultural practices, make use of the term.

cultural knowle dge and musical ontology

unregulated. In general, the habit of objectifying aspects of musical practice is
omnipresent; in consequence, the problem of musical ontology, while most properly
one that ar ises only among those who believe in an objective status for the work
of music, also has applications to situations where no such object is posited.
In this chapter, I propose that solutions to the problems of cultural knowledge in
general, and of musical ontology in particular, can be found in what we now know
about processes of categorization. In brief, knowing that Beethovens quartet is
music or that anything else is music, for that matter means knowing how to
categorize sequences of sound-events in accordance with conceptual models shared
with other members of a culture. More specically, each work of music or
whatever other unit of cultural cur rency we choose to focus on constitutes a
Type 1 category of the sort discussed in chapter 2.5 On the one hand, the transparency and immediacy of the conceptual models that guide the process of categorization explain why we describe works of music in the same language we use for
objects: we speak of the Beethoven quartet just as we speak of the red cup, even
though the for mer (as sounding music) is a phenomenon completely unlike the latter.6 On the other hand, because membership in Type 1 categor ies is by degree, a
certain amount of uidity is possible, varying in degree with the cultural context
relative to which the conceptual model for the categor y is framed. We might
include a wide variety of performances in the categor y for the Beethoven quartet some more, some less competent, even some in an ar rangement for four-hand
piano but we could well draw the line at a version for pitched percussion and
automobile hor ns.
In what follows, I explore the idea that musical works can be thought of as Type
1 categor ies and that such categor ies are a manifestation of cultural knowledge. I
approach this issue by taking a close look at two popular songs from the early twentieth century: I Got Rhythm and Bye Bye Blackbird. Songs like this challenge
our ideas about what constitutes a work of music: while they are almost always
associated with a score (as is a Beethoven quartet), they are also part of a performance tradition that involves some level of improvisation (introducing the uidity
typical of many non-Western perfor mance traditions).
I Got Rhythm provides a good example of this situation, for there were multiple competing versions of the song almost from the moment it appeared on
Broadway late in 1930. In the rst section below, I consider a number of recordings
5. Of course, a work of music could be regarded as a Type 2 category were necessary and sufcient
conditions for categor y membership specied. This is how I would inter pret Nelson Goodmans proposal that the score is the absolute deter minant of musical identity. Goodman wr ites, If we allow the
least deviation [from the score], all assurance of work-preservation and score-preservation is lost; for by
a ser ies of one-note er rors of omission, addition, and modication, we can go all the way from
Beethovens Fifth Symphony to Three Blind Mice. Thus while a score may leave unspecied many features of a performance, and allow for considerable variation in others within prescr ibed limits, full compliance with the specications g iven is categor ically required. Nelson Goodman, Languages of Art: An
Approach to a Theory of Symbols, 2nd ed. (Indianapolis: Hackett Publishing, 1976), 186 87.
6. There are, of course, aesthetic perspectives that support this habit (such as that associated with the
work-concept, as descr ibed by Lydia Goehr in The Imaginary Museum of Musical Works, 89 119) and
technological innovations that give it credence (like music notation and sound recording), but these are
not preconditions for the notion of a musical object.

203

204

analysis and theory

of the song made dur ing the 1930s and 1940s, explore how the model for the song
changed and how it stayed the same, and show that both the uidity and the stability of this model are straightforward consequences of our use of categor ization to
organize our understanding of music.
Bye Bye Blackbird presents a somewhat different problem, which takes us
deeper into the relationship between categor ization and cultural knowledge.Written as a popular song dur ing the 1920s and quite successful when it rst appeared,
it never really entered the standard repertoire until a recording by Miles Davis in
1956 transfor med the song into a jazz tune.7 Because there was no tradition of gradual change, it was possible for radically different models for the song to coexist and
come into conict. In the second section below, I consider this sort of conict in
light of what Mikhail Bakhtin called double-voiced discourse and what Henr y
Louis Gates Jr. calls Signifyin(g). I attempt to show that a fuller account of how we
categorize musical works requires that we go outside of the relatively parochial concerns of musical perfor mance and consider the context in which musical performance takes place.
In the third section, I offer some concluding thoughts on cultural knowledge,
musical ontology, and the cognitive processes through which we structure our
understanding of the world. I end the chapter with some reections about how the
perspective on musical ontology I offer here can change the way we view music
theory.

categorization, musical ontology, and

cultural knowledge
A Multiplicity of Rhythms, 1930 1946
I Got Rhythm was wr itten by George and Ira Gershwin for their musical Girl
Crazy, which opened on Broadway on 14 October 1930. The song, intended to be
a show stopper, appeared near the end of the rst act, and its refrain was repeated a
number of times: to accompany a tap dance chorus after the nal vocal refrain; as an
encore for Ethel Mer man, whose success with the song became legendar y; as part
of the mostly instrumental entracte before the second act; and, sung by the whole
company, as the shows concluding number.
The song itself, the score for which is given in example 5.1, comprises an introduction, verse, and refrain. The melody of the verse begins with an eight-measure
phrase in G minor (featur ing a prominent b5 at its midpoint), which is given a var-

7. I borrow the distinction between a song and a tune from Richard Crawford, who wr ites, in the
jazz tradition, we usually speak of tunes, not songs. A jazz tune is dened rst and foremost by its structure: by the patter n of repetition and contrast in its melodic phrases and the har monic framework underlying them. Second, it is dened by its ethos: by the mood it projects and the tempo at which it is played.
Only third does its melody come into play, for in the jazz tradition the melody is often little more than
an entre into the perfor mance; after being heard, it is usually discarded for free melodic invention by the
performers. Richard Crawford, George Gershwins I Got Rhythm (1930), in The American Musical
Landscape (Berkeley: University of Califor nia Press, 1993), 221, 225.

cultural knowle dge and musical ontology

example 5.1 Version of George and Ira Gershwins I Got Rhythm used in act 1 of
Girl Crazy. I Got Rhythm, words and music by George Gershwin and Ira Gershwin,
1930 WB Music Cor p. (Renewed). All rights reserved. Used by permission. Warner Bros.
Publications U.S. Inc., Miami, FL 33014.

(continued)

ied repetition. This is followed by a four-measure phrase, which is also repeated

(again, with variations), after which a two-measure transition leads to the refrain.
The refrain is in the AABA form typical of popular songs of this period, but it
includes an ascent to F5 near the end of the last A section. This climactic r ise
requires a two-measure tag to br ing the melody back to Bb4 and yields a 34measure AABA' form.
I Got Rhythm was only rarely recorded as it appeared in the or iginal show.
Although it contr ibuted signicantly to her professional career, Ethel Mer man did
not record it until 1947. Many amateurs and semi-professional musicians must have
performed the song from the sheet music, which did reproduce the music as it

205

206

analysis and theory

example 5.1

(continued)

appeared in Girl Crazy. This was not the case, however, with professional musicians.
Even in the earliest recordings (made within a few weeks of the shows opening),
the verse loses its introductory (and nar rative) function and is used instead as an
interlude between repetitions of the refrain. For example, the recording made on 23
October 1930 by Red Nichols and His Five Pennies (a group drawn from the pit
orchestra for Girl Crazy) begins with an instrumental version of the or iginal refrain,

example 5.1

(continued)

(continued)

208

analysis and theory

example 5.1

(continued)

followed by the refrain with vocal.8 Only after this does the music for the verse
enter, but without vocals. The recording concludes with two more statements of
the refrain, also without vocals. A very similar ar rangement was used by Fred Rich
and a studio band in their recording of 29 October 1930. The overall for m of their
rendition consists of three refrains (the second with vocal), followed by an instrumental verse, followed by one complete and one partial statement of the refrain (the
latter using only the B and A' sections to conclude the perfor mance).9
In the two years following the opening of the show, almost all of the recordings
that appeared made use of vocals, although none of which I am aware used the
words for the verse.10 With the instrumental version recorded by Don Redman and
8. Red Nichols and His Five Pennies, I Got Rhythm, in Rarest Brunswick Masters, 1926 1931: Red
Nichols and His Five Pennies (MCA Records, [LP] MCA 1518, 1982).
9. Fred Rich and His Orchestra, I Got Rhythm, in 1930 Youre Driving Me Crazy: Portrait of a
Year in Music (Phontastic, [CD] PHONT CD 7618, 1992).
10. A fairly comprehensive list of recordings of I Got Rhythm from 1930 to 1942, based on Br ian
Rusts The American Dance Band Discography 19171942 (New Rochelle, N.Y.: Arlington House, 1975), is
included in Crawfords George Gershwins I Got Rhythm, 222 25.

cultural knowle dge and musical ontology

His Orchestra on 30 June 1932, however, we encounter a new approach to the
song.11 In this ar rangement, the verse was eliminated completely, as was the role of
the vocalist. The result is something very much like the version of I Got Rhythm
that can be found in jazz fake books, such as shown in example 5.2.12 After 1932,
the number of recordings that include vocals drops off precipitously, and I Got
Rhythm became a tune used by dance bands for instrumental improvisation a
jazz standard as may be observed in recordings by Red Norvo and His Swing
Sextette in 1936 or Chick Webb and His Little Chicks in 1937.13 Seen from the
perspective of these recordings, George Gershwins own perfor mance of I Got
Rhythm as a set of variations for piano and orchestra is somewhat anomalous, for
there is little if any sense of improvisation, nor is this music intended for dancing.14
From 1932 to 1942, recorded perfor mances are remarkably consistent in their
approach to I Got Rhythm: all use the same thirty-four-measure form, all make
example 5.2 Jazz fake book version of George and Ira Gershwins I Got Rhythm.
I Got Rhythm, words and music by George Gershwin and Ira Gershwin, 1930 WB
Music Cor p. (Renewed). All rights reserved. Used by permission. Warner Bros. Publications
U.S. Inc., Miami, FL 33014.

11. Don Redman and His Orchestra, I Got Rhythm, in Don Redman and His Orchestra, 19311933
(Classics, [CD] 543, 1990).
12. Jazz fake books typically contain hundreds of songs. Generally, as is true of the piece in ex. 5.2,
each is given as a melody, along with chord symbols that outline a basic har monization and some minimal instructions for perfor mance; the words for the songs may or may not be included. These mater ials allow players familiar with jazz perfor mance practice to give an acceptable rendering of a song even
if they have never played it before that is, they can fake their way through a perfor mance.
13. Red Norvo and His Swing Sextette, I Got Rhythm, in Red Norvo (Time-Life Records, [LP]
STL-J14, 1980), originally recorded 16 March 1936; Chick Webb and His Little Chicks, I Got Rhythm,
in Spinnin the Webb: Chick Webb and His Orchestra (GRP Records, [CD] 513678L, 1994), originally
recorded 21 September 1937.
14. George Gershwin, I Got Rhythm, in Gershwin Performs Gershwin: Rare Recordings 19311935
(Musical Her itage Society, [CD] 512923A, 1991), originally recorded 19 February 1934.

209

210

analysis and theory

some reference to the or iginal melody, and all preserve the affect and tempo of the
original. However, an ar rangement recorded by the Jimmy Dorsey Orchestra on
28 November 1942 signaled yet another approach to I Got Rhythm, for after
employing the customar y thirty-four-measure format for the head, the band
switched to a thirty-two-measure format for the solos.15 In truth, this was not the
rst time the musical mater ials of I Got Rhythm had been associated with a
thirty-two-measure form. Early on, the song had been used as the basis for a number of other compositions.16 These der ived tunes typically retain the har monic
framework, overall for m, and affect of the or iginal song but substitute a new
melody and g ive the whole a new title. In the case of I Got Rhythm, derived
tunes allowed jazz players to keep what was useful from the song its fast tempo,
lively affect, and straightforward harmonic structure yet escape the limitations of
its simple, predominantly pentatonic melody. The earliest recorded tune der ived
from I Got Rhythm, Sidney Bechet and the New Orleans FeetwarmersShag
(recorded 15 September 1932), did retain the thirty-four-measure form of I Got
Rhythm, with its two-measure tag at the end of the nal A section.17 However,
Fletcher Hendersons Yeah Man, which appeared the following year, did away
with the two-measure tag, yielding a thirty-two-measure form in which all of the
A sections were of equal length.18 There were two advantages to this change. First,
the excision of the two-measure tag made the connection between Yeah Man
and I Got Rhythm more obscure: Yeah Man could sound a bit more like a
brand new tune. Second, jazz practice of this per iod was beginning to embrace
longer solos. The 34-measure for m of I Got Rhythm, it will be recalled, ends
with a two-measure tag added to its nal A section. A musician who wanted to
take two choruses in succession would consequently nd the regular ow of
eight-measure phrases in the backg round har monic structure inter rupted in midstream with the appearance of the ten-measure A' section. When a musician took
two choruses on the 32-measure form, however, the succession of eight-measure
phrases was constant: the improvisation could stretch seamlessly across the two
refrains, allowing the player to build longer melodic lines and to shape the dynamic
structure of the solo over larger spans of music. Although after Yeah Man the
tunes der ived from I Got Rhythm were invariably based on the thir ty-two15. Jimmy Dorsey and His Orchestra, I Got Rhythm, in Jimmy Dorsey and His Orchestra,Vol. 2:
194244 (Hindsight Records, [LP] HSR-153, 1980). For perfor mances with an emphasis on improvisation that are guided only by a relatively informal ar rangement, the practice among jazz perfor mers is to
state the tune at the beg inning and end of the perfor mance. In these cases, the tune functions as the
head. The head is typically followed by solos over the har monic framework of the tune, which are, in
turn, followed by a concluding statement of the tune.
16. Richard Crawford calls such der ived tunes contrafacts, borrowing the ter m from one of several medieval practices to which it is linked. The specic practice invoked is one in which new texts are
grafted on to preexistent music, which was itself freely adapted. See Crawford, George Gershwins I
Got Rhythm, 225, 228 29.
17. Sidney Bechet and the New Orleans Feetwarmers,Shag, in Sidney Bechet (Time-Life Records,
[LP] STL-J09, 1980).
18. Fletcher Henderson and His Orchestra, Yeah Man, in Fletcher Henderson, Developing an American Orchestra, 19231937 (Smithsonian Collection, [LP] R 006, 1977), originally recorded 18 August
1933.

cultural knowle dge and musical ontology

measure form, it was only around 1942 that this form was brought to I Got Rhythm
itself.
At rst, the thirty-two-measure form was used only for solos the head
retained its thirty-four-measure form, complete with the two-measure tag (as in the
Jimmy Dorsey arrangement). By the mid-1940s, however, even the tag for the head
had fallen away.When Charlie Parker recorded I Got Rhythm with a group of
all-star perfor mers on 22 April 1946, the thirty-two-measure form was used for the
head, as well as for the twenty-three choruses that followed.19 The transfor mation
of I Got Rhythm into a platfor m for improvisation is even more evident in a
recording made two years earlier by Lester Young and the Kansas City Six. In this
performance, not only is the thirty-two-measure harmonic patter n used throughout, but also the melody of the song is never stated.20 In this for m that is, as a
thirty-two-measure, AABA harmonic structureI Got Rhythm came to be one
of the basic elements of the jazz players musical vocabulary: the Rhythm changes
shown in example 5.3.21
example 5.3 Scheme for Rhythm changes (based on the har monic structure of the
thirty-two-measure for m of I Got Rhythm). I Got Rhythm, words and music by
George Gershwin and Ira Gershwin, 1930 WB Music Cor p. (Renewed). All rights
reserved. Used by permission. Warner Bros. Publications U.S. Inc., Miami, FL 33014.
1

bmaj7 7
B
G

m7

bmaj7 7
B
G

m7

bmaj7 7
B
G

m7

b
F m7 B 7

m7

b
F m7 B 7

bmaj7

b
E7 B maj7 G7

bmaj7

b
E7 B maj7

m7

bmaj7 7
B
G

bmaj7
B

17

25

bmaj7 7
B
G

m7

bmaj7 7
B
G

m7

b
F m7 B 7

bmaj7

b
b
E7 B maj7 F7 B maj7

19. Charlie Parker and others, I Got Rhythm, in Bird: The Complete Charlie Parker on Verve
(Verve/Polygram Classics, [CD] 837 142 2, 1988).
20. Lester Young with the Kansas City Six, I Got Rhythm, in The Tenor Sax: Lester Young, Chu
Berry and Ben Webster: The Commodore Years (Atlantic, [LP] SD 2-307, 1973), originally recorded 28
March 1944. That the perfor mance by Youngs group resulted from an ar rangement employed for successive performances rather than happenstance is suggested by an alter nate take of the same tune
recorded the same day (included on this album), which uses precisely the same ar rangement.
21. The ter m changes refers to a conceptual structure that consists of a succession of harmonies
correlated with a metr ic and hypermetric framework. Knowing this structure knowing where and
when the har monies change is essential for creating convincing improvisations.
On the importance of Rhythm changes as a basic har monic prototype, see Paul F. Berliner, Thinking in Jazz: The Innite Art of Improvisation, Chicago Studies in Ethnomusicology (Chicago: University of
Chicago Press, 1994), 76 78.

211

212

analysis and theory

From the evidence provided by recordings, then, what counted as I Got Rhythm
changed from 1930 to 1946 as the song was transfor med from a Broadway or popular song, to a jazz standard, to a platfor m for improvisation. This prcis is not quite
accurate, however, for even as new conceptions of the song emerged, older ones
remained in circulation. Benny Goodman, for instance, continued to use the thirtyfour-measure form of the refrain even after others had adopted the thirty-twomeasure form;22 productions of the musical (including the 1943 movie Girl Crazy,
starring Mickey Rooney and Judy Garland) continued to use the entire song after
most other perfor mances had omitted the verse.23 From almost the moment of its
introduction, there were multiple Rhythms in circulation, and what belonged
under the rubric I Got Rhythm depended on where you stood in the stream of
American popular culture.

Categorizing I Got Rhythm

Collected in table 5.1 are all of the versions of I Got Rhythm discussed thus far.
How might someone go about categorizing these various versions were they encountered in the course of daily life rather than as a pre-selected list? That is, assuming one
were not so fortunate as to have each perfor mance identied as a perfor mance [or
recording, or rendering] of I Got Rhythm, what would be the basis for including
any or all of these in the category performances of I Got Rhythm? As shown in chapter 1, categories like this are organized around a conceptual model that sets up an idealized representation of what counts as a typical member of the category.24
As a place to begin, let us consider a conceptual model focused on the attr ibutes
of the song as it rst appeared in Girl Crazy (leaving aside the variants created as the
song was reprised over the course of the musical). The model (which I shall call the
pop-music model, for reasons that will become clear below) is diagrammed in
gure 5.1 and consists of ve cor related conceptual elements, character ized as the
things necessar y for an adequate perfor mance of the song. Two aspects of this
model should be clar ied. First, although the verses of popular songs were often
omitted in perfor mance, I take the fact that composers still bothered to wr ite verses
and that the verse was included, in some for m or another, in early recordings as evidence of its importance to the identity of the song. Second, my distinction between
variation and improvisation is based on the extent rather than the presence of musical mater ials not specied by the score: if these mater ials are relatively limited, I
would tend to regard them as a product of variation; if they are extensive, I would
tend to regard them as a product of improvisation. Where to draw the line involves
another set of conceptual models for perfor mance practice; nonetheless, the distinction is relatively clear in early recordings.

22. Benny Goodman Sextet, I Got Rhythm, in Benny Goodman (CBS Records, [LP] P5 15536 CP,
1981), originally recorded 18 September 1945.
23. Although the plot line for the movie is completely different from that of the musical, the musical numbers are the same.
24. By categor ies like this I mean not categor ies involving musical works specically, but the categories we use in daily life generally.

cultural knowle dge and musical ontology

table 5.1. Select Performances of I Got Rhythm
Performers

Date of Recording

Red Nichols and His Five Pennies

Fred Rich and His Orchestra
Don Redman and His Orchestra
George Gershwin
Red Norvo and His Swing Sextette
Chick Webb and His Little Chicks
Jimmy Dorsey and His Orchestra
Lester Young with the Kansas City Six
The Benny Goodman Sextet
Charlie Parker and others

23 October 1930
29 October 1930
30 June 1932
19 February 1934
16 March 1936
21 September 1937
28 November 1942
28 March 1944
8 September 1945
22 April 1946

Applying this model to the perfor mances listed in table 5.1 yields the typicality
effects character istic of Type 1 categor ies. Most typical would be the perfor mances
by Red Nicholss and Fred Richs groups; performances from the or iginal production of the musical (which are not included in the list of recordings) would of
course also be reckoned as highly typical.25 By contrast, the perfor mance led by
Gershwin, which drops the verse and lyr ics, would not confor m as well with the
model. Even less typical would be the perfor mances led by Redman, Norvo,Webb,
and Goodman, for all these add a substantial emphasis on improvisation. Finally,
on the fr inges of the category would be the perfor mances led by Dorsey, Young,
and Parker, for these not only omit the verse and lyr ics but also use a thirty-twomeasure form of the refrain as a platfor m for extended improvisations.
As we see in this example, conceiving of various renditions of a song as a Type
1 category organized around a conceptual model offers a way to explain the peculiar ontological status of musical works. Our sense that there is only one thing that
should be called I Got Rhythm is supported by the idealization basic to the
model and to the models role as the conceptual anchor for the categor y. At the
same time, our ability to accept as members of the category things that do not fully
correspond to this model shows how the model functions as a guide to understanding the structure of the category as a whole. Although Charlie Parkers I Got
Rhythm is not like Ethel Mer mans I Got Rhythm, it shares enough attr ibutes
with the latter that we will not want to exclude it from the category outr ight. Further, the model can provide a basis for descr ibing why we would grant Parkers performance a liminal rather than a central status within the category we might say
something along the lines of He uses too much of that crazy improvisation, so he
really isnt playing I Got Rhythm.
25. There was no recording of the original production of Girl Crazy. However, there was a painstaking recreation of the 1930 production undertaken by Elektra Nonesuch in 1990, and Lorna Lufts performance there (in the part of Kate Fothergill or iginated by Merman) agrees in most respects with the
version of I Got Rhythm recorded by Merman in 1947. See George Gershwin and Ira Gershwin, Girl
Crazy (Elektra Nonesuch, [CD] 79250-2, 1990), recorded 26 28 February 1990.

213

214

analysis and theory

figure 5.1

Pop-music model for the category performances of I Got Rhythm

Useful as this approach to musical ontology is, it demands at least a bit of further elaboration: we need to nd a way to descr ibe how the conceptual model for
a musical work changes over time. After all, while Parkers 1946 perfor mance might
have been regarded as unusual by some, it is clear from the musical practice of his
contemporar ies that it was far from exceptional. Here Richard Crawfords account
of three different approaches taken to I Got Rhythm during the years 1930 1942
provides a helpful guide.26 The rst approach, in which I Got Rhythm was
treated as a popular song, is similar to that assumed by the pop-music model diagrammed in gure 5.1. The second approach descr ibed by Crawford is one in
which I Got Rhythm is treated as a jazz standard. A conceptual model that ts
this approach is diagrammed in gure 5.2 and consists of ve things necessary for an
adequate perfor mance.27 Adopting this model would, of course, change what
counted as a typical version of the tune. Now the perfor mances led by Redman,
Norvo, Webb, and Goodman would be central to the categor y, and the performances led by Dorsey, Young, and Parker only a bit less typical. Far more atypical
would be the perfor mance led by Gershwin, with its relative lack of emphasis on
improvisation, and the perfor mances by Red Nicholss and Fred Richs groups, with
their inclusion of the verse.
The third approach to I Got Rhythm noted by Crawford was as a musical
structure that served as the basis for various der ived tunes. I would like to generalize this somewhat to accommodate the slightly expanded time frame represented by
the recordings collected in table 5.1. After 1942, I Got Rhythm served not only
as the basis for der ived tunes but also as a framework for improvised perfor mance

26. Crawford, George Gershwins I Got Rhythm, 218.

27. Explicit mention of the melody in this conceptual model is intended as a hedge against confusing I Got Rhythm with any of the tunes der ived from it.

cultural knowle dge and musical ontology

figure 5.2

Jazz-standard model for the category performances of I Got Rhythm

a framework embodied in the Rhythm changes of example 5.3. This leads us to a

third model for I Got Rhythm, which is diagrammed in gure 5.3. This improv
model, conceived as a skeletal vehicle for improvisation and composition, consists of
four things necessary for an adequate perfor mance. Once again, what would count
as a typical version of the tune changes. From the perspective provided by the
improv model, the perfor mances led by Young and Parker would be most typical,
with that led by Dorsey following close behind. The perfor mances led by Redman,
Norvo,Webb, and Goodman, with their allegiance to the or iginal for m and melody
of Gershwins song, would now be regarded as less typical, and the perfor mances
led by Gershwin, Nichols, and Rich would have to be considered as thoroughly
atypical.
Let us now return to the matter of how the conceptual model for a musical
work changes over time. If we take these three models as an histor ical sequence
the pop-music model followed by the jazz-standard model, which is followed by
the improv model we can see that each successive conceptual structure retains
some of the elements of the previous structure and either modies or drops others.
The jazz-standard model retains from the pop-music model the elements
concer ning affect and the mater ials of the refrain. It changes those relating
to function and lyr ics, and it adds an element specifying the presence of the
melody. (The latter is implicit in the materials of the refrain inherited
from the pop-music model, but it is made explicit here).
The improv model retains from its jazz-standard counter part the element
concer ning affect. It changes the components of the model associated with
the refrain, melody, and function, and it drops any requirement for the lyr ics.
As is evident from table 5.1, these are not the only models possible, nor should my
descriptions of the model be taken as comprehensive or nal. One could easily

215

216

analysis and theory

figure 5.3

Improv model for the category performances of I Got Rhythm

imagine idiosyncratic or transitional models that blend features of those outlined

here. Thus a model based on the perfor mance led by George Gershwin, for
instance, would bor row some features from the pop-music model and some from
the jazz-standard model; a model based on the ar rangement played by the Jimmy
Dorsey band would incor porate some aspects from the jazz-standard model and
some from the improv model. And were such models taken as the basis for our
determinations as to what belonged in the category performances of I Got Rhythm,
there would of course be changes in what counted as a typical or atypical member
of the category.
When people share the conceptual model for a song (realizing that all such sharing is approximate), they will tend to make similar judgments about what counts as
a typical or an atypical rendering of the song (or whether a succession of sounds
should even be counted as an instance of the song). As I proposed in chapter 3, sharing knowledge in this way is one of the bases of culture. This becomes especially
clear when we consider other roles played by the model for a musical work: as a set
of cognitive resources for perfor mance, as a basis for negotiations on how musical
practice should proceed, and as a focus for how musical practice can be regulated.

Conceptual Models and Performance

Up to this point, we have considered music that is part of a tradition of performance
at least partially regulated by the score. But what of performance traditions in which
scores play no part that is, music that is part of an aural tradition? Where scores (or
similar artifacts) are absent, performers must rely on a store of knowledge about
music and music making, knowledge often gathered through a long and painstaking process that starts in earliest childhood and continues for many years.
Constantin Brailoiu was perhaps the rst to suggest that musical perfor mance
within aural traditions relies on stored cognitive constructs. As a way of explaining
the source of the divergent perfor mances folk musicians gave of the same song,
Brailoiu proposed that such perfor mances were ephemeral incar nations of an ideal

cultural knowle dge and musical ontology

archetype stored in the mind of the musician.28 In his work on African music,
Simha Arom adopted Brailoius idea, but he called the archetype a model.29 James
Cowdery, whose research has focused on Ir ish music, proposed a similar cognitive
basis for perfor mance practice, but one der ived from the tradition of inquiry centered on the tune family.30 Cowdery called the construct a tune model and conceived of it as a generating patter n in the mind of the perfor mer or musical community.31 Finally, James Porter has developed detailed descr iptions of a succession of
conceptual perfor mance models for the chang ing versions of the song My Son
David sung by the Scottish singer Jeannie Robertson between 1953 and 1960.32
For Porter, the conceptual perfor mance model represented a holistic, self-contained
concept of a particular song in the mind of the singer.33 This model enabled the
singer to g ive successive but nonidentical perfor mances of the same song on a
number of occasions, but the model was nonetheless subject to change over time
that is, the same song gradually became a different song as the conceptual performance model changed.
The idea behind all of these approaches is that perfor mers who work within
aural traditions base their perfor mance of a given tune on a cognitive construct that
is stored in memory and that represents essential features of that tune. In their basic
features, such constructs are essentially the same as the conceptual models I
described in chapter 3. I have used these same models in this chapter to explain how
individuals deter mine whether a particular perfor mance is of a specic tune and
how typical that perfor mance is. The difference that distinguishes my models from
those of Brailoiu, Arom, Cowdery, and Porter is that the sort of constructs they pro-

28. The relevant texts in translation and the or iginal are as follows: Lacking an unchallengeable
[musical] text, we must admit that we never collect more than variants and that latent in the singers
minds lives an ideal archetype of which they offer us ephemeral incar nations. [A dfaut de tout texte
irrcusable, force nous est dadmettre que nous ne recueillons jamais que des variantes et que, dans
lesprit des chanteurs, vit, dune vie latente, un archtype idal, dont ils nous offrent des incar nations
phmres.] Constantin Brailoiu, Le Folklore musicale, in Musica aeterna: La Vie et la production musicales de tous les temps et de tous les peuples, en tenant compte particulirement de la Suisse, de la Belgique, de la
France et de la musique de nos jours, ed. Gottfr ied Schmid (Zur ich: M. S. Metz, 1949), 2: 319.
29. Simha Arom, African Polyphony and Polyrhythm: Musical Structure and Methodology, trans. Martin
Thom, Barbara Tuckett, and Raymond Bond (Cambr idge: Cambridge University Press, 1991), 148.
30. The idea of a tune family was rst proposed by Samuel P. Bayard in his classic essay Prolegomena to a Study of the Pr incipal Melodic Families of British-Amer ican Folk Song, Journal of American
Folklore 63 (1950): 1 44.
31. James R. Cowdery, The Melodic Tradition of Ireland (Kent, Oh.: Kent State University Press, 1990), 33.
32. James Porter, Jeannie Robertsons My Son David: A Conceptual Performance Model, Journal
of American Folklore 89 (1976): 7 26; and idem, Context, Epistemics, and Value: A Conceptual Performance Model Reconsidered, in Selected Reports in Ethnomusicology: Issues in the Conceptualization of Music,
7, ed. James Porter and Ali Jihad Racy (Los Angeles: University of California, Department of Ethnomusicology, 1988), 69 97. Porters thoughtful appraisal of what might be involved in a conceptual performance model goes well beyond the pr imitive models I sketch here and should be taken into account
in any further application of theories of categor ization to ethnographic research.
It should be noted that while Arom, Cowdery, and Porter all use the ter m model for their hypothetical cognitive constructs, they do so without reference to any of the cognitive science research on
models or related constructs.
33. Porter, Context, Epistemics, and Value, 71.

217

218

analysis and theory

pose are used not for categor izing musical perfor mances, as I attempt to do, but for
producing them. In truth, we have already seen conceptual models used in this way,
for Jeff s conceptual models for Twinkle not only helped him organize the
Montessor i bells in particular patter ns but also guided his actions when he performed the song. In the present context, however, the network of knowledge with
which we are concer ned is far more extensive than that with which Jeff was
engaged. The skills required for an adequate perfor mance of a tune like I Got
Rhythm (relative to the perfor mance tradition descr ibed above) are extensive: a
trumpet player, for instance, must know where to nd the notes for the tune on her
instrument, how to produce these notes in the r ight order, how to shape this production to create a convincing version of the tune, how to coordinate her playing
with that of other members of an ensemble, and so on.34 For perfor mers, then, there
will be a number of conceptual models related to the various challenges of creating musical sound, as well as the conceptual model for a particular piece.
The complexity of this network of knowledge emphasizes once again the
importance of the global models relative to which local models are framed. With
regard to the conceptual model for a musical work, the global model speeds the
process of learning the conceptual model for a new tune and also provides a framework for discourse among musicians. For instance, among perfor mers the jazzstandard model for I Got Rhythm (which was diagrammed in g. 5.2) is framed
relative to a global model for jazz practice, which includes the ve elements diagrammed in gure 5.4. A comparison of gure 5.4 with gure 5.2 shows that each
generic element in the for mer has its more specic cor respondent in the latter.
Confronted for the rst time with a tune like I Got Rhythm, a musician who has
internalized the global model will already have a start on assembling a local conceptual model to use as a guide in perfor ming the tune. The global model also
serves as a frame for discourse (both verbal and musical) about local models. That is,
approaching I Got Rhythm as a song that belongs to the perfor mance practice
of jazz means that features such as its har monic structure, melody, and character are
what are relevant, rather than whether there is a denitive score on which to base
performance.
Conceptual models, then, are used not only to structure our understanding of
34. Of necessity, this network of knowledge will include embodied knowledge: what it feels like to
make musical sounds and to coordinate this activity with other physical activities. Richard Powers
touches on the importance of such knowledge in his novel Galatea 2.2, which centers on a hypothetical exper iment to replicate aspects of human intelligence on a vast and spreading neural network, implemented via parallel processing on a massive supercomputer. At one point, the implementation, named
Helen, asks How do you sing? The narrator responds by singing a fragment of a song he lear ned as a
child: Bounce me high, bounce me low, bounce me up to Jericho. To his amazement, the implementation not only absorbs this song but can later be heard singing it, over and over, like a child trying out
her voice.Yet sing is not quite the r ight word, for Powerss nar rator realizes that there is something
missing: Helen did not sing the way real little girls sang. Technically, she almost passed. Her synthesized
voice skittered off speechs earth into tentative, tonal Kitty Hawk. Her tune sounded remarkably limber, given the scope of that mechanical tour de force. But she did not sing for the r ight reasons. Little
girls sang to keep time for kickball or jump ropes. . . . Helen didnt have a clue what keeping time meant,
never having twirled a jump rope, let alone seen one. Richard Powers, Galatea 2.2 (New York: Farrar
Straus Giroux, 1995), 205.

cultural knowle dge and musical ontology

tunes have a basic
harmonic pattern and a
clearly articulated form
tunes have melodies
which function as the
head of the tune

the jazz tune

tunes often have words,

but these are not essential
for the identity of a tune

tunes have various functions,

chief among which is to serve
as a vehicle for improvisation

tunes have an
identifiable character

figure 5.4

Partial representation of the global model of jazz practice

what we hear but also to create the things we hear. In both cases, the model is part
of a network of knowledge that, through being shared by the members of a musical community and realized in sound, constitutes musical culture.

Negotiating Conceptual Models

Musical practice is not a steady state, as may be seen in the different versions of I
Got Rhythm collected in table 5.1. Over time, conceptual models will be
modied in response to changes in cultural pr iorities. Rarely are such modications
the result of creative at by one individual (although one such case is discussed
below, in connection with Bye Bye Blackbird). More commonly, they are the
product of a gradual ser ies of negotiations within a musical community.
The basic outlines of the process of negotiating the structure of conceptual
models can be illustrated by consider ing two highly simplied situations. Each
involves a musician who initiates the process of negotiation by suggesting changes
to the constituent elements of a given conceptual model. In the rst case, the negotiations occur within a g roup of musicians occupied with perfor ming the same
work; in the second case, the negotiations occur between the musician and an audience of non-perfor mers.
The rst situation can be conceived of as taking place in a rehearsal or a performance of the version of I Got Rhythm associated with the jazz-standard model.
Instead of playing the usual har monic progression for the refrain, the pianist, without comment, introduces a completely new ser ies of harmonies for the A section
and truncates the overall for m to thirty-two measures. The other musicians may
accept this change (by adjusting their perfor mance to the new harmonies), reject it

219

220

analysis and theory

(by resolutely hewing to the structure of the or iginal model), or suggest further
modications (by offering their own inter pretation of the har monic progression).
This process could also play out with the aid of language (the changed har monic
progression could be announced in advance) or symbols (a chord chart for the
changed refrain could be supplied to the other musicians), or through various combinations of musical perfor mance, language, and symbols by any or all participants.
The second situation can be conceived of as taking place dur ing a perfor mance
of I Got Rhythm by a solo pianist, playing for an audience of nonperfor mers.
Again, instead of the usual har monic progression for the refrain, the pianist introduces a new ser ies of harmonies and truncates the overall for m to thirty-two measures. The audience may respond by accepting the change (through applause or a
similar activity), rejecting it (by shouting catcalls or throwing bottles), or suggesting further modications (by shouting More! or Too much!). (Of course, it is
quite probable that an audience would not even notice changes of this sort, but for
the sake of the argument, I shall assume that it did.)
In each of these imaginary situations, some elements of the conceptual model for
the tune are changed and some are retained. The conceptual models thus for m a
framework for negotiation that is, negotiation is conducted in ter ms of the conceptual model for the tune. Series of such negotiations then lead to the more substantive changes in conceptual models of the sort suggested by my account of how
such models might be linked in a histor ical succession. Of course, the imaginary situations I have described are highly simplied: as the works of Ruth Finnegan, Paul
Berliner, and Ingrid Monson have demonstrated, actual negotiations among musicians are much more complex than those outlined here.35 Musical discourse is r ich
and creative, and it includes comedy and irony, as well as meaningful exchanges
about the business at hand. Nonetheless, what makes this communication possible
is the knowledge shared among musicians, knowledge that is organized by conceptual models.

Limits on Negotiation
The mutability of conceptual models under the social and cultural pressures of
negotiation has the potential to destabilize cultural knowledge by transfor ming it
until it becomes alien and unfamiliar. Under certain circumstances, the interconnectedness and dispersion character istic of cultural knowledge can counteract destabilization, a process noted by Mantle Hood.36 Under other circumstances, however,
the challenge to stability may be signicant. In the case of musical practice, such
challenges might come from a number of sources: from the innovations of inuen-

35. Ruth Finnegan, The Relation between Composition and Performance: Three Alternative
Modes, in The Oral and the Literate in Music, ed.Yoshihiko Tokumaru and Osamu Yamaguti (Tokyo:
Academia Music, 1986), 73 87; Berliner, Thinking in Jazz; Ingrid Monson, Saying Something: Jazz Improvisation and Interaction, Chicago Studies in Ethnomusicology (Chicago: University of Chicago Press,
1996).
36. Mantle Hood, The Reliability of Oral Tradition, Journal of the American Musicological Society 12
(1959): 201 09.

cultural knowle dge and musical ontology

tial individuals, from contact with other cultures, from a sudden increase in the
complexity of musical practice, or from changes in technology. Faced with this sort
of challenge, a musical community may develop strategies to eliminate or control
the instability of the conceptual models upon which the practice of that community is based.
One common strategy is to represent key structural elements of a local conceptual model through artifacts such as scores or recordings. Although the score illustrations used for I Got Rhythm carry the aura of author ity, they actually follow
from the conceptual models for the song: a score is an artifactual manifestation of
the elements of the conceptual model deemed most relevant to the musical practice
of which the model is a part, created as a means of stabilizing the model. The score
for the or iginal version of I Got Rhythm (ex. 5.1) can thus be seen as a reection
of the conceptual model for pop songs (g. 5.1). Such songs invariably had a verse,
even if it was only rarely perfor med. Since the songs were often perfor med by musicians with only limited score-reading abilities, chord symbols (which often do not
match the chords indicated by the music notation) are almost always included, and
they serve to indicate the minimal mater ials necessary to accompany the voice. The
jazz fake book rendering of I Got Rhythm (ex. 5.2) does away with the verse and
any specics about how the accompaniment is to be realized; it reects the essential features of the jazz-standard model (g. 5.2). And the Rhythm changes of
example 5.3 preserve only the bare essentials, cryptic to anyone not schooled in jazz
but more than enough to serve musicians as a starting point for improvisation, provided they are familiar with the concepts and relations embodied in the improv
model (g. 5.3). Thus, both in its production and inter pretation, musical notation
reects (rather than generates) the local and global conceptual models that constitute musical practice.37
The idea that musical notation could serve as a way to control (rather than simply enable) musical practice is not a new one, having been proposed by Charles
Seeger over fty years ago.38 With respect to Western practice, the limits on negotiation that follow from this reliance on musical notation operate on two levels.
First, any given instance of musical notation, as an embodiment of the work of
music, limits negotiation to those aspects of music deemed nonessential to the
work (such as the specics of dynamics, timbre, and tempo): these are the things left
relatively ambiguous by the score.39 Second, the terms for negotiation become artifactual: any aspect of musical practice that would vie for deter mining the identity of
a musical work must be such that it can be expressed artifactually. Arguments over
37. Leo Treitler has made similar points about the role of notation in medieval music. See Treitler,
Transmission and the Study of Music History, in Report of the Twelfth Congress, Berkeley 1977, ed. Daniel
Heartz and Bonnie Wade (Kassel: Brenreiter-Verlag; Philadelphia: The American Musicological Society, 1981), 202 11; and idem, History and the Ontology of the Musical Work, Journal of Aesthetics and
Art Criticism 51 (1993): 483 97.
38. Charles Seeger, Oral Tradition in Music, in Funk and Wagnalls Standard Dictionary of Folklore,
Mythology, and Legend, ed. Maria Leach and Jerome Fried (New York: Funk and Wagnalls, 1950), 2: 825 29.
39. Nelson Goodman observes that where the score attempts to capture what I have called the
nonessential aspects of music it ceases to be, in his ter ms, notational. See Goodman, Languages of Art,
183 86.

221

222

analysis and theory

musical identity (such as those that play out in discussions about the Urtext for a
score) are arguments about the status and inter pretation of artifacts, and any competing account of musical identity must be expressed in artifactual ter ms.
It bears mention that the author ity of notation is not monolithic: most musicians
accept notation as something to be inter preted and consider the way it is to be used
to create a musical perfor mance a matter for negotiation. And even where notation
has been introduced with the explicit intent of eliminating variation within musical practice, it may be circumvented. As Ruth Davis has shown, the modied Western notation applied to Tunisian art music beginning in the 1930s for the pur pose
of standardizing its melodic tradition has been tur ned into a creative tool for reinterpreting and redening that tradition.40 Notation has thus slowed, but not eliminated, the uidity of Tunisian musical practice.

Categorization, Musical Ontology, and

Cultural Knowledge
By the late 1940s, as we have seen, there were multiple versions of George and Ira
Gershwins I Got Rhythm in circulation. There was the show stopper and popular tune from the Broadway musical, the hot jazz tune played by any number of
dance bands, and the relatively abstract platfor m for improvisation used as the basis
for any number of derived tunes by jazz composers and for extended improvisations
by smaller ensembles. Confronted with this multiplicity of I Got Rhythms, how
might a listener discr iminate between them? Based on what we now know about
processes of categor ization, we can say that our listener would almost certainly recognize that, while all these versions belonged in the category performances of I Got
Rhythm, not all would represent the category equally well. Depending on the conceptual model around which the category was organized, some versions would be
regarded as quite typical of the category, others as less typical, and still others in danger of being excluded from the category altogether. Following Richard Crawfords
description of three different approaches to I Got Rhythm, I proposed three
models the pop-music model, the jazz-standard model, and the improv model
which would yield different inter pretations of typicality when applied to a range
of performances of the tune. The knowledge represented by such models, when
combined with knowledge about musical technique and framed by global conceptual models applicable to an entire repertoire, can also serve as the basis for musical
performance within aural traditions. Because they are retained by individual members of a musical community, who may wittingly or unwittingly change the model
as they use it, conceptual models are subject to change over time. When such
changes are brought to the attention of other members of the musical community,
a ser ies of negotiations about what constitutes the model may ensue. At times,
changes to conceptual models may be perceived as a threat to the integ rity of a
musical community, and the community may then seek either to halt or to retard
change by placing limits on what can be negotiated.
40. Ruth Davis, The Effects of Notation on Performance Practice in Tunisian Art Music, The
World of Music 34 (1992): 111.

cultural knowle dge and musical ontology

Given the various conceptual models we have considered and the ways these
models are inter preted and cor related with other types of knowledge, it is clear that
the network of knowledge within which any musical work (or similar construct) is
embedded is a complex one. And yet we have thus far restricted ourselves to an
almost purely musical world. Matters only stand to become more complicated
when we contemplate how this world might connect with the cultural context
within which musical practice is itself embedded. In the next section, we encounter
some of these complications and gain a sense of the richer understanding of musical practice to which they lead, as we explore some of the different ways Bye Bye
Blackbird was conceived during the middle of the twentieth century.

bye bye blackbird, signifyin(g), and

cultural knowledge
On 5 June 1956, Miles Davis, John Coltrane, Red Garland, Paul Chambers, and
Philly Joe Jones went into Columbias 30th Street Studio in New York and recorded
three tunes: Tadds Delight, Dear Old Stockholm, and Bye Bye Blackbird.41
Although the recording session took place at a time when Davis was actively shaping the sound and repertoire of his quintet, the choice of Bye Bye Blackbird was
nonetheless somewhat odd. Unlike the other two tunes, Bye Bye Blackbird had
never been a part of the jazz repertoire, but was simply an upbeat popular song usually played as a foxtrot. Recorded by Davis as a ballad, it acquired a melancholy and
an irresistible groove that were thoroughly exceptional.
The choice was made odder still by the racial overtones associated with the tune.
First published in 1926, Bye Bye Blackbird played a role in the racially charged
election campaign for the mayoralty of Chicago in 1927. Big Bill Thompson, a
Republican who had successfully cultivated the African American vote to become
mayor of Chicago in 1915, returned in 1927 to challenge Democratic mayor William
E. Dever. As part of their strategy to ght off this challenge, the Democrats appealed
to racial prejudice, voiced specically through Bye Bye Blackbird. The ploy ultimately backred, for it galvanized African American support for Thompson:
When some of the Dever supporters sent calliopes through the streets piping the
strains of Bye Bye Blackbird, and spreading a circular which displayed a trainload of
Negroes headed from Georgia with Thompson as pilot of the train, and the caption:
This train will start for Chicago, April 6, if Thompson is elected, the obvious answer
of the colored leaders was: Elect Big Bill or its going to be bye-bye blackbirds in
Chicago.42
41. These three tunes, together with Round Midnight, Ah-leu-cha, and All of You, were
released on Miles Davis, Round about Midnight (Columbia, [CD] CK 40610, 1987), originally recorded
27 October 1955 and 5 June and 10 October 1956. There are extensive transcr iptions and a discussion of
this recording of Bye Bye Blackbird in Berliner, Thinking in Jazz; see esp. pp. 678 88.
42. Harold F. Gosnell, Negro Politicians: The Rise of Negro Politics in Chicago, with an introduction by
James Q.Wilson (Chicago: University of Chicago Press, 1966; originally publ. in 1935, with an introduction by Robert Park, Social Science Studies, 35), 54 55. There is a discussion of a similar cartoon
from around the same time that invokes the phrase Bye Bye Blackbird in Lloyd Wendt and Her man
Kogan, Big Bill of Chicago (Indianapolis: Bobbs-Mer rill, 1953), 256.

223

224

analysis and theory

In a similar vein, Jack Chambers reports that dur ing the 1930s, the few African
Americans who participated in activities dominated by whites, like the semiprofessional National Football League, had the song sung to them by the all-white
spectators.43
What seems to be reected in Daviss adoption of the tune is the sort of doublevoiced discourse rst descr ibed by Mikhail Bakhtin: on one level of discourse,
Daviss Bye Bye Blackbird is an aging pop song, with racial overtones that tur n
against its perfor mers; on another level, it is a deeply felt rewriting felt both in
terms of melancholy and groove that makes a mockery of the simplistic messages
of the song, not least by rising above them. This process of repetition (in the sense
of returning to an earlier text) and revision is associated with a kind of intertextuality Henr y Louis Gates Jr. called Signifyin(g), which he argued is fundamental and specic to African American rhetor ical practice.44 While Gatess primary focus was on literature, he extended his argument to other domains, music
among them, and his arguments have proved compelling to musicologists. Adopting
one of Gatess examples, Ingrid Monson was able to show in some detail how John
Coltranes 1960 recording of Richard Rodgers and Oscar Hammersteins My
Favorite Things (written in 1959) accomplishes this sort of Signifyin(g). Coltranes
interpretation completely revises the song: the focus is changed from the verse to
the interludes, Rodgers and Hammersteins waltz is transfor med into a polyrhythmic extravaganza, and what was an upbeat tune becomes a brooding one.45
There is, however, a problem with claiming, in any simple way, that Davis is Signifyin(g) on Bye Bye Blackbird, and that problem is proled vividly by a recording of the tune in 1955 by Peggy Lee. In her perfor mance, Lee adopts an exceptionally slow tempo slower than Daviss and completely vanquishes the upbeat
sentiments of the or iginal song.46 Lee, too, seems to have engaged in double-voiced
discourse, but would we want to say that she is Signifyin(g)? Indeed, this sort of performance practice is so common in jazz within and without the African American community that Signifyin(g), present everywhere, would seem to signify
nothing.47
Two sets of issues, then, are raised by Daviss recording of Bye Bye Blackbird.
First, what is the nature of the double-voiced discourse of Signifyin(g)? Is it in any
way distinct from double-voiced discourse as a whole? Second, what part does the
reinterpretation of musical works or, from the perspective developed in the pre-

43. Jack Chambers, Milestones I: The Music and Times of Miles Davis to 1960 (Toronto: University of
Toronto Press, 1983), 237.
44. Henry Louis Gates, Jr., The Signifying Monkey: A Theory of African-American Literary Criticism
(New York: Oxford University Press, 1988), chap. 2. The parenthetical g in Signifyin(g) is meant to signal African American pronunciation, as well as to distinguish it from the more general process of
signication; for discussion, see The Signifying Monkey, 46.
45. Monson, Saying Something, 106 21.
46. Peggy Lee, Bye Bye Blackbird, in Songs from Pete Kellys Blues (Decca, [LP] DL 8166, 1955),
recorded 6 May 1955.
47. The practice of revisiting the work of other composers and musicians was also common in the
music of the previous century: we thus have Beethoven Signifyin(g) on Diabelli in his Op. 120 variations
and Brahms Signifyin(g) on Paganini in his Op. 35 variations.

cultural knowle dge and musical ontology

ceding section, revisions to the conceptual model for a musical work play in Signifyin(g)? In what follows, I begin by developing an approach to double-voiced discourse in ter ms of a cultural theory of information exchange proposed by the linguist Eve Sweetser. I proceed then to explore some of the conceptual models for
Bye Bye Blackbird in circulation dur ing the per iod 1926 1955, and, nally, I
return to Daviss inter pretation of the tune.

Double-Voiced Discourse and a Cultural Theory

of Information Exchange
Mikhail Bakhtin developed his notion of double-voiced discourse with reference to
the novels of Dostoevsky, going so far as to claim that this rhetor ical mode was essential to Dostoevskys style.48 Double-voiced discourse is actually the third of three
types of discourse distinguished by Bakhtin. The rst type is direct discourse,
which is essentially the unmediated disquisition of the author. The second type
Bakhtin called objectied or represented discourse; its most common for m is the
direct speech of characters. In double-voiced discourse (the third type), the author
takes the direct discourse of someone else and infuses it with the authors own intentions and consciousness while still retaining the or iginal speakers intentions. Thus
two discourses and two consciousnesses are present at the same moment.49
With this taxonomy in place, Bakhtin further distinguished three types of double-voiced discourse. In the rst, the discourse of the author and that of the other
person essentially cooperate the author inects, but does not substantially change,
the discourse of the other person. The second type of double-voiced discourse is
associated with parody and irony (this is essentially the sort of discourse Monson
deals with in her analysis of Coltranes My Favorite Things). Here the discourse
of the author is opposed to that of the other person; as Bakhtin puts it, the second
[that is, authorial] voice, once having made its home in the others discourse, clashes
hostilely with its pr imordial host and forces him to serve directly opposing aims.
Discourse becomes an arena of battle between the two voices.50 In the third type
of double-voiced discourse, the other persons discourse remains outside the
purview of the authors, which nonetheless becomes inuenced by that of the other
person. The result is often a hidden polemic: the authors discourse represents a
reaction to the discourse of the other person, but the latter can be heard only by
implication it is never directly represented.
Each of these three types of discourse and, most important, the three types of
double-voiced discourse can be analyzed with reference to a cultural theor y of
48. Mikhail Mikhailovich Bakhtin, Problems of Dostoevskys Poetics, ed. and trans. Caryl Emerson, with
an introduction by Wayne C. Booth, Theory and History of Literature, 8 (Minneapolis: University of
Minnesota Press, 1984), chap. 5.
49. The premise of double-voiced discourse is that the two domains of discourse remain separate
and identiable; it is thus different from the process of conceptual blending discussed previously in chap.
2 (in that blending is concer ned with how elements from two distinct domains come together in a third
domain), but it may be seen as a related phenomenon.
50. Bakhtin, Problems of Dostoevskys Poetics, 193. I shall follow Bakhtins practice and speak of the rst
voice in double-voiced discourse as that of the other and the second voice as that of the author.

225

226

analysis and theory

information exchange developed by Eve Sweetser. Sweetser used the theor y to

explain the results of an exper iment by Linda Coleman and Paul Kay that was
designed to show that the English word lie (in the sense of to prevaricate) functioned like a Type 1 categor y. That is, rather than following a clear-cut denition,
some statements were regarded as typical lies, others as atypical. Coleman and Kay
hypothesized that a statement is judged to be a more-or-less good example of a lie
on the basis of three conditions. First, the actual statement must be false (a condition that is commonly taken as necessar y and sufcient when lie is construed as
a Type 2 category). Second, the speaker must believe the statement to be false.
Third, in utter ing the statement, the speaker must intend to deceive the listener.51
To test this hypothesis, Coleman and Kay constructed eight stor ies that satised all,
only some, or none of these conditions. They then asked sixty-seven subjects to
determine whether, in the case of each story, a lie had been told. In support of their
hypothesis, the results indicated that lie functioned as a Type 1 categor y: judgments of what constitutes a lie were not made comprehensively, but, rather, could
be made by degree.
Although they were able to prove their hypothesis, Coleman and Kay were not
able to explain satisfactor ily what it was about the word lie, or about the conditions sur rounding its proper use, that made it function as a Type 1, rather than as a
Type 2, category. Sweetser, for her part, proposed that judgments about lies were
made relative to a cultural theor y of information exchange.52 The theory, as formulated by Sweetser, consists of three relatively straightforward propositions, which
can be stated as rules:
1. Try to help (do not harm). This rule is combined with the belief that knowledge is helpful and that misinfor mation is har mful (which leads to the
next rule).
2. Give knowledge (do not misinform). Sweetser argues that what counts as
knowledge, in everyday terms, is what we believe. All things being equal,
we assume that what we believe matches some state of affairs or possible
state of affairs: we take what we believe to be true, to count as knowledge.
Using this understanding of knowledge together with (1) and (2), we
arrive at the nal rule.
3. Say what you believe (do not say what you do not believe). If what we believe
counts as knowledge, and if we wish to be helpful, we must give knowledge.
This chain of entailments represents a cultural theory of information exchange that
allows us to reach the somewhat sur prising conclusion that if a speaker (S) utters a

51. Linda Coleman and Paul Kay,Prototype Semantics: The English Word Lie, Language 57 (1981):
26 27.
52. Eve E. Sweetser, The Denition of Lie: An Examination of the Folk Models Underlying a
Semantic Prototype, in Cultural Models in Language and Thought, ed. Dorothy Holland and Naomi
Quinn (Cambr idge: Cambridge University Press, 1987), 43 66. Sweetsers models have the same function as the theor ies descr ibed in chap. 3; I have simplied some aspects of her account in the interests
of concision.

cultural knowle dge and musical ontology

proposition (P), then that proposition is true. The causal chain can be represented
schematically:
a. S said P. [as part of an infor mation exchange; from rules (1) and (2)]
b. S believes P. [(a) combined with rule (3)]
c. Therefore, P is true. [(b) combined with the understanding that we believe
what is true and do not believe what is false]
This is, without a doubt, a greatly simplied perspective.What is important is that
the model of information exchange set out by the theory works quite a bit of the
time, and when it does not work, the theory provides a way to inter pret where
things went wrong.
To illustrate, let us consider two situations. In the rst, I claim that I saw a fr iend
of mine walking down the street. On closer inspection, it turns out to be someone
who just looked like my friend. According to the simplest denition, my claim is a
lie a statement that is not true. However, since my behavior as a whole ts with
the cultural theor y of information exchange (my intent was to provide helpful
information), my claim will most likely not be judged a lie. In the second situation,
I see someone walking down the street whom I mistake for my friend. Remembering that I owe my friend money and being a cheap sort, I claim that the person
I see is not my friend. According to the simplest denition, this is not a lie, since my
claim is true: the person walking down the street is not, in fact, my friend. Nonetheless, since I deliberately said what I thought to be misinfor mation, I have violated
the theory of information exchange and my claim might well be judged a lie.
Of course, this straightforward theory of information exchange does not operate
in isolation but is inter preted through other cultural theor ies to explain types of discourse that seem to contravene the theory. Jokes, tall tales, and ction all may have
limited truth-value, but their lies are permitted as part of a cultural theor y that
interprets certain types of exceptional behavior as entertainment. Social lies are
based on the recognition that not all for ms of knowledge are helpful: rarely when
someone remarks How are you? do they really want to know how we are. The
cultural theory of information exchange is thus ltered through a theory that says
maintaining social order through polite behavior is more important than saying
what you believe to be true.53
Applying Sweetsers example of a cultural theor y of information exchange to
Bakhtins rst two types of discourse is relatively straightforward. All things being
equal, we assume that direct discourse the voice of the author involves the
truth; we similarly give the benet of the doubt to the objectied discourse of the
authors characters.54 With double-voiced discourse, however, the application of the
53. For a related view of how information exchanges contr ibute to and constitute social order, see
Michael B. Bakans account of how various Balinese musicians reported the results of a gamelan competition in Music of Death and New Creation: Experiences in the World of Balinese Gamelan Beleganjur,
Chicago Studies in Ethnomusicology (Chicago: University of Chicago Press, 1999), chap. 5.
54. The genre of ction, which is the focus of Bakhtins analysis, requires an intr iguing nesting of
theories about infor mation exchange. On the level of everyday life, ction is generally recognized as
nonfactual (except by the delusional).Within a given work of ction, however, we usually apply the theory of information exchange outlined by Sweetser: we assume that characters say what is true in the con-

227

228

analysis and theory

theory is a bit more complicated. With the rst type of double-voiced discourse
where the discourse of the author and the other person essentially cooperate the
idea that both voices speak the truth is easily accommodated. What results is a sort
of heterophony of discourse: two voices saying the same thing in different ways.
With the second and third types parody and hidden polemic the contradictions
and tensions between the two voices suggest that one of the voices must be violating the pr inciples of the cultural theor y of information exchange. The traditional
resolution of the problem, implicit in Bakhtins analysis of discourse as a whole, is
that in double-voiced discourse, the voice of the author is to be pr ivileged: it is the
voice of author ity.55 This resolution requires modication of Sweetsers cultural
theory of information exchange by the addition of one further proposition:
4. The voice of authority is always helpful.
On the one hand, it is difcult to ascr ibe this inter pretation to Bakhtin without a
crushing sense of irony, given his personal history. But on the other hand, it is certainly the case that, in his analyses of double-voiced discourse, the rst (nonauthor ial) voice is the one he most often distrusts.
Given this analysis, equating Gatess notion of Signifyin(g) with Bakhtins double-voiced discourse brings us to two conclusions. First, the voices of black rhetor ic
become voices of authority: they are the helpful voices that speak the truth. Second,
Signifyin(g) is by no means unique to the African American community indeed,
Dostoevsky and countless other authors are also Signifyin(g). But these conclusions are hardly satisfying, for they trivialize the points Gates wishes to make about
African American literature and culture and miss the ways language can be used
against oppressors.
The solution to these problems lies in two important adjustments to the cultural
model of information exchange adumbrated here. First, a distinction must be made
between author and author ity. For the African American community, the rst
voice in double-voiced discourse was almost always the voice of author ity. Since
experience had shown that the voice of authority was almost always unhelpful, it
was a voice to be distrusted. Rather than trusting one of the voices of a doublevoiced discourse, it was important actively to distrust one of those voices; by default,
the nonauthor itative voice of the author was to be regarded as the helpful one. Second, it did not particularly matter if the statements of the author were true or not.
Indeed, as Gates points out, one of the common African American ter ms for Signifyin(g) is lies.56 What mattered more was whether the author, in some way,
adhered to the rst principle of information exchange: to be helpful. And it is clear
that, to an oppressed people, speech that co-opts, mocks, or indirectly challenges the
voice of authority, whether the statements made in the course of that speech are

text of the story.When a character says something that is not true within the context of the story (even
though it may be true in the everyday world), we then have to fall back on this theory or its elaborations
to make sense of the utterance.
55. The pun is not sur prising, since authority and author share the same Latin root (auctor).
56. Gates, The Signifying Monkey, 56, 71.

cultural knowle dge and musical ontology

true or not, is singularly helpful.With these adjustments, distinctions between doublevoiced speech and Signifyin(g) become clear: Signifyin(g) relies on a modied theory of information exchange that assumes that the rst voice in double-voiced discourse is the voice of an extracultural author ity and that this author ity is not to be
believed. Under these circumstances, double-voiced discourse is, as Bakhtin characterized it, an arena of battle between the two voices. However, it is an arena in
which the author is at a distinct disadvantage and must rely on whatever combination of truths, half-truths, and lies is necessary to win the day.
My intent with this analysis is to show how Bakhtins notion of double-voiced
discourse functions relative to a highly simplied, yet quite serviceable, rendering of
a cultural theory of information exchange and how the double-voiced discourse of
Signifyin(g) relies on a modied version of this theory. In applying this perspective
on discourse to music, I restrict my discussion to the level of the conceptual model
for a work. Musical discourse, understood in this way, is thus concer ned with the
realization and projection of the conceptual model for a work. In ter ms of the cultural theor y of information exchange (extended to music), our assumption is that
the perfor mer intends to project this model and that it is valid for the piece. My
decision to impose this restriction is based on practical reasons. It will keep discussion focused on the main topics of this chapter, it will greatly simplify the analyses,
and it requires no strong analogy between linguistic and musical discourse. There is,
however, no reason that, mutatis mutandis, this treatment could not be extended to
more nely grained levels of musical discourse.

Conceptualizing Bye Bye Blackbird, 1926 1955

Bye Bye Blackbird was wr itten in 1926 by the song-wr iting team of Mort Dixon
and Ray Henderson. The score, given in example 5.4, reveals a song typical of the
thousands wr itten dur ing the days of Tin Pan Alley. It consists of an eight-measure
introduction (taken from the concluding phrase of the refrain); a two-measure vamp
to allow the vocalist to get ready; a thirty-two-measure verse in AABA' form; and
a thirty-two-measure refrain in AA'BA" form (where A' presents the mater ial of the
A section a diatonic step higher, and A" changes the or iginal A section at midphrase
so that it ends on the tonic).57
The general topic of the song centers around a nar rator in alien or unhappy surroundings, symbolized in the rst verse by the trope of winter, the blackbirds doleful pronouncements, and the use of minor mode. The nar rator rejects these conditions and looks to where theres sunshine galore, a promise held out by the refrain:
a waiting loved one, a familiar bed, and a light burning in the window, all intro-

57. At least two other scores for Bye Bye Blackbird have been in circulation. Both of these are in
F major, and both shorten the verse to a sixteen-measure form by omitting the inter nal AB sections. One
score, with the copyright renewed and assigned to the Remick Music Cor poration, has a short, fourmeasure introduction; the sixteen-measure verse; and the refrain. The other, copyrighted in 1948 by the
Remick Music Cor poration, was printed by Francis, Day and Hunter, London. It gives the refrain rst,
followed by the introduction and verse. I have found no recordings of performances based on either of
these scores.

229

230

analysis and theory

example 5.4
Music Group

Ray Henderson and Mort Dixons Bye Bye Blackbird, Fred Ahlert

duced against the backdrop of a spr ightly G major. The second verse (which
replaces the blackbird with a bluebird) reects the hope offered by the refrain, conditioned nonetheless by the return to minor mode and references to longing, fading like a ower, and hours that are like one long tear.
Recordings of the period indicate typical treatments of the song. In their version
of 22 April 1926, for example, Bennie Kruegers Orchestra begins with a short
introduction (which bears no resemblance to the published one) and proceeds to an

cultural knowle dge and musical ontology

example 5.4

(continued)

(continued)

instrumental version of the refrain, verse, and refrain.58 (As with the early recordings
of I Got Rhythm, the Krueger ar rangement uses the verse as an interlude
between the refrains.) This is followed by a sung refrain and one nal instrumental
refrain. The tempo of the recording (h = 116 120) is a bit faster than the Moderato indicated by the score, but it is perfectly suitable for a foxtrot.59 Gene Austins
58. Bennie Kruegers Orchestra, Bye Bye Blackbird (Brunswick, [78] no. 3186, 1926).
59. The song is in fact identied as a foxtrot on the label of the record, which also includes the
rubric for dancing.

231

232

analysis and theory

example 5.4

(continued)

recording, made a week later, follows the score rather more closely.60 Although he
adds his own introduction and piano accompaniment and varies the melody in the
third phrase of the verse, his perfor mance otherwise confor ms to the score given in
example 5.4, even adher ing to a somewhat slower tempo than did Kruegers group
(h = 96 100). By most standards, Bye Bye Blackbird was a hit: it was recorded at
least fourteen times between March and June of 1926, and Austins recording was
no. 1 on the Billboard charts of 4 September of that year.

60. Gene Austin, Bye Bye Blackbird (Victor, [78] no. 20044-B, 1926), originally recorded 29 April.

cultural knowle dge and musical ontology

example 5.4

(continued)

The various perfor mances of Bye Bye Blackbird recorded in 1926 could be
collected into a category organized around a conceptual model quite similar to that
used for early versions of I Got Rhythm. The model, which is diag rammed in
gure 5.5, consists of ve cor related conceptual elements, character ized as the
things necessary for an adequate perfor mance of the song. However, where the popmusic model for I Got Rhythm continued to have some cur rency in the years
just after the rst appearance of the song (in part because of its association with the
musical), the model for Bye Bye Blackbird appears to have gone dor mant for a

233

234

analysis and theory

figure 5.5

Model for the category performances of Bye Bye Blackbird

time: there are few signs of the song having been recorded after 1926 and throughout the 1930s.61
New recordings of Bye Bye Blackbird did begin to appear in the 1940s. There
are a number of indications, however, that by then the conceptual model for the
song had changed. The novelty recording by Freddie Schnickelfritz Fisher and
His Orchestra (made on 22 February 1941), with its rapid succession of unlikely
instruments (including solo whistling and bar itone saxophone) and its shave-anda-haircut ending, is just short of an all-out joke.62 In a 1941 Ger man propaganda
recording by Charlie and His Orchestra, the words of the song are altered to celebrate early Axis victor ies in the Second World War and to taunt the Br itish (Bye
Bye Blackbird becomes Bye Bye Empire).63 And in a tr ibute to the songs of
1926 recorded on 27 May 1942 by the piano duo of Marlene Fingerle and Arthur
Schutt, Bye Bye Blackbird is the nal song in a medley that includes I Know
That You Know and Baby Face.64 In each of these recordings, only the refrain
of the song is used, and the tempo varies from the moderate swing of Charlie and
His Orchestra (h = 84) to the Fisher Orchestras foxtrot (h = 120 126).While there
are some signs the song has entered the repertoire of standards its verse has

61. Recording, of course, is not the only measure of popular interest in a song: sales of records, piano
rolls, and sheet music; programming of the song in live performances; and the playing of recordings on
radio broadcasts and in juke boxes would also provide an indication of continued interest in the song.
Unfortunately, no infor mation on these indicators is presently available for Bye Bye Blackbird.
62. Freddie Schnickelfritz Fisher and His Orchestra, Bye Bye Blackbird (Decca, [78] no.
25357/3788, 1941).
63. Charlie and His Orchestra, Bye Bye Blackbird, in Charlie & His Orchestra: German Propaganda
Swing 19411942 (Harlequin, [CD] HQ CD 03, 1990).
64. Marlene Fingerle and Arthur Schutt, Medley:I Know That You Know,Baby Face,Bye Bye
Blackbird, in Song Hits of 1926, Songs of Our Times (Decca, [LP] DL 5170, 1950).

cultural knowle dge and musical ontology

the musical materials
for the refrain

the lyrics for the

refrain

Bye Bye Blackbird

(pop song, ca. 1941)

intended for entertainment or

dancing; performers use variation
but little improvisation

performed with
a lively affect

acknowledgment (explicit or
implicit) that the song was written
for a previous generation

figure 5.6

Model for the category performances of Bye Bye Blackbird, ca. 1941

fallen away, and it is readily available for parody it is also clearly marked as an
old song. The conceptual model for the categor y performances of Bye Bye Blackbird, ca. 1941, then, would consist of ve cor related conceptual elements (which are
diagrammed in g. 5.6).65
As a consequence of a growing interest in the 1920s as an era, this model was, if
anything, reinforced dur ing the postWorld War II per iod. Following the phenomenal success of The Jolson Story in the fall of 1946,Warner Brothers lms contracted
in 1947 for the lming of a similar story with (and about) Eddie Cantor. Although
Warner Brothers ultimately dropped the project, Cantor and Sidney Skolsky (producer of The Jolson Story) eventually lmed The Eddie Cantor Story independently,
and it was released in 1953.66 As part of the soundtrack for the movie, Cantor
recorded Bye Bye Blackbird in something approximating the style of the 1920s,
although using only the refrain of the song.67 A similar approach can be heard in
65. I do not mean to suggest, with the idea that Bye Bye Blackbird was written for a previous
generation, that the fteen years between 1926 and 1941 actually ll the span of a human generation.
However, I do think this per iod is long enough to mark a signicant change in popular tastes, which is
what this portion of the conceptual model aims to capture.
66. Herbert G. Goldman, Banjo Eyes: Eddie Cantor and the Birth of Modern Stardom (New York:
Oxford University Press, 1997), 283 84.
67. Eddie Cantor, Bye Bye Blackbird, in Songs by Eddie Cantor from The Eddie Cantor Stor y
(Capitol, [LP] L-46, 1954). There is no evidence that Cantor recorded Bye Bye Blackbird during the
1920s or 1930s, although he might well have performed it in live shows; in fact, Goldman character izes
it as one of the standards of Cantors repertory from the 1920s. By most accounts, The Eddie Cantor Story
was a op; its failure is generally attr ibuted to the inadequacies of Keefe Brasselle (who played the role of
Cantor) and to a scr ipt that was little more than a remake of The Jolson Story. Of the soundtrack, however, Goldman wr ites: The musical ar rangements were rst rate, and the vocals better than might be

235

236

analysis and theory

Johnny Maddoxs recording of the song on piano (accompanied by a rhythm section), released in 1955, which inter prets it in a quasi-ragtime style as part of a collection of 1920s songs intended to accompany tap dancing.68
During this same per iod the general fascination with the character and music of
the 1920s was given a specic focus when there was a resurgence of interest in early
jazz, variously called New Orleans or Dixieland jazz.69 Partly, this came about
because of a str ike by musicians against the recording industry from 1942 to 1944,
which led record companies to reissue older recordings. Around the same time,
there emerged an audience who felt the swing era had passed but who were not
attracted to be-bop, the new jazz style that emerged dur ing the Second World War.
In 1950, the actor Jack Webb (now chiey remembered for his role as Sergeant Fr iday on the Dragnet radio and television ser ies) developed a radio ser ies around a jazz
musician in 1920s Kansas City, reecting contemporary interest in the roaring 20s
and Webbs own enthusiasm for early jazz.70 The radio ser ies was eventually developed into the movie Pete Kellys Blues, which Webb directed and in which he also
starred. In the movie, Pete Kelly is a bandleader threatened by a gangster who is
attempting to extract protection money from musicians. Kelly initially tr ies to resist
but eventually capitulates. As part of his agreement with the gangster, Kellys band
has to take on the gangsters girlfriend, Rose Hopkins (played by Peggy Lee), as a
singer. Rose is an alcoholic but can still render a moving ballad. A tragic character,
Rose is eventually reduced to the mentality of a ve-year-old by alcoholism and by
a beating from the gangster.
As part of the promotion of Pete Kellys Blues as a jazz movie, Lee, together
with Ella Fitzgerald (who played a blues singer working in a roadhouse), recorded
an album of songs associated with the movie on 10 May 1955. Although Bye Bye
Blackbird appears in the movie, neither Lee nor Fitzgerald sings it. Instead, it is
rendered by a drunken chorus of band members returning from a late-night party.
On the album, Lee retains Roses persona from the movie and sings the song (consisting only of the refrain) as a slow ballad (q = 72 76), with slight allusions to Billie Holidays distinctive style of inection.71 The arrangement makes use of complicated har monies typical of the treatment of ballads in 1940s and 1950s jazz and

expected, considering that Cantor had suffered a major heart attack less than three months earlier and
that his voice had lost its for mer brilliance after World War II. Goldman, Banjo Eyes, 284.
68. Johnny Maddox, Bye Bye Blackbird, in Tap Dance Rhythms (Dot, [LP] DLP 3008, 1955).
69. Chip Deffaa, Traditionalists and Revivalists in Jazz, Studies in Jazz, 16 (Metuchen: Scarecrow Press;
Institute of Jazz Studies, Rutgers, The State University of New Jersey, 1993), 132. In the following I refer
to this style of jazz as traditional jazz, retaining the quotation marks to highlight the function of the
term as a rubric rather than a judgment about the place of this music within any histor ical account of
jazz styles. See also Ber nard Gendron, Moldy Figs and Moder nists: Jazz at War (1942 1946), in Jazz
among the Discourses, ed. Krin Gabbard (Durham, N.C.: Duke University Press, 1995), 31 56.
70. Charles Emge, Meet Jazzman Jack Webb: Swing Era Helped Make Him Music-Conscious,
Downbeat 22 (10 August 1955): 8.
71. The character of Rose seems to have had a profound personal effect on Lee. In her autobiography, she speaks of Rose in the third person, speculates on whether a visit from James Dean was intended
for her or for Rose, and wr ites of moving out of Roses trailer [on the set] once the lm was completed. See Peggy Lee, Miss Peggy Lee: An Autobiography (New York: Donald I. Fine, 1989), 165 67.

cultural knowle dge and musical ontology

adds words for the second pass through the refrain that signal a new perspective on
the song:
A while ago when I was young, I heard a song and heard it sung
[male chorus sings as distant echo] Bye bye blackbird

I dont know why it makes me sad, a happy song should make me glad

[male chorus] Bye bye blackbird

Singing in her persona as a dissolute, tragic chanteuse, Lee evokes a melancholy

that, up to this point, had not been associated with Bye Bye Blackbird.
Pete Kellys Blues was, at best, a mixed success cr itically, and Lees treatment of
Bye Bye Blackbird might well have been forgotten (especially since she did not
sing the song in the movie). Fate, however, intervened: the singer and pianist Bobby
Short heard and liked Lees recording and decided to approach Bye Bye Blackbird
in a similar fashion, nding reective sadness in the refrain rather than joyful anticipation.72 In his recording of 15 September 1955, Short also adopts a slow ballad
tempo (q = 84) and fashions his accompaniment out of rich, progressive harmonies.
To provide an introduction, he starts with the B section of the refrain, the one section that retains the sense of alienation generated by the or iginal verse: No one
here can love and understand me . . . The effect is to cast a shadow over the
remaining lyr ics of the refrain, emphasizing what is negative in them the nar rators care and woe, distant loved one, and need to travel late into the night rather
than the pleasant things that wait at home. Short also adds his own lyrics for the second pass through the refrain (with words adapted from the song Im a Little Blackbird Looking for a Bluebird):73
Ive been all over the east and the west, in search of clover to feather my nest,
Bye bye blackbird
I have a daydream the same as you do, why would a daydream have to be hoo-doo?
Bye bye blackbird

No one here can love and understand me,

Oh, what hard luck stor ies they all hand me. (etc.)74

With these two recordings by Lee and Short, Bye Bye Blackbird was transformed from an upbeat popular song in a fast tempo, designed for dancing a
hop tune, as Short has called it75 into a melancholy jazz ballad in a slow tempo,
designed for listening. The motivation for this change was not, strictly speaking,
musical, for most of the music in Pete Kellys Blues kept to a traditional or Dixieland style.76 Instead, the change was initially motivated (that is, in Peggy Lees

72. Personal communication from Bobby Short, 8 March 2000.

73. Im a Little Blackbird Looking for a Bluebird was wr itten by Clarke Grant, Arthur Johnston,
George W. Meyer, and Roy Turk. It was recorded by both Louis Armstrong and Sydney Bechet in the
mid-1920s.
74. Bobby Short, Bye Bye Blackbird, in Bobby Short (Atlantic, [LP] 1230, 1956), recorded 15 September 1955. As will be noted, Shorts revised lyr ics return to Dixons or iginal refrain at midpoint (with,
No one here can love and understand me).
75. Personal communication from Bobby Short, 8 March 2000.
76. The soundtrack for the movie (with pr iceless nar rations by Jack Webb) consisted exclusively of

237

238

analysis and theory

figure 5.7

Jazz ballad model for the category performances of Bye Bye Blackbird

recording) by a romanticized notion of the female jazz singer, itself derived from
any number of nineteenth-century fantasies about perfor mers.77
The new conceptual model for Bye Bye Blackbird suggested by Lees and
Shorts recordings, and diagrammed in gure 5.7, consists of four elements necessary for an adequate perfor mance. Gone is the sense of history that had freighted
approaches to the song since the early 1940s; it is replaced by a sense of immediacy
generated by the necessar ily reective nature of the perfor mances.
It may be useful at this point to consider differences in the way conceptions of
I Got Rhythm and Bye Bye Blackbird changed over roughly the same per iod
in time. Because I Got Rhythm was an endur ing part of the popular music and
jazz repertoire, its conceptual model changed only g radually, through a process of
slow but relatively constant negotiation. In contrast, no such process is in evidence
for Bye Bye Blackbird. After its rst blossoming, the tune, and its conceptual
model, faded from popular consciousness. When it was recalled dur ing the early
1940s, it was as part of a larger process of revisiting, through music, an earlier era:
Bye Bye Blackbird was simply emblematic of a popular tune from the 1920s.
When Peggy Lee introduced her version, the change to the conceptual model was
sudden and dramatic, by at of her character in the movie.
traditional arrangements. See Pete Kelly and His Big Seven (Matty Matlock and others), Pete Kellys
Blues (BMG Special Products, [CD] DRC12081, 1998), originally recorded 7, 9, 11 June 1955.
77. For a recent, and stimulating, discussion of some of these fantasies, see Susan Ber nstein, Virtuosity of the Nineteenth Century: Performing Music and Language in Heine, Liszt, and Baudelaire (Stanford, Calif.:
Stanford University Press, 1998), chap. 3.
My character ization of the persona developed by Lee in connection with Pete Kellys Blues should
not obscure the powerful presence of African American women who sang the blues and jazz, or the very
real tr ials to which these women were subjected. (For a discussion of the role of these women in shaping the blues in particular, see Susan McClary, Conventional Wisdom: The Content of Musical Form [Berkeley: University of California Press, 2000], 42 49.) I do believe, however, that Lee, as a white entertainer
in the United States of the 1950s, was able to choose this persona (much as she chose to emulate Billie
Holidays voice in the course of her performance of Bye Bye Blackbird) and that part of the framework for this choice was a romanticized notion of the suffer ing perfor mer.

cultural knowle dge and musical ontology

Thus by the end of 1955 there were two conceptual models for Bye Bye Blackbird in circulation. The rst conceptualized the tune as a somewhat dated but
nonetheless upbeat pop song; the second conceptualized the tune as a slow, melancholy ballad. There may have also been still circulating a third, and more sinister
model, derived from the rst, that inter preted the song in racial ter ms a signal for
African Americans to be on guard, that they were not wanted here. It was into the
context provided by these models that Miles Davis introduced his version of Bye
Bye Blackbird.

Daviss Bye Bye Blackbird, Double-Voiced

Discourse in Music, and Signifyin(g)
The recording that Davis made with Garland, Chambers, Jones, and Coltrane was
strictly instrumental, at a tempo slower than the hop version of the tune (q =
120), but notably faster than the tempos adopted by Lee and Short. After a br ief
introduction by Garland, Chambers, and Jones that establishes a loose, almost loping, groove, Davis enters playing the refrain. From his rst notes, Davis announces
the song as a jazz ballad, using a languid approach to the melody and muted trumpet to create an understated, slightly distant reection on the song. After the head,
Davis takes two choruses, followed by two choruses from Coltrane that occasionally
hint at a double-time feel. Garland takes a chorus and a half , and, at the br idge of
Garlands second chorus, Davis reenters with the melody and guides the performance to a conclusion.
Daviss principal contr ibutions are to infuse Bye Bye Blackbird with an ir resistible groove and to establish it as a venue for instrumental commentar y and
expansion. Although these contr ibutions are by no means insignicant Daviss
recording virtually established Bye Bye Blackbird as a jazz standard, something it
had never been before one can also hear the inuence of Lees and Shorts recordings in his version. There is, obviously, the slower tempo and hint of melancholy
adopted from the jazz-ballad model (although the g roove Davis sets up tends to
move the tune back toward the pop-music model). There is also, however, the same
sense of immediacy that marked Lees and Shorts perfor mances: Bye Bye Blackbird is not, in Daviss hands, about nostalgia.78 Davis has modied the conceptual
model for the tune used by Lee and Short, but he has also retained enough of it that
we can still recognize its main elements.
The immediate relationship between the jazz-ballad model for Bye Bye Blackbird and the model Davis set out in his recording suggests that double-voiced discourse can be accomplished in music when one conceptual model for a musical
work is understood to make reference to another conceptual model for the same
78. The element of nostalgia has continued to cling to Bye Bye Blackbird as a popular tune (as
distinct from a jazz standard) and has been exploited in a number of lms. In Jonathan Demmes Melvin
and Howard (1980), Bye Bye Blackbird is the one song Howard Hughes is able to recall when Melvin
Dummar presses him on pain of being thrown out of Dummars truck to sing something, and it
becomes a deeply felt recollection of Hughess childhood. In Nora Ephrons Sleepless in Seattle (1993),
Bye Bye Blackbird is cast as the song the boys deceased mother used to sing to him when he couldnt
sleep.

239

240

analysis and theory

work. If we know Peggy Lees recording of Bye Bye Blackbird, hear Miles Daviss,
and understand the latter as making reference to the for mer, musical doublediscourse takes place.79 Musical double-discourse thus requires, minimally, a listener
who knows at least two conceptual models for a given work. Connections between
the models may be drawn out by the perfor mer, but they may be more oblique and
constructed mostly by the listener. Each of the three types of double-voiced discourse descr ibed by Bakhtin can be created: an inection of the rst conceptual
model by the second (which is what Davis does); a parody of the rst conceptual
model by the second (which is what Monson descr ibes Coltrane as doing); or tacit
reference to the rst model by the second (which is one way to conceive how
derived tunes operate).
But what of Signifyin(g)? Signifyin(g) would require that the rst conceptual
model be recognized as the authoritative one and that its author ity is not to be
believed: that is, it is not a valid model for the work. Validity is then g ranted by
default to the conceptual model used by the perfor mer. A consequence is that the
validity of this model depends on the rejection of the model associated with authority; it is not granted independently. Under these circumstances, a case can be made
that, in his recording of My Favorite Things, Coltrane is Signifyin(g) on Rodgers
and Hammersteins song: the validity of the Broadway-show-tune model for this
song has been rejected, and it is Coltranes model of the tune that is accorded validity. It is much less clear, however, whether Davis is Signifyin(g) on Bye Bye Blackbird. It is difcult to ascr ibe author ity of a sort that one should be skeptical of to
either Lees or Shorts recordings, and it is unclear whether Davis was aware of, or
cared about, the pop-music model for the tune (with or without its overlay of
racism). This does not mean that a listener might not hear the recording by Davis as
Signifyin(g) on Bye Bye Blackbird, only that the relationship of Daviss rendition
of the song to the concept of Signifyin(g) is much more equivocal.
Under what circumstances might we hear Davis as Signifyin(g) on Bye Bye
Blackbird? First, Davis could, in the course of a performance (recorded or otherwise), make an obvious reference to one of the earlier models for the tune, perhaps
by playing a fragment of the verse or by evoking another nostalg ic tune from the
1920s. The other model for Bye Bye Blackbird would thus be brought to our
attention (even if only as something anomalous in Daviss performance) to be then
regarded as a voice of author ity under challenge. Second, we could br ing to Daviss
performance a knowledge or awareness of one of the earlier models, hear that he
reverses certain of the key elements of this model, and inter pret this as subverting
authority. Third, Davis could, through either word or deed, situate himself as in dialogue with what he regarded as an author itative version of the tune.
With this analysis of Daviss recording of Bye Bye Blackbird I do not mean to
suggest that Signifyin(g) is a rare occur rence in music, but only that it involves a
complex matr ix of cultural knowledge and that the listener must understand the
process of Signifyin(g) as taking place. When this matr ix of knowledge is not in
place, what may be left to hear is only double-voiced discourse, or unmediated
79. Although my focus here is on recordings, the same situation would obtain were our exper ience
with live, as opposed to recorded, performances.

cultural knowle dge and musical ontology

authorial discourse. (Indeed, all music could be heard as involved in double-voiced
discourse of a quite general kind, in that the conceptual model for each work is in
dialogue with the conceptual models for other works.) Double-voiced discourse,
as well as Signifyin(g), becomes interesting when tensions between the models
become obvious that is, when the tensions become as important as the models
themselves.

Signifyin(g) and Cultural Knowledge

Given the cognitive complexity of the process of Signifyin(g), it is well worth asking why the notion has become a trope among those who work on jazz and popular music. Three reasons come to mind. First, Signifyin(g) offers a clear alter native
to the Saussurean model of meaning construction: in Signifyin(g), meaning is something that is not xed but is created by the listener.80 Second, Signifyin(g) provides
a way of interpreting the cultural practice of jazz and popular music in ter ms developed to account for the unique prole of one of its most important inuences,
African American culture. Third, applying Signifyin(g) to music allows music cr itics to avail themselves of the power and prestige of literary criticism.
Of these three, it is perhaps only the second that has much relevance to music,
as opposed to the practice of writing about music. Musical Signifyin(g) is a complex
process, distinct from linguistic Signifyin(g) and thoroughly implicated in the dense
matrix of cultural knowledge. To explain how it works when it works requires
not only a theory of Signifyin(g) but also a way to account for musical ontology, an
explanation of the framework for discourse provided by cultural theor ies of information exchange, and a theory of double-voiced discourse.With these in place we
can start to see and hear how Signifyin(g) shapes musical discourse.

cultural knowledge and

musical ontology
Clifford Geertz was skeptical about using descr iptions of individual cognitive
processes to build an account of culture, and rightly so. Cognitive structure is
incredibly complex, and the relationship between individual knowledge and cultural knowledge is not a simple one. Nonetheless, one can discer n patter ns of consistency in the features selected for dening and evaluating cultural products, patterns that are indicative of cultural knowledge. The approach to cultural knowledge
I have developed in this chapter is meant to capture these consistencies and to facilitate a discussion of how cultural knowledge is organized, sustained, and changed. It
is not meant as a literal rendering of what someone has in mind in evaluating particular cultural products. The approach is thus an inter pretation of culture, but of a
80. This is essentially the approach taken by Robert Walser in Out of Notes: Signication, Interpretation, and the Problem of Miles Davis, in Jazz among the Discourses, ed. Krin Gabbard (Durham,
N.C.: Duke University Press, 1995), 165 88.While my approach to Signifyin(g) is broadly consonant
with this treatment of meaning construction, my argument is made from the perspective of research in
cognitive science rather than structuralism.

241

242

analysis and theory

somewhat different sort than those called for by Geertz, for it assumes that culture
is what we know, rather than the specic things we do, make, or say.
As members of a culture, one of the things we know is how to categor ize
sequences of sound events as music (or whatever other ter m is operational in our
culture). Knowing that a Beethoven str ing quartet is music (to return to Geertzs
example) means that it confor ms to our conceptual model for music, no more and
no less. Similarly, John Cages 4'33" is music if it confor ms to our conceptual model
for music; if it does not, we may simply regard it as a rather strange r itual in which
a group of people sit in silence for a specied per iod of time.81 The conceptual
models around which we organize our categor ies of music are also the means by
which we negotiate or resist cultural change, guide perfor mance, and engage in
complex patter ns of musical discourse. They are, quite simply, the stuff of musical
ontology.
Quite obviously, what I offer here is only a beginning. In particular, the account
of the conceptual models appropriate to cultural knowledge needs to be rened
through further study. Coleman and Kays study of the meaning of lie, together
with Sweetsers interpretation of that study, offers an interesting example of how
relevant empir ical research might be conducted. If conceptual models truly have
value as a tool for descr ibing cultural knowledge, then it should be possible to
develop exper imental protocols that test whether and how members of a culture
use such models. What will result is not a simple picture of what someone has in
mind when reasoning about a given situation (which is at best a dubious goal) but
a better understanding of the elements of knowledge that make it possible to function as a member of a culture.
Another obvious extension of the work presented in this chapter is to other
repertoires, both contemporar y and histor ical. In some cases (for instance, when
dealing with a remote histor ical per iod), the application may be so speculative as to
be useful chiey in providing a fresh perspective, but in others it will yield sur prising results. There are processes every bit as complex and challenging as Signifyin(g)
that operate in other repertories for instance, the intr icate patter ns of inuence
and innovation that abound in the music of Gustav Mahler and Arnold Schoenberg patterns that await only our interest and efforts to discover them.
Perhaps the most intr iguing consequence of the argument put forth in this chapter is that, if musical works are cognitive categor ies, and if music is a cognitive category, then music theory is about the study of categor ies or more typically the
conceptual models around which musical categor ies are organized. Understood this
way, music theory immediately escapes the g ravitational force eld of the text
that has at times kept the literary criticism of the previous generation earth-bound
(even while remaining relentlessly recondite). Thus liberated, music theorys own
traditions of high abstraction and unabashed pragmatism may yet nd a place in
contemporary intellectual discourse.
81. Although I have not tr ied (or wanted) to develop an overarching conceptual model for music
here, it is worth noting in this connection that Cages four-and-a-half minutes of silence has a score and
has been presented in concert numerous times. If these things are important elements of someones conceptual model of music, they would most likely regard Cages piece as music.

chapter six

words, music, and song:

the nineteenth-century lied

n 15 December 1822, the poet Wilhelm Mller noted the publication of a

group of songs that took his poems as their texts. In a letter to the composer
of the songs, Bernhard Klein, Mller observed, indeed, my songs lead . . . only half
a life, a paper-life, black upon white . . . until music breathes the breath of life into
them, or at least, when it slumbers within, calls it out and wakens it.1
That music could breathe life into poetr y is a sentiment also expressed in a
eulogy for a composer Mller chose not to address in the foregoing letter, but
whose settings have kept Mllers poems from disappear ing into the obscure
recesses of nineteenth-centur y German poetr y. Speaking of Franz Schuberts
Lieder, Josef von Spaun wrote:

That which moved the poets breast Schubert rendered true and transgured in tone
in each of his songs as none had done before him. Every one of his songs is actually a
poem on the poem that he set to music and not only for sentiment, which is doubtless at home in songs, but also for all the magic of fantasy, for its super natural char m
and its noctur nal terrors, could he nd the perfect moment yes, even for the lofty
seriousness of reection he had a language.2

These two remarkably similar views of the relationship between poetr y and
music are, without a doubt, as much a testament to the intellectual and artistic life
1. Quoted in Carl Koch, Bernhard Klein (17831832): Sein Leben und seine Werke (Leipzig: Oscar
Brandstetter, 1902), 34, n. 8. The complete text runs as follows: In der That fhren meine Lieder, die
zu einem declamator ischen Vortrage, wenige ausgenommen, durchaus nicht geeignet sind, nur ein halbes
Leben, ein Papierleben, schwarz und wei(Ach, wie traur ig sieht in Letter n Schwarz auf Wei das Lied
mich an!) bis die Musik ihnen den Lebensodem einhaucht, oder ihn doch, wenn er dar in schlummert,
herausruft und weckt.
2. Was des Dichters Brust bewegte, hat Schubert in jedem seiner Lieder treu und verklrt in Tnen
wiedergegeben, wie keiner vor ihm. Jede seiner Lieder-Compositionen ist eigentlich ein Gedicht ber
das Gedicht, das er in Musik setzte; und nicht fr Empndung nur, die im Liede wohl daheim ist, sondern auch fr jeden Zauber der Fantasie, ihren ber irdischen Reitz und ihre nchtlichsten Schrecken,
wute er den eigensten Augenblick zu nden, ja selbst fr den hohen Er nst des Gedankens hatte er eine
Sprache. Till Gerrit Waidelich, ed., Franz Schubert: Dokumente, 1817 1830,Verffentlichungen des Internationalen Franz Schubert Instituts, 10 (Tutzing, Germany: Hans Schneider, 1993), 516 17.

243

244

analysis and theory

and ideals of early nineteenth-century Germany and Austria as they are clear-eyed
proposals for a regard of song.3 And yet they capture something important about
the possibilities that ar ise when words and music come together: under certain circumstances the result is far more than simply the sum of its parts.
The topic of song has fascinated wr iters on music at least for the last two centuries, and a study of the study of song could be its own book. In a 1992 essay, Ko
Agawu offered a survey of some of this work (as well as his own contr ibution to it)
and identied four basic models for the analysis of song, of which three were fairly
well entrenched in the literature on song.4 The rst model der ives from the work
of Suzanne Langer and operates on the premise that when words and music come
together in song, the music wholly absorbs the words. As Langer put it, Song is not
a compromise between poetry and music, though the text taken by itself be a great
poem; song is music.5 The second model proposes that there is an ir reducible relationship between words and music; as Lawrence Kramer has put it, A poem is
never really assimilated into a composition; it is incorporated, and it retains its own
life, its own body, within the body of music.6 The third model inter prets song as
a compound structure in which the words carry the primary semantic content and
the music colors or enhances what the words signify. Agawu observed that this
model is limited because it provides no independent account of the meaningfulness
of music.
In reviewing these three models, Agawu was rather too polite, resisting the
temptation to note that, within the tradition of song analysis, there have been relatively few attempts to come to ter ms with how words and music combine to create the phenomena called song. There are thorough and painstaking analyses of
song texts, with cursory mention of the music that sets the text; there are carefully
crafted analyses of musical structure, supported at tur ns by isolated instances of how
the words buttress or anticipate this structure; but there are few comprehensive
analyses that tr y to capture how the meaning constructed by the words and the
meaning constructed by the music come together to create the meaning associated
with a particular song.7 As an improvement, Agawu proposed a fourth model in
which song is construed as a conuence of three independent but overlapping systems: music and words and song. However, he almost immediately withheld full
endorsement of this model, noting that it provided no account of a concrete iden3. Both must also be situated within the more personal agendas of their authors. Mllers letter is
part of an attempt to get the composer to set more of his poems; Spauns eulogy is part of an attempt to
preserve what he viewed as the true legacy of his departed friend.
4. Victor Ko Agawu, Theory and Practice in the Analysis of the Nineteenth-Century Lied, Music
Analysis 11 (1992): 3 36.
5. Susanne K. Langer, Feeling and Form: A Theory of Art Developed from Philosophy in a New Key (New
York: Charles Scr ibners Sons, 1953), 152.
6. Lawrence Kramer, Music and Poetry: The Nineteenth Century and After (Berkeley: University of California Press, 1984), 127.
7. One analysis that develops a comprehensive account of a song is David Lewins Auf dem Flusse:
Image and Background in a Schubert Song, in Schubert: Critical and Analytical Studies, ed.Walter Fr isch
(Lincoln: University of Nebraska Press, 1986), 126 52. However, Lewins main interest in this essay is in
developing an account of certain remarkable features of this particular song; his study, thus, did not set
out to delineate a broad methodological approach to the analysis of song.

word s, mu s i c, and s ong

tity for song but, instead, allowed song an identity only by default. This led him to
muse,Perhaps, then, what the model points to is song as process, not product.What
is interesting, in other words, is not what song is, but what it becomes in its per petual str iving for a concrete mode of existence.8
Nicholas Cook was able to steer a course between the Scylla of emphasizing the
text at the expense of the music and the Charybdis of emphasizing the music at the
expense of the text, as well as to give an account of a concrete identity for song, by
approaching song as an instance of multimedia: words were one medium, music
another.9 However, because Cook dened multimedia as the perceived interaction of
different media, he tended to regard songs as successful to the extent that the syntax
of their words and music remained separate indeed, in a state of contest. The
approach is effective for many songs, and the breadth offered by integrating song
analysis into the analysis of multimedia as a whole makes for a very comprehensive
perspective. Nonetheless, what is missing is a sense of how, despite their differences,
words and music work together in song how music can breathe the breath of life
into poetry or create a poem on a poem.
In this chapter, I explore further the nature of song and also the process of conceptual blending. As noted in chapter 2, conceptual blending is a pervasive and
often transparent process in which elements from two cor related mental spaces
combine in a third.10 And, as suggested by the examples used in chapter 2, the mental spaces basic to conceptual blends can be set up by music as well as by language.
In what follows, I expand on this idea, working on the assumption that music can
set up complete discourse structures, which can be cor related with the discourse
structure of a text to give rise to a unique domain for the imagination. What I am
concer ned with, then, are not isolated or very general cor relations between music
and text but a conceptual blend produced by an entire song.
In the rst section that follows, I analyze two settings of Wilhelm Mllers poem
Trockne Blumen (Withered Flowers), taken from the cycle of poems he called
Die schne Mllerin (The Beautiful Miller Maid). The rst is by the Berlin composer
Bernhard Klein, whom Mller addressed in the letter quoted above; the second is
by Franz Schubert. These analyses will allow me to show how different musical
environments interact with the same text to produce two quite different conceptual
blends, and thus two different songs. In the second section, I turn to Robert Schumanns Im Rhein (In the Rhein), from his song cycle Dichterliebe (Poets Love). In
Im Rhein, the music makes possible a conceptual blend whose emergent structure includes a nar rative considerably more extensive than that of the text alone. In
the third section, I consider two settings of a poem by Joseph von Eichendorff, In
der Fremde (In Foreign Lands), the rst by Schumann, the second by Johannes

8. Agawu, Theory and Practice, 7 8.

9. Nicholas Cook, Analysing Musical Multimedia (Oxford: Clarendon Press, 1998); see esp. pp. 98 106
and Cooks analysis of Robert Schumanns Wenn ich in deine Augen seh from Dichterliebe on pp.
136 43.
10. A focus on the possibilities for meaning construction that ar ise in this blended space is one of
the fundamental differences between the idea of conceptual blending and both Cooks conception of
multimedia and Bakhtins notion of double-voiced discourse.

245

246

analysis and theory

Brahms. Again, two different songs result, but, in this case, the reasons lie with the
gener ic spaces that structure each conceptual blend.

trockne blumen
Wilhelm Mllers Trockne Blumen
Mllers Die schne Mllerin had its genesis in the diversions of a group of young
artists and intellectuals who gathered at the house of the privy councilor Fr iedrich
August von Stgemann dur ing the autumn and winter of 1816 17.11 There Mller
(1794 1827) together with the Stgemanns daughter Hedwig (1799 1891); the
artist Wilhelm Hensel (1794 1861); his sister, the poetess Luise Hensel (1798
1876); the publisher Fr iedrich Frster (1791 1868); and others put together an
informal drama structured around the story of a pretty millers maid and her suitors.
Each member of the group took a part, for which they wrote a verse scr ipt. In the
enactment of the resulting Liederspiel, the participants would recite their verses,
using whatever gestures or changes in vocal inection the situation required in
order to tell their part of the story.
The evidence of the poems and anecdotes that survive from the Stgemann circle suggests that its Liederspiel had a thoroughly tragic ending. In what can be
reconstructed of the story they developed, the millers daughter, named Rose
(played by Hedwig von Stgemann), was courted not only by the young miller
(played by Mller) but also by a country squire (played by Frster), a hunter (played
by Wilhelm Hensel), and a gardener (played by Luise Hensel). The miller, after prevailing br iey over the other suitors, falls into despair when the hunter wins the
miller maids affections, and commits suicide. The maid, lled with remorse, then
joins him in death.
During the three years after the meetings of the Stgemann circle, Mller, apparently intr igued by the possibilities of the story of the miller maid and her suitors,
went on to create his own version of the stor y, consisting of twenty-ve poems
under the title Die schne Mllerin (Im Winter zum lesen), which was rst published
in 1821.12 The cycle consists of a verse prologue (in the voice of the poet), twenty11. My account of Die schne Mllerin follows Susan Youenss excellent work on the or igins and historical context of the cycle, which is reported in three of her publications:Youens, Behind the Scenes:
Die schne Mllerin before Schubert, 19th Century Music 15 (1991): 3 22; idem, Schubert: Die schne Mllerin, Cambridge Music Handbooks (Cambr idge: Cambridge University Press, 1992); and idem, Schubert,
Mller, and Die schne Mllerin (Cambridge: Cambridge University Press, 1997). Another important
source is Graham Johnsons copious notes for his recording of the cycle. See Johnson, Die schne Mllerin, The Poems and Their Backg round, in Die schne Mllerin, D795 (Hyperion, [CD] CDJ 33025,
1996).
Although the works I discuss here, both literary and musical, use trockne in their titles, the proper
word is trockene, which may be contracted to trockne. For the sake of consistency, I shall use
trockne throughout.
12. The cycle initially appeared as part of the fancifully titled Siebenundsiebzig Gedichte aus den hinterlassen Papieren eines reisenden Waldhornisten. The edition of Die schne Mllerin upon which I have relied
and that appears below is from Wilhelm Mller, Vermischte Schriften, ed. Gustav Schwab (Leipzig: F. A.
Brockhaus, 1830), 1: 1 56.

word s, mu s i c, and s ong

three poems proper to the drama, and a verse epilogue (again in the voice of the
poet), together intended to take about an hour to perfor m.13 In poems 1 5, we are
introduced to the miller, the millstream, the miller maid, her father, and the group
of young workers in his charge. In poems 6 11, we become aware of the millers
growing infatuation with the maiden and the depths of his unrequited love. In
poems 12 14, the miller becomes convinced that the miller maid is, as he says in
poem 12, Mine! Despite this belief, there is absent from the poetic nar rative any
explanation for what changed his perspective from one of dreamy mooning to a
self-assured sense of certain love; the poems provide no evidence that the miller
maid has actually returned his love. In poems 15 17, the hunter suddenly enters the
scene and, in the confused understanding of the miller, steals the miller maid away.
Just as suddenly, the bliss of the preceding poems is replaced by anger, jealousy, and
vindictiveness. In poems 18 21, the miller, torturing himself with visions of the
hunter and miller maid in fond embrace, gradually begins to dwell more and more
on death. It is in poem 21, Trockne Blumen, that the miller decides to commit
suicide, bringing to culmination the thoughts of death and the grave that came to
dominate the preceding poems. Poem 22 is a sad dialogue between the miller and
the stream, and poem 23 is given to the stream alone, as it gently and eer ily caresses
the drowned miller. The cycle concludes with an epilogue, which is cur iously
detached from the passion of the preceding poems, and which attempts to build a
bridge from Mllers intense world of fantasy back to the world of the everyday.
The basic stor y told in Trockne Blumen, the text and a translation of which
are given here, can be seen to proceed in three stages.14 In the rst three stanzas, the
miller addresses a collection of withered, moribund owers that are unaccountably
wet. In the second three stanzas, he contemplates the ineffectuality of his tears (the
tacit cause of the wetness of the owers) and the insignicance of his existence in
the face of love that has died. This group ends with a stanza that echoes the rst
stanza of the poem, both in its return to the image of the faded owers and in the
recurrence, in reverse order, of the rhyming pair gab and Grab. Together, these
suggest a circular return that closes off the rst six stanzas. The nal stage of the
story commences with the last two stanzas of the poem. These speak of a future in
which the miller lies dead in his g rave, in which the miller maid recognizes the
truth of his love, and in which owers bloom once more. The poem thus involves
sharp contrasts: dying owers set against blossoming owers, winter set against
spring, and dead love set against true love.
trockne blumen

withered flowers

Ihr Blmlein alle,

Die sie mir gab,

All you owers

That she gave to me,

13. Mller wr ites in the prologue, As its wintertime I expect you wont regret a br ief hour here
in the countryside (as it is depicted in the cycle of poems). [Erhoffe, weil es grad ist Winterzeit, / Thut
euch ein Stndlein hier im Grn nicht Leid.] Mller, Vermischte Schriften, 1: 1.
14. There are other ways to parse the poem, as we shall see later in this discussion. The analysis I
offer here is simply for the pur poses of drawing out some of the aspects of the poem important for a
number of its musical settings.

247

248

analysis and theory

Euch soll man legen

Mit mir ins Grab.

You shall be laid

With me in the grave.

Wie seht ihr alle

Mich an so weh,
Als ob ihr wtet,
Wie mir gescheh?

Why do you all

Look at me so sadly,
As if you knew
What happened to me?

Ihr Blmlein alle,

Wie welk, wie bla?
Ihr Blmlein alle,
Wovon so na?

All of you owers,

Why so faded, so pale?
All of you owers,
Why are you so moist?

Ach, Trnen machen

Nicht maiengrn,
Machen tote Liebe
Nicht wieder blhn.

Ah, tears do not make

The greenness of May,
They do not make dead love
Bloom anew.

Und Lenz wird kommen

Und Winter wird gehn,
Und Blmlein werden
Im Grase stehn.

And spring will come

And winter will go,
And owers will
Grow in the grass.

Und Blmlein liegen

In meinem Grab,
Die Blmlein alle,
Die sie mir gab.

And owers lie

In my grave,
All the owers
That she gave to me.

Und wenn sie wandelt

Am Hgel vorbei,
Und denkt im Herzen:
Der meint es treu!

And when she walks

Past the mound,
And ponders in her heart:
His love was true!

Dann Blmlein alle,

Heraus, heraus!
Der Mai ist kommen,
Der Winter ist aus.

Then all you owers,

Come out, come out!
May is here,
Winter is over.

One thing immediately evident is that this poem relies, as do most, on conceptual blends.When the miller addresses the owers, for instance, it is not as plants but
as anthropomorphized beings, creatures that can not only look upon the miller but
do so with pathos: Wie seht ihr alle / Mich an so weh, / Als ob ihr wtet, / Wie
mir gescheh? This, of course, is another example of the anthropomorphic blends
discussed in chapter 2. Another blend is set up in the fourth stanza, where Mller
combines concepts related to the seasons and to romantic love. Thus maiengrn
becomes both a harbinger of spring and a sign of loves rst bloom.Yet a third blend
crops up in the seventh and eighth stanzas, when concepts related to the seasons
combine with those of life and death: the consummatory Der Mai ist kommen /
Der Winter ist aus represents both the future tur n of seasons the miller envisions

word s, mu s i c, and s ong

and his rebirth as the eter nal lover of the miller maid. The poem is thus a r ich discourse structure in its own r ight, one with intr iguing possibilities for pair ing with
the discourse structures of music.

Bernhard Kleins Trockne Blumen

The Berlin composer Ber nhard Klein (1793 1832) apparently discovered Mllers
poetic cycle shortly after it was published and set Trockne Blumen as the last of
his ve Lieder und Gesnge mit Begleitung des Piano Forte. This collection was brought
out by E. H. G. Christiani in 1822,15 and the inclusion of Trockne Blumen is all
the more remarkable in that the texts for the rst four songs involve rather conventional stor ies of wanderlust and relatively innocent romantic encounters.
Trockne Blumen, as Klein clearly recognized, is a text of quite another order.
Kleins setting, the score for which is given in example 6.1, is guided by his division of the text into four strophes, each compr ising two stanzas of the poem.
Although the music for each strophe is different, so much mater ial is reused that the
song seems more strophic than through-composed, as the for mal outline g iven in
table 6.1 indicates.
The song begins with a br ief piano introduction that sets out the mater ial used
to begin each of the rst three strophes. This mater ial, labeled A1 in table 6.1, consists of three iterations of a motion from dominant to tonic, culminating in a passage
in parallel sixths that returns to the dominant. The whole is couched in a spare twoand three-voice texture, with a repetitive melody built up from a compact succession of ascending and descending gestures that circulate around C5 and A4.
Although A as tonic is scarcely in doubt, it falls short of attaining a full measure of
stability, given the absence of root-position tonic tr iads in strong rhythmic positions
and the emphasis on the dominant har mony at both the beginning and end of each
A1 section. Sections B1 and B2 consist entirely in prolongations of the dominant,
either by temporar ily transfor ming it into a leading-tone har mony (as in mm.
11 12 and mm. 20 24) or through a pathos-drenched passage that introduces an
unprepared VII # of IV that leads through IV back to VII, and then to V (as in
mm. 13 16 and mm. 25 28). In each of the B sections, the melody struggles
upward to D5 three times, only to fall back either to B4 or CS5. Of these two sections, B2 is slightly more intense. In it, a repeated Fn4 replaces the repeated E4 that
signaled the opening of the B1 section, giving rise to a VII that lasts from m. 20 to
m. 24.
In the third strophe, Klein lightens the mood of the music by shifting to the par15. Mller is the author of all of the poems of Kleins Lieder und Gesnge of 1822; Trockne Blumen
is the only poem Klein selected from Die schne Mllerin. It is unclear what motivated Kleins choice of
poet. In his letter of 15 December quoted above, Mller praised Kleins settings and asked if he would be
interested in setting more of his poetry. Given the tone of the letter, it would seem unlikely that Mller
was an active participant in the selection of texts for the 1822 collection of songs. I should note, as well,
that another Berlin composer, Ludwig Berger, published a setting of Trockne Blumen in 1818; analysis of this work, as well as the context for its publication, can by found in Lawrence Zbikowski, The
Blossoms of Trockne Blumen: Music and Text in the Early Nineteenth Centur y, Music Analysis 18
(1999): 307 45.

249

250

analysis and theory

example 6.1

Bernhard Klein, Trockne Blumen

Andantino

A1

Voice

Ihr Blm lein

Piano

B1

al

le die sie mir gab,

euch soll man le gen mit mir ins Grab.

Wie

seht

ihr

10

al

le

mich

an

so

weh

als wenn ihr w tet

wie mir ge

A1

15

scheh?

Ihr Blm lein al

le wie welk, wie bla?

Ihr Blm lein

allel major. The repeated notes that lent an air of intensity or obsession to the dominant prolongations of the preceding B sections now reappear to keep the music
from stalling on the rst and second beats of each bar of section A2. Section B3
begins with another prolongation of the dominant, but instead of being transformed into a leading-tone har mony in m. 36, as it had at the analogous juncture in
sections B1 and B2, it now remains a dominant, leading initially to a I in m. 38, and
then to the rst cadence on A thus far. With the ar rival on the I of m. 38, the
melody nally reaches E5, preparatory to its ar rival on A4 in m. 40.

example 6.1

(continued)
B2

19

al

le wo von so

na?

Ach,

Tr

nen

ma

chen

nicht

23

mai

29

en

grn,

ma chen to

te

Lie

be

nicht wie der

blhn!

A2 Poco pi lento

Der Lenz wird kom

men, und Win ter wird gehn

und Blm lein

B3

32

wer den im Gra se stehn,

und

Blm

lein

lie gen

37

in

mei

nen

Grab

die Blm lein al

le

die

sie

mir

gab.

Und

wenn

sie

(continued)

example 6.1

(continued)

42

wan

delt

am

gel

vor

bei,

und

denkt im

Her

zen

A3

47

der meint es

treu!

dann Blm lein al

le her aus her

men

ter

aus!

der

51

Mai

ist

kom

der Win

is

aus

55

Mai

59

ist

kom

men

der

Win

ter

ist

aus.

der

word s, mu s i c, and s ong

Table 6.1.

Formal Outline of Bernard Kleins Trockne Blumen

Section and Key

Introduction
Strophe 1

Strophe 4
Postlude

Musical Material and Melody

Measures

Stanza 1
Stanza 2
Stanza 3
Stanza 4
Stanza 5
Stanza 6
Stanza 7
Stanza 8

A1 , a
A1, a
B1, b
A1, a
B2, b
A1, a
B3, b'
C, c
A2, a'
A2, a' (fragment)

04
48
816
1620
20 28
2933
3340
40 48
4858
58 64

A minor

Strophe 2
Strophe 3

Text Stanza

A major

The C section beg ins another prolongation of the dominant, one whose
repeated notes are strongly reminiscent of sections B1 and B2. Absent, however, is
the relatively active bass line that opened each of the latter sections; it is replaced
here by a pedal on E3. The melody, too, has changed, now peacefully circulating
around B4, with unimpeded ascents to E5. After m. 45, the intensity of the section
increases with the cessation of the repeated notes and the introduction of the chromatic second alto line of mm. 46 47. In m. 48, however, the melodic line that
descends from E5 once again comes to rest on a clear dominant har mony, which
leads immediately to the reprise of the A mater ial.
Although section A3 opens with the same gure in parallel tenths that began the
song (adjusted to t the major mode), the gure now leads toward a dominantseventh in m. 50 and then to a relatively extensive prolongation of the tonic in
mm. 51 55. This prolongation prepares the way for the cadence of mm. 56 58
with which the vocal part concludes. A brief piano postlude follows; it is based on
the opening of the A3 section, but it includes a motion from Fn to E that recalls the
minor mode of the rst half of the song.
Thus, Kleins music falls into two distinct parts. The rst part, through m. 28, sets
out an intense and unstable A minor. Relatively lengthy and obvious prolongations
of large-scale har monic dissonances (such as those associated with the dominant
and leading-tone) and sudden inter ruptions mark the texture, as do persistent
repeated notes. The second part, from m. 29 onward, disperses the intensity of A
minor and replaces it with a goal-or iented and (for the most part) orderly A major.
Here there are also prolongations of the dominant, but of a sort that seem to afr m
rather than challenge the viability of tonic.What had been difcult in the rst part
becomes easy in the second; what had been troubled becomes calm.
This music, together with Mllers text, sets up the conceptual integration network (CIN) diagrammed in gure 6.1. Two observations should be made about the
network represented by this diag ram, observations that extend to all of the CINs
discussed in this chapter. First, as was the case in chapter 2, the diagram represents
a kind of analytical snapshot of the elements and relations that make up the network, and that contr ibute to the emergent structure of the conceptual blend. A

253

254

analysis and theory

Generic space
contrasting
ontological
states

Music space
A minor: diminishedsevenths, D5 ceiling,
lack of cadence
A major: cadences,
E5, F#5, limited
chromaticism

Text space
present: dying flowers,
tears, hopelessness
future: spring,
blooming flowers, hope
for the future

Blended space
obsessive, mercurial
present
orderly, calm future

figure 6.1

CIN for Kleins Trockne Blumen

fuller account of the way concepts combine within this network would require a
series of such snapshots. Second, what I am concer ned with here is the overall discourse set up by the text and the overall discourse set up by the music. Although
there are interesting details at more local levels, I am most intr igued by what conceptual blending can tell us about song and by what song can tell us about conceptual blending. Let me expand on this for just a moment before returning to Kleins
Trockne Blumen.
The theory of conceptual blending assumes that there are structural invariances
between the input spaces of a blend: these invariances, encapsulated in the elements
and relations of the gener ic space, are what make conceptual blends possible. In the
case of songs, the invariances are between the mental space set up by the text and
the mental space set up by the music. Put another way, the fact that combinations of
text and music can give rise to conceptual blends suggests that there are syntactic
correspondences between linguistic and musical discourse. Although the syntax of
most nineteenth-century German poetry is usually clear at or below the level of the
stanza, the syntax of songs from the same milieu needs to be conceived in ter ms of
the song as a whole, which was construed as a single, if often extended, musical
utterance. Because of this, I nd it most useful to focus on conceptual blends that
encompass the entirety of a given song that is, blends that ar ise from the overall
discourse set up by the text and the overall discourse set up by the music.16
16. This perspective also leads me to regard the poem set by the composer as a text, rather than a
poem. Susan Youens, for different reasons, advocates a similar approach; see Youens, Retracing a Winters

word s, mu s i c, and s ong

What nineteenth-centur y songs can tell us about conceptual blending stems
from the character istic of musical syntax I have just descr ibed. The blend set up by
a song (assuming that there are structural cor relates between the mental spaces set
up by the text and the music) spans at least two or three minutes, involves two different media (with two different, if related, kinds of syntax), and engages us on both
emotional and intellectual levels. All these features have the potential to add immeasurably to our understanding of conceptual blending as a general cognitive
process.17
With these thoughts in mind, let us return to the CIN set up by Kleins song.
The gener ic space for the CIN is structured around the notion of two linked but
contrasting ontological states. Such states are often temporally indexed (as past and
present or as present and future), but they can be more abstractly represented as
theme and variation, or prototype and copy. The delineation of two ontological
states often provides the basis for a more extensive interpretive framework. This
framework then fur nishes an account of the cause for the change of state or an
explanation for why one state is to be preferred over another.
Contrast between the two ontological states is supported in Mllers poem by an
emphasis on the present in the rst four stanzas and an emphasis on the future in
the second four stanzas. There is also far less activity descr ibed in the rst four stanzas than in the second four. Of course, there are exceptions to both of these generalizations. In the rst half of the poem, the future is summoned at the end of the
rst stanza and the past is alluded to in the fourth; the gaze of the owers and the
tears of the miller suggest that the scene is not wholly devoid of activity. In the second half of the poem, the past is evoked briey in the last two lines of the sixth
stanza, the rst two lines of which call up the embodiment of lifelessness.
In the mental space of the music, the linked but contrasting ontological states of
the gener ic space are set up by the A minor and A major music of the two halves
of the song. These states are linked less by temporal index than by conceptual frame:
the A minor music functions as the theme, as it were, and the A major music as the
variation. Nonetheless, the A major music rewrites the character and syntax of the
A minor music. It exchanges the third and sixth scale steps of minor for those of
major, provides the rst cadences on tonic, avoids the fully diminished seventh
chords omnipresent in the rst half of the song, and allows the melody to ascend
to E5 and FS5. As a result, it enacts a specic shift of time frame by placing the A
minor music r mly in the past and announcing the A major music as the new present. This shift is underlined by the mir rored processes that occupy each half of the
song: over the course of the rst half of the song, the music becomes more intense

Journey: Schuberts Winterreise (Ithaca, N.Y.: Cornell University Press, 1991), 307 12. As will also become
clear, I am typically concer ned with the story told by the text as a whole, a parabolic level discussed by
Mark Turner in The Literary Mind (New York: Oxford University Press, 1996), esp. chap. 1.
17. Although it should be clear from the foregoing, let me emphasize that not all songs will set up
conceptual blends. Among those songs that do set up blends, some blends will be more successful (that
is, lead to more extended elaborations) than others. One example of the latter situation is provided by
the third phrase of Brahmss In der Fremde; see the discussion of the song later in this chapter.

255

256

analysis and theory

and darker; over the course of the second half, the music becomes less intense (with
the exception of the poignant recollection of mm. 46 47) and br ighter.
Correlations between these idealized representations of contrasting ontological
states provide an opportunity for the combination of concepts in the blended space,
which draws upon elements of the input spaces to create a portrayal of the millers
transfor mation from a tortured soul into a hopeful and ecstatic one. As was the case
in chapter 2, the emergent structure of conceptual blends (including the present
one) can be descr ibed in ter ms of the processes of composition, completion, and
elaboration.

composition The combining of elements from the text-space and the

music-space is quite straightforward. The enervated present of the rst four stanzas
and the A minor music combine to create a representation of one ontological state
in the blended space: a stark image of the present as dominated by obsessive
thoughts, scant opportunity for rest, and a mercurial play of emotions. The active
future of the second four stanzas and the A major music combine to create a representation of the second ontological state: what was shadowy has become clear;
there is an orderly progress toward goals; and the sur roundings are, almost without
exception, calm and soothing. The transfor mation of one state into the other occasions a strange conceptual inversion in this part of the song: the owers that blossom after the change to A major in m. 29 seem to be the same ones that the miller
predicts will lie in his grave (cf. mm. 31 33 and mm. 34 40), even though the latter event, according to the logic of the poem, must precede the for mer.
completion

Building on the context provided by this blend of concepts, we

can imagine outward signs of the millers torment in the rst half of the song a
sluggish, depressed body language, accompanied by sudden expressions of anguish
as his agony increases and how, in the second half of the song, these give way to
gestures and behavior associated with peacefulness and happiness as the miller contemplates his future.We can also imagine that, overall, the miller grows happier and
more at peace with himself as he becomes more distant from his earlier tor ment.
This inference is supported in part by the music of the concluding A3 section,
where the melody breaks through to the high FS5 of m. 49, a moment followed by
the long prolongation of tonic in mm. 51 55 that prepares the singers nal
cadence.

e laboration

Elaboration of the blend suggests that the miller will act in

order to make the future he imagines a reality. His Heraus, heraus! thus calls forth
not only the futures owers but also the future itself. Elaboration may also provide
an explanation for how the state of affairs portrayed in the rst half of the song is
transfor med into the state of affairs portrayed in the second half . Here the music
provides a useful guide, for it suggests that the present can be transfor med into the
future by reinterpreting it, just as the musical mater ials from A minor were reinterpreted in A major. Thus, the dead love of the present will become the eter nal love
of the future; the dying blossoms will become living ones. The strange celebration
that is enacted in mm. 34 40, as the miller contemplates the owers that will be laid

word s, mu s i c, and s ong

in his grave, then becomes more comprehensible, for in the second half of the song

the withered owers that he holds are transfor med into symbols of his eter nal love.

Franz Schuberts Trockne Blumen

Although the exact circumstances of Franz Schuberts encounter with Mllers Die
schne Mllerin are not known, it appears that Schubert came upon the poetic cycle
sometime in the rst part of 1823 and set twenty of the twenty-ve poems dur ing
the second half of the same year. The completed cycle was published the following
year by the r m of Sauer and Leidesdorf.Trockne Blumen, the score for which is
given in example 6.2, is the eighteenth song in the cycle. Schuberts setting divides
the eight stanzas of the poem into a 3 + 3 + 2 grouping. The music in E minor for
the rst three stanzas is repeated for the second three, after which there is a shift to
the parallel major for the nal two stanzas. Schubert does not repeat any text in his
setting of stanzas 1 6, but he does repeat the last two stanzas in the patter n 7 8,
7 8, 8.
The song begins with the most minimal of piano introductions: two bars of a
repeated E minor chord, played piano, each eighth-note chord followed by an eighth
rest. This patter n continues as the voice enters with a bare melody outlining E
minor. With the shift of the bass from E3 to D3 in m. 5, the E minor chord is transformed into a G chord in position. The latter leads to a cadence on G on the second beat of m. 6, coincident with the conclusion of the rst stanza of text. The
music of mm. 3 6 is then repeated for the second stanza in mm. 7 10, sustaining
the spare effects created by the minimal melodic and har monic mater ials, but
adding the barest hint of forward motion when the voice closes on B4 instead of on
G4, as it had in m. 6.
In m. 11, the music becomes more active and its har monic mater ial somewhat
more variegated: the rhythmic patter n of the accompaniment is livelier; the harmonies change every beat; and the melody, as it follows these har monic changes,
becomes more animated. The initial har monic strategy for the setting of the third
stanza is, in compressed for m, similar to that used for the rst two stanzas: E minor,
now preceded by its dominant in m. 11, leads again to G in m. 12. Measure 11 is
repeated in m. 13, but this time it is followed by an Italian augmented-sixth chord
that leads to the dominant of E minor, while the singer reaches past E5 for the rst
time to grasp a high FS5. The piano then echoes this gesture, sustaining the dominant chord for a half cadence on B that lasts through m. 16. Schuberts setting of
stanzas 4 6 in mm. 17 29 is virtually identical to the music he provided for the rst
three stanzas; the only signicant change occurs in mm. 28 29, where the E5-DS5FS5 of the melody is restated as FS5-E5-DS5.
With the shift to the parallel major in m. 30, the accompaniment becomes yet
more active: Schubert adds a lightly tr ipping bass melody to the animated rhythmic patter n of the right hand. In m. 31, the singer takes up the bass melody, in
modied for m.With the momentar y shift toward CS minor in mm. 33 34, however, the melody becomes streamlined once again. The CS minor chord in m. 34
yields in the following measure to another augmented-sixth chord. This leads (in
m. 37) to a cadential , which prepares the rst cadence on E in the song. Schubert

257

example 6.2

Franz Schubert, Trockne Blumen, from Die schne Mllerin, D. 795/18

Ziemlich langsam
Voice

Ihr Blm lein al

le, die sie

mir

gab,

euch

le mich an

so

weh,

als

Piano

soll

man le gen mit mir

ob

ihr w

ins

Grab. Wie seht

ihr al

tet, wie mir ge

scheh? Ihr Blm lein

al

le, wie welk, wie bla? ihr

13

Blm lein al

le, wo von

so na?

Ach Tr

nen ma chen nicht

18

mai en grn, ma chen to

te Lie be nicht wie der blhn, und Lenz wird kom men, und

258

example 6.2

(continued)

22

Win ter wird gehn, und Blm lein wer den im Gra

se stehn, und Blm lein lie

gen in

26

mei nem Grab,

die Blm lein

al

le, die sie

mir gab!

30

Und wenn

sie wan

delt am H

gel vor bei

und

33

denkt im Her zen, der meint es treu! dann Blm lein al

le, her aus,

her aus,

der

37

Mai

ist kom men, der Win ter ist aus.

Und wenn

sie wan delt am

(continued)

259

example 6.2

(continued)

41

gel vor bei

und denkt

lein al

le, her aus,

im Her

zen, der meint

es treu!

dann

44

Blm

her aus,

der Mai

ist kom men, der

47

Win

ter ist aus,

dann Blm

lein al

le, her

aus,

50

Mai

ist kom men, der Win

ter ist aus.

53

dim.

her aus,

der

word s, mu s i c, and s ong

immediately repeats the whole of mm. 30 38 in mm. 39 47 and then, without
pause, repeats the last four bars, with a change of bass and a slight thickening of texture. A six-bar piano postlude, reprising the accompanimental gure of m. 30, concludes the song. In the course of the postlude, the register drops an octave, and the
mode shifts to minor for the last four bars.
Although this descr iption cannot do it justice, the effect of the relatively spare
musical mater ials that Schubert employs in this setting is profound. A somewhat
more nuanced view of the overall musical discourse can be gained through a middleground sketch of the voice-leading of mm. 1 16 and 30 38, given in example
6.3. As the sketch shows, in the rst portion of the song the repeated motions
toward G major simply prolong E minor and the initial B4 of the Anstieg. It is only
in m. 14 that the music nally begins to move, wrenched from its enervated reiteration of tonic (or the tonic-substitute mediant) by the augmented-sixth chord that
leads to the dominant. The augmented-sixth chord also provides the impetus for
the continuation of the Anstieg, which is temporar ily left hanging on the FS5 over
the dominant. Given the long delay of the continuation of the Anstieg, almost the
entirety of mm. 1 29 remains stalled on a prolongation of tonic and B4. Only at the
half cadences of mm. 15 16 and mm. 28 29 does the music move away from B4
and momentar ily struggle upward toward the Kopfton. With the shift to the parallel major in m. 30 the oppressive harmonic stasis of the rst half of the song lifts, and
the music begins to move forward, pulled along by the extended dominant preparation of mm. 33 37. Finally, in m. 37, the voice achieves the Kopfton, but as GS5
rather than Gn5. This accomplishment is followed immediately by the descent of
the fundamental line, supported by the rst cadence on tonic of the entire song.
Schuberts music falls into two distinct parts in a way that parallels Kleins setting of Mllers poem. The rst part, up to m. 29, presents a minimal, washed-out

example 6.3
Blumen

Middleground sketch of mm. 1 16 and 30 38 of Schuberts Trockne

11

30

35

10

10
5

8
6
4

7
5
3

261

262

analysis and theory

Generic space
contrasting
ontological
states

Music space
spare, static E minor,
lurch to half cadence,
arrival on F#5
flowing, directed E
major, completion of
fundamental line

Text space
present: dying flowers,
tears, ineffectuality,
inevitability of seasons
future: Spring,
blooming flowers,
transcendance through
death

Blended space
miller 1numb,
hopeless
miller 2fevered,
active

figure 6.2

CIN for Schuberts Trockne Blumen

E minor inter rupted twice by rather overwrought cadences on the dominant. The
second part, from m. 30 on, moves forward, and without so much as a backward
look culminates in author itative and emphatic cadences on E major.
This music, together with Mllers text, sets up the CIN diag rammed in gure
6.2. The gener ic space, like that for Kleins Trockne Blumen, is structured around
the notion of two linked but contrasting ontological states. However, where in
Kleins setting the emphasis is on connections between the ontological states, in
Schuberts setting the emphasis is on their incommensurability. Kleins approach
takes the point of view of an observer: the undesirable situation represented by the
rst ontological state can be improved by reconstructing it in ter ms of the second.
Schuberts approach takes the point of view of a participant: recognizing that the
rst ontological state offers no hope, it is simply abandoned in favor of the second.
The change of emphasis in the gener ic space reects a reinterpretation of the
elements of the textual space, prompted by Schuberts division of the text into a
group of six stanzas followed by a group of two stanzas (which are subsequently
repeated). While the ontological states can still be character ized as present and
future, the index is now emotional rather than temporal. The present is still occupied with dying owers, thoughts of the grave, tears, and dead love, but it is also suffused with a sort of numbness, born of the millers inability to br ing about any
change in his circumstances. He barely realizes that the tears that moisten the dying
owers are his own; he is powerless to stop them, just as he is powerless to stop the

word s, mu s i c, and s ong

coming of the seasons, or of his own death. Thus the future of Und Lenz wird
kommen / Und Winter wird gehn is contained within the present, a leaden emotional domain circumscr ibed by the echoing rhymes of the rst and sixth stanzas.
By contrast, the future state called forth by the last two stanzas of the poem is
devoid of such dark thoughts: here the miller maid wanders past the resting place of
the miller; here spring has truly arrived, and the winter of despair has been banished.
Contrast between ontological states is also apparent in the musical space. Schuberts E minor music is, for the most part, spare and static. After m. 11, it gradually
pulls itself into an agitated state, lurches into a half cadence, and then returns to its
barren beginning. His E major music, in contrast, exploits and adds to the quickening of pulse that precedes the half cadences of the E minor music, and it dr ives
toward a consummatory cadence that effaces the equivocation and ir resolve of the
rst half of the song. This is music of possibility and completion, of action.
Correlations between these idealized representations of contrasting ontological
states provide an opportunity for the combination of concepts in the blended space,
which draws on elements of the input spaces to create a portrayal of two very different personae for the miller.

composition Composition provides the concepts basic to this portrayal.

The E minor music and the rst six stanzas of text combine to summon a miller
who sits in numb contemplation of what seems to be a hopeless predicament.When
the gravity of his situation seizes him, he stirs for a moment (responding to the
augmented-sixth chords), but even the intensity of torment is denied him and he
sinks back into his tor por. The E major music and the last two stanzas portray a
quite different miller who, with growing condence, reaches out toward goals
most particularly, achieving immortality through true love beyond the grasp of
the millers rst persona.
completion Through completion we can build on what we know about
human behavior in such extreme states: we can imagine that in his rst persona the
miller is virtually inert until, at the moment of his anguish, he half rises, a look of
panic on his face; similarly, we can imagine the expansive gestures of the miller in
his second persona and perhaps also see the somewhat fevered look in his eyes as
he proclaims Der Mai ist kommen / Der Winter ist aus.
elaboration While the emergent structure produced by composition and
completion offers a wealth of material to the imag ination, it is elaboration that
offers the most str iking inter pretive possibilities. Elaboration of the blend suggests
that the transition from the rst persona to the second marks a decisive turn in the
millers story.With this transition, the miller has moved beyond merely entertaining
the thought of death and has now accepted the fact of it. The miller will overcome
his ineffectuality, he will transfor m himself, and he will do so through suicide. Supporting this inter pretation are the almost frenetic repetitions of the second half of
the song (as though the miller still needs to convince himself of the necessity of his

263

264

analysis and theory

course of action) and the sad tur ning toward minor of the postlude as the enor mity
of his decision begins to sink in.18

Two Settings, Two Millers

These two settings of Mllers Trockne Blumen give us two rather different
millers.We rst come upon Kleins miller tur ning over the shards of his shattered
love affair and becoming more distraught as he realizes the hopelessness of his
predicament. Then there is a sudden shift, and what was dark and foreboding
becomes light and promising. The millers burden has been lifted, for he has found
a way to breathe life into dead owers and dead love: he will welcome death rather
than fear it. In contrast, Schuberts miller has two str ikingly different personae. The
rst is emptied out, nearly paralyzed by despair, numbly aware of his misery and his
inability to escape it. The second is effusive, expansive in welcoming a future in
which the miller is no more. Schuberts miller is, in short, psychotic: at rst immobile in the face of death and then feverishly and joyously embracing it.
Are both millers present in Mllers poem? Yes and no.Yes, in that his poem can
support these different inter pretations, as the CINs above have shown. And yes, in
that, from the perspective of the poetic cycle as a whole, the miller is a truly complex character, a quiet psychopath destined for destruction. But no, in that it
requires the agency of the music to br ing forth these specic millers.
The differences between these two millers suggests that a review of the different compositional strateg ies that brought them into being is warranted so that we
may better understand the role of musical syntax in breathing them to life. The similarities between these two settings of Mllers poem are str iking: both employ a
shift from minor to major mode; both avoid any denitive cadence on the tonic in
the rst portion of the song; and both make use of relatively delicate textures, especially in the rst half of the song.19 More distinctive are Kleins and Schuberts
respective treatments of dominant and tonic. Klein, through extended prolongations, employs the dominant as a tension-generating device, and in consequence the
dominant receives more emphasis than the tonic. By contrast, Schubert is spar ing
with the dominant of his home key, reserving it as a device to pull the music temporarily out of its stasis (in the rst half of the song) or using it as part of a nor mative cadence (in the second half of the song). Tonic, by compar ison, is given quite
a bit of emphasis, so much so that the rst portion of the song becomes threatened
with immobility.
The conceptual blends created by Kleins and Schuberts songs, then, are not sim18. Further evidence for the moment of this decision is provided by the poetic cycle as a whole. In
all of the poems up to Trockne Blumen, the millers presence is marked by the use of the rst person.
After the shift to the third person in Trockne Blumen, the miller never appears in the rst person
again.
19. In addition to the similar ities between Kleins and Schuberts settings of Mllers poem, both also
share features with Ludwig Bergers 1818 setting of the poem, which was mentioned in n. 15; the common features are discussed in Zbikowski, The Blossoms of Trockne Blumen, 335. Although such
similar ities are provocative, no research of which I am aware traces specic patter ns of inuence between
Berger, Klein, and Schubert.

word s, mu s i c, and s ong

ply a matter of words and music happening at the same time (and thus being combined by dint of sharing the same temporal frame). For blending to take place, the
mental spaces set up by the text and the music must have a shared topography,
reected in part in commonalities between the syntactic structures of the two
media. This shared topography does not exhaust the resources of the input spaces.
Indeed, the difference in syntax important to Nicholas Cooks conception of multimedia (discussed at the beginning of this chapter) is no less important in conceptual blending, for it ensures that the stor ies told by both music and text are compelling in themselves. Nonetheless, common structural features (which may obtain
at a different level of structure than those that lead to contest between media) are
necessary for conceptual blending to take place.
In the settings of Trockne Blumen, words and music work together to tell stories that enhance and expand on structural features inherent in the source spaces for
each conceptual blend. I now tur n to a song by Robert Schumann in which the
mental space set up by the music makes possible a conceptual blend whose emergent structure includes a nar rative considerably more extensive than that suggested
by the text alone.

im rhein
In 1840, having established himself as a composer of works for the piano, Robert
Schumann tur ned his energies to song. And for midable energies they were, for by
the time the year had ended, he had wr itten over 140 songs. Schumann quite obviously worked at speed: the twenty songs in the or iginal version of his Dichterliebe
cycle (four were cut before its publication) were written from 24 May to 1 June.
Nonetheless, among the songs wr itten that year are some of the most profound and
haunting of the nineteenth century, in part because Schumann found ways to combine words and music to create fantastic worlds that extend far beyond those that
can be summoned by either words or music alone.

Heines Im Rhein
Im Rhein, the text and translation of which are given here, is the sixth of sixteen songs that make up Schumanns Dichterliebe, rst published by C. F. Peters in
1844.20 The text for all of the songs was taken from Heinr ich Heines Lyriches Intermezzo, a collection of sixty-ve poems published in 1823. Although usually ter med
a song cycle, Dichterliebe is held together not so much by common nar rative as by
theme: all of the poems deal with romantic love, which is transfor med, splintered,
or distorted by reminiscence or overheated emotions.
im rhein

in the rhine

Im Rhein, im heiligen Strome,

Da spiegelt sich in den Welln,

In the Rhine, in the holy r iver,

There, reected in its waves,

20. The text for Im Rhein given here is from the song as it appears in Clara Schumanns edition
of the Lieder of Robert Schumann, but with text repetitions omitted.

265

266

analysis and theory

Mit seinem groen Dome

Das groe, heilige Kln.

With its great cathedral,

Is the great holy city, Cologne.

Im Dom, da steht ein Bildnis,

Auf goldenem Leder gemalt;

In meines Lebens Wildnis

Hats freundlich hineingestrahlt.

In the cathedral there hangs a portrait

Painted on gilded leather;
Into my lifes wilder ness
It has shone down kindly.

Es schweben Blumen und Englein

Um Unsere Liebe Frau,

Die Augen, die Lippen, die Wnglein,

Die gleichen der Liebsten genau.

Flowers and cherubs oat

Around Our Dear Lady.
Her eyes, her lips, her cheeks
Are exactly like those of my love.

The three stanzas of Im Rhein can be divided into two parts. The rst part,
which compr ises the rst stanza, summons a broad view of Cologne centered on its
cathedral and reected in the waves of the Rhein. There is a sense of vastness to the
perspective (suggested by the repetitions of gro ), and of the holy (provided by the
cathedral and the descr iption of Cologne as the holy city).21 The second part of
the poem, which compr ises the second and third stanzas, moves inside the cathedral
to focus on a portrait within. It is at this point that the presence of a nar rator is
made explicit and the shifting perspective of the poem acquires an anchor and a
rationale.We are here in this city, inside this cathedral, to view a painting of singular importance to the nar rator, for its radiance has proved a balm to the troubles of
his soul. In the third stanza, the poetic lines become longer, and the nar rator takes
us inside the painting itself . He begins by describing a heavenly vision of the
Madonna, but as he takes us deeper into the painting, the image of the Madonna
becomes ever more worldly, until it becomes apparent that the nar rator savors this
image because it is the exact likeness of his beloved. The poem ends with this sudden shift from the spir itual to the car nal, and we hang in mid-air, suspended both by
and within the nar rators vision.
The elegance and power of Heines unsettling poem stem from two parallel
processes. The rst is perspectival: we start from the largest scale the city of
Cologne and its cathedral and then nar row the scope, rst to the inter ior of the
cathedral, then to a painting within, and then to specic features of the face portrayed in the painting. The second process is conceptual and affectual: we begin
with ideas about vastness and holiness; proceed to the active solace provided by the
contemplation of a religious work of art; and then are suddenly confronted with an
intense, carnal passion for a lover.

Schumanns Im Rhein
Schumanns setting of the poem Im Rhein, the score for which is given in example 6.4, is in ABA' form. The rst sixteen measures, which make up the A section,
present a forceful, leaden E minor. Beginning forte, the voice (accompanied by the
21. Schumann provided additional emphasis on the holy by changing the end of Heines rst line
from im schnen Strome to im heiligen Strome.

Robert Schumann, Im Rhein, from Dichterliebe, Op. 48

example 6.4

Ziemlich langsam
Voice

Im Rhein,

im hei

li gen Stro

me,

da

spie

gelt

Piano

sich

in den Welln,

me,

das

mit

sei

li ge

Kln.

nem gros

sen

Do

12

gros

se,

hei

Im

Dom,

da

18

steht

ein Bild

mei

nes Le

niss, auf gol de nem Le der ge malt;

in

23

bens Wild

niss hats freund lich hin ein ge strahlt.

(continued)

example 6.4

(continued)

28

Es schwe ben Blu

men und

33

Eng

lein um un

sre

lie

be Frau;

38

die Au

gen, die Lip

ritard.

Lip

pen, die Wng

lein,

die glei

chen der Lieb

ritard.

sten

43

48

53

ritard.

ge

nau.

pen,

die

word s, mu s i c, and s ong

left hand of the piano in octaves) makes a stepwise ascent from E4 to G4 in m. 4,
answered by a stepwise descent from C5 back to E4. At this point, voice and piano
part ways.While the voice repeats its melodic mater ial almost exactly (save for the
crucial change of the penultimate FS4 for Fn4), the piano continues the descent initiated in m. 4, sinking to the G1-G2 octave of mm. 11 13. In the second half of
m. 13, the bass echoes the voices descent from C to G of mm. 12 13 but, prompted
by the important Fn4 in the voice, it arrives on a G1S-G2S octave instead. This
arrival coincides with the nal E4 in the voice part and with the completion of the
rst stanza of the poem. The right hand of the piano lls out, if minimally, the bare
texture created by the voice and the left hand; it does so with an accompaniment
made up of cascading gestures in a halting patter n of alternating eighth- and dottedquarter-notes.
The dominant of mm. 15 16, with which the A section concludes, leads to the
A minor chord and piano of m. 17 that starts the B section. Here, the str iding bass
of the left hand disappears, replaced by a more uid and g ravitationally liberated
adaptation of the eighth/dotted-quarter patter n of the right-hand accompaniment.
In mm. 17 21, the voice also abandons its relentless descents and oats gently
between D5 and G4; both voice and piano momentar ily cadence on G major in
m. 21. Although there is a br ief struggle in mm. 22 25 as the voice counters the
pianos descending gures with an ascent from A4 to E5, once the E5 is achieved
the air clears, and the equanimity of the preceding phrase is restored, an effect
conr med by another cadence on G major in m. 27. There follows a br ief piano
interlude that beg ins a chromatically har monized and rhythmically displaced
descent through the sixth G2 -B1 (the voice-leading of which is shown in ex. 6.5).
The voice joins the piano in m. 31, and the completion of the sixth-descent in
m. 34 is followed by a cadence on A minor in the following measure. An A minor
chord, prolonged through a chromatically inected voice-exchange (as shown in
ex. 6.5), prepares the way for the return to E minor effected by the ar rival on the
B chord of mm. 42 43.
The return to E minor in m. 44 also marks a return to the mater ial of the A section, which (other than the mezzo forte dynamic) is exactly the same as in its rst
appearance until the left hand of the piano reaches the FS2-FS3 octave in m. 49 (as
it did in m. 6 of the A section). At this point, the rhythmic displacements of
mm. 27 41 reappear, and the bass continues, now in augmentation, to the A1-A2
octave of m. 54. The A1-A2 octave then leads, through AS, to the dominant of
mm. 56 57 and the nal cadence.
example 6.5
27

Linear analysis of mm. 27 42 of Schumanns Im Rhein

notated
F

32

35

42

6
4
3

269

270

analysis and theory

Schumanns music thus enacts a threefold process. The rst part of this process
(in the A section of mm. 1 16) establishes an austere, weighty E minor. It is well to
note that the tonality is established with minimal reference to the sort of harmonic
progressions and disjunct bass motions that are typical of tonal music of the per iod:
there is no cadence on E minor in the entire section, and, moreover, the dominant
of E minor is completely absent, its place taken by the leading-tone seventhchord.22 The second part of this process (in the B section of mm. 17 43) moves
completely, if only temporar ily, away from E minor. Here there are denitive
cadences, but on G major and A minor. The large-scale instability of this section (a
product of its avoidance of the tonal center so unequivocally established in the A
section) becomes apparent only g radually. Ultimately, E minor cannot be avoided
forever, and the section concludes (in mm. 42 43) with the very rst appearance
of the dominant of E minor in the entire song. The third part of the process (in the
A' section of mm. 44 58), initiated by the denitive return to E minor, reprises the
opening music. However, the music is now colored by rhythmic displacements that
recall the latter half of the B section and that dr ive inexorably forward to a nal
cadence in the lowest practicable register.
This music, together with Heines text, sets up the CIN diag rammed in gure
6.3. The gener ic space for the CIN is structured around the notion of exterior and
interior domains. In the textual space, the exter ior domain is that summoned in the
rst stanza of the poem: the city of Cologne, with its r iver and cathedral. The interior domain is that summoned in the second two stanzas: the cathedral and the
painting with which the nar rator is obsessed. In the musical space, the exter ior
domain is that of E minor, with its stark octaves and str iding phrase rhythm. The
interior domain is that of the contrasting (or not-E-minor) music of the B section, with its more owing accompanimental patter n, softer dynamic, rhythmic displacements, and large-scale instability.
Correlations between these idealized representations of exter ior and inter ior
spaces provide an opportunity for the combination of concepts in the blended
space, which draws on elements of the input spaces to create a r ich portrayal of the
actions of the nar rator and the environment within which they take place.

composition Composition combines the exter ior view of the river, city,
and cathedral with the austere E minor music of the A section and suggests a
sweeping and somewhat foreboding perspective, surveyed with a deliberative and
relentless gaze. In turn, the inter ior view of the cathedral and the painting within
combines with the soothing music of the B section to create a sense of refuge, a
surcease from the harsh realities without.

22. It is interesting to speculate on whether Schumann himself would have differentiated between
the dominant and the leading-tone chord. The distinction is of more concer n to theor ists of the period
than to composers. Nonetheless, Schumann would most certainly recognize the shar per dissonance created by the leading-tone chord (with or without an added seventh), and it is worth noting that in the E
minor portions of Im Rhein he favored these over the relative consonance of the dominant chord
until m. 42.

word s, mu s i c, and s ong

Generic space
exterior and
interior domains

Music space
E minor music of A
section (austere,
foreboding)
not-E-minor music of B
section (soothing,
consolatory)

Text space
exterior view of river,
city, and cathedral
interior view of
cathedral, painting,
portrait of Madonna

Blended space
gray, cold (perhaps
threatening) exterior
intimate, secure
(if fragile) interior

figure 6.3

CIN for Schumanns Im Rhein

completion

Completing the image, we can imagine details of each domain:

an exter ior world gray and cold and in some measure threatening; an inter ior world
intimate and secure, if somewhat frag ile. We can also imag ine something of the
physical and emotional life of the nar rator. Outside the cathedral, he is almost certainly self-absorbed, a disheveled gure hurrying to escape the demons that tor ment
him. Inside, he stands rapt, even as the tumult in his mind gradually increases to the
point of becoming unbearable.

elaboration Our elaboration of the conceptual blend produced by Schumanns setting of Heines poem might well seek to develop a nar rative that could
explain the forces that dr ive the nar rator into the cathedral: a personality or perspective that makes life in the outside world intolerable, or a deep obsession with
the beloved that renders her both an object of worship and something to be owned
and controlled. However, we could also follow the lead of Schumanns music,
which, in response to the denouement of the poem, returns to the E minor music
of the opening. We are thus returned to the world outside the cathedral and there
can see the nar rator, now more disturbed than ever, stagger ing under the weight of
his obsession with his lover, just as the music staggers with the rhythmic displacements of mm. 49 55. This furtherance of the nar rative by the music suggests that
it is the psychic tor ment of the nar rator that is behind the sense of foreboding summoned by the exter ior domain, as well as for the ir reality of the inter ior domain.
It is an oppressive torment that absorbs all, and among whose res the nar rator

271

272

analysis and theory

wanders, knowing no rest. The elaboration of the blend, then, suggests a profoundly
troubled mind to whom the Rhein represents more than a mir ror for a g reat and
holy place: it also offers the means to extinguish the res that consume him.

Summary
The dark twists and tur ns of love explored in Heines Lyriches Intermezzo are, in their
way, not so very far removed from those Mller captured in his Die schne Mllerin.
However, where Mller develops the topic and its ramications over the course of
twenty-three poems, it is Heines gift to be able to accomplish this in the space of just
one. This suggests a mastery of psychological detail both in understanding his subjects and in what can be asked of his reader that is altogether impressive. Equally
impressive is Schumanns ability to exploit this level of detail by combining Heines
words with music that supports, expands on, and even offers a completion for the
story told by the poem. It might be argued that Schumanns setting str ips some of the
ambiguities from Heines poem, ambiguities that offer even more possibilities for
interpretation than those provided by the song. The point is well taken, for the conceptual blend set up by the song emphasizes certain aspects of the topography of the
text space at the expense of others. And yet the gains are considerable, for the conceptual world that emerges from syntactic invariances between Heines text and
Schumanns music is one r ich in detail and str iking in its immediacy. The challenges
and rewards of creating such worlds no doubt fueled the burst of creativity that led
to Schumanns amazing output of songs in 1840 and that produced a transfor mation
in the way composers thought about how to organize their musical mater ials.

in der fremde
The similar ities between Kleins and Schuberts settings of Trockne Blumen were
striking in that Schubert apparently did not know Kleins setting, and Schubert was
occupied with developing a song cycle as opposed to creating a single Lied: not
only are the musical details of each setting similar, but the gener ic spaces behind the
conceptual blends that each song sets up are almost exactly the same. In Schumanns
and Brahmss settings of Joseph von Eichendorff s In der Fremde, nearly the exact
opposite set of conditions holds. Brahms knew and respected Schumanns setting,
but he seems to have wanted to show his fr iend and mentor what he could do with
the same mater ial. The result is two songs every bit as different as those of Klein and
Schubert, but whose divergences stem from the gener ic spaces that structure the
blends produced by each song.

Eichendorff s In der Fremde

In der Fremde, the text and translation of which are given here, rst appeared as
the text of an untitled song introduced in the course of Eichendorff s 1833 novella
Viel Lrmen um nichts.23 The scene in the novella is a huge and gloomy hall within
23. Joseph Freiherr von Eichendorff , Viel Lrmen um nichts, ed. Gerhart Baumann and Siegfr ied
Grosse, Neue Gesamtausgabe der Werke und Schr iften (Stuttgart: J. G. Cottasche Buchhandlung Nachf.,

word s, mu s i c, and s ong

a castle; outside, a storm rages, and, tucked away in a window niche, a young woman
sings to her own accompaniment on the guitar. Her song is a deeply sad one, concerned with loneliness and thoughts of a solitary death far from home. The song
over, however, she springs up, full of life, to meet her beloved and his fr iends, who
have chanced upon her musings. The scene within which this poem rst occurs is
thus one of sharp contrasts between the cold and forbidding inter ior of the castle, which makes for scant refuge from the stor m outside, and the blushing and lively
young woman who rushes to embrace her fr iends.
in der fremde

in foreign lands

Aus der Heimat hinter den Blitzen rot

Da kommen die Wolken her,
Aber Vater und Mutter sind lange tot,
Es kennt mich dort keiner mehr.
Wie bald, wie bald kommt die stille Zeit,
Da ruhe ich auch, und ber mir
Rauscht die schne Waldeinsamkeit,
Und keiner mehr kennt mich auch hier.

From home, behind the red lightning

There come clouds,

But Father and Mother are long dead;

No one there knows me anymore.

How soon, how soon will that quiet time come,

When I too shall rest, and over me

Shall rustle the forests beautiful solitude,

And no one here shall know me anymore.

The poem itself offers images and impressions rather than any specic nar rative
structure. Dominant are the images of nature (in the for m of the red lightning,
approaching clouds, and promise of the forests loneliness). Of equal importance is
the nar rators profound sense of abandonment (her father and mother are long dead,
and she is unknown both in her homeland and in her nal place of rest). However,
there are also within the poem shar p contrasts similar to those that accompanied its
introduction in the novella: the comforting images of the homeland, parents, and sylvan setting are juxtaposed with the ambiguous here of the poem, solitude of the
forest, and oppressive anonymity of the present sur roundings. Nonetheless, connections between the stor ies told by the poem and the novella were obscured, and some
of the ambiguities of the poem were muted, when it was placed by Eichendorff s
friend Adolf Schll in the Totenopfer section of an 1837 volume of Eichendorff s
collected poetry, at which time it acquired its present title.24

Schumanns In der Fremde

It was in Schlls 1837 edition of Eichendorff s poetry that Schumann (or his
ance, Clara Wieck) came upon the poem, which he began to set to music on 4
May 1840; the score is given in example 6.6.25 Although there are indications Schu1957), 478. The text provided here is the same as that in the novella; some of Schumanns departures from
Eichendorff s poem are discussed in the next section.
24. For a discussion of Schlls role as editor of Eichendorff s work, see Ansgar Hillach and KlausDieter Krabiel, Eichendorff-Kommentar (Munich: Winkler, 1971), 1: 47 48.
25. For a stimulating discussion of the songs from Schumanns Op. 39, their ordering in the cycle,
and their relationship to the 1837 edition of Eichendorff s poetry, see Patrick McCreless, Song Order in
the Song Cycle: Schumanns Liederkreis, Op. 39, Music Analysis 5 (1986): 5 40. Dates of the composition
of the songs are given in Robert Schumanns Liederbcher, which are reproduced in Viktor Er nst Wolff,
Lieder Robert Schumanns in ersten und spteren Fassungen (Berlin: H. S. Hermann, 1913), 12 19. The same

273

example 6.6

Robert Schumann, In der Fremde, from Liederkreis, Op. 39 No. 1

Nicht schnell

Voice

Aus der Hei

mat hin

ter den Bli

tzen rot

da

Piano

Mit Pedal
4

kom men die Wol

ken her.

A ber Va

ter und Mut

ter sind

lan

ge tot,

es

kennt mich dort kei

ner mehr.

Wie

10

bald,

ach wie bald

kommt die

stil

le

Zeit,

da

12

ru

he

ich

274

auch,

da

example 6.6

(continued)

14

ru

he

ich

auch

und

ber mir rauscht die

17

sch

ne

sch

ne Wald

Wald

ein

sam

keit,

die

20

ein sam keit,

und

kei

23

hier,

und

kei

ner kennt mich mehr hier.

26

275

ner kennt mich mehr

276

analysis and theory

mann intended In der Fremde to be the opening song of his Op. 39 Liederkreis,
when the cycle was rst published in 1842 by Tobias Haslinger, Der frohe Wandersmann (later published as Schumanns Op. 77 No. 1) appeared in its stead.26 It
was only with the second edition of the Eichendorff cycle, published in 1849 by
Friedrich Whistling, that In der Fremde made its appearance as the opening song.
In revisions to the song, Schumann changed Eichendorff s poem slightly to improve
his melody, adding the ach in m. 10 and changing the last line of the poem from
Und keiner mehr kennt mich auch hier to Und keiner kennt mich mehr hier.27
Schumanns setting begins with a owing accompaniment that continues throughout the song. Although there is no evidence that Schumann referred to Eichendorff s novella in the course of setting this poem, some character istics of the music
seem to t the image portrayed there of a woman singing to her own accompaniment on the guitar. The running ar peggios in a relatively low register are typical of
guitar music, and the melody is folk-like in its simplicity. The rst ve measures
present a thoroughly straightforward FS minor, using only tonic and dominant.
When the melody is repeated in mm. 6 9, the har monization becomes a bit r icher
with the fully diminished seventh chords of mm. 6 and 7, but the basic har monic
framework established in the opening measures remains intact. Beginning in m. 10,
however, there is a shift toward A major, and the sound br ightens considerably as the
higher register of the piano is called into service for the r inging B4-FS4 fth of
mm. 10 11. A proper ar rival on the I of A in m. 13 is thwarted by the melodys
temporary escape to E5, and when the outer-voice sixths are retraced in mm. 14 15
(as shown in ex. 6.7), the har mony at the point of arrival is changed to the dominant of B minor (that is, FS major) rather than the tonic of A.28 This not only

sources are cor related with the manuscripts at the Deutsche Staatsbibliothek Berlin in Reinhold
Brinkmann, Schumann und Eichendorff: Studien zum Liederkreis Opus 39, Musik-Konzepte, 95 (Munich:
Edition Text + Kr itik, 1997), 80 86. See also Herwig Knaus, Musiksprache und Werkstruktur in Robert
Schumanns Liederkreis: Mit dem Faksimile des Autographs, Schriften zur Musik, 27 (Munich: Emil
Katzbichler, 1974).
26. Der frohe Wandersmann and why Schumann may have included it in the 1842 edition of the
Liederkreis are discussed in Jon W. Finson, The Intentional Tourist: Romantic Irony in the Eichendorff
Liederkreis of Robert Schumann, in Schumann and His World, ed. R. Larry Todd (Princeton, N.J.: Princeton University Press, 1994), 156 70.
27. A compar ison of the two versions is given in Wolff, Lieder Robert Schumanns, 56 57; the revisions
are not given in the facsimile provided in Knaus, Musiksprache und Werkstruktur in Robert Schumanns
Liederkreis.
28. In my analysis, I have emphasized the sense of departure evoked by mm. 10 14 that is, the
rm motion away from minor and the implied cadence on A major. In a compelling analysis of In der
Fremde, David Ferris discusses Rufus Hallmarks reconstruction of one of Schumanns earlier versions
of the song; in this version Schumann does indeed cadence on A major. See Ferris, Schumanns Eichendorff Liederkreis and the Genre of the Romantic Cycle (New York: Oxford University Press, 2000), 111 13.
Working from a strongly Schenkerian perspective, Ferris also makes a convincing argument that, in ter ms
of large-scale tonal structure, the move to A major in the nal version of the song is, in fact, illusory and
is subordinate to an expansion of the subdominant of FS minor (pp. 98 106). Although Ferriss reading
makes sense of Schumanns overall har monic and contrapuntal plan, it does not preserve the sense that
with the ar rival on FS major the promised cadence on A major has been subverted. It is this sense of subversion that my analysis, focused as it is on a more surface level of harmonic and contrapuntal activity,
attempts to capture.

word s, mu s i c, and s ong

example 6.7

Linear analysis of mm. 12 15 of Schumanns In der Fremde

12

14

10

returns the music to minor but also starts the return to FS, accomplished in the latter half of m. 19 when a B minor chord is transfor med into the leading-tone chord
of CS. The cadence on FS minor in m. 21 is followed by eight measures with an FS
pedal in the bass, over which there are double chromatic neighbors to FS (Gn and
ES), as well as a major third (AS).
This music, together with Eichendorff s text, sets up the CIN diag rammed in
gure 6.4. The gener ic space, like that for Kleins and Schuberts settings of
Trockne Blumen, is structured around the notion of two contrasting ontological
states. Lacking, however, is the notion that there is a smooth or ready connection
between the two states: although there is a sense that they are linked in some larger
scheme, they are portrayed as basically incommensurate.
Contrast between ontological states in Eichendorff s poem is supported by a
marvelously subtle shift of conceptual metaphor.When reference is made to the past
or the present, a spatial mapping is used: the past is there (the dort of the fourth
line), the present is here (the hier of the eighth line).When reference is made to
the future, an affective mapping is used: the future is quiet (the stille of the fth
line) and partakes of forest-solitude (the Waldeinsamkeit of the seventh line).
Each of these mappings is further associated with somewhat more diffuse emotions.
The spatial mappings of the past and present are connected with the poignant
memory of departed parents and the sense of abandonment that comes with disappearing from memory, with the confused or uncomfortable feelings summoned by
the red lightning and approaching clouds, and with the loneliness of an anonymous
existence. The affective mappings of the future are connected with a sense of rest
and of being in the presence of beauty.
Contrast between ontological states can also be seen in Schumanns music. The
FS minor music is, for the most part, resolutely secure, yet it is, nonetheless, tinged
with chromaticism. Thus, in mm. 1 9, regular cadences appear, but also the fully
diminished seventh chords of mm. 6 and 7. Similarly, mm. 20 28 offer the secur ity
of the pedal point, as well as the double chromatic neighbors to FS, which render
ambiguous the assertion of major in the nal measures. The music of mm. 10 19
tends to be br ighter (note the higher register in the piano accompaniment that is
especially audible in the r ight hand of mm. 10 15) and less involved with chromaticism (with the important exceptions of the melodic deection to AS in m. 15
and the fully diminished seventh chord of m. 17). However, in this same passage,
convincing cadences are conspicuous only for their absence.

277

278

analysis and theory

Generic space
contrasting
ontological states

Text space
past, presentspatial
mappings: abandonment
loss, threatening weather
futureaffective
mappings: forest-solitude,
quiet rest

Music space
secure F# minor,
tinges of chromaticism
brighter not-F#minor music, higher
register, lack of cadence

Blended space
matter-of-fact,
melancholy present,
temporarily displaced by
animated, hopeful
future, only to return to
melancholy present

figure 6.4

CIN for Schumanns In der Fremde

Correlations between these idealized representations of contrasting ontological

states provide an opportunity for the combination of concepts in the blended space.
The combination draws on elements of the input spaces to create a portrayal of two
worlds, not totally unlike those of Im Rhein, but framed in emotional rather than
architectural ter ms.

composition

Composition combines the FS minor/major music with the

spatially mapped past and present. Evidence that Schumann was aware of the
importance of maintaining a consistent musical environment for the rst four lines
of the poem is provided by his manuscr ipt, for after wr iting a more varied accompaniment for mm. 6 10 he scratched it out and continued the patter n established
in mm. 1 5.29 The affectively mapped future is rst combined with A major and
then, with the voice still near the top of its range for the song, B minor.With the
repetition of die schne Waldeinsamkeit, however, the melody drops, and we
return to FS and the spatially mapped present.

completion Following the structure of the blend and our knowledge that
a single rst-person voice is being projected, we could imagine a singer who begins
with almost matter-of-fact observations but who gradually becomes more engaged
29. See the unpag inated Faksimile des Autographs appended to Knaus, Musiksprache und Werkstruktur in Robert Schumanns Liederkreis.

word s, mu s i c, and s ong

and more melancholy. This melancholy is then rather abruptly displaced (in m. 10)
by more animated and hopeful emotions, leading the singer to a higher register and
to sing with a fuller voice (in keeping with the profusion of longer note values in
the melody). The hopeful attitude cannot be sustained, however, and solitude once
again becomes loneliness, the present that is reasserted in m. 21 effectively vanquishing the br ight promise of the future.

elaboration

The imaginary world of Schumanns song is one quite different from that of Eichendorff s original poem. The song sung by the young woman
in the window niche is almost folk-like, its regular meter and stylized emotions disturbed only once, by the metr ical awkwardness of the sixth line. But this momentary shadow of emotional depth passes, and it is possible to imagine the singer
jumping up, the guitar and song ung aside, to greet her fr iends. Schumanns singer
will not be so easily released. The almost for mulaic melancholy of the songs opening reveals itself in the end to be inexorable, pulling the energy and hope from the
singer and leaving her hollow and forlor n. Schumanns song leaves us with a singer
for whom eter nity now stretches out as at and colorless as her present, her spirit
struggling as the walls of anonymity close around it.

Brahmss In der Fremde

Johannes Brahms was but nineteen when, in November of 1852, he made his setting of In der Fremde, published the next year by Breitkopf and Hrtel as the fth
of his Sechs Gesnge, Op. 3; the score is given in example 6.8.30 In der Fremde was
the rst of the Op. 3 songs that he composed, and it shows clear debts to Schumann:
Brahms adopted Schumanns version of the text, the key of FS minor, and some
important motivic details. Nonetheless, the song is clearly Brahmss own, and
nowhere is this more apparent than in the conceptual blend it sets up.
After a short introduction that anticipates the vocal melody without directly stating it (by presenting only elements of the piano accompaniment for the melody),
the song proper commences, laid out in four phrases. These phrases have almost
identical beginnings, both in har mony and melody, although each ends rather differently. The rst touches back upon the tonic at its midpoint (m. 6) and then
deects away to A major. Arrival on A (m. 8) is unsettled by the entrance of the
dominant on the second beat (over an A pedal), which supports an echo of the last
three notes of the voice part. The second phrase moves at its midpoint to D major
(m. 12); its cadence (on the downbeat of m. 13) is as unsettled rhythmically as that
of mm. 7 8, but it gains a bit of stability when the fourth line of the poem is
repeated (mm. 13 15). Deection toward A major occurs again in the third phrase
(m. 19), but the completion of the cadence is delayed by two beats and the vocal
line continues, ultimately eliding with the beginning of the fourth phrase in m. 22.
The fourth phrase breaks off momentar ily at its midpoint (which, as in all the
30. Brief discussions of the song can be found in Lucien Stark, A Guide to the Solo Songs of Johannes
Brahms (Bloomington: Indiana University Press, 1995), 15 16; and Er ic Sams, The Songs of Johannes
Brahms (New Haven, Conn.:Yale University Press, 2000), 36 37.

279

example 6.8

Johannes Brahms, In der Fremde, from Sechs Gesnge, Op. 3 No. 5

Poco agitato

Voice

Aus der
a tempo

poco rit.

Piano

Hei

mat hin

ter den Bli tzen rot, da

kom men die Wol ken

her.

ber

Va

ter und Mut ter sind lan ge

tot, es

rit.

13

mehr, kennt

mich dort kei

ner

mehr.

kennt mich dort kei ner

a tempo

Wie

rit.

word s, mu s i c, and s ong

example 6.8

(continued)

17

bald, ach wie bald kommt die stil

le

21

Zeit, da

ru

he ich auch und

a tempo

poco rit.

ber mir rauscht

die sch ne Wald ein sam keit,

poco rit.

25

und

sostenuto

dim. e rit.

kei

ner

kennt mich mehr hier, kei

ner

kennt mich

mehr

hier.

dim. e rit.

preceding phrases, is marked by an ar rival on the tonic), and the mode suddenly
changes to major on the last beat of m. 23. Brahms then completes the phrase by
recalling the melody of m. 12, but this time har monized in FS rather than D.Within
this context, the Gn in the melody at m. 25 recalls the end of Schumanns setting,
an association made stronger when the repetition of the last line of the poem in
(mm. 26 28) is set to a melody similar to that which appeared in m. 26 of Schumanns piano accompaniment.
Besides the obvious melodic repetitions engendered by this overall compositional scheme, two additional features bear note. First, the compass of the melody
is relatively restricted (from GS4 to G5, the smallest range of any of the Op. 3 songs).
Second, the melody is constantly brought back to A4: each phrase begins on A4, it

281

282

analysis and theory

Generic space
discomfort of
foreign surroundings

Music space
rhythmic disruptions
of phrase and cadence,
oppressiveness of
F# minor, prevalence
of melodic A4

Text space
alienation: mysterious
red lightning,
anonymity, orphaned
by parents and
the world

Blended space
restlessness,
grudging, uneasy
acceptance of
predicament

figure 6.5

CIN for Brahmss In der Fremde

marks the ar rival on the midpoint of each phrase, and the second and third phrases
end on it (in the latter case, this br ings about the elision of the end of the third
phrase with the beginning of the fourth).
This music, together with Eichendorff s text, sets up the CIN diag rammed in
gure 6.5. The gener ic space is structured around the feeling of discomfort attendant on being in alien sur roundings. This space is rather unlike those for the other
CINs discussed in this chapter in that it does not lend itself as readily to a nar rative
framework. However, discomfort, by its nature, suggests a dynamic situation (to the
extent that we want to return to a state of comfort). This may not quite constitute
a nar rative in its own right, but it creates an opportunity at least for a minimal story.
Eichendorff s poem supports this sense of discomfort with the ominous portents
of the clouds and red lightning, as well as with the isolation and anonymity noted
in the fourth and eighth lines. Discomfort is also summoned, if negatively, through
the thought that rest is to be found elsewhere.
The discomfort of alien sur roundings is suggested musically by the repeated
attempts in mm. 4 21 to escape the pitch A4 and the key of FS minor. (The accompaniment patter n for the FS minor sections also makes use of an insistent CS4 that
starts as a simple bit of rhythmic activation in the introduction, but soon begins to
cloy.) Each attempt fails, of course, as it is undercut by insecure cadences and the
relentless pull of A4. When a sort of escape is fashioned, at the songs conclusion,
one senses that the forces of oppression have been held in abeyance more than
defeated. Although the mode changes to major, the Gn5 summons the b2 that is

word s, mu s i c, and s ong

more typical of minor; although the melody ends away from A4, it has not accomplished a structural close on FS4.31
Correlations between these idealized representations of discomfort at alien surroundings provide an opportunity for the combination of concepts in the blended
space that draws on the elements of the input spaces to create a portrayal of life in
der Fremde.

composition

Composition combines the images of isolation and the alien

in the rst four lines of the poem with two successive musical phrases that begin in
the same place but move away to end at a distance from the tonic and with the
downbeat destabilized. The combination of the last line of the poem and the concluding music is slightly more tenuous: the return of the voice to the high end of its
range and the recall of the end of phrase 2 reinforce the portrayal of discomfort, but
the sonorous FS major chords and pedal point also suggest a sort of reconciliation
outside of the basic mapping that structures the CIN.32
Although I have not emphasized the failure of elements from the text-space and
music-space to combine in my discussions of the process of blending, it is particularly noticeable in the case of lines 5 7 of Eichendorff s poem and mm. 17 23 of
Brahmss setting. At this point, the topography of the two spaces is not unifor m
more specically, the text at this point (concer ned as it is with contemplation of
peaceful rest in the embrace of the forest) does not confor m to the structural
premises of the gener ic space. Although words and music certainly co-occur here,
they do not combine from the perspective of the CIN diagrammed in gure 6.6.
This lack of correspondence may be one reason Brahmss song seems weak here: his
text-painting of ber (as a long, relatively high note that is manifestly over the
accompaniment) offers rather insufcient compensation for the failure of words and
music to work together on a larger scale.

completion It is through completion of the conceptual blend that

Brahmss song takes on a dynamism that extends beyond the rhythmic disruptions
and key-area juxtapositions of the music. As noted, we know that states of discomfort cry out for resolution, whether that is provided by shifting position in our chair
or moving away from someone who makes us uneasy.When the music of the rst
half of each phrase is followed by the contrasting music in the second half, it is easy
to hear this as a moving away, a physical adjustment to unpleasant circumstances.
The sudden shift to major and return of the higher vocal tessitura at the end of
the song suggest not so much a nal resolution as an uneasy truce: the struggle of
the rst twenty-odd measures has yielded results, but they can hardly be called
denitive.
31. From a Schenkerian perspective, the lack of a clear descent to 1 (through the melodic progression AGSFS) poses interesting, although by no means ir resolvable, problems.
32. It is worth noting that the two lines of the poem most clearly associated with the discomfort of
abandonment and alienation lines 4 and 8 are also the only two that Brahms chose to repeat, thus
providing an opportunity to emphasize the composition of elements from the text-space and musicspace.

283

284

analysis and theory

elaboration

Entering further into the world of Brahmss song, we can see

that Brahmss singer is thoroughly within the foreign lands as she sings, and, although her deections toward major keys suggest an attempt to make herself comfortable in these alien sur roundings, she remains an anonymous stranger she cannot escape the iron grip of FS minor, just as she cannot escape her isolation.
Nonetheless, Brahmss singer does not despair at least not yet. Instead, she struggles against her fate, crying out against her anonymity in her highest register,
attempting to use her predicament as a lever to gain recognition and a life in this
foreign land. Given this text and this music, we can imagine her shaking her st at
fate; if her eyes ll with tears, they are tears of a struggle against destiny.

Summary
With his setting of In der Fremde Brahms clearly wanted to demonstrate his
independence as a composer, as well as to pay homage to an important fr iend and
mentor.Yet, two other factors are important. The rst is raw youth: Brahms at nineteen was no stranger to feelings of alienation and restless energy, and it is hardly surprising that he drew these from Eichendorff s poem. The second are the compositional advances made by Schumann in the Liederjahr of 1840: the year bore witness
not only to an incredible outpour ing of song but also to a new understanding of
how musical syntax could be shaped to respond to the possibilities of poetry.
Brahms lear ned the lessons of Schumanns Liederjahr, as well as the many other
lessons taught by the music of his century, and he built on them.Where cadence was
still a pr imary articulator of the musical argument for the Schumann of 1840, it
became something innitely more malleable for the Brahms of 1852. Indeed, over
the course of Brahmss In der Fremde, there is not one clear cadence; instead,
there are cadences disrupted, so understated as to be almost invisible (as the cadence
of m. 23), or assembled out of the bits and pieces of traditional cadential for mulae
(which is how Brahms br ings the song to a conclusion). Brahms, of course, knew
when and how to exploit traditional cadential for mulae, and he did so with g reat
success in the years that followed, but his setting of In der Fremde shows that,
when his understanding of a text demanded it, he could create a new set of syntactic possibilities for musical discourse.

conceptual blending and song

As part of his anthropological fantasy on the or igins of language, Jean-Jacques
Rousseau imagined the pr imordial world that gave birth to the languages of the
temperate south. Around the water wells necessary for life in these ar id lands people gathered, and young men met young women for the rst time.
Beneath aged oaks, conquerors of years, an ardent youth gradually forgot its ferocity,
gradually they tamed one another; through endeavoring to make themselves understood, they learned to explain themselves. There the rst festivals took place, feet
leaped with joy, eager gesture no longer sufced, the voice accompanied it with passionate accents; mingled together, pleasure and desire made themselves felt at the same

word s, mu s i c, and s ong

time. There, nally, was the true cradle of peoples, and from the pure crystal of the
fountains came the rst res of love.33

As Rousseau would have it, the language that developed around these aqueous
gathering places was not like moder n language, but a fusion of words and music:
Around the fountains of which I have spoken, the rst discourses were the rst songs;
the periodic and measured recurrences of rhythm, the melodious inections of accents
caused poetry and music to be bor n along with language; or rather, all this was nothing but language itself in those happy climates and those happy times when the only
pressing needs that required anothers help were those to which the heart gave rise.34

In the myth Rousseau offers, music and language dont simply cohabitate; they
commingle their syntaxes: the per iodic and measured recurrences of rhythm and
the inections of accent are combined in an Ur-language, which, only later, after
the Fall or its equivalent, separated into music and language.
In the end, Rousseaus fantasy has more to tell us about the role of language and
music in the anthropological speculations of the eighteenth centur y than it does
about the possibility that the two media were once one.35 But the conviction that
music represents a communicative medium to equal language a conviction Rousseau shared with his old antagonist Rameau36 and the idea that the sum of music
and language might be g reater than its parts are concepts wr iters like Mller and
von Spaun believed were demonstrated, in fairly specic ways, by the nineteenthcentury Lied. While it is possible to dismiss such beliefs as products of overheated
rhetor ic, the fascination composers, audiences, and cr itics have had with song cannot be ignored. Conceptual blending offers a way to explain why this fascination
has been so endur ing: under certain circumstances, elements and relations from the
mental space set up by the music of a song blend with elements and relations from
the mental spaces set up by the text to create a new world for the imagination a
world of supernatural char m, noctur nal ter ror, and much, much more.
For conceptual blending to occur, the mental spaces that serve as inputs to the
blend must have a shared topography. For the songs analyzed in this chapter I have
located this shared topography in commonalities between the complete discourse
33. Jean-Jacques Rousseau, Essay on the Or igin of Languages (c. 1763), in Essay on the Origin of
Languages and Writings Related to Music, trans. and ed. John T. Scott, Collected Writings of Rousseau, 7
(Hanover, N.H.: University Press of New England, 1998), 314. In my comments on Rousseaus essay, I
am more concer ned with the idyllic state he imagines than with the overall context relative to which the
image is framed. For a probing discussion of this context, see Lydia Goehr, The Quest for Voice: Music, Politics, and the Limits of Philosophy (Berkeley: University of California Press, 1998), 98 106.
34. Rousseau, Essay on the Or igin of Languages, 318.
35. For more on the role of language and music in speculations about human or igins by French
writers of the eighteenth century, see Downing A. Thomas, Music and the Origins of Language: Theories
from the French Enlightenment, New Perspectives in Music History and Cr iticism (Cambr idge: Cambridge
University Press, 1995), chap. 2.
36. Although Rousseau and Rameau both argued that music was the equal to language, for
Rousseau this meant melody and for Rameau it meant har mony. This difference of opinion (among
other things) was the basis for their debate about musical expression, discussed at the end of chap. 3 in
this volume.

285

286

analysis and theory

structure set up by the music and the complete discourse structure set up by the
text. That such commonalities might exist suggests deep cor respondences between
musical and linguistic syntax. To be sure, these cor respondences are, in most
instances, of a general sort, outlining as they do constructs like contrasting ontological states (as in the settings of Trockne Blumen and in Schumanns In der
Fremde), exterior and inter ior domains (as in Im Rhein), or discomfort with
foreign surroundings (as in Brahmss In der Fremde). And yet, such highly abstract
constructs still have their power, for they can nd realization in manifold concepts
associated with both the emotions and the intellect.
Agawu ended his provocative 1992 essay with these words: We await the development of a syntax of song.37 The preceding analyses suggest that this syntax is to
be found not in the song proper, but in the way the constituent media of song
words and music relate to one another. The syntax of song, then, is a syntax of
conceptual blending, through which music breathes life into poetr y and poetry
breathes life into music, and together they proceed as song.

37. Agawu, Theory and Practice, 30.

chapter seven

competing models of music:

theories of musical form

and hierarchy

idway in her jour ney through Wonderland, Lewis Car rolls Alice takes leave
of the Cheshire Cat and makes her way toward the house of the March Hare.
There she nds a tea party under way a mad, perpetual tea party, attended by the
March Hare, the Mad Hatter, and the somnolent Dor mouse. As the party proceeds,
Alices patience begins to wear thin, tried by turnabout word play and abrogated
manners, and is stretched to the limit when she nds out that the Hatters riddle
Why is a raven like a wr iting desk? is without solution.

Alice sighed wearily. I think you might do something better with the time, she said,
than wasting it in asking r iddles that have no answers.
If you knew Time as well as I do, answered the Hatter, you wouldnt talk about
wasting it. Its him.
I dont know what you mean, said Alice.
Of course you dont! the Hatter said, tossing his head contemptuously. I dare
say you never even spoke to Time!
Perhaps not, Alice cautiously replied; but I know I have to beat time when I
learn music.
Ah! That accounts for it, said the Hatter.He wont stand beating. Now, if you only
kept on good terms with him, hed do almost anything you liked with the clock.1

Exchanges like this are common enough in Wonderland, a place inhabited by

creatures whose logic is often as fantastic as their appearance. Nonetheless, the game
Carroll is playing at here is less a matter of labyrinthine logic and more a matter of
different ways to structure what is putatively the same domain. Alice uses two different cross-domain mappings to reckon with time: time as something to be saved
or wasted (mapping from the domain of commodities or substances onto the
domain of time); and time as a set of regular motions (mapping from the domain
of physiologically dened space onto the domain of time). For the Hatters cockeyed contr ibution, Carroll exploits a traditional character ization of time as a human
1. Lewis Carroll, The Annotated Alice: Alices Adventures in Wonderland and Through the Looking Glass,
intro. and with notes by Martin Gardner, illustrated by John Tenniel (New York:W.W. Norton, 2000), 72.
The tea party is the topic of chap. 7 of Alices Adventures in Wonderland.

287

288

analysis and theory

agent (as in Father Time, or the New Year as a diapered baby), an anthropomor phic
mapping that exploits conceptual blends like those discussed in the latter portion of
chapter 2. The mer ry confusion of mappings that results is part and parcel of the
imaginary world Car roll constructed indeed, Wonderland is a maze of interliated conceptual blends. However, the scene at the tea party also throws into relief
a fundamental conict between Alices world and that of Wonderland, which comes
to a head in the nal chapter. There, Alice responds to the Queen of Heartss
peremptory Off with her head! (Sentence rst verdict afterwards) by dismissing the entire court and courtroom with Who cares for you? Youre nothing
but a pack of cards! At this, Wonderland disappears, and Alice wakes to nd herself lying on the bank of the river, with her head in the lap of her sister.2
A confusion of mappings similar to the one apparent in Alices dialogue with the
Hatter often ar ises in music theory. The result, while not quite as entertaining as a
turn in Wonderland, can have some of the same avor. Theorists appear to be talking about the same thing, but they structure their understanding of it in very different ways. Two examples are taken up in this chapter. The rst involves musical
form, the second musical hierarchy.
During the eighteenth century, the study of musical for m emerged as an important basis for discourse about music. Form was regarded as dealing with the very
essence of musical works: how musical mater ials were arranged and cor related over
the course of a piece of music. As the study of form developed and was continued
through the nineteenth century, theorists worked with two basic and seemingly
opposed models of musical for m, one static, the other dynamic. Musical for m,
viewed from a static perspective, is reminiscent of architecture (a parallel all the
more ironic, given Friedrich von Schellings character ization of architecture as
frozen music)3 and typically consists of either a framing structure clad with musical mater ial or relatively abstract containers lled with musical events. Musical for m
from a dynamic viewpoint is processive and a bit unpredictable: the musical work
emerges over the course of time, and musical mater ials are both the substance of and
raison dtre for this emergence. As a further complication, the cross-domain mappings that activated these two models changed dur ing the nineteenth century. Consequently, there are two levels of mapping involved, both of which can breed confusion about just what is meant by musical for m: a gener ic level associated with the
two basic models (one inter preting for m as static, the other as dynamic); and a
specic level associated with the cross-domain mappings that activate the gener ic
models. In the rst section that follows, I explore static and dynamic models of
musical for m that emerged in the eighteenth centur y and the way these models
were transfor med by changes in ideology and music education that took place
beginning late in that century.
2. Carroll, The Annotated Alice, 124.
3. Architecture, as the music of the plastic arts, thus necessar ily follows ar ithmetical relationships.
Since it is music in space, however, in a sense, solidied music, these relationships are simultaneously geometric relationships. Friedrich Wilhelm Joseph von Schelling, Philosophy of Art, ed., trans., and intro.
Douglas W. Stott, with a foreword by David Simpson, Theory and History of Literature, 58 (Minneapolis: University of Minnesota Press, 1989), 165. See also, on p. 177, the passage that begins If architecture
is indeed solidied music . . .

competing mode ls of mu s i c
The two models of musical hierarchy I deal with in the second of the two sections that follow are a somewhat different matter, for these models were rst developed to apply to domains other than music. It was only in the early nineteenth century that they were applied to music, and then to two rather different aspects of
musical organization. The rst model views hierarchy as a matter of control: each
level in a hierarchy (with the exception of those at the extremes) controls the nextlower level and is itself controlled by the next-higher level. This model stretches
back at least to the Middle Ages and is most often used to account for tonal organization. The second model of hierarchy relies on a more componential approach:
the elements of level A of the hierarchy combine to make up individual elements at
the next-higher level (level B of the hierarchy); the elements of level B then combine to make up individual elements at the next higher level (level C of the hierarchy), and so on, until the account of structure is exhausted. This model emerged
during the seventeenth century and is most often used to explain musics metr ical
structure.
When confronted by a maze of mapping like those I have just outlined, mappings that yield contradictor y accounts of quite basic musical mater ials, our inclination may be to throw up our hands and shout the equivalent of Who cares for
you? Youre nothing but a pack of cards! But problems of musical understanding
are rarely so easily dismissed. Instead, we must enter into the mad tea party that is
music theory, where music theor ists from different nations and different histor ical
periods gather around the table to discuss and dispute musical concepts. If we are to
make sense of the conversation there, we need to understand how cross-domain
mapping shapes this discourse indeed, as I suggested in chapter 2, how it makes
music theory possible.

two models of musical form

There was a certain enviable simplicity to the study of music in the n de sicle
Anglo-Amer ican musical world, due in no small part to its near-complete subjugation to the Ger man musical world.4 Music teachers in Anglo-Amer ican circles
were either Ger man or trained in Ger many, the repertoire studied was almost
entirely German, and the theoretical models developed exclusively with respect
to German musical practice had been for mulated by German music theor ists of
the nineteenth century. Thus the beginning Anglo-Amer ican music student of the
late nineteenth and early twentieth century learned a tradition of musical study that
was almost entirely Ger man. This is not to say, however, that the tradition they
assimilated was completely unied. Consider, for instance, the account of musical
motives offered by Ebenezer Prout, a prominent music educator in England during the late nineteenth century. Having distinguished between phrases (which end
with a cadence) and sections (subphrase units that do not necessar ily end with a
cadence), Prout makes one further subdivision:
Every section contains at least two [metric] accents, and can be separated into smaller
parts, containing only one accent each. Such parts are called motives, and cor respond
4. By German musical world, I mean the Ger man-speaking countr ies of Europe.

289

290

analysis and theory

exactly to the feet in poetry. Here we have what (to bor row a scientic ter m) we may
call the protoplasm the germ out of which all music springs. A thorough knowledge of the nature of the motive is therefore indispensable to anyone who would understand the fundamental pr inciples of musical for m.5

Important for Prouts character ization of a motive are the dynamic context provided by metrical accentuation he will later dene motive as composed of a
strongly accented note preceded by at least one unaccented note6 and the somewhat more abstract notion of a motive as the seed of a musical organism.
Rather different is the approach to motive taken some ten years later by Percy
Goetschius, born in New Jersey but educated in Stuttgart, where he rst taught.
Goetschius situates the motive in an explicitly grammatical context:
The smallest unit in musical composition is the single tone. The smallest cluster of
successive tones (from two to four or ve in number) that will convey a denite musical impression, as miniature musical idea, is called a Figure. Assuming the single tone
to represent the same unit of expression as a letter of the alphabet, the melodic gure
would be dened as the equivalent of a complete (small) word;pursuing the comparison further, a series of gures constitutes the melodic motive, equivalent to the
smallest group of words (a subject with its article and adjective, for example); and two
or three motives make a Phrase, equivalent to the complete, though comparatively
brief, sentence (subject, predicate, and object).7

In contrast with Prouts dynamic account of musical for m, Goetschius relies on a

highly schematized g rammar of musical elements. So formalized is this g rammar
that it ultimately reduces to a matter of counting measures: Melodic motives differ in length from one to four measures; by far the most common extent, however,
is two measures, and the student will do wisely to accept this dimension and analyze
accordingly, unless there is unmistakable evidence to the contrary.8
What we have in these excer pts, then, are two markedly different approaches to
musical for m, voiced through character izations of motive.9 Prouts is the more
dynamic: music is likened to some sort of organism, and the motive is the seed from
which it spr ings. Goetschiuss approach, by contrast, is rather more static: music is
but prose in tones, and the motive is simply the smallest intelligible syntactic unit.

5. Ebenezer Prout, Musical Form (London: Augener, 1893), 26.

6. Prout, Musical Form, 31.
7. Percy Goetschius, Lessons in Music Form: A Manual of Analysis of All the Structural Factors and
Designs Employed in Musical Composition (Bryn Mawr, Penn.: Oliver Ditson, 1904), 19.
8. Goetschius, Lessons in Music Form, 23.
9. Although I have chosen to highlight the differences between Prouts and Goetschiuss approaches
to musical for m, Robert Gjerdingen has reminded me that the two theor ists knew each others work and
exchanged dedications of their texts. Moreover, not all theor ists within the Anglo-American tradition
agreed on the importance of motive. For instance, Stewart Macpherson, whose book on for m supplanted
that of Prout and who was mightily suspicious of overly thorough analyses, wrote, As an example of a
too mathematical conception of analysis may be instanced the stress laid by some theor ists upon an arbitrary dissection of the music into certain metr ical fragments called motives, a dissection of which usually
conveys little or nothing to which the musical sense can respond. Macpherson, Form in Music: With Special Reference to the Designs of Instrumental Music (London: Joseph Williams, 1908), 39. A footnote makes
clear that in referring to some other theor ists, Macpherson actually meant Prout.

competing mode ls of mu s i c
However, two related factors complicate these seemingly neat constructions. First,
neither Prout nor Goetschius eschews entirely the approach of the other. Prout is
perfectly capable of rendering a crushingly static account of musical for m, and
Goetschius knows how to enliven his account to make music seem more like an
organism than a dr y recitation. Second, both of these models of form originate in
mappings between language and music (a mapping more obvious in the excer pt
from Goetschius, but evident as well in Prouts reference to poetic feet). To untangle matters, we need to begin with theor ies of musical for m developed dur ing the
Enlightenment. From there we must proceed to explore how these theor ies were
subsequently reformulated in the rst half of the nineteenth centur y under the
inuence of political and ideological pressures that transfor med the way music was
taught.

Grammar, Rhetor ic, and Musical Form

in the Eighteenth Century
Mappings between music and language go back at least to the Middle Ages. For
instance, the ninth-century Musica enchiriadis opens with an analogy that anticipates
Goetschius by over a thousand years: Just as the elementar y and indivisible constituents of speech (vox articulata) are letters, from which syllables are put together,
and these in tur n make up verbs and nouns, and from them is composed the fabric
of a complete discourse, so the roots of song (vox canora) are phthongi [musical notes],
which are called soni in Latin. The content of all music is ultimately reducible to
them.10 In the eighteenth centur y, mappings between language and music ourished and as an almost inevitable consequence became more manifold and more
complicated. Enlightenment preoccupations with the or igins of language and the
place of music within systems of human communication led to a number of proposals for how language and music did, or did not, relate to one another.11 The
continued growth of print culture made it increasingly common for music and language to intersect as score and text on the printed page, a phenomenon exploited
by author-pr inters such as Samuel Richardson.12 Finally, a growing middle class
fueled a demand for music instruction books, and, within these books, mappings
between language and music provided a ready frame of reference for the explanation of fundamental musical concepts. During this per iod, two mappings from the
domain of language onto the domain of music became particularly important for
accounts of musical for m: the rst mapped from grammar onto music, the second
mapped from rhetor ic onto music.

10. Musica enchiriadis and Scolica enchiriadis, ed. Claude V. Palisca, trans. Raymond Erickson, Music
Theory Translation Ser ies (New Haven, Conn.:Yale University Press, 1995), 1.
11. For a discussion centered on eighteenth-century France, see Downing A. Thomas, Music and the
Origins of Language: Theories from the French Enlightenment, New Perspectives in Muisic History and Cr iticism (Cambr idge: Cambridge University Press, 1995).
12. Richardsons insertion of a musical score into the text of a novel is discussed by Janine Barchas
in The Engraved Score in Clarissa: An Intersection of Music, Narrative, and Graphic Design, EighteenthCentury Life 20 (1996): 1 20.

291

292

analysis and theory

g rammar and music

Mappings between grammar and music were based

on the assumption that music was language-like in its organization that is, it had
structural features analogous to those of language and a nor mative syntax. The mapping also relied on a familiar ity with and respect for grammar acquired through the
thorough grounding in the subjects of the Classical tr ivium rhetoric, grammar,
and logic that was a common feature of early education in the eighteenth century. Just as every verbal construct, whether spoken or wr itten, can be divided up
into clauses, sentences, and paragraphs, the thinking went, so can musical works.
This was the basis of Johann Matthesons account of musical grammar, rst proposed in his Kern melodischer Wissenschaft of 1737 and subsequently expanded in his
Der vollkommene Capellmeister of 1739.13 Mattheson began his discussion with vocal
music, showing how the grammar of music had to be in rapport with that of the
text: each of the signicant divisions in the text had to be supported by analogous
divisions in the music. This led Mattheson to a discussion of what things made for
effective musical divisions (typically, cadence was the main articulative event) and
how such divisions were graded from the barely noticeable to the emphatic.14
In a subsequent chapter of Der vollkommene Capellmeister, Mattheson argued that
knowledge of this sort was even more important for knowing how to wr ite instrumental music. He illustrated his discussion with the minuet shown in example 7.1.15
Mattheson analyzes the whole as a parag raph, made up of two per iods (or Stzen;
mm. 1 8 and 9 16), which are divided into a clause (marked by the colon below
m. 4), a subclause (marked by the semicolon below m. 12), and phrases (marked by
the commas below mm. 2, 6, and 10). Lacking, however, is any explanation for how
these analytical decisions were made why, for instance, mm. 1 4 constitute a
clause, while mm. 9 12 constitute a subclause. The reader is left with a clear sense
of the importance of grammatical structure to musical works that grammar is, in
fact, a self-evident property of musical works but with only a few hints as to how
the grammatical divisions in music are deter mined.
A more complete methodology for analyzing musical grammar was provided
some fteen years later by Joseph Riepel in his Grundregln zur Tonordnung insgemein
of 1755 (the second volume of his Anfangsgrnde zur musicalischen Setzkunst). Using
a thoroughly schematic and comparatively systematic approach to musical organization (wr itten in the dialogue style common in pedagogical treatises of the day),
Riepel made explicit connections between grammatical components and musical
passages, illustrated through hundreds of examples drawn from or applicable to
13. Johann Mattheson, Kern melodischer Wissenchaft (Hamburg: C. Herold, 1737; repr. Hildesheim:
Georg Olms, 1976); idem, Der vollkommene Capellmeister (Hamburg: C. Herold, 1739; repr. Dokumenta
musicologica, 1. Reihe, Druckschr iften-Faksimiles, 5, ed. Margarete Reimann [Kassel: Brenreiter,
1954]).
14. Mattheson, Der vollkommene Capellmeister, 180 95; Ernest C Har riss, Johann Matthesons Der vollkommene Capellmeister: A Revised Translation with Critical Commentary, Studies in Musicology, 21 (Ann
Arbor, Mich.: UMI Research Press, 1981), 380 404.
15. Mattheson, Der vollkommene Capellmeister, 223 25; Harriss, Matthesons Der vollkommene
Capellmeister, 451 53. For a discussion and analysis of this passage, in a slightly different context, see Ian
Bent, ed., Fugue, Form and Style (vol. 1 of Music Analysis in the Nineteenth Century), Cambridge Readings
in the Literature of Music (Cambr idge: Cambridge University Press, 1994), 5 6.

competing mode ls of mu s i c
example 7.1 Minuet analyzed by Johann Mattheson in Der vollkommene Capellmeister,
part 2, chap. 13

Fine

D.C. al fine

11

instrumental music.16 Riepel based his analysis of the grammatical function of a

musical passage on har monic structure and on the assumption that subdivision of
the phrase into even-numbered units was the nor m in instrumental music.
Although his references to the grammatical structure of speech are explicit, they are
also relatively few in number. Subphrase units (which Riepel called Einschnitte, and
which were usually less than four measures in length) were marked by incomplete
harmonic progressions that is, they ended inconclusively. In contrast, the phrase
(the Satz or, more frequently, the Absatz) was a four-bar unit that provided a complete har monic progression: it had a conclusive ending. Each Absatz was further
character ized on the basis of whether it ended on tonic, on the dominant, or in
some other key.17 By this means, Riepel was able to give an exhaustive (and at times
exhausting) account of the different sorts of grammatical structures that might be
evinced by music, based on the ways various components could be t together and
modied.
Riepels approach to musical grammar was adopted by Heinr ich Chr istoph
Koch in his Versuch einer Anleitung zur Composition (1782 1793), but Koch began his
account by returning to the model of speech. After noting the importance of certain more or less noticeable resting points of the spir it to those ne arts that attain
their goal through speech namely, poetry and rhetor ic Koch then draws the
connection to music.
Speech . . . breaks down into various sentences [Perioden] through the most noticeable
of these resting points of the spir it; through the less noticeable the sentence, in turn,

16. Joseph Riepel, Grundregln zur Tonordnung insgemein (vol. 2 of Anfangsgrnde zur musicalischen Setzkunst) (Regensburg: Johann Leopold Montag, 1755), 32 56; idem, Smtliche Schriften zur Musiktheorie, ed.
Thomas Emmer ig,Wiener musikwissenschaftliche Beitrge, 20 (Vienna: Bhlau, 1996), 138 62.
17. For further discussion of Riepels approach, see Justin London, Riepel and Absatz: Poetic and
Prosaic Aspects of Phrase Structure in 18th-Centur y Theory, Journal of Musicology 8 (1990): 505 19;
Nola Reed Knouse, Joseph Riepel and the Emerg ing Theory of Form in the Eighteenth Centur y,
Current Musicology 4 (1986): 46 62; and John Walter Hill, The Logic of Phrase Structure in Joseph Riepels Anfangsgrnde zur musikalischen Setzkunst, Part 2 (1755), in Festa musicologica: Essays in Honor of
George J. Buelow, ed. Thomas J. Mathiesen and Benito V. Rivera, Festschrift Ser ies, 14 (Stuyvesant, N.Y.:
Pendragon Press, 1995), 467 87.

293

294

analysis and theory

breaks down into separate clauses [Stze] and parts of speech [Redetheile]. Just as in
speech, the melody of a composition can be broken up into per iods by means of analogous resting points, and these, again, into single phrases [Stze] and melodic segments
[Theile].18

However, Koch also saw limits to the applicability of mapping from the grammar of
speech onto the grammar of music. After a long and thoughtful passage in which he
pursues the notion that musical phrases are organized around subjects and predicates
as are those of speech, Koch tur ns away from the model. He writes, I am abandoning this compar ison because, as already mentioned, subject and predicate cannot
be differentiated enough in melodic sections. We must content ourselves with
learning how to deter mine, through feeling, the presence or absence of resting
points of the spir it and the nature of sections with regard to their completeness.19
As did Riepel, Koch assumes subdivision of the phrase into even-numbered units to
be the nor m. He uses har monic structure specically, the har mony with which
each phrase ends20 to deter mine the divisions of musical grammar. And he builds
larger musical structures by modifying and concatenating musical phrases.
Mappings between grammar and music of the sort evident in the work of
Mattheson, Riepel, and Koch allowed writers in the eighteenth centur y to give
detailed accounts of musical organization and to indicate how composers manipulated musical mater ials to create unique musical statements that nonetheless conformed to certain basic syntactic pr incipals.21 The mapping yields a character ization
that emphasizes the static rather than the dynamic what Koch came to call the
mechanical aspect of music in the introduction to the second volume of his Versuch.22 Each unit of musical grammar can be contemplated and discussed separately,

18. Heinrich Chr istoph Koch, Introductory Essay on Composition: The Mechanical Rules of Melody, Sections 3 and 4, trans. Nancy Kovaleff Baker, Music Theor y Translation Ser ies (New Haven, Conn.: Yale
University Press, 1983), 1. The source from which Bakers translation is drawn is the second volume of
Kochs Versuch einer Anleitung zur Composition (Leipzig: Adam Fr iedrich Bhme, 1787; repr. Hildesheim:
Georg Olms, 1969). I have adapted this and the following quotation to restore Kochs somewhat
unwieldy locution resting points of the spir it (Ruhepuncte des Geistes), which preserves the subjective aspect of determining music-grammatical divisions.
19. Koch, Introductory Essay on Composition, 6, translation adapted. In a note to this passage, Koch
explains that he abandoned str ict compar isons between the phrases of speech and melody because his
treatise is intended for beg inning musicians who wish to lear n composition, and these seldom have
either grammatical knowledge of speech or familiar ity with that part of logic which explains the different types of phrases and their closures.
20. Koch wr ites, The essential difference between an inter nal phrase and a closing phrase depends
on nothing more than the essential difference of ending for mulas. Koch, Introductory Essay on Composition, 7.
21. As noted in chap. 2, it is through mappings such as these that theor ists br ing an integrated system of terms and structural relations to bear on particular problems of musical understanding. In chap. 2,
the problems discussed had to do with ways to character ize tonal organization. Here the problems relate
to how musical mater ials should be organized to create meaningful utterances.
22. Nancy Kovaleff Baker and Thomas Chr istensen, eds. and trans., Aesthetics and the Art of Musical
Composition in the German Enlightenment: Selected Writings of Johann Georg Sulzer and Heinrich Christoph
Koch, Cambridge Studies in Music Theor y and Analysis, 7 (Cambridge: Cambridge University Press,
1995), 140.

competing mode ls of mu s i c
and each can be t together with other such units to create a complete statement
at the level of a paragraph or higher. The perspective, then, is inherently constructivist: mapping grammar onto music isolates the building blocks of music and
describes how they t together to create coherent structures.
The account of musical for m produced by this mapping is one that tends to
emphasize complete, free-standing structures, as well as a synoptic view of such
structures.23 What is incomplete is ungrammatical; what is ungrammatical is incomplete. This is a view of basic mater ials well suited to pedagogy mappings from
grammar to music are invariably connected with lessons in composition but one
that has little to say about the process of constructing grammatical statements that
are both complete and compelling.

rhetoric and music

One of the principal uses of mappings from

rhetoric to music dur ing the eighteenth century was to descr ibe how to go about
composing music and the laying-out of musical mater ials in particular. Mattheson,
in his Das neu-erffnete Orchestre of 1713, offers the following analysis: A [musical]
composition compr ises three things: Inventio (Erndung), Elaboratio (Ausarbeitung),
[and] Executio (Ausfhrung or Auffhrung), which seem to display a close relationship to oratory or rhetor ic. The last two things can be lear ned; the rst has no competent master.24 Of course, Mattheson is being somewhat disingenuous here, since
the three techniques he mentions and, for that matter, the very term composition are all der ived from the teaching of Classical rhetor ic and were only secondarily applied to music. Twenty-six years later, in Der vollkommene Capellmeister,
Mattheson expanded this plan to include a total of ve stages: Inventio (Erndung),
Dispositio (Einrichtung), Elaboratio (Ausarbeitung), Decoratio (Schmckung or Zierde),
and Executio (Ausfhrung or Auffhrung).25 In this more extensive application of
rhetoric to music, Mattheson also drew on the structural categor ies of rhetor ical
theory to character ize the parts of a musical work (as distinct from the process of
creating the work). These he identied as the Exordium (Eingang), Narratio
(Bericht), Propositio (Antrag), Conrmatio (Bekrfftigung), Confutatio (Wiederlegung),
and Peroratio (Schlu). Mattheson then used these categor ies to analyze the music
of a da capo aria by Benedetto Marcello, but the result was at best a mixed success:

23. As I understand it, this synoptic view of the musical work does not recover what Michel Foucault calls the simultaneity of thought: musical works, whether paragraphs or chapters, must still
unfold in the temporal ser ies specied by their grammar. The synoptic function instead reects the taxonomy of forms basic to grammatical systems. There is an immediacy to such taxonomic units be they
phrases or parag raphs, books or symphonies that is independent of the time it takes to traverse the
individual syntactic components they compr ise. On the simultaneity of thought, see Foucault, The Order
of Things: An Archaeology of the Human Sciences (New York:Vintage Books, 1973), 83.
24. Es gehren sonst zu einen Composition dreyerley: Inventio, (Die Erndung) Elaboratio, (Die
Ausarbeitung) Executio, (Die Ausfhrung oder Auffhrung) welches eine ziemliche nahe Verwandschafft
mit der Orator ie oder Rhetor ique (Rede-Kunst) an den Tag legt; Die beyden lezten Stcke knnen
erlernet werden; zum ersten hat sich noch kein tchtiger Maitre. Johann Mattheson, Das neu-erffnete
Orchestre (Hamburg: Printed for the author, 1713; repr. Hildesheim: Georg Olms, 1993), 104.
25. Mattheson, Der vollkommene Capellmeister, 235; Harriss, Matthesons Der vollkommene
Capellmeister, 470.

295

296

analysis and theory

the ar ia simply did not provide an adequate target for mappings from the traditional
structure of oratory.26
Where Matthesons mappings between grammar and music pointed to a fruitful
model for character izing musical organization, his proposed mappings between
rhetoric and music never our ished in the same way. One problem was that the
structural model of an oration used in Classical rhetor ic was simply too r igid for
music indeed, there was every sign that it was coming to be regarded as too r igid
for speech as well.27 Another problem was that the exibility of musical semantics
made it difcult to develop hard and fast rules for what counted as a Propositio, a
Conrmatio, or any of the other elements of an oration: it was clear to Mattheson
and other theor ists that music had meaning, but just how this meaning cor responded with the gures and processes of Classical rhetor ic proved difcult to
specify.28
The account of the process of composition afforded by rhetor ic was given new
promise when it was incor porated into the theor y of aesthetics promulgated by
Johann Georg Sulzer in his Allgemeine Theorie der schnen Knste, rst published in
two large volumes in 1771 and 1774. Sulzer drew on Classical rhetor ic to develop
a general character ization of the creative process, reecting his belief that art shared
with rhetor ic the goal of moral improvement. According to Sulzer, the artist rst
discovers the ideas necessary to a work of art through invention. After the ideas have
been discovered, the artist then realizes them through a three-stage process: Anlage
(layout), Ausfhrung (realization), and Ausarbeitung (elaboration).29 Of these, the latter two stages are identical with those identied by Mattheson in Das neu-erffnete
Orchestre. Sulzers rst stage cor responds with the second stage Dispositio (Einrichtung) of the expanded plan presented in Der vollkommene Capellmeister. (Of
Sulzers treatment of Matthesons rst stage Inventio more presently.)
The approach to composition outlined by Sulzer, while promising, was also very
26. Matthesons analysis is in Der vollkommene Capellmeister, 237 39 (pp. 472 76 of the Harriss translation). For discussion, see Mark Evan Bonds, Wordless Rhetoric: Musical Form and the Metaphor of Oration,
Studies in the History of Music, 4 (Cambridge, Mass.: Harvard University Press, 1991), 82 90; and Peter
Hoyt, Review of Mark Evan Bonds, Wordless Rhetoric, Journal of Music Theory 38 (1994): 127 29.
27. For further discussion of the limitations of rhetorical models, see Hoyt, Review of Wordless
Rhetoric, 130 33.
28. The problem is, if anything, more intractable at the level of individual rhetor ical gures, as is evident in the complicated and occasionally redundant system set up by Joachim Bur meister in the early
seventeenth century; see Bur meister, Musical Poetics [1606], trans. Benito Rivera, Music Theory Translation Series (New Haven, Conn.:Yale University Press, 1993). For his part, Mattheson attempts no formalization of rhetor ical gures, preferring to descr ibe a few of them, indicate the applicability of many
more, and then move on; see Der vollkommene Capellmeister, 243 44 (pp. 482 84 of the Harriss translation).
Although locating the specics of rhetor ical structure in musical discourse is challenging, the importance of rhetor ic for all manner of thought and expression even beyond the eighteenth centur y
should not be underestimated. For a discussion of the importance of rhetor ic in scientic discourse, for
instance, see Jeanne Fahnestocks recent Rhetorical Figures in Science (New York: Oxford University Press,
1999).
29. My summary of Sulzers aesthetic theory and translation of the Ger man ter ms for the stages in
the compositional process are taken from Thomas Chr istensens introduction to Sulzers writings in
Baker and Chr istensen, Aesthetics and the Art of Musical Composition, 3 24.

competing mode ls of mu s i c
abstract, and it lacked specics where music was concer ned. These were supplied by
Koch in the second and third volumes of his Versuch, in which he endeavored to
show how Sulzers three-stage process could be implemented with musical mater ials.30 Where Sulzer was vague, having attempted to give an account of composition
adequate for all the ne arts, Koch was specic, providing a detailed account of just
what was involved at each stage of the compositional process, which he illustrated
with an analysis of an ar ia from Carl Heinr ich Grauns highly popular cantata Der
Tod Jesu of 1755.31 Making Sulzers framework for the compositional process
specic for music also provided a further justication for the technical language of
musical grammar developed by Koch, for Koch believed it was by the conscious
manipulation of this grammar that the beg inning composer could create an aesthetically satisfying nal work.
Mappings between rhetor ic and music allowed music theor ists like Mattheson
and Koch to descr ibe the compositional process in ter ms of distinct stages, which
made it possible to give specic advice about how the novice composer should proceed along the way. Inasmuch as this process was construed as analogous to that
undertaken in other art forms, music was placed on a parallel with sister arts to
which it had for merly been but a poor cousin. Mapping from rhetor ic to music also
allowed music more specically, untexted music to be regarded as a medium
with persuasive powers similar to that of orator y. Although the trope of music as
oratory was common at least since the Renaissance, preoccupations with the relationship between language and music dur ing the Enlightenment revivied the analogy and put into relief the dynamic aspect of music: here was a type of oratory that
not only unfolded over time but also was conducted in the pure language of
tones what Mattheson called Klang-Rede (sound-speech).
The view of the compositional process and its product afforded by mapping
from rhetor ic to music also highlighted the essential mystery of musical creation.
This is plainest in the quotation from Matthesons Das neu-erffnete Orchestre
invention is the process that knows no competent master but it is also evident
in Sulzers discussion of invention in the Allgemeine Theorie and in Kochs adaptation
of that discussion in his Versuch. At rst Sulzer maintains that invention is a capacity that can be improved through diligent and careful practice: The power of
invention, like the power of judgment, is a natural and inbor n faculty that all men
possess in proportion to their own genius. And just as one may turn to reason to aid
ones power of judgment, so can ones talent for invention be augmented, in as much
as it might be treated, like logic, as a part of philosophy.32 Sulzers initial suggestion
for accomplishing this augmentation is to tur n to rhetor ic, urging the artist to
exploit the rules and methodology that rhetor icians had developed for the express

30. Kochs realization of Sulzers plan is discussed in Ian Bent, The Compositional Process in Music
Theory 1713 1850, Music Analysis 3 (1984): 29 55; and by Nancy Kovaleff Baker in her introduction
to Kochs wr itings in Baker and Chr istensen, Aesthetics and the Art of Musical Composition, 111 43. Koch
was apparently unaware of Sulzers work when he prepared his rst volume, published in 1782.
31. Kochs analysis of Grauns ar ia can be found in translation in Baker and Chr istensen, Aesthetics
and the Art of Musical Composition, 163 76.
32. Baker and Chr istensen, Aesthetics and the Art of Musical Composition, 56.

297

298

analysis and theory

purpose of discovering useful ideas. But as attractive as such an orderly process is for
the ever-moralizing Sulzer, he ultimately realizes that it is not the most productive
where invention is concer ned. Better to be constantly occupied with ones art, for
one will surely stumble upon good ideas this way. As Sulzer ruefully admits,
The most important inventions probably do not ar ise through the rst deductive
manner descr ibed above, but rather by the second way: the main subject appears only
dimly at rst to the artist; he recognizes its importance and takes time to think about
its contents so it can be set in its proper light. This is how a famous composer told me
he worked. He had more than once saved mater ial that he heard by chance to use as
a theme or subject for a composition. He never could have invented anything as good
had he decided ahead of time to look for something having the identical character of
expression.33

In this Koch agrees, writing, Theory will never be able to invent a truly effective means to indicate just how the beg inning composer should contr ive that a
beautiful melody ar ises in his soul.34 However, once inspiration str ikes and invention has commenced, the composer must work with haste to capture the whole in
a plan of the sort Koch developed for Grauns aria. The plan in place, the composer
may then proceed to the more mechanical side of musical composition.
The mystery surrounding invention, together with the focus on purely musical
meaning afforded by mapping the structures of rhetor ic onto music, suggests an
independence for music adumbrated by Kochs reections on the value of theory.
Music, now more than simply mater ials to be manipulated, is something that ar ises
in the soul of the composer. Although the full owering of the idea that music was
an autonomous medium would have to wait until the tur n into the nineteenth century, we can see its seeds in eighteenth-centur y mappings from rhetor ic to music.
These provided a view of music as a powerful and highly dynamic medium with
the potential to shape the thought of those who listened to it.35

g rammar, rhetoric, and musical form It has been traditional

and convenient to think of grammar and rhetor ic as separate subdomains of language. Of the two, grammar is the simpler and more basic, necessar y for proper
communication but not the stuff to tur n heads. Grammar, reckoned this way, is thus
pure syntax, a careful if static ordering of linguistic units. By compar ison, rhetor ic is
more complex and subtle, persuasive because it has true semantic force. As the
means by which the passions can be moved, it is truly dynamic.36 Nonetheless, the
33. Baker and Chr istensen, Aesthetics and the Art of Musical Composition, 61.
34. Baker and Chr istensen, Aesthetics and the Art of Musical Composition, 186.
35. For a discussion of the emergence of the idea that music was an autonomous art form, see Lydia
Goehr, The Imaginary Museum of Musical Works: An Essay in the Philosophy of Music (Oxford: Clarendon
Press, 1992), chap. 6; and Bellamy Hosler, Changing Aesthetic Views of Instrumental Music in 18th-Century
Germany, Studies in Musicology, 42 (Ann Arbor, Mich.: UMI Research Press, 1981).
36. When Michel Foucault sought to character ize the distinct functions of rhetor ic and grammar,
he seemed rather deliberately to invert this traditional perspective, arguing that rhetor ic is spatial and that
grammar distr ibutes that spatiality in time that is, rhetor ic is static and grammar is dynamic (Foucault,
The Order of Things, 84). However, spatial here refers to the simultaneity of thought to its essential
and ongoing polyphony. Grammar distr ibutes this spatiality in time only by freezing fragments of it so

competing mode ls of mu s i c
distinctions between the two are hardly as clear-cut as this traditional perspective
would have it, and nowhere is this more evident than when they are mapped onto
music.
Returning briey to Riepel, his classication of cadence types provides a case in
point. In distinguishing between an Absatz that ends with a garbled cadence (verstmmelter Schlu ) and one that ends with a complete cadence (vollstndiger Schlu ),
Riepel uses the rhetor ical ter m enthymeme for the rst, and syllogismus for the
second.37 Descending momentar ily into the labyrinth of rhetorical ter ms, we can
discover that enthymeme refers to an incomplete syllogism, one in which either the
major or minor premise of the syllogism is left implied. A syllogismus, by contrast,
calls on the audience to draw an obvious conclusion, the major and minor premise
having been stated.38 Although the mapping between these two terms and the
musical phenomena to which Riepel would have them apply is less than perfect,
what is important is the implication that each of the grammatical components is
imbued with semantic value syntax has its semantic function. Indeed, as I argued
in chapter 4, the manipulation of musical syntax is one of the principal means by
which musical semantics is created.
Careful attention to the grain and texture of discourse about music is thus necessary to deter mine which of these metaphor ical mappings is used to structure a
given theory of music. Generally speaking, eighteenth-century writers used grammar to emphasize things that do not change in music suitably dened, an Einschnitt is always an Einschnitt, an Absatz always an Absatz. The same wr iters used
rhetor ic to emphasize how music functions in time, either in the challenges it presents a composer attempting to deploy musical mater ials or in the reaction of a listener persuaded by a musical argument. Grammar gave a view of musical for m that
was static, and rhetor ic gave a view of musical for m that was dynamic. Signicant,
however, is the context relative to which these character izations were made. As the
dynamic view offered by rhetor ic began to be more important for those who
thought about music, and as the cultural context for thinking about music changed,
mappings from language began to give way to other possibilities for conceptualizing music.

Formal Theory in the Nineteenth Century

Over the course of the nineteenth centur y, the two basic models of musical for m
that emerged in the eighteenth century remained in place. Musical works could be
regarded as manufactured things that compr ised discrete and inherently static
components, or as instances of quasi-abstract communication that unfolded dynamically in time. However, explicit mappings from language to music faded as they

that they can be ordered, much as musical mater ials, isolated from their becoming, are transfor med into
basic building blocks for the beginning composer.
37. Riepel, Grundregln zur Tonordnung insgemein, 59; idem, Smtliche Schriften zur Musiktheorie, 153.
38. These denitions are drawn from Gideon Burtons online Silva Rhetoric, at http:/humanities.
byu.edu/rhetor ic/silva.htm.

299

300

analysis and theory

were absorbed into musical nomenclature or replaced by other mappings that

served a similar pur pose. Instrumental in this process were the context within which
music was taught and the intellectual framework relative to which music was conceptualized. Beginning in the nal years of the eighteenth centur y, music instruction was taken over by institutions concer ned as much with the advancement of the
state as with education. In consequence, the move toward a complete and comprehensive curriculum that arose in furtherance of the for mer objective brought with
it the development of integrated courses of study that employed both static and
dynamic accounts of music. At around the same time, music came to be thought of
in organicist ter ms, and, as a result, even the most basic and isolated of musical materials were imbued with dynamic character istics. Hence, the conceptual metaphors
for musical for m established in the eighteenth century form as static, or form as
dynamic remained in place, but they were activated by a number of different linguistic metaphors.

the paris conservatoire and the berlin university

During the eighteenth century, music education in the French- and Ger man-speaking
countr ies was car ried out in a diffuse manner. Students typically studied with a local
music master or at a school associated with a church, and there were few if any
efforts at centralized music instruction. As a result of various political upheavals, this
situation began to change in the late eighteenth centur y, and music education,
together with instruction in other topics, began to be more closely controlled by
the state.
The circumstances of music education changed most rapidly in France. After the
1789 revolution, numerous educational reforms were proposed, focused largely on
efforts to create a utilitar ian educational system and to promote deductive reasoning for the arts and sciences; some of these were at least partially enacted.39 As a
consequence of the diminishing role of the church after the revolution, the system
of matrises the music schools associated with the major cathedrals collapsed,
leaving two principal options for music instruction: the School of the National
Guard (a product of the revolution) and the Old Regimes School of Singing
(founded in 1784 with the intent of providing a reliable supply of singers for the
Acadmie Royale de Musique, known familiarly as the Opra). To realize the goal
of conserving a French national music that was free from the inuence of Italy and
Germany, the Republic in 1795 consolidated the faculty and power of these two
schools and for med the Conservatoire National de Musique et de Dclamation.
The institution, as or iginally envisioned, was to provide free training for musicians
(especially those destined for militar y service or the Opra). It was also to serve as
ofcial arbiter of music used for state functions or music that was to be sanctioned
by the state. The Conservatoire also became the source of the rst national music
theory, for its or iginal charter included the provision that the texts used therein
would be wr itten by its faculty. Because of the political power that eventually
39. A thorough account of changes in music education after the French revolution is given in Cynthia M. Geselle, The Institutionalization of Music Theory in France: 1764 1802 (Ph.D. diss., Princeton University, 1989), 235 306.

competing mode ls of mu s i c
accrued to the Conservatoire, the texts that it adopted became the core of a national
music cur riculum.
The rst texts adopted by the Conservatoire were models of practicality, simplifying those notions that were deemed unnecessar ily complex and str iving toward
uniformity of method. As time went on, however, the faculty went beyond the
highly utilitar ian education envisioned by the or iginal reformers and developed a
sophisticated pedagogy that embraced virtually all aspects of musical thought. As an
example, the pedagogy created by Antoine Reicha, who was appointed professor of
counter point at the Conservatoire in 1817, extended from his Trait de mlodie of
1814 through the Cours de composition musicale (?1816 18) to the Trait de haute composition musicale of 1824 26.40 In these texts, Reicha, with the zeal of a geometer
and the passion of an artist, strove to provide a systematic approach that extended
from the most basic aspects of musical structure expressed through a grammar of
music analogous to that of Koch to a full account of the largest for ms. Thus was
institutionalized an approach to music that began with isolated components and
culminated in large for ms conceived as dramas enacted in sound.41 Within this
course of study, musical for m was presented as both static and dynamic. In the early
stages of the pedagogical plan, form was static, consisting of various components
that could be assembled to create musical compositions. In the later stages, it
became thoroughly dynamic, as this perspective yielded to one that accommodated,
even if it did not completely sanction, the artistic license necessary to create works
such as Mozarts Overture for the Marriage of Figaro.42
Changes to the Ger man universities were not as sudden as those that led to the
formation of the Conservatoire, but they were motivated by pragmatic and political
concer ns similar to those that shaped Frances new educational system. There had
developed in Ger many over the course of the eighteenth centur y the perception
that a university education served no useful pur pose. This notion, together with
efciencies imposed by French rule and the growing nancial burden presented by
universities whose traditional sources of nancial support had begun to erode, contributed to the disappearance of over half the Ger man universities dur ing the
Napoleonic era.43 As a way of counter ing this general disintegration, a delegation of
professors from the University of Halle, which had been suspended by Napoleon in
1806, approached King Fr iedrich Wilhelm III in the late summer of 1807 and asked
40. Reichas Cours de composition musicale ultimately replaced Charles-Simon Catels Trait dharmonie,
which had been approved by the General Assembly of the Conservatoire in 1801 as the Conservatoires
main har mony text.
41. For a discussion of afnities between Reichas account of the grande coupe binaire (now conventionally called sonata for m) and drama, see Peter Hoyt, The Concept of Dveloppement in the Early
Nineteenth Centur y, in Music Theory in the Age of Romanticism, ed. Ian Bent (Cambr idge: Cambridge
University Press, 1996), 141 62.
42. For a discussion of the range of Reichas analytical perspective, which included the Overture
from the Marriage of Figaro, see Ian Bents commentar y on one of the analyses from Reichas Trait de
haute composition musicale, in Music Analysis in the Nineteenth Century, 1: 152 53.
43. Theodore Ziolkowski, German Romanticism and Its Institutions (Princeton, N.J.: Princeton University Press, 1990), 228. My account of the history of German universities owes much to chap. 5 of this
book, which is given over to a consideration of the institution of the university dur ing the time of German Romanticism.

301

302

analysis and theory

that their university be reestablished in Berlin. The king was apparently receptive to
the idea and is reported to have responded that the state must replace through
intellectual powers what it has lost in the way of physical powers.44 Loosely modeled on the ideals associated with the University of Jena when Fr iedrich Schiller,
Johann Gottlieb Fichte, and Fr iedrich von Schelling lectured there, the University
of Berlin was conceived as an institution combining faculties in the sciences, medicine, and the humanities. It was to be unied by a central faculty of philosophy
professing an encyclopedic approach to knowledge. The university for mally
opened in October 1810 with a faculty that included Fichte and Fr iedrich Schleiermacher.
By the time the composer and music theor ist Adolph Ber nhard Marx was
offered a chair at Berlin in 1830, much of the spirit of the Romantic university had
faded to a memory.45 What remained were a faculty that included Hegel in its ranks
(who began lectur ing at Berlin in 1818) and the notion that the university was an
institution by, of, and for the state. Marxs encyclopedic approach to music theory
resembled Reichas sweeping attempt to provide a complete account of musical
knowledge in his three treatises. In 1837, Marx published his four-volume Die Lehre
von der musikalischen Komposition, praktisch-theoretisch. This comprehensive tract,
which went through six full or partial editions dur ing Marxs lifetime, presented a
graduated course of study that began with the simplest of musical mater ials and
forms and proceeded to the most complex.46 However, where the pedagogy of
Reichas treatises was conceptualized around the models of language and drama,
Marx adopted a different model, one closely associated with the same Romantic
idealism that had shaped the structure of the University of Berlin. This was the
model of organicism.

organicism As has often been noted, organicist models, which map features
from principally vegetative domains onto inorganic or articial domains, ourished
among early Romantic philosophers and wr iters.47 The model was, however, also in
use dur ing the late eighteenth century. Our friend Sulzer, as part of his more practical remarks on invention, advises the artist faced with an intractable situation to
put the work to one side for a short while. Gradually, solutions will present themselves, until a way is found where none was thought to exist.
44. Ziolkowski, German Romanticism, 287.
45. An account of Marxs place at the University of Berlin can be had in Scott Bur nhams introduction to Adolph Ber nhard Marx, Musical Form in the Age of Beethoven: Selected Writings on Theory and
Method, ed. and trans. Scott G. Burnham, Cambridge Studies in Music Theor y and Analysis, 12 (Cambridge: Cambridge University Press, 1997), 5 6.
46. For overviews of the content of Marxs Die Lehre and discussions of the educational context
within which it was embedded, see Burnhams introduction to Musical Form in the Age of Beethoven, 7 11;
and Mark Spitzer, Marxs Lehre and the Science of Education: Towards the Recuperation of Music
Pedagogy, Music and Letters 79 (1998): 489 526.
47. A thorough account of the role of organicist models in theor ies of invention dur ing the eighteenth and nineteenth centur ies (framed in ter ms of the memorable phrase vegetable genius) can be
found in Meyer Howard Abrams, The Mirror and the Lamp: Romantic Theory and the Critical Tradition
(Oxford: Oxford University Press, 1953), 198 225.

competing mode ls of mu s i c
It is one of many remarkable secrets of psychology that apparently clear thoughts can,
when one tur ns to them for deeper contemplation, refuse to be developed or comprehended in a clear way. But when they are left alone they will by themselves grow
in greater clar ity, much as that per iod in which plants ger minate unnoticed and all at
once burst into full bloom. Some concepts will gestate little by little in our mind, so
to speak, and extract themselves from the mass of obscure ideas into the clear light.
Every artist must rely upon such fortunate moments of genius, and if he cannot always
nd what he diligently seeks, he must await with patience that moment when his
thoughts r ipen.48

Here the mystery of invention, which Sulzer also connected with rhetor ic, is
explained in ter ms of an organicist model: ideas are seeds, and, given time and a little care, they will put forth complete and fully articulated intellectual constructs
the owers of invention that could never be produced by more mechanical
means.
The perspective evident in Sulzers account of unconscious psychology had a
profound effect among Ger man thinkers of the late eighteenth and early nineteenth
centur ies. These proponents of nature philosophy, including Fichte, Schelling,
Jean Paul Richter, and August Wilhelm Schlegel, embraced the newly emerging
biological sciences, nding in them a source of concepts that could replace those
derived from Cartesian or Newtonian mechanics. It was this model that Marx
absorbed and adapted for the character ization of musical for m.
In the rst volume of Die Lehre, after a few preliminar ies, Marx begins the journey that would culminate in a consideration of the most exalted compositional
forms of his day with a br ief contemplation of the relationship between a major
scale and its tonic. He observes that no note within the scale provides the sense of
repose that the tonic does: pausing on any other note is unsatisfying and demands
further motion. This observation leads Marx to a pr inciple basic to all music, the
main ter ms of which are placed in typographical relief on the page:
In this we have found the opposition that thoroughly per meates all of music:
Rest and Motion
Tonic and Scale.49
Marx extended this opposition to apply to all aspects of musical for m, formulated as
the dynamic impulse rest motion rest. For Marx, this basic patter n is truly comprehensive: it governs the modest four-measure phrase, as well as the entire expanse
of sonata for m.50

48. Baker and Chr istensen, Aesthetics and the Art of Musical Composition, 63.
49. Hierin haben wir einen durch die ganze Tonkunst wirksam hindurchgehende Gegensatz,
Ruhe und Bewegung
Tonika und Tonleiter
gefunden. Adolph Bernhard Marx, Die Lehre von der musikalischen Komposition, praktisch-theoretisch, 4th
ed. (Leipzig: Breitkopf and Hrtel, 1841 47), 1: 27.
50. This character ization is from Bur nhams introduction to Marx, Musical Form in the Age of
Beethoven, 9.

303

304

analysis and theory

It was in locating a source for this impulse that Marx tur ned to an organicist perspective that resonated not only with his era but also with the character ization of
invention that had been offered by Sulzer some sixty-ve years earlier. However,
where for Sulzer the idea is the seed, for Marx it is the motive: The motive is the
primal conguration [Urgestalt] of everything musical, just as the ger minal vesicle,
that membranous sac lled with some uid element (or perhaps with solid bodies),
is the pr imal conguration of everything organic the true primal plant or pr imal
animal. The motive, this conjunction of two tones or some other unities, simply
is.51 Marxs mapping from the organic domain to the musical domain makes possible an account of form that extends from little more than a pair of notes to entire
symphonies. Evident throughout is an approach to musical for m that is not only
dynamic (through the progression rest motion rest ) but also motivated by the
forces latent within the most basic of musical mater ials. According to Marx,
the motive, as the essential organic entity, offers something that can connect with
the composers spirit. Should the spir it adopt the motive, the composers intellect
will then come into play, shaping the motive to create works of art.52 The organicist perspective that Marx employs thus serves as a way to character ize the dynamic
aspect of musical for m and as a launching point for the pedagogy of composition. It
also fullls the promise of eighteenth-centur y theor ists who rst suggested that
music was an independent domain of meaning and expression: in the organicist perspective of the early nineteenth centur y, the fully autonomous work of music is
born.53
As heady as was the intellectual environment associated with the organicist perspective basic to his conception of form, Marxs Die Lehre was, as its full title
announced, both theoretical and practical. In his discussion of the basic elements of
musical for m, Marx notes a certain confusion regarding the notion of a per iod,
which is for med from two fundamental elements (the Gang and the Satz) derived
from the motive: If, in my compositional method, I associated the per iod with the
fundamental for ms, it was only for reasons of method, in order to keep this important and ubiquitously hard-working for m constantly before the students eyes from
the very beginning.54 Marxs pedagogy, shaped by the practical exigencies of a system of education intended to work the will of the state, had its static building
blocks, as well as its dynamic idealism. Although Marx could always trace these
building blocks back to their organic or igins, they nonetheless represented basic
components of a students education and, at least for the beg inner, of the musical
work.
51. Marx, Musical Form in the Age of Beethoven, 66. This passage is from Marxs 1856 essay Form in
Music.
52. Marx, Musical Form in the Age of Beethoven, 66 67.
53. Although my inter pretation of the or igin of the fully autonomous work of music is somewhat
different from that of Lydia Goehr, this is due mostly to the somewhat different focus of my approach.
I am concer ned here not simply with the idea of a fully autonomous work but with the complex of ideas
represented by organicism that allowed music theor ists to articulate the notion of an autonomous work
on both a practical and a theoretical level. Goehrs interpretation can be found in The Imaginary Museum
of Musical Works, 159 75.
54. Marx, Musical Form in the Age of Beethoven, 68.

competing mode ls of mu s i c
That the high abstraction of Marxs theory caused problems, at least for some
theorists, is suggested by the approach to motive taken by Hugo Riemann. Riemann, as editor of the ninth edition of Marxs Die Lehre (which appeared in 1887),
was well acquainted with Marxs conception of a motive. Nonetheless, in his MusikLexikon of 1882, Riemann began his denition of a motive by mapping from yet
another domain: Motive means in music, as it does in architecture, the most basic
character istic parts of an artistic structure.55 Riemann goes on to distinguish
between rhythmic motives, melodic motives, and har monic motives, all conceived
in architectural ter ms. Although the dynamic impulse basic to Marxs conception of
motive would appear to have vanished, in other of his works it is evident that Riemann has simply transfer red it to metr ic structure, which he construed as having,
as its essence, the impulse upbeat-downbeat.56 This construal gave Riemann a way to
represent the musical dynamism that Marx had developed through an organicist
metaphysics in more concrete ter ms. To the extent that the student or reader understands the basic exper ience of musical rhythm, so will he or she understand the
dynamic impulse basic to music.

models of musical form

To all appearances, Ebenezer Prout and

Percy Goetschius are a bit like Alice and the Mad Hatter (although I shall not venture to say which theor ist takes which role), seeming to talk about the same thing
but doing so in very different ter ms. But where Alice and the Hatter are engaged
with the elusive topic of Time, Prout and Goetschius are dealing with the rather
more humble matter of Motive. And where the Hatter is as surely a denizen of
Wonderland as Alice is not, Prout and Goetschius draw their ideas about motive and
form from the same pool of German music scholarship.
As we can now see, Prouts approach is infor med by the work of Marx and Riemann: he conceives of motive in ter ms of both metr ical accentuation and organicism, although in the end it is the for mer that is foregrounded in his theory while
the latter remains safely in the backg round. Goetschius, in contrast, conceives of
motive in almost exclusively grammatical ter ms. His approach resembles, rst and
foremost, that of Reicha (whose three treatises were translated into Ger man by Carl
Czerny and published in a four-volume French-Ger man bilingual edition in 1832).
To a lesser extent, it resembles that of Koch (whose work was not as widely known
in the nineteenth centur y). Thus, Goetschiuss method is shaped not only by a
grammatical approach to musical for m but also by a pedagogy that was thoroughly
pragmatic. Just as Reichas theory of form was part of a larger program of complete
instruction in music for a nation intent on building its musical identity, so
Goetschiuss manual was part of a ser ies of texts intended to fulll the same function in the New World.
Beyond being able to sort out the lineages of Prouts and Goetschiuss ideas
55. Motive nennt man in der Musik wie in der Architektur die letzten charakter istischen Glieder
eines Kunstgebildes. Hugo Riemann, Musik-Lexikon (Leipzig: Verlag des Bibliographischen Instituts,
1882), 605.
56. Hugo Riemann, System der musikalischen Rhythmik und Metrik (Leipzig: Breitkopf and Hrtel,
1903), 13 18.

305

306

analysis and theory

about motive and for m, however, we now also have a better grasp on the different
sorts of conceptual work done by the mappings they employ. Mappings that yield
a static account of musical for m put into relief the explicitly conceptual side of
music: they rely on our ability to isolate musical events or ser ies of musical events
and put them into categor ies. Members of the categor ies can then be compared
with one another, as can the categor ies themselves. The view that results is of music
as mater ial that can be shaped and manipulated, whose meaning is neatly contained
in discrete packages. Mappings that yield a dynamic account of musical for m put
into relief the exper iential side of music: music as something to be lived through,
which may move us or not, but which in all cases resists packaging into neat conceptual containers. Here meaning is at best only imperfectly contained and is in
constant danger of overowing the crude boundar ies of ontological necessity.
As the examples drawn from the ways musicians character ized for m in the eighteenth and nineteenth centur ies show, the cross-domain mappings we use in our
theories of music are more than simple cur iosities of language, for these mappings
have everything to do with our understanding of how music is organized, along
with what counts as music in the rst place.

two models of musical hierarchy

As I attempted to show in the preceding pages, descriptions of musical for m have
occasioned a number of cross-domain mappings, operating on both the specic and
generic level. The situation with musical hierarchy is somewhat different, in part
because hierarchy was not deemed an attr ibute of musical works as such but was
instead used to character ize the organization of musical mater ials. Consequently,
explicit discussions of hierarchy rarely appear in accounts of works of music but
instead for m the context for those accounts. As an example, let us consider three
contemporary analyses of the same passage: the rst eight measures of the theme
from the opening movement of Mozarts Piano Sonata in A major K. 331. All three
analyses deal with the rhythmic structure of the passage (the score appears in ex. 7.2).
example 7.2
mm. 1 8

W. A. Mozart, Piano Sonata in A major K. 331, rst movement, theme,

Andante grazioso

competing mode ls of mu s i c
Edward Cones analysis of the theme appeared in his Musical Form and Musical
Performance of 1968. The analysis comes just after the introduction of a str iking
metaphor for the melodic and har monic shape of a musical phrase: Cone likens the
shape of a phrase to a game of catch:
If I throw a ball and you catch it, the completed action must consist of three parts: the
throw, the transit, and the catch. There are, so to speak, two xed points: the initiation
of the energy and the goal toward which it is directed; the time and distance between
them are spanned by the moving ball. In the same way, the typical musical phrase consists of an initial downbeat (/), a period of motion ( ), and a point of arrival marked
by a cadential downbeat (\).57

This metaphor then provides the framework for his analysis of the rhythm of
Mozarts theme, a portion of which is given in example 7.3. Cone rejects a simple
alternation of strong and weak measures of the sort posited by Riemann on the basis
of harmony. Cone asserts, rather, that the rst and fth measures of the period should
be considered strong because of the r m statement of tonic. The fourth and eighth
measures should also be deemed strong because of the emphasis provided by the
cadences. A consideration of the motivic structure of the per iod enables Cone to
rene this analysis. The rst phrase consists of two individual sequential measures followed by a two-measure unit. The shape of the two-measure unit duplicates the
shape of the phrase in miniature: it consists of two half-measure units followed by a
full measure. As we see in example 7.3, the internal dynamic of the two-measure unit
is the same as that of the complete phrase. The second phrase begins with the same
sequential measures as the rst, but in m. 7 the compression that prepares the closing cadence br ings the r ise from A4 to CS5 into prominence. Cones hearing of this
rise as a third member of the sequence prompts him to assign it the same rhythmic
symbol as the second member of the sequence, heard in m. 6.Voice-leading concerns
and the forward energy of the sforzando cause Cone to group the nal eighth note of
m. 7 with the cadential mater ial of m. 8; this constitutes the closing downbeat.
Although Leonard Meyers analysis of this theme appeared after Cones (he presented it rst in one of the Ernest Bloch lectures he gave at the University of California at Berkeley in 1971), his methodology is essentially the same as that which he
had developed with Grosvenor Cooper over a decade earlier. Cooper and Meyer
understood rhythmic structure to be perceived as an organic process in which smaller
rhythmic motives, while possessing a shape and structure of their own, also function
as integral parts of a larger rhythmic organization.58 Rhythmic relationships are analyzed as patter ns of beats, in which a stable accent and one or more weak beats are
grouped together in different ways; these low-level, foreground patter ns combine
with one another in various ways to for m more extended rhythmic groupings. In this
way the musical surface gives rise to a hierarchy of rhythmic groupings.59
57. Edward T. Cone, Musical Form and Musical Performance (New York:W.W. Norton, 1968), 26 27.
58. Grosvenor W. Cooper and Leonard B. Meyer, The Rhythmic Structure of Music (Chicago: University of Chicago Press, 1960), 2.
59. Leonard B. Meyer, Explaining Music: Essays and Explorations (Berkeley: University of California
Press, 1973; repr. Chicago: University of Chicago Press, 1978), 27 28. Meyers Bloch Lectures of 1971
form the basis for Explaining Music.

307

308

analysis and theory

example 7.3 Edward Cones analysis of the opening theme of Mozarts Piano Sonata K.
331, mm. 1 8

Meyers analysis of the opening per iod of Mozarts theme is given in example
7.4. The rst level of the analysis starts at the half-measure level of the music, impor- patterns of the rst two measures of each phrase do
tant for Meyer because the d
much to deter mine the grouping and accentual patter n of the complete bar. In
mm. 3 and 4, Meyers analysis on level 2 cor responds closely with Cones reading:
although the symbols are slightly different (Meyers indication for a retrospectively
weak accent replacing Cones symbol for an initial downbeat), on this level as well
as the third and fourth, the analyses for the most part agree. The main difference
between the two is that Cone starts his analysis at what for Meyer is the third level
of a rhythmic hierarchy.
In his article The Theory and Analysis of Tonal Rhythm (1978), Robert Morgan noted that, although both Cones and Meyers work reected the inuence of
Schenkers theory of tonal music, their analyses of Mozarts theme were still
beholden to certain aspects of metrical dogma. For his part, Morgan proposed considering the theme from a more consistently Schenkerian viewpoint. Morgan noted
that Schenkers theory supplied a method for locating points at which structural
motions or iginate and ter minate; his idea was to cor relate these points of structural
origin and ter mination with the points of principal accent through which the larger
rhythm of a passage is regulated. In Morgans analysis, the rst phrase of the theme

competing mode ls of mu s i c
example 7.4 Leonard B. Meyers analysis of the opening theme of Mozarts Piano
Sonata K. 331, mm. 1 8
A

antecedent

m'

n
p

Andante grazioso

p'

1.
2.
3.
4.

consequent
m

a'
m'

n'

is governed by the overall motion from CS5 and A3, in m. 1, to B4 and E3 in m. 4

(see ex. 7.5, level 3). This overall motion is articulated on the next level (level 2 of
ex. 7.5) by a motion down to A4 and FS3 in m. 3. Although the A could be seen as
the completion of a third span from CS5, the bass FS3 and the acceleration away
from A4 to B4 on the second beat of m. 3 prevent closure.
Up to this point, Morgans analysis agrees, in most respects, with the analyses by
Cone and Meyer. Morgan does not, however, believe the rst beat of m. 4 should be
heard as accented relative to the second beat. The acceleration of m. 3 continues
into m. 4, in his view, carrying the music forward to the ar rival of the bass on the
dominant on the second beat. This ar rival, according to Morgans analysis, serves as
the rhythmic goal and closing accent of the phrase. From Morgans perspective,
then, the rhythm of this phrase is not a matter of strong and weak beats or strong
and weak measures; it is created rather by an articulated tonal process that moves
between the two points dened by the outer-voice frame.60
The acceleration in the rst phrase is answered by an even greater acceleration
60. Robert P. Morgan, The Theor y and Analysis of Tonal Rhythm, Musical Quarterly 64 (1978):
446.

309

310

analysis and theory

example 7.5
Robert Morgans analysis of the opening theme of Mozarts Piano Sonata
K. 331, mm. 1 8

in the second, for the consequent phrase must go through the dominant to ar rive
on the tonic. That ar rival gives a different character to the closing accent of the second phrase; for Morgan, it is the difference in character between the two closing
accents, and not some analogy to two large beats, that is responsible for the upbeatdownbeat effect created by the pair ing of antecedent and consequent. Implicit in
this analysis is the notion that accent refers to a point of emphasis, which also leads
Morgan to reject Cones and Meyers reading of the rhythm of mm. 3 and 4 as a
diminution of the overall rhythm of mm. 1 through 4. If mm. 3 and 4 supply the
downbeat for mm. 1 and 2, they must represent an ar rival, but in fact the point of
arrival occurs only at the end of the unit. Morgan does not believe m. 3 can simultaneously be both an ar rival and a departure.
Although these three analyses converge on a number of points, their divergences
reect two different models of musical hierarchy. The rst model, which gives r ise
to what I call chain-of-being hierarchies, does a good job of capturing the har monic
and scale-degree dependencies typical of the har monic organization of instrumental music of the late eighteenth centur y. The second model, which gives rise to
what I call atomistic hierarchies, does a good job of capturing the regularity and nesting of accentual patter ns typical of musical meter dur ing the same per iod. Because
both models are associated with what appears to be a common set of structural relations called hierarchywe might be misled into thinking that they are simply
variations of each other. Such is not the case, however, for these models developed
at different times in history, and with reference to different domains. These differences are reected in the structure of the models and also in the ways they have
been applied to music.

competing mode ls of mu s i c
Histor ical Models of Hierarchical Organization

The principal mark of chain-of-being hierarchies is a conception of the universe as

a precisely ar ranged system of interdependent levels or degrees of existence extending from a Supreme Being down to the lowliest organism. Although this concept
is part of an incredibly rich collection of ideas one wr iter has suggested that the
history of chain-of-being hierarchies is, in reality, a history of Occidental thought61
only a few central aspects need concer n us here.
The ingredients for this complex of ideas came from Plato and Aristotle; Aristotles connection of powers of the soul with levels of being in De anima proved
particularly inuential. It was in Neoplatonism, however, that these ideas rst
appeared fully organized into a coherent general scheme.62 Crucial to this organization was Plotinuss idea that the perfection of the One included a superabundance
which, overowing, was the source of the Many. According to Plotinus, this generation of the Many from the One could not come to an end as long as any possible
variety of being in the descending ser ies was left unrealized.
One of the principal ways the Neoplatonic cosmology was transmitted to the
Middle Ages was through Macrobiuss Commentary on the Dream of Scipio (which
dates from the late fourth to early fth centur ies a.d.). In a passage expanding on
the revelation presented to Scipio that minds are given to man out of the eter nal
and divine re of the stars, Macrobius wr ites:
Accordingly, since Mind emanates from the Supreme God and Soul from Mind, and
Mind, indeed, forms and suffuses all below with life, and since this is the one splendor lighting up everything and visible in all, like a countenance reected in many mirrors arranged in a row, and since all follow on in continuous succession, degenerating
step by step in their downward course, the close observer will nd that from the
Supreme God even to the bottommost dregs of the universe there is one tie, binding
at every link and never broken. This is the golden chain of Homer which, he tells us,
God ordered to hang down from the sky to the earth.63

61. C. A. Patrides, Hierarchy and Order, in Dictionary of the History of Ideas: Studies of Selected Pivotal Ideas, ed. Philip P.Wiener (New York: Charles Scr ibners Sons, 1973), 2: 434.
62. Concer ning the complex of ideas associated with the g reat chain of being and Neoplatonism,
see Arthur O. Lovejoy, The Great Chain of Being: A Study of the History of an Idea (Cambridge, Mass.: Harvard University Press, 1936), 61; concer ning De anima, see p. 58.
63. Ambrosius Aurelius Theodosius Macrobius, Commentary on the Dream of Scipio, ed. and trans.
William Har ris Stahl (New York: Columbia University Press, 1990), 145. It is also in Macrobius (pp.
185 200 in the Stahl translation) that we nd a detailed discussion of the har mony of the spheres, as well
as the legend of Pythagoras and the blacksmiths shop. The connection between har mony and order
found in Macrobius was an important one, and it may have contr ibuted to the application of a
metaphor ical chain-of-being hierarchy to music. However, it is also str iking that the image of a ladder
is used in both domains. In the Middle Ages, the hierarchical organization summoned by the Great
Chain of Being was commonly represented by Jacobs Ladder; see Paul G. Kuntz, A Formal Preface and
an Infor mal Conclusion to The Great Chain of Being: The Universality and Necessity of Hierarchical
Thought, in Jacobs Ladder and the Tree of Life: Concepts of Hierarchy and the Great Chain of Being, ed. Marion Leathers Kuntz and Paul Gr imley Kuntz, American University Studies, Ser.V, Philosophy, 14 (New
York: Peter Lang, 1987), 6. Also, some early mnemonic diagrams show the gamut as notes on a ladder; see
Karol Berger, Musica Ficta: Theories of Accidental Inection in Vocal Polyphony from Marchetto da Padova to
Gioseffo Zarlino (Cambridge: Cambridge University Press, 1987), 9 10.

311

312

analysis and theory

The notion of a Great Chain of Being proved a concise and powerful model through
which humans could give order to their universe. It also provided a model through
which the distr ibution of power could be accomplished and justied. The model
is quite evident in ecclesiastical hierarchies: power ows from the top down, in the
same way that being and soul ow from the Godhead down through the various
levels of being, from highest to lowest.
Inherent in the Neoplatonic scheme is a tension between the self-sufciency of
the One and the multitudinous abundance of creation it gives rise to: it is a paradox
that the One needs nothing else and yet produces everything else. Giordano Bruno,
writing in the sixteenth centur y, reveled in this paradox.64 Yet it was also Bruno
who, by combining the concept of atomism with the comprehensive organization
provided by the Neoplatonic scheme, inspired a way of thinking about hierarchical organization that would eventually compete with the Great Chain of Being.
For Bruno, atomism was a metaphysical pr inciple that provided the basis for a
demonstration of the underlying unity of all nature.65 This view was especially
inuential in England, where Bruno lived from 1583 to 1585 and where two factors
combined to transfor m his vague, metaphysical scheme into a compelling account
of order. The rst factor came in the for m of a challenge to the model for the distribution of power provided by the application of a chain-of-being hierarchy to the
social sphere. In a social chain-of-being hierarchy, the power of any individual is
linked, as an essential property, to the level occupied by that individual within the
system: the higher the level, the greater the power over others. A refutation of this
model can be seen in Thomas Hobbess Leviathan of 1651. Hobbes argued that all
humans were basically equal in the faculties of body and mind: all had approximately equal power over others. Because of this equality, the natural state of individual humans was one of competition. Only by imposing a social order on humans
from without, in the for m of a common power to keep them all in awe, was it
possible to overcome the natural state of virtual or actual war.66 Thus the properties
ascribed to each member of a society (which result in equal measures of power for
all members) are different from those according to which societal order should be
established (which result in unequal measures of power): the properties of the individual are independent of the properties of the system.
The second factor important in the development of an alter native view of hierarchical organization was the increased presence of mechanical devices in peoples
lives.67 These machines provided thinkers of the time with practical models of the
organization of nature. Among the most important of these was the mechanical
clock. As Gideon Freudenthal has observed: In the construction of the mechanical clock it was possible to produce complicated movements by an appropriate disposition of gears and a dr iving force. The task of science [in the seventeenth and
eighteenth centur ies] can be inter preted as the attempt to discover in a limited area
64. Lovejoy, The Great Chain of Being, 120.
65. Robert H. Kargon, Atomism in the Seventeenth Centur y, in Dictionary of the History of Ideas:
Studies of Selected Pivotal Ideas, ed. Philip P.Wiener (New York: Charles Scr ibners Sons, 1973), 1: 133.
66. Thomas Hobbes, Leviathan, or the Matter, Forme, and Power of a Common-Wealth Ecclesiasticall and
Civill (London: Andrew Crooke, 1651), chap. 13.
67. Kargon, Atomism in the Seventeenth Century, 1: 132.

competing mode ls of mu s i c
and to a limited extent the pr inciples of construction of the divine clock.68 The
model of a clock is one that appears consistently in the dispute between Isaac Newton and Gottfr ied Wilhelm Leibniz on the ultimate nature of matter. Of particular
interest are the different methodological approaches used by Newton and Leibniz.
Leibniz held to what had become, by the early eighteenth centur y, a traditional
Baconian methodology: scientic inquir y should seek to disassemble the divine
clock, piece by piece, never speculating beyond what could be observed. Although
Newton was in sympathy with this methodology, it was inadequate for a purely
mechanical account of physics, since the ultimate workings of the clock its
atomic structure could not be observed. Newtons solution, given in the nal
Query appended to the fourth edition of his Opticks, was a practical one that
divided scientic theor izing into two halves.69 The rst half is analysis: This Analysis consists in making Exper iments and Observations, and in drawing general Conclusions from them by Induction, and admitting of no Objections against Conclusions, but such as are taken from Exper iments, or other certain Truths.70 The
second half is synthesis or composition, in which the various inductive generalizations produced by analysis are organized into a logical system based on a limited
number of essential pr inciples. In the case of the divine clock, synthesis of necessity played a larger role than analysis (which led Leibniz to accuse Newton of arguing from hypothesis rather than from observation). However, Newton believed this
exception to his overall methodology was justied by the potential results:
To derive two or three general Pr inciples of Motion from Phnomena, and afterwards
to tell us how the Properties and Actions of all cor poreal Things follow from those
manifest Pr inciples, would be a very great step in Philosophy, though the Causes of
those Pr inciples were not yet discoverd: And therefore I scruple not to propose the
Principles of Motion above-mentiond, they being of very general Extent, and leave
their Causes to be found out.71

In other words, the scientist simply had to assume that the general pr inciples produced by synthesis held, even on levels that could not be directly observed: the
mechanical workings of the divine clock were assumed to be consistent throughout.
When combined with the independence of elemental and systemic properties
proposed in the social sphere, this approach developed into a new conception of
hierarchical order, which I call an atomistic hierarchy. In contrast to a chain-ofbeing hierarchy, an atomistic hierarchy can be built from the bottom up, its organization governed by a set of general pr inciples that obtain no matter what the

68. Gideon Freudenthal, Atom and Individual in the Age of Newton: On the Genesis of the Mechanistic
World View, trans. Peter McLaughlin, Boston Studies in the Philosophy of Science, 88 (Dordrecht: D.
Reidel, 1986), 63. Sulzer also used the analogy of a clock, but to character ize the unity of a work of art;
see Baker and Chr istensen, Aesthetics and the Art of Musical Composition, 43 46.
69. R. I. G. Hughes, Reason and Exper iment in Newtons Opticks: Comments on Peter Achinstein, in Philosophical Perspectives on Newtonian Science, ed. Phillip Br icker and R. I. G. Hughes (Cambridge, Mass.: MIT Press, 1990), 177.
70. Isaac Newton, Opticks, or a Treatise of the Reections, Refractions, Inections and Colours of Light, 4th
ed. (London: W. Innys, 1730; repr. New York: Dover, 1952), 404.
71. Newton, Opticks, 401 02.

313

314

analysis and theory

specic attr ibutes of a given level. It is a concise and concrete account of organization, and one that, through applications in physics, chemistry, and biology, has had
an enor mous inuence on how humans view their world.

Structural Features of Chain-of-Being and

Atomistic Hierarchies
Each of these models of hierarchy offers a different view of how the world is organized. The fundamental conceit of chain-of-being hierarchies is to regard a domain
as a huge organism pervaded by a force, the or igins of which are mysterious.72
Being and political power are both examples of such forces. The mysterious
force behind any chain-of-being hierarchy manifests itself as properties instantiated
in varying degrees by the elements of the domain. In a traditional Great Chain of
Being, for instance, a rock has the property of substance, a tree has the properties of
substance and life, and an insect has the properties of substance, life, and mobility. The
domain can be organized into a hierarchy in which any given level is distinguished
by a specic set of properties that embody aspects of being. These properties include
all of the properties that distinguished lower levels, plus properties unique to the
given level: trees include substance and add life; insects include substance and life and
add mobility. Inclusion within the hierarchy is thus of properties, but not elements.
The fundamental conceit of atomistic hierarchies is to regard a domain as an
extended mechanism operating according to a limited set of general pr inciples. The
laws of classical physics are an example of one such set of principles. All of the
actions and properties of elements within the domain, from the smallest to the
largest, and from the most simple to the most complex, follow these general pr inciples. Thus subatomic particles and planets both confor m to the same pr inciples.
Classes of elements may be distinguished according to common confor mance with
the general pr inciples of the system. The actions and properties of electrons, protons, and neutrons on the atomic level are by no means identical, but they can be
distinguished from the actions and properties of atoms on the molecular level. In
addition, all of the elements within a given class have the shared property of combining into units: electrons, protons, and neutrons combine to for m atoms, and
atoms combine to for m molecules. The domain can be organized into a hierarchy
in which each level is a confor mance class whose elements combine into units that
constitute the elements of the next higher level in the hierarchy. This process continues recursively until the limits of the system are reached. Inclusion within the
hierarchy is thus of elements, but not properties.
The unique features of each of these models of hierarchy can be character ized in
terms of relationships that are established between the general, abstract property of
energy and the container schema. Energy takes for m as either the mysterious
72. While the parallels with the ideology of organicism that blossomed in the late eighteenth century are intriguing, it should be remembered that the organicism associated with chain-of-being hierarchies dates back at least to the Middle Ages. Accordingly, the mysterious or igins of the power behind
chain-of-being hierarchies is somewhat different from the mystery associated with Romantic organicism
since the for mer is an essential property of the model where the latter would be better descr ibed as accidental.

competing mode ls of mu s i c
force of a chain-of-being hierarchy or the power that dr ives the mechanism of an
atomistic hierarchy. For both models, the container schema provides the basic
internal structure of the levels that make up the hierarchical system.
In a chain-of-being hierarchy, the energy proper to the system (that is, being
or political power) is distr ibuted unequally throughout the system. Systemic organization is consequently based on the amount of energy (inter preted as properties
of being or extent of political power) each element of the system has. The containers (or hierarchical levels) that articulate this organization g roup elements
according to their amount of energy (or systemic properties). Although chain-ofbeing hierarchies often offer vivid and dynamic accounts of organization, hierarchical structure (based on the contents of the containers that make up the hierarchy) is, in fact, relatively abstract.
In an atomistic hierarchy, energy is distr ibuted throughout the system according
to a xed set of general pr inciples. Systemic organization is consequently based on
the way different elements of the system manifest these pr inciples. The containers
(or hierarchical levels) that articulate this organization group elements according to
common confor mance with the general pr inciples. The generality of the organizing principles of atomistic hierarchies means that the pr inciples can seem somewhat
removed from actual exper ience, although hierarchical structure (based on the contents of the containers that make up the hierarchy) is, in fact, relatively concrete.
Given these structural attr ibutes, each model has a slightly different potential as
a source domain. A chain-of-being hierarchy will map these predicates onto a target domain:
The target domain is pervaded by a mysterious force.
This force manifests itself as properties instantiated to varying degrees by
the elements of the domain.
The domain can be organized into a hierarchy, in which each level is distinguished by a unique set of properties.
Each level of the hierarchy includes all of the properties of the next lower
level and adds new properties unique to that level.
The higher levels of the hierarchy more completely manifest the mysterious force (and are thus more perfect) than the lower levels of the hierarchy.
Because chain-of-being hierarchies have often been used to structure natural and
human domains, applying the model to other domains may suggest that elements of
that domain have natural or anthropomorphic aspects, even though these are not
necessary entailments of the mapping.73
An atomistic hierarchy will map these predicates onto a target domain:
The target domain is analogous to an extended mechanism.
This mechanism operates according to a limited set of general pr inciples.
All of the actions and properties of elements within the domain follow
these general pr inciples.
73. As the analyses at the end of chap. 2 suggest, factor ing anthropomorphism into the application
of chain-of-being hierarchies to different domains would quickly lead to a dense and highly interconnected network of correlated domains.

315

316

analysis and theory

Classes of elements may be distinguished according to common conformance with the general pr inciples and the shared property of combining
into units.
The domain can be organized into a hierarchy, in which each level represents a confor mance class.
The elements on each level of the hierarchy combine to for m units; these
units constitute the elements of the next higher level in the hierarchy.
The recursive combination of elements into units continues until the limits of the system are reached.
Because atomistic hierarchies have often been used in scientic explanations, applying the model to other domains may suggest that a scientic approach is being
used, even though this is not a necessary entailment of the mapping.
Given the different structural attr ibutes each model of hierarchy will map onto
a target domain, it follows that each provides a different model for reasoning about
the target domain. Nowhere is this clearer than in the rst thoroughgoing applications of these models to the domain of music dur ing the early nineteenth century.

Mapping Models of Hierarchy onto Music

chain-of-being hierarchies and tonal theory Early in his

Cours complet dharmonie et de composition of 1806, Jrme-Joseph de Momigny presents a C major scale ar ranged as two symmetr ical tetrachords:
G

C
C

Momigny reminds his readers that the three main notes of the scale are G, C, and
F. He goes on to stress the centrality of tonic, providing a telling image in the
process: The tonic tr iumphs over all the other notes. It is [tonic] who plays the premier role in le Ton. It is the center of gravity, the purpose of all purposes, the end
of all ends; in a word, it is to [tonic] that the scepter of the musical empire is
entrusted.74 At the beginning of the chapter in which this passage occurs,
Momigny linked his concept of Ton to that of hierarchy;75 with his image of tonic
holding the scepter, he leaves no doubt as to what sort of hierarchy he has in mind.
74. The complete passage is as follows: Jai dj dit que ce sont l les trois Notes pr incipales du Ton.
La tonique lemporte sur toutes les Notes. Cest elle qui joue le premier rle dans le Ton. Elle est le centre de gravit, le but de tous le buts, la n de toutes les ns; en un mot, cest elle que le sceptre de
lempire musical est con. Jrme-Joseph de Momigny, Cours complet dharmonie et de composition,
daprs une thorie neuve et gnrale de la musique (Paris: Chez lauteur, 1806), 1: 47. The image is used again
on p. 81 of the same volume.
Although the early nineteenth-century notion of Ton can be considered to be approximately equivalent to the modern notion of key, there are also differences between the two concepts; specically, Ton has
a dynamic quality lacking in key. For this reason, I shall not use the latter as a translation of the former.
75. Ton is the hierarchy, the order established between the notes of a genus and a mode. [Le Ton
est la hirarchie, lordre tabli entre les Notes dun genre et dun mode.] Momigny, Cours complet
dharmonie et de composition, 1: 47.

competing mode ls of mu s i c
A similar, if less colorful, image may be found in the fuller explanation of order in
music that he offered twelve years later in the second music volume of the Encyclopdie mthodique. This order, which Momigny claims is purely metaphysical, consists in a natural hierarchy of seven notes that are arranged according to the authority of the one called tonic.76 As Renate Groth has noted, this is essentially the same
denition of order given over two decades later by Franois-Joseph Ftis; the most
important difference is that Ftis assigned it the name tonalit.77
By mapping a chain-of-being hierarchy onto the domain of music, Momigny
creates a very specic notion of Ton. The origin of Ton is mysterious Momigny
never gets more specic than claiming that it is metaphysical.78 The notes organized by Ton occupy different levels, according to their directive force; Momigny
establishes these levels through various musical propositions. Perhaps most important, notes are regarded as agents with powers proportional to their status within the
hierarchy: the properties character istic of Ton increase as one ascends toward tonic.
Thus, even though Momigny identies dominant as the generator of the scale
(based on the presence of a minor seventh in the overtone ser ies), he argues that
it must be regarded as subservient to tonic, for the author ity of tonic cannot be
challenged.79
The conception of tonal organization produced by mapping a chain-of-being
hierarchy onto music was a powerful one.We see it not only in Momignys early
formulation of the concept of Ton but also in the persuasive account of tonal organization presented in the rst and third volumes of Heinrich Schenkers Neue
musikalische Theorien und Phantasien. As presented in Schenkers Harmonielehre of
1906, the mysterious force that pervades the domain of music is that of nature, manifested through the overtone ser ies. Earlier, in chapter 3, in a discussion of conceptual models as the basis for theor ies, I quoted a central passage from the Harmonielehre on Schenkers organicist approach to the overtone ser ies and pursued its
application to his approach to musical analysis. The same passage and related discussion is directly relevant to our present concer ns, and we turn to them again here.
In Schenkers conception, every tone within the musical domain is possessed of the
urge to produce endless overtones, much as living beings have procreative urges.80
Beginning with a single tone, Schenker creates a community of tones ar ranged in
76. Jrme-Joseph de Momigny, Musique, in Musique, Series 36 of Encyclopdie mthodique, ou par
ordre de matires, ed. Nicolas tienne Framry, Pierre Louis Ginguen, and Jrme-Joseph de Momigny
(Paris: Chez Madame veuve Agasse, 1818), 2: 179.
77. Renate Groth, Die franzsische Kompositionlehre des 19. Jahrhunderts (Wiesbaden: Franz Steiner,
1983), 58.
78. On this use of metaphysical (which can also be seen in the work of Ftis), see Thomas Chr istensen, Ftis and Emerging Tonal Consciousness, in Music Theory in the Age of Romanticism, ed. Ian Bent
(Cambridge: Cambridge Univesity Press, 1996), 37 56.
79. Momigny, Cours complet dharmonie et de composition, 1: 49 51, 47 48.
80. It is self-evident that the urge to produce unending generations of overtones belongs to every
tone in equal measure. One might also compare this urge to that of animals, for it appears in fact to be
in no way infer ior to the procreative urge of a living being. [Es ist selbstverstndlich, da den Trieb,
Generationen von Obertnen ins Unendliche zu zeugen, jeder Ton in gleichem Mae besitzt. Man darf,
wenn man will, auch diesen Trieb einem Animalischen vergleichen, denn er scheint in der Tat dem
Fortpanzungstr ieb eines Lebewesens durchaus nicht nachzustehen.] Heinr ich Schenker, Harmonielehre,
Neue musikalische Theor ien und Phantasien, 1 (Stuttgart: J. G. Cottasche, 1906), 14, p. 42.

317

318

analysis and theory

hierarchical order, with the notes closest to the source-pitch having priority over
those more remote.81 The tonal system that results is a manifestation of the desire of
the source tone to rule over others and to extend this rule as far as possible.82
In Der freie Satz, from 1935, Schenker retained this conception of tonal organization and, by emphasizing the organicism latent in this conception, built from it an
account of the coherence of entire works of music. He did this by proposing that
the transfor mation of the musical source (now the Naturklang, or chord of nature)
proceeded through discrete levels from background to foreground. The background
represented an abstract, initial working-out of the tendencies of the chord of nature
through the Ursatz, or fundamental structure; the foreground represented the musical work as we conventionally think of it.83 Schenker makes clear that this foreground emanates from and is controlled by the background.84 Thus the hierarchical organization of music is not restricted to the ordering of single tones into a
scalar system, but manifests itself as a ser ies of transfor mations that steadily enr ich
musical content. In Schenkers vision of the relationship between structural levels,
a version of the Neoplatonic cosmogony central to the or iginal Great Chain of
Being continues to echo: Between fundamental structure and foreground there is
manifested a rapport much like that ever-present, interactional rapport which connects God to creation and creation to God. Fundamental structure and foreground
represent, in terms of this rapport, the celestial and the ter restrial in music.85

81. Schenker, Harmonielehre, 17 18, pp. 54 55.

82. But now, what does relationship mean in the life of the tone and what in this context would
be implied by the intensity of living life to the fullest? The relationships of a tone are its system. Once
the egoism of the tone manifests itself in such a way that the tone (rather like a human) might rule over
its fellow-tones rather than be ruled by them, then these very systems provide the means for satisfaction
of this egoistic love of power. A tone rules over others when it subjugates them to the relationships outlined in the systems above. [Nun aber, was soll denn Beziehung heien im Leben des Tons und was
soll hier wohl Intensitt des Sichauslebens bedeuten? Beziehungen des Tones sind seine Systeme. uert
sich der Egoismus des Tones dar in, da er, hierin einem Menschen hnlich, lieber ber seine Mittne
herrscht, als da er von ihnen beher rscht wird, so sind ihm zur Befr iedigung dieser egoistischen
Herrschsucht eben in den Systemen die Mittel zur Her rschaft geboten. Ein Ton her rscht ber die
andern, wenn er sich dieselben nach den in den Systemen geschilderten Verhltnisse unterwirft.]
Schenker, Harmonielehre, 38, pp. 106 07.
83. Robert Snarrenberg has recently pointed out what I would character ize as competing conceptual
models in the usual translations of Schenkers terms (e.g., fundamental structure for Ursatz); see Snarrenberg, Competing Myths: The Amer ican Abandonment of Schenkers Organicism, in Theory, Analysis and Meaning in Music, ed. Anthony Pople (Cambr idge: Cambridge University Press, 1994), 45 52.
84. All the foreground diminutions, including the apparent keys arising out of the voice-leading
transfor mations, ultimately emanate from the diatony in the background. Heinrich Schenker, Free Composition (Der Freie Satz), ed. and trans. Ernst Oster, New Musical Theor ies and Fantasies, 3 (New York:
Schirmer Books, 1979), 4, p. 11. [Ist doch alle Vordergrund-Diminution, einschlielich der scheinbaren
Tonarten aus den Stimmfhrungsvervandlungen, zuletzt eben aus der Diatonie im Hinterg rund
erossen.] Heinr ich Schenker, Der freie Satz, Neue musikalische Theor ien und Phantasien, 3 (Vienna:
Universal-Edition A. G., 1935), 4, p. 32.
85. Schenker, Free Composition, 160 (Passage E). Aehnlich wie von Gott zum Geschpf zu Gott eine
Fhlungnahme waltet, stets ineinanderlaufend, stets gegenwrtig, wirkt sich eine Fhlungnahme auch
zwischen Ursatz und Vordergrund aus als gleichsam einem Jenseits und Diesseits in der Musik.
Schenker, Der freie Satz, 18.

competing mode ls of mu s i c

atomistic hierarchies and metrical theory

A rather different view of musical order is presented by Gottfr ied Weber in his Versuch einer geordneten Theorie der Tonkunst, which appeared in a number of editions in the early
nineteenth century. In the course of introducing this particular view,Weber focused
not on the organization of pitches but on the organization of musical rhythm. For
Weber, rhythm depends on regularity, and the mechanical nature of this regularity
is revealed in his breathless denition:
If . . . a symmetrical measured division of the times is found in a piece of music, i.e. the
time is distr ibuted into exactly equal general divisions, and these are again divided into
equal parts, and the latter are farther separated into equal smaller quotas or proportional parts, &c. and the duration of the tones in relation to one another is exactly
measured according to such divisions of time, so that a general division always appears
as a symmetr ically ar ranged group of several smaller portions of time, and these taken
together as a smaller subordinate group of yet smaller parts, and the accent is also symmetr ically apportioned amongst all these divisions of time then the music is measured and rhythmical.86

This mechanical notion of rhythm is accompanied by a mechanical notion of timekeeping: Weber emphasizes the importance of inventions like the metronome and
devotes a lengthy footnote to instructions for constructing a pendulum that will
accomplish the same thing.87
Weber follows his digression into the mechanics of time-keeping with a detailed
account of the rules by which musical measures are constructed, rst dividing the
measure into its constituent parts and then building measures according to symmetrical patter ns of accent. He then proceeds to further levels of rhythmic organization:
But there is still a higher symmetry than this. That is to say, as parts of times taken together for m small groups, so also can several groups taken together be presented as
parts of a larger group, of a greater or a higher rhythm, of a rhythm of a higher order.
We may go still farther, and to such a greater rhythm we may annex moreover a second and a third, so that these two or three together constitute again a still higher rhythm.
Thus, e.g. in the following passage, two measures taken together constitute a small
rhythm, two of these, taken together, constitute again a rhythm of a higher species, and
again two of the latter taken together, constitute a capital or pr incipal rhythm.88

The construction of the members of the larger rhythms is a symmetry proceeding

more by the great; it is perfectly similar to that involved in the structure of measures,
except that it is simply on a larger scale. As a measure consists of two or three parts,
86. Gottfr ied Weber, Theory of Musical Composition, Treated with a View to a Naturally Consecutive
Arrangement of Topics, ed. John Bishop, trans. James F.Warner (London: Robert Cocks, 1851), 1: 62 [47].
87. Weber, Theory of Musical Composition, 1: 68 71 [51].
88. Weber, Theory of Musical Composition, 1: 85 [68].

319

320

analysis and theory

so two or three measures form the parts of a greater rhythm, and several such rhythms
are again parts of a still higher group.
Hence the measures are distinguished from one another in such higher rhythms, in
respect to their g reater or less inter nal weight or accentuation, in the same way the
parts of the measure are distinguished among themselves; i.e. the heavy or accented
measures assume a prominence above the lighter, as do the heavier parts of the measure
above the lighter.89

What Weber has descr ibed, of course, is an atomistic hierarchy of metrical groups:
beats combine into measures, measures combine into groups of measures, and so on
up through a ser ies of hierarchical levels.
Mapping an atomistic hierarchy onto the domain of musical rhythm emphasizes
the regularity of accentual patter n often associated with the concept of meter. By
turning rhythm into a machine,Weber can descr ibe its organization with three basic
principles: (1) all units of rhythm are either accented or unaccented; (2) these units
are grouped into cyclic patter ns of accent that for m symmetr ical groups; (3) the
groups of one level constitute the units of the next higher level. All of the objects
within this domain must confor m to these pr inciples, but not necessar ily in exactly
the same way. Although accent or nonaccent applies to both beats and g roups of
measures, with beats, the property lasts for but an instant, whereas with groups of
measures, the property lasts for the duration of the group. No limit is specied for
the hierarchy that results, although Weber does not develop levels beyond those
shown in his example.
The approach to musical rhythm proposed by Weber became standard during
the nineteenth century, although subsequent theor ists often added interesting complications. Moritz Hauptmann, for instance, imposed a Hegelian inter pretation on
metrical groupings, which per mitted him to frame the basics of meter and har mony
in the same ter ms.90 By the time of Riemanns System der musikalischen Rhythmik
und Metrik, the hierarchical structure represented by Webers capital or pr incipal
rhythm was presented as the nor mative framework (normatives Grundschema) of
rhythmic organization: beats for med measures, which in tur n formed duple groups,
which combined into two half-phrases of four measures each, which joined to create the eight-measure phrase.91

Reasoning about Music through Chain-of-Being

and Atomistic Hierarchies
Chain-of-being and atomistic hierarchies developed at different points in the history of Western thought to explain two different domains of human exper ience.
These differences are reected in the basic structure of each type of hierarchy, and
they result in two distinct models for reasoning about the world.When these mod89. Weber, Theory of Musical Composition, 1: 87 [69].
90. Moritz Hauptmann, The Nature of Harmony and Metre, trans. W. E. Heathcote, 2nd ed. (London:
Sonnenschein, 1893; repr. with a new forward by Sigmund Levarie, New York: Da Capo Press, 1991),
189 91.
91. Riemann, System der musikalischen Rhythmik und Metrik, 196 98.

competing mode ls of mu s i c
els are mapped onto music they yield different inferences about how musical materials are organized, which lead to different ways to analyze music. If our reasoning is
guided by a chain-of-being hierarchy, we infer that each musical work constitutes
a domain pervaded by a mysterious force. Musical analysis consists in explicating the
way different musical elements embody this mysterious force. By this means, the
proper place of these elements in the musical hierarchy can be deter mined and an
account of musical structure can be produced. If our reasoning is guided by an
atomistic hierarchy, we infer that each musical work operates like an extended
mechanism. Musical analysis consists in identifying the basic components of this
mechanism and explicating the pr inciples according to which it works. The view of
musical structure that results is of a hierarchy of distinct yet inter related substructures that, combined, constitute the musical work.
With these two models for reasoning in mind, let us now return to the three
analyses of the rhythmic organization of Mozarts little theme discussed at the opening of this section and consider in more detail how these two models of hierarchy
condition and constrain musical analysis.

Conceptual Models and Music Theory

According to the Schenkerian paradigm Robert Morgan adopts, tonality is characterized by the property of control: within a tonal composition, every pitch controls
or is controlled by other pitches. Pitches located at the foreground have only the
property of being controlled. Pitches located at middleground levels control pitches
at lower levels and are themselves controlled by pitches at higher levels. Pitches
located at the backg round level are governed only by restrictions placed on the
Ursatz by the chord of nature.
Morgans cor relation of rhythmic organization with this account of tonal organization has two important entailments for the conception of musical rhythm. First,
rhythm becomes a matter of motion between structurally signicant events. The
initiation and completion of each rhythmic motion is marked by the appearance of
the pitches that control a given span; the character of the rhythmic motion will
reect relationships between these controlling pitches. Thus the rst rhythmic
motion of Mozarts theme begins with the Kopfton in m. 1 and ends with the
appearance of 2 in m. 4; the motion has the character of an antecedent because a
complete motion from 3 to 1 has not yet been achieved (see ex. 7.5). Second, the
type of accent assigned a g iven musical moment will tend to be unequivocal.
Emphasis, as an articulation of rhythmic motion, follows from hierarchical structure:
emphasized pitches will be those located at higher levels of the tonal hierarchy.
Because the location of a given pitch within the hierarchy is not equivocal, the
emphasis it receives will not be equivocal. Thus m. 3 cannot have both an upbeat
accent and a downbeat accent at the same time but must be given a rhythmic reading that matches the overall tonal structure.
Although Leonard Meyers treatment of Mozarts theme certainly takes account
of tonal structure, his rhythmic analysis begins, as noted, at the musical surface. The
accentual g roups of the rst (surface) level of the hierarchy become the elements
of the second level of the hierarchy, the accentual g roups of the second level

321

322

analysis and theory

become the elements of the third level, and so on up through four levels of hierarchical organization (see ex. 7.4).92 From Meyers perspective, rhythmic organization,
insofar as it is metr ical, is independent of tonal organization, since the elemental
patter ns of metrical accent can be specied in the absence of pitch. The overall
rhythmic character of a passage results from an interaction between tonal organization and patter ns of metrical accent occur ring on a number of hierarchical levels. The emphasis accorded a given musical event can only be reckoned in context.
On level 2 of Meyers hierarchy, m. 3 is (as a whole) accentually weak; on level 3, it
is one component of a strong accent; and on level 4, it is a subcomponent of a weak
accent. The property of accent is consequently different on each hierarchical level,
since the accent proper to each level pertains to different spans of musical events.
The differences between these two accounts of the rhythmic structure of
Mozarts theme follow in part from the way each model of hierarchy constrains reasoning about music. From the perspective of a chain-of-being hierarchy, the musical surface is the least important of all levels. Given the way musical events are
assigned to hierarchical levels, any polysemous reading of such events is little more
than analytical dither ing. From the perspective of an atomistic hierarchy, there can
be no grand plan of musical organization that does not or iginate with the simplest
and most readily apprehended of musical mater ials. Nonetheless, there is a bit more
to the story. The constraint these models impose on our reasoning goes beyond the
inferential possibilities provided by their relational structure and penetrates to the
very core of what counts as relevant phenomena for theor izing and analysis. Indirect evidence for this constraint is supplied by the way each model of hierarchy is
mapped onto music. Chain-of-being hierarchies have been mapped almost exclusively onto the domain of musical pitch they have generally not been mapped
onto the domain of musical meter. Atomistic hierarchies, on the other hand, have
been most often mapped onto the domain of musical meter mapping them onto
the domain of pitch has not met with wide acceptance.93
Explaining how this constraint works requires that we briey revisit the procedures of cross-domain mapping discussed in chapter 2, with particular attention to
the Invariance Pr inciple. According to this pr inciple, for those portions of the
source and target domains involved in a cross-domain mapping, the mapping preserves the image-schematic structure of the target domain, and it imports as much
of the image-schematic structure from the source domain as is consistent with this
preservation. Mapping a model of hierarchy onto music will preserve aspects of our
musical exper ience and structure these by importing the relationships between
92. In the complete analysis Meyer actually adds two additional levels of hierarchy, although the
sixth and nal level is somewhat speculative; see Meyer, Explaining Music, 39.
93. Eugene Nar mours recent attempts to map an atomistic hierarchy onto pitch have met with only
limited success, and the application has not been widely adopted. See the in-depth treatment of the issue
in Nar mour, Some Major Theoretical Problems Concer ning the Concept of Hierarchy in the Analysis of Tonal Music, Music Perception 1/2 (1983 84): 129 99; Narmour discusses mm. 1 8 of the rst
movement of Mozarts Piano Sonata in A, K. 331, on pp. 160 98. In some recent analyses, Narmour
reads musical structure as only quasi-hierarchical and thus better dealt with in ter ms of networks. See
Narmour, The Analysis and Cognition of Basic Melodic Structures: The Implication-Realization Model (Chicago: University of Chicago Press, 1990).

competing mode ls of mu s i c
energy and the container schema embodied by this model. When we take a
chain-of-being hierarchy as the source domain, we preserve our sense that some
aspects of music are more important than others, and we import a model of systemic organization in which energy is distr ibuted unequally and hierarchical levels
group elements according to their amount of energy. This mapping works well for
the asymmetr ical pitch relations typical of tonal music but less well for the measured
domain of metrical rhythm. When we take an atomistic hierarchy as the source
domain, we preserve our sense that some aspects of music are regular and recurring,
and we import a model of systemic organization in which energy is distr ibuted
equally and hierarchical levels group elements according to common confor mance
with general pr inciples. This mapping works well for the measured domain of metrical rhythm but less well for the asymmetr ical pitch relations typical of tonal music.
Thus the image schemata that g round our conceptual models also facilitate mapping these models onto novel domains, subject to the limitations of the imageschematic structure of both domains.
Additional evidence for the part played by image schemata in conceptual models and cross-domain mapping is suggested by Edward Cones analysis of Mozarts
theme, which combines aspects of the approaches demonstrated by Morgans and
Meyers analyses. For Cone, as for Morgan, musical rhythm involves motion
between points of tonal stability. The rst and fth measures of the theme should be
strong because of the r m statement of tonic; the fourth and eighth measures
should also be strong because of the emphasis provided by arrivals on important
harmonies. However, Cone, like Meyer, is also interested in the contr ibution of the
musical surface to our understanding of musical rhythm as a whole. Cones overall
interpretation of the rhythmic structure of the opening eight measures is thus
shaped by meter, motivic structure, and surface accents. What guides Cones analysis is his concept of musical energy it was this that prompted his compar ison of
a musical phrase to a game of catch. The path of energy that marks the transit of the
ball in the game of catch is the analogue for the path of musical energy constituted
by a given passage in a work of music. Starting with the concept of energy rather
than with a model of hierarchy allows Cone to tap into both models of hierarchy
and draw from each what he needs. By this means, he can forge a single account of
musical rhythm that recognizes both hierarchic paradigms but which is restricted to
neither. Thus, Cones analysis demonstrates a conceptual blend, the gener ic space of
which focuses on the notion of energy common to both models of hierarchy.
According to the approach I have developed here, the discrepancies between
these three analyses of Mozarts theme reect the role of conceptual models in theories of music. At work are two different models of hierarchical organization. One
maps a chain-of-being hierarchy onto music and highlights one set of aspects of the
musical domain. The other maps an atomistic hierarchy onto music and highlights
a different set of musical aspects. Two of the analyses those by Robert Morgan
and Leonard Meyer rely on relatively straightforward mappings, one from each
model of hierarchy.Where the structure of the models of hierarchy diverge, there
the musical analyses based on these models will also diverge. The third analysis, by
Edward Cone, blends the two mappings by focusing on commonalities between
their image-schematic structure. By focusing on these commonalities, Cone puts

323

324

analysis and theory

some distance between his inter pretation and notions about hierarchical organization and produces yet a third analysis of Mozarts theme. The discrepancies between
the three analyses are thus not the result of different opinions about what constitutes
the theme of the rst movement of Mozarts sonata, nor are they a consequence of
analytical inconsistencies. Rather, they follow from the structure of these two conceptual models and the way they are employed in accounts of music.

competing models of musical form

and hierarchy
The foregoing represents an attempt to sort out why musicians reasonable, highly
experienced musicians disagree about such things as musical for m and musical
hierarchy. As I have endeavored to show, such disagreements can be traced to competing models of musical structure. Much as Alice and the Hatter have different
ideas about what time is, so music theor ists have different ideas about the properties
of musical for m and the structure of hierarchical relationships. The analyses I offer
should help explain why this is so. If one is thinking about musical for m in ter ms of
grammar or building blocks, it may not be immediately evident how music achieves
its rhetor ical effects or seems to g row from within. If one is thinking about musical relationships in ter ms of dominance and control, musical meter may appear to
be something of a cipher, for meter provides only weak demonstrations of these
properties.
More interesting, however, is the way these models reect back on the theor ists
who employ them, much as Alices models of time reect the world of a seven-yearold girl in Victorian England, while the Hatters model of time reects the world of
Wonderland. Theorists who focus on a static model of musical for m are almost
invariably occupied with a very basic level of music pedagogy. Dynamic models
only start to make real sense once a level of artistry has been achieved that is, once
the student can speak and understand the language of music with some uency. Not
surprisingly, then, the two models begin to overlap and intertwine in confusing and
interesting ways in the early nineteenth century, for the institutionalization of music
education required a bit of both perspectives. Nation building required the mechanisms of rationalized instruction and of artistic exegesis, the rst to guarantee a
future, the second to celebrate a past. Similarly, theorists who focus on atomistic
hierarchies are convinced that we enter music from its surface and only later can
penetrate to its depths by understanding the general laws that connect them both.
Theorists who make recourse to chain-of-being hierarchies, on the other hand, are
convinced that surfaces are illusions the real truth of music lies hidden in mysterious powers that guide compositional creation like an unseen hand. These summaries are too reductive, of course, but they make in br ief the point toward which
this chapter has argued at length: the models we choose for our character izations
of music tell us at least as much about ourselves as they do about music.

conclusion

cognitive structure, theory,

and analysis

et us return to Prousts A la recherche du temps perdu and to the encounters

between listener and music descr ibed within. The introduction to this volume
was framed around the moment when Charles Swann rst hears the Andante of a
sonata for violin and piano byVinteuil at the bourgeois salon of the Verduins. As that
moment unfurled, Swann moved from vague and unfor med impressions to active
understanding: he began to conceptualize music. Now, a year later, Swann again
encounters the same ctitious composition, this time at a soire at the Marquise de
Saint-Euvertes. Over the course of this year, Swann has become obsessed both with
the Andante especially as represented through the fragment he calls the little
phrase and with Odette, the courtesan whom he had also rst come to know
at the Verduins. At the moment we nd him, Odette has left him, and he nds
himself surrounded by people who, if they knew of his love for her, would only
trivialize it. Then, unexpectedly, the musicians commence the sonata, and Swann
nds himself enraptured once more. Where in his rst encounter the sonata had
amounted to not much more than pleasant sensations, it is now a full-edged discourse, speaking in a beautiful and paradoxical language: Ever since, more than a
year before . . . the love of music had, for a time at least, been bor n in him, Swann
had regarded musical motifs as actual ideas, of another world, of another order, ideas
veiled in shadow, unknown, impenetrable to the human mind, but none the less
perfectly distinct from one another, unequal among themselves in value and
signicance.1 The ideas that enthrall Swann distinct yet shadowy, perfectly clear
yet impenetrable are musical concepts; the discourse to which he attends is
musical discourse. What began as a sketchy, provisional transcr ipt has blossomed
into a theory of musical organization that is robust enough to reach out to the transcendental.
The process that led to this transfor mation (as descr ibed in chap. 1) begins with
categories of musical events, of which the motifs (or motives) upon which Swann
focuses are a good example. Musical motives show a number of similar ities to basic-

1. Marcel Proust, Swanns Way (vol. 1 of Remembrance of Things Past), trans. C. K. Scott Moncr ieff
and Terence Kilmartin (New York:Vintage Books, 1981), 379 80.

325

326

conc lu s i on

level categor ies, offering an effective compromise between efciency and infor mativeness. The variability of motives, as they are shaped to the exigencies of musical
context and compositional process, matches perfectly the variability of Type 1 categories; our impression that there is a motive reects the conceptual models around
which such categor ies are organized. Of course, motives of the sort typical in late
nineteenth-centur y music, of which the little phrase is intended to be a paragon,
represent only one possibility for musical categor ies. Categories might also be organized around a har monic structure or a g round bass, around a two-voice contrapuntal patter n or a persistent rhythmic gure. As the analysis of atonal music has
suggested, musical categor ies might even involve relatively abstract relationships,
which are obscure at rst but which emerge with repeated listenings.2
As I proposed in chapter 1, we can construe musical concepts as equivalent to
musical categor ies. Although such concepts may be afliated with language, language is not necessary for their creation or maintenance. The paradox presented by
Proust that musical ideas are ideas veiled in shadow, impenetrable to the human
mind, but none the less perfectly distinct from one another is simply a reection
of the limits of language. Musical concepts are of another world, another order,
because they extend into a domain that is beyond words.
One of the principal ways musical concepts are related to other concepts
musical and otherwise is through the process of cross-domain mapping. A bit
before the passage quoted above, the little phrase is rst a kind of superhuman agent
transfor med into sound (whose presence is itself likened to the perfume of a
ower), then the words this spir it speaks, then an almost cor poreal lover, and nally
once again a comforting spir it:
Swann felt its presence like that of a protective goddess, a condante of his love, who,
in order to be able to come to him through the crowd and to draw him aside to speak
to him, had disguised herself in this sweeping cloak of sound. And as she passed, light,
soothing, murmurous as the perfume of a ower, telling him what she had to say, every
word of which he closely scanned, regretful to see them y away so fast, he made involuntar ily with his lips the motion of kissing, as it went by him, the har monious,
eeting for m. He felt that he was no longer in exile and alone since she, who addressed herself to him, was whisper ing to him of Odette.3

This dense eld of overlapped mappings cannot, of course, quite summon the
acoustic reality of the little phrase (which exists, after all, only in Prousts imagination), but it does select certain aspects of the musical domain as targets. The little
phrase is distinct enough from the sur rounding mater ials that it can be likened to
a person emerg ing from a crowd; it has a measure of delicacy about it, both in its

2. I am thinking especially here of the work of David Lewin, who has a knack for discovering
intriguing musical relationships, character izing them through impressively abstract for mulations, and
managing to create analyses of compelling immediacy. On the possibility of hearing these for mulations,
Lewin comments, One should not ask of a theory, that every formally true statement it can make about
musical events be a perception-statement. One can only demand that a preponderance of its true statements be potentially meaningful in sufciently developed and extended perceptual contexts. Lewin,
Generalized Musical Intervals and Transformations (New Haven, Conn.:Yale University Press, 1987), 87.
3. Proust, Swanns Way, 378.

cog n i t ive st ruc ture, th e ory, and analys i s

overall aesthetic shape and in its lack of assertiveness; despite this delicacy, the substance of its communicative force is such that it evokes response not simply in words
but in deeply intimate gestures. If, after this descr iption, the little phrase still remains
elusive, we can nonetheless generate a fairly clear sonic picture of what it involves.
Prousts mappings have thus constructed an image of this imaginary music, selecting
musical concepts vague and ill for med, to be sure, but convincing nonetheless
from the general eld of possible musical constructs.
It is through similar, if more prosaic, mappings that we structure our understanding of music, building up the integrated systems of terms and relations through
which we descr ibe our exper ience with and conceptions of music. Such crossdomain mappings have their ultimate g rounding in embodied exper ience (as we
observed in chap. 2). In tur n, our exper iences with music allow us, on occasion, to
map from the musical domain onto other domains. For instance, in the passage just
cited, music (as represented by Vinteuils sonata) provides a model according to
which a succession of eeting impressions the movement of a persons body, the
scent of a ower, the brush of a kiss, a scatter ing of forms can be construed as
constituting an indivisible whole. Just as the sonic mater ials proper to a musical
work appear and disappear and yet always remain part of the work, so the eeting
impressions noted by Swann are, in truth, but some of the disparate threads out of
which his love for Odette is woven.4
The possibility of mapping structure both to and from the musical domain contributes to the process of conceptual blending, in which elements from two cor related mental spaces are blended together in a third mental space. This offers a further inter pretation of Swanns encounter with the little phrase, for what appears
before him is a blend of features drawn from the human and sonic domains a protective goddess who has disguised herself in a sweeping cloak of sound. Just such a
blend seems to have been intended by Proust, who goes on to wr ite that Swann
had no longer, as of old, the impression that Odette and he were unknown to the
little phrase. Had it not often been the witness of their joys? True that, as often, it
had warned him of their frailty.5 As the story continues, the little phrase becomes
fully personied, engaging in dialogue with Swann, offering sage advice, and leading him to transports of the spir it.What Proust is relying on, of course, is the sort
of anthropomorphic blend he also employed in his descr iption of Swanns lilac
trees, and which Tovey employed so strategically in his analysis of Beethovens Sixth
Symphony (both discussed in this volume in the latter portion of chap. 2).
The musical concepts produced by categor ization and the structural relationships produced by cross-domain mapping come together in conceptual models,
which provide the basis for reasoning about music. Swann relies on two models. Of
these, the rst yields a view of music as a type of discourse structure: musical
motives are regarded as ideas that are shaped into a communicative medium by the
way they are organized in time and with respect to other musical ideas. The sec4. This is particularly apparent because Prousts allusion to the perfume of a ower is anticipated
a few pages earlier by Swanns detailed recollection of the chrysanthemum Odette gave him early in their
romance.
5. Proust, Swanns Way, 378.

327

328

conc lu s i on

ond model yields a view of the nonlinguistic as super natural. Words, as proper to
human discourse, are necessar ily mortal; those things that escape words are consequently immortal; things that are immortal exist only outside the natural world.
These models are then combined into a theor y of music appropriate to Swanns
moment of clairvoyance, following the perspective developed in chapter 3. According to this theor y, musical discourse is a suprahuman mode of communication,
something evident as the two musicians conclude the perfor mance that had summoned the little phrase and Swanns rapture.
Swann dared not move, and would have liked to compel all the other people in the
room to remain still also, as if the slightest movement might imper il the magic presence, supernatural, delicious, frail, that was so soon to vanish. But no one, as it happened, dreamed of speaking. The ineffable utterance of one solitary man . . . breathed
out above the rites of those two hierophants, sufced to ar rest the attention of three
hundred minds, and made of that platfor m on which a soul was thus called into being
one of the noblest altars on which a super natural ceremony could be perfor med.6

This is a theor y of music rather different from the others I have discussed, not
the least because where specics are wanted we nd only sweeping generalities.
Nonetheless, the theory of music that Swann has developed in his efforts to understand the effect of Vinteuils sonata is in truth but a manifestation of a portion of the
broader theory of aesthetics Proust sought to develop in A la recherche du temps perdu.
Within this theor y, music was to occupy the same place accorded it by Schopenhauer: as a pure expression of the will.7 Necessary to this perspective is the conceit
that absolute music somehow exists outside the domain of human cognition.
The perspective relative to which I have framed this book, of course, is that
music,absolute or otherwise, is wholly within the domain of human cognition. In
making this argument, I have focused not on the most basic aspects of musical cognition the level where raw data is taken in by the auditory processing system and
transfor med into electro-chemical infor mation within the brain but on a relatively high level of cognition accessible to our conscious thoughts. My emphasis
has been on how general cognitive capacities (like those for categor ization, crossdomain mapping, and the construction and maintenance of conceptual models and
theories) are specied for music. Indeed, understanding something as music as
opposed to the hammer ing of a woodpecker, or an ag reeable if not particularly
interesting succession of sounds, or just noise is intimately bound up with directing these general cognitive capacities to the specic task of making sense of patterned sound.
This approach to music cognition makes it possible to give a fairly detailed account
of how M. Swann (and M. Proust, for that matter) conceptualize music. It also provides a perspective on broader problems of musical understanding, four of which I
treated at length in the second part of this book.

6. Proust, Swanns Way, 383.

7. On this point, see Jean-Jacques Nattiez, Proust as Musician, trans. Derrick Puffett (Cambr idge:
Cambridge University Press, 1989), chap. 4.

cog n i t ive st ruc ture, th e ory, and analys i s

musical syntax

The idea that musical works could be woven from relatively limited bits of musical mater ials motives was one that gained cur rency
around the time that composers became increasingly interested in the resources of
purely instrumental music that is, around the middle of the eighteenth centur y.
As such, the idea that motives are important for musical discourse is hardly a new
one, even if it did not assume prominence in music theory treatises until the
midnineteenth century.Viewing a motive as a cognitive category has the practical
benet of explaining why we construe diverse motive forms as the same, despite
their manifold differences. The perspective becomes really interesting, however,
when we start to consider relationships among the members of a category of
motives, for these offer a way to explain the way the musical syntax specic to a
given work emerges over time. As the composer deploys now more, now less typical members of the categor y, we not only develop a more complete sense of the
structure of the category but also come to see how typicality cor relates with overall compositional design. This syntactic strand could be woven together with those
spun from har monic and for mal syntax to create a r ich, multileveled discourse (as
we saw in the analyses of works from the eighteenth century presented in chap. 4).
The same approach also provides a way to explain how the musical topics so
important in eighteenth-centur y instrumental music emerged. As we saw in the
rst movements of two Beethoven str ing quartets, categor ies of motive forms were
distinguished by how they were used in the course of a movement. In the case of
a motivic categor y used only for beg innings, for instance, this property then becomes part of the conceptual model for the categor y of motive forms: in addition
to its other attr ibutes, the category now acquired the attr ibute used only for beginnings. This change has the consequence of giving the categor y a semantic value:
within the context of this movement, the motive means this is a beginning. As a
result, a musical topic, even if of a very local and restricted sort, has been bor n. For
eighteenth-centur y music, this process is of interest chiey for answering the r iddle of how musical topics were created in the rst place and for understanding the
interplay between locally generated signs and established topics. For music with a
less-clearly dened tradition, such as that of the twentieth century, the process is of
rather more interest, for it offers a way to explain how each individual piece generates a semiotic system. Indeed, it is in part through establishing and manipulating
categor ies of musical events that composers of the twentieth centur y create musical meaning. We can also see the part this process plays in helping us understand an
unfamiliar repertoire: by attending to the specics of musical syntax, we can begin
to understand the process of meaning construction of which this syntax is a part.
In a manner of speaking, music teaches us how we should listen to it.
The dynamic approach to musical syntax made possible by consider ing relationships among the members of musical categor ies (and among categor ies themselves) is one that is well suited to developing an account of large-scale rhythm
that is, rhythm at the phrase level or above. The challenge of theorizing large-scale
rhythm has been to develop a model that is more than simply an adaptation of metric theory or a reinterpretation of large-scale har monic theory. One way to develop
such a model is to construe relationships among categor y members in dynamic
terms. For instance, if we return to the compositional strategies discussed in chapter

329

330

conc lu s i on

4, we could use cross-domain mapping to cor relate the typicality of category members with stability, and atypicality with instability. Applied to the opening measures
of the Allegro from Mozarts Dissonance Quartet, what emerges is a large-scale
pattern that begins with stability (in mm. 23 26), moves to instability (beginning in
m. 27), returns to stability (in mm. 31 34), moves somewhat more quickly to instability (beginning in m. 35), returns (after a br ief hiatus) to stability (in m. 44), and
then nally dissolves into instability (after m. 48).8 The patter n could be further
applied to each subsequent episode involving the motive, producing a dynamic
account of the movement as a whole.

musical ontology That an exploration of musical ontology might yield

signicant insights into the structure and character of cultural knowledge might at
rst seem sur prising, for as common as music is, it is still not the central business of
human culture. However, because music is resolutely ephemeral, knowledge of it is,
in most cases, preserved only in human cognitive structure. We can know what
Greek potter y of the fth centur y b.c. was like, because some of it has been preserved into the present era. The same cannot be said of Greek music. Theoretical
and other wr itings give us hints about the character of Greek music, but direct
exper ience of it is impossible. Music, where it is not preserved through some artifactual means, exists solely in the conceptual models and theor ies of the members
of a musical community. In consequence, giving an account of the basis for determinations of musical identity what it means for something to be considered an
instance of music means giving an account of the cultural knowledge upon
which such deter minations are based.
Although it is easy to hold up the Western preoccupation with the score as the
acme of musical objectication, the notion of a musical object is not a consequence
of artifacts like scores but is instead an offshoot of the conceptual models we use to
structure our understanding of what we hear. Because these models are, for the most
part, completely transparent, we take them for the way things are: conceptual models are part and parcel of what counts as reality (as noted in chap. 3). Our impression
that music is an objective category, and that there are musical objects, is simply a
consequence of the cognitive efciency conceptual models represent.
In that the conceptual model relative to which we make deter minations of musical ontology captures the things our culture deems most important about patter ned
sound (or the place patter ned sound occupies in our culture), our theor ies of music
are consequent to these models: what counts as music determines what counts as
music theory. One prominent example of this was provided by the Greek music
theories of antiquity discussed in the introduction. However, more subtle examples
are available as well: for Rameau, music meant French music, and it was relative
to that national tradition of musical expression that his theory was framed. Similarly,
8. The changing patter n of stability and instability I have outlined very much resembles the patter n
that occurs at the level of the beat, specically in the context of what Harald Krebs calls metr ic dissonance. The principal difference is that the patter n as I have described it is expanded to encompass entire
musical phrases. This theory of metric dissonance is most fully worked out in Krebs, Fantasy Pieces: Metrical Dissonance in the Music of Robert Schumann (New York: Oxford University Press, 1999).

cog n i t ive st ruc ture, th e ory, and analys i s

for Schenker music meant music from the Ger man-speaking world, and it was to
this repertoire that his theor y was directed. This conception of musical ontology
can be extended to other repertoires, like jazz (as noted in chap. 5). There we
observed that certain jazz musicians of the 1950s and 1960s regarded the har monic
possibilities of a tune (that is, the potential for the har monic template of the tune to
be reconstituted through the modication or substitution of harmonies) and its
overall affect as more important than the bare pitch structure of its melody or the
basic syntax established by its har monies. In consequence, the theory of music that
issued from this practice a theory woven throughout the interviews and quotations that for m the fabric of Paul Berliners Thinking in Jazz was centrally concerned with ways to create and layer the multiple musical discourses made possible
by a particular har monic patter n.9 For this music, it was far less important to
develop an explanation of the imperatives of melodic or har monic syntax, since
these were assumed to be almost totally dependent on local context.

music and othe r domains

According to at least some denitions,

language and music are quite separate domains. The distinction is eminently sensible: when we want to know if supper is ready, we expect a response in words, not
a musical phrase ending with a deceptive cadence. Nonetheless, inasmuch as both
language and music are products of human cognitive processes, it also seems sensible that they will rely on some of the same cognitive capacities. For example, evidence from conceptual blending suggests that they both employ similar discourse
strategies. Admittedly, these strategies, as represented by the gener ic space of a conceptual integration network, are rather abstract, so elemental as to be almost simplistic. And yet there is strong evidence that such simple stor ies are basic to human
cognition and communication indeed, as we observed in chapter 2, Antonio Damasio has argued that they are involved with consciousness itself .10 In situations
where linguistic discourse and musical discourse are simultaneous and structured by
the same basic strategy, concepts from the mental spaces set up by each can combine
in a third mental space: a conceptual blend. Situations of this type were demonstrated by the analyses of songs in chapter 6. And, as we saw with Brahmss In der
Fremde, if a unifor m topography between the constituent spaces in a conceptual
integration network is lacking, conceptual blending will not occur. The combination of words and music in such cases may be pleasant enough, but it will not realize the full potential of these conjoined media.
A basic conceptual blend involves four coordinated mental spaces (as noted in
chaps. 2 and 6). However, as Toveys analysis of Beethovens Sixth Symphony shows,
conceptual blending may involve any number of mental spaces. It is fairly easy to
imagine songs that might involve multiple spaces for instance, if the song makes
9. Paul F. Berliner, Thinking in Jazz: The Innite Art of Improvisation, Chicago Studies in Ethnomusicology (Chicago: University of Chicago Press, 1994).
10. For an overall account of the importance of abstract discourse structures to human cognition, see
Mark Turner, The Literary Mind (New York: Oxford University Press, 1996). The approach Antonio
Damasio takes in his The Feeling of What Happens: Body and Emotion in the Making of Consciousness (New
York: Harcourt Brace, 1999) is notable for the empir ical evidence and level of neurophysiological detail
that he br ings to bear.

331

332

conc lu s i on

reference to a well-known myth, or if, as in the case of Schuberts Trockne Blumen, there are strong parallels between the situation descr ibed in the words and
details of the composers life.11 Even more intr iguing are the possibilities offered by
opera, which can easily engage and often revels in simultaneous musical, linguistic, theatr ical, gestural, and scenic discourse.Various kinds of musical multimedia music and lm, music and dance, music and visual images offer further possibilities for explor ing conceptual blends involving multiple spaces. In each case,
where a unifor m topography between the spaces set up by these media obtain, conceptual blending can occur.What is potentially intr iguing about the blends that can
be set up by the coordination of these diverse media are the views of discourse
structure that will result, so that the syntax of music will shed light on the syntax
of other media, and they on music.While differences between the media are essential to the exper ience of multimedia, as Nicholas Cook has suggested, the deep
commonalities between the discourse structure of various media offer new ways to
think about musical structure.

competing models of musical structure

The theory of music

developed by Charles Swann, while useful for making a reconciliation with what he
imagined to be the end of his love affair, offers hardly any explanation for why the
little phrase works the magic that it does or how Vinteuils sonata can hold three
hundred Parisian socialites enthralled. Where we long to discover more about its
syntax, or its place within a tradition of similar works, or the means by which it
sprang forth from the mind of its creator, Swanns theory gives us only obscurantism
and effusion. That this should be so is a direct consequence of the conceptual models his theory employs. Given the assumption that music is a type of discourse structure (rather than, say, some sort of organism or mechanism), Swanns theory places
an emphasis on the message rather than the medium and on temporal constraint as
an essential element of musical exper ience. Music has things to tell us, but the disappearance of these messages is practically coincident with their emergence. Given
the assumption that nonlinguistic discourse is super natural (rather than something
humans make use of regularly), we cannot expect to make much sense of the things
music tells us; we can only be grateful that we have been able to hear them at all. Of
course, Swanns and Prousts theory of music is a product of a particular cultural and histor ical moment, one that sought to explain how humans could participate in the divine even while they were forever shut off from it. Other cultural and
historical moments including our own will seek other explanations, employing
the categor ies, cross-domain mappings, conceptual models, and theor ies that resonate with them most deeply and constructing a vision of music that best ts with
their vision of the world as a whole.
Theories of music, then, are not so much about truth as about a particular perspective on what aspects of patterned sound are rewarding or important to attend
to; they are equally concer ned with the integration of this perspective into a larger

11. Although this biographical cor relation does not gure in my consideration in chap. 6 of Schuberts song, I touch on these parallels in Lawrence Zbikowski, The Blossoms of Trockne Blumen:
Music and Text in the Early Nineteenth Century, Music Analysis 18 (1999): 340.

cog n i t ive st ruc ture, th e ory, and analys i s

view of human society and culture. Similarly, music does not offer us an unmediated
vision of the truth despite what M. Swann believes but a truth assembled from
the possibilities of patterned sound and human cognitive capacities.
Most of my discussion has focused on cognitive structure and theory, and I have had
relatively less to say about analysis especially analysis as the object of a separate
domain of inquiry. The reason for this apparent imbalance, as mentioned in the
introduction to this volume, is that while analysis is essential for making claims
about musical structure explicit, it is not independent of theory: every analysis relies
on some theor y of music.12 For that matter, I am somewhat doubtful about making a meaningful distinction between descr iption (as a more or less passive reporting of the relevant features of a work) and analysis (as a thorough contextualization
of aspects of musical structure), since decisions about what counts as relevant features and musical structure rely equally on some theoretical understanding about
music. It seems more promising to distinguish between ad hoc theor ies and extended theor ies, as I did at the end of chapter 3. This distinction is based on the
kind of problems a theor y engages, its place within a tradition of inquiry, and the
degree to which it is rendered systematic. Generally speaking, descriptions or analyses based on extended theor ies are more interesting than those based on ad hoc theories, for extended theor ies draw much more deeply on our cognitive capacities
and cultural knowledge.
Although analysis has been a bit of a dark horse, it has hardly been out of the
running. Indeed, it has been present throughout the course traced by the preceding pages: in the analyses by Schoenberg, Tovey, Rameau, Schenker, Mattheson,
Meyer, Cone, and Morgan I have discussed, and in the analyses I myself have presented. In the case of earlier wr iters, I have been most interested in what their analyses have to say about their theor ies of music. My own analyses, of course, are based
on the theor y that our understanding of music is shaped by our cognitive capacities and that these are visible at every turn in our encounters with music. Through
our theor ies and analyses, we thus shed light not just on music but on the fascinating mysteries of human thought.
The abiding mystery I have sought to explain is the one embodied in M.
Swanns rst encounter with Vinteuils sonata: how it is that we can make some sort
of sense of an unfamiliar music almost immediately and yet have this same work
provide continued reward in our further exper iences with it. Music, as revealed by
Prousts vignette, is thus simultaneously immediate and complex, transparent and
opaque. The solution I have offered is that our understanding of music relies on
cognitive capacities that we use throughout our daily lives. The cognitive processes
behind categor ization, cross-domain mapping, conceptual models, and theor ies
would all be operational had we never heard a note of music. Doubtless, our lives
would be immeasurably impoverished had this unlikely tur n of events transpired,
12. Among the rst to draw this distinction was Edward T. Cone in Analysis Today, Musical Quarterly 46 (1960): 172 88. This essay can also be found in Paul Henry Lang (ed.), Problems of Music Analysis: The Princeton Seminar in Advanced Musical Studies (New York:W.W. Norton, 1960), 34 50; and Cone,
Music, a View from Delft: Selected Essays (Chicago: University of Chicago Press, 1989), 39 54.

333

334

conc lu s i on

but not only because we would have never been exposed to what Proust characterized as another world, another order, ideas veiled in shadow, impenetrable to the
human mind, but none the less perfectly distinct from one another. We would also
have never had these cognitive capacities stretched and transfor med by the challenge of listening to the play of patterned sound. Nor would we have taken what
we learned in our encounters with that domain back to our understanding of the
world as a whole. And we would never have had the opportunity to extend our
conceptualization of music into theor ies of music theories that, in the webs of
relations they build both within and without music, offer their own topic for intellectual investigation. Conceptualizing music, seen in this way, is thus both a celebration and an exploration of the means by which our minds go out to meet the
world.

bibliog raphy

Books and Articles

Abelson, Robert P., and John B. Black. Introduction. In Knowledge Structures, edited by
James A. Galambos, Robert P. Abelson, and John B. Black. Hillsdale, N.J.: Erlbaum Associates, 1986.
Abrams, Meyer Howard. The Mirror and the Lamp: Romantic Theory and the Critical Tradition.
Oxford: Oxford University Press, 1953.
Agawu,Victor Ko. Playing with Signs: A Semiotic Interpretation of Classical Music. Princeton,
N.J.: Princeton University Press, 1991.
. Theory and Practice in the Analysis of the Nineteenth-Century Lied. Music Analysis 11 (1992): 3 36.
Aldwell, Edward, and Carl Schachter. Harmony and Voice Leading. 2nd ed. San Diego: Harcourt Brace Jovanovich, 1989.
Ames, David W., and Anthony V. King. Glossary of Hausa Music and Its Social Contexts.
Evanston, Ill.: Northwestern University Press, 1971.
Aristotle. On the Sublime. Transated by W. Hamilton Fyfe. Loeb Classical Library, 199. Cambridge, Mass.: Harvard University Press, 1995.
Aristoxenus. Elementa harmonica. In Harmonic and Acoustic Theory.Vol. 2 of Greek Musical Writings. Edited and translated by Andrew Barker, 126 84. Cambridge Readings in the Literature of Music. Cambridge: Cambridge University Press, 1989.
Armstrong, Sharon Lee, Lila R. Gleitman, and Henry Gleitman. What Some Concepts
Might Not Be. Cognition 13 (1983): 263 308.
Arom, Simha. African Polyphony and Polyrhythm: Musical Structure and Methodology. Translated
by Martin Thom, Barbara Tuckett, and Raymond Bond. Cambr idge: Cambridge University Press, 1991.
. New Perspectives for the Descr iption of Orally Transmitted Music. World of Music
23 (1981): 40 58.
Attali, Jacques. Noise: The Political Economy of Music. Translated by Brian Massumi, with a foreword by Fredric Jameson, with an afterword by Susan McClary. Theory and History of
Literature, 16. Minneapolis: University of Minnesota Press, 1985.
Bailey, Robert, ed. Prelude and Transguration from Tristan and Isolde. New York:W.W. Norton,
1985.
Bakan, Michael B. Music of Death and New Creation: Experiences in the World of Balinese Gamelan Beleganjur. Chicago Studies in Ethnomusicology. Chicago: University of Chicago
Press, 1999.

335

336

b i bl i og raphy

Baker, Nancy Kovaleff, and Thomas Chr istensen, eds. and trans. Aesthetics and the Art of Musical Composition in the German Enlightenment: Selected Writings of Johann Georg Sulzer and
Heinrich Christoph Koch. Cambr idge Studies in Music Theory and Analysis, 7. Cambr idge:
Cambridge University Press, 1995.
Bakhtin, Mikhail Mikhailovich. Problems of Dostoevskys Poetics. Edited and translated by Caryl
Emerson, with an introduction by Wayne C. Booth. Theory and History of Literature,
8. Minneapolis: University of Minnesota Press, 1984.
Bamberger, Jeanne. Cognitive Issues in the Development of Musically Gifted Children. In
Conceptions of Giftedness, edited by Robert J. Sternberg and Janet E. Davidson, 388 413.
Cambridge: Cambridge University Press, 1986.
. The Mind behind the Musical Ear: How Children Develop Musical Intelligence. Cambridge, Mass.: Harvard University Press, 1991.
Barbera, C. Andr. The Consonant Eleventh and the Expansion of the Musical Tetractys:

A Study of Ancient Pythagoreanism. Journal of Music Theory 28 (1984): 191 223.

Barchas, Janine. The Engraved Score in Clarissa: An Intersection of Music, Narrative, and

Graphic Design. Eighteenth-Century Life 20 (1996): 1 20.

Barker, Andrew, ed. Harmonic and Acoustic Theory.Vol. 2 of Greek Musical Writings. Cambridge
Readings in the Literature of Music. Cambr idge: Cambridge University Press, 1989.
Barsalou, Lawrence W. Cognitive Psychology: An Overview for Cognitive Scientists. Hillsdale, N.J.:
Erlbaum Associates, 1992.
. Deriving Categor ies to Achieve Goals. The Psychology of Learning and Motivation,
Advances in Research and Theory 27 (1991): 1 64.
. Flexibility, Structure, and Linguistic Vagary in Concepts: Manifestations of a Compositional System of Perceptual Symbols. In Theories of Memory, edited by A. F. Collins,
S. E. Gathercole, M. A. Conway, and P. E. Morris, 29 101. Hillsdale, N.J.: Erlbaum Associates, 1993.
. Frames, Concepts, and Conceptual Fields. In Frames, Fields, and Contrasts: New
Essays in Semantic and Lexical Organization, edited by Adrienne Lehrer and Eva Feder
Kittay, 21 74. Hillsdale, N.J.: Erlbaum Associates, 1992.
. The Instability of Graded Structure: Implications for the Nature of Concepts. In
Concepts and Conceptual Development: Ecological and Intellectual Factors in Categorization,
edited by Ulric Neisser, 101 40. Cambridge: Cambridge University Press, 1987.
. Intraconcept Similar ity and Its Implications for Interconcept Similar ity. In Similarity and Analogical Reasoning, edited by S.Vosniadou and A. Ortony, 76 121. Cambridge: Cambridge University Press, 1989.
. Perceptual Symbol Systems. Behavioral and Brain Sciences 22 (1999): 577 609.
Barsalou, Lawrence W., Karen Olseth Solomon, and Ling-Ling Wu. Perceptual Simulation
in Conceptual Tasks. In Cultural, Psychological, and Typological Issues in Cognitive Linguistics: Selected Papers of the Bi-Annual ICLA Meeting in Albuquerque, July 1995, edited by
Masako K. Hiraga, Chris Sinha, and Sher man Wilcox, 209 28. Amsterdam: John Benjamins, 1999.
Barsalou, Lawrence W.,Wenchi Yeh, Barbara J. Luka, Karen L. Olseth, Kelly S. Mix, and LingLing Wu. Concepts and Meaning. In Chicago Linguistics Society 29: Papers from the
Parasession on the Correspondence of Conceptual, Semantic, and Grammatical Representations,
edited by Kathar ine Beals, Gina Cooke, David Kathman, Sotaro Kita, Karl-Er ik McCullough, and David Testen. Chicago: University of Chicago, Chicago Linguistics Society,
1993: 23 61.
Bayard, Samuel P. Prolegomena to a Study of the Principal Melodic Families of BritishAmerican Folk Song. Journal of American Folklore 63 (1950): 1 44.
Belzner,William. Music, Modernization, and Westernization among the Macuma Shuar. In

b i bl i og raphy
Cultural Transformations and Ethnicity in Modern Ecuador, edited by Norman E. Whitten,
Jr., 731 48. Urbana: University of Illinois Press, 1981.
Bent, Ian. The Compositional Process in Music Theor y, 1713 1850. Music Analysis 3
(1984): 29 55.
, ed. Fugue, Form and Style.Vol. 1 of Music Analysis in the Nineteenth Century. Cambridge Readings in the Literature of Music. Cambridge: Cambridge University Press,
1994.
Berger, Karol. Musica Ficta: Theories of Accidental Inection in Vocal Polyphony from Marchetto da
Padova to Gioseffo Zarlino. Cambridge: Cambridge University Press, 1987.
Berlin, Brent. Ethnobotanical Classication. In Cognition and Categorization, edited by
Eleanor Rosch and Barbara B. Lloyd, 9-26. Hillsdale, N.J.: Erlbaum Associates, 1978.
. Ethnobiological Classication: Principles of Categorization of Plants and Animals in Traditional Societies. Princeton, N.J.: Princeton University Press, 1992.
. How a Folkbotanical System Can Be Both Natural and Comprehensive: One
Maya Indians View of the Plant World. In Folkbiology, edited by Douglas L. Medin and
Scott Atran, 71 89. Cambridge, Mass.: MIT Press, 1999.
. Speculations on the Growth of Ethnobotanical Nomenclature. Language in Society 1 (1972): 51 86.
Berlin, Brent, Dennis E. Breedlove, and Peter H. Raven. Principles of Tzeltal Plant Classication:
An Introduction to the Botanical Ethnography of a Mayan-Speaking People of Chiapas. New
York: Academic Press, 1974.
Berliner, Paul F. Thinking in Jazz: The Innite Art of Improvisation. Chicago Studies in Ethnomusicology. Chicago: University of Chicago Press, 1994.
Bernard, Jonathan W. Messiaens Synaesthesia: The Cor respondence between Color and
Sound Structure in His Music. Music Perception 4 (1986): 41 68.
Bernstein, Susan. Virtuosity of the Nineteenth Century: Performing Music and Language in Heine,
Liszt, and Baudelaire. Stanford, Calif.: Stanford University Press, 1998.
Biederman, Irving. Recognition-by-Components: A Theory of Human Image Understanding. Psychological Review 94 (1987): 115 47.
Blacking, John. Music, Culture, and Experience. In Music, Culture, and Experience: Selected Papers of John Blacking, edited by Reginald Byron, with a foreword by Bruno Nettl, 223 42.
Chicago Studies in Ethnomusicology. Chicago: University of Chicago Press, 1995.
Boethius, Ancius Manlius Severinus. Fundamentals of Music. Edited by Claude V. Palisca. Translated by Calvin Bower. Music Theory Translation Ser ies. New Haven, Conn.: Yale University Press, 1989.
Bohlman, Philip V.Ontologies of Music. In Rethinking Music, edited by Nicholas Cook and
Mark Everist, 17 34. Oxford: Oxford University Press, 1999.
Bonds, Mark Evan. Wordless Rhetoric: Musical Form and the Metaphor of Oration. Studies in the
History of Music, 4. Cambridge, Mass.: Harvard University Press, 1991.
Brailoiu, Constantin. Le Folklore musicale. In Musica aeterna: La Vie et la production musicales
de tous les temps et de tous les peuples, en tenant compte particulirement de la Suisse, de la Belgique, de la France et de la musique de nos jours, edited by Gottfr ied Schmid, vol. 2,
277 332. Zurich: M. S. Metz, 1949.
Brinkmann, Reinhold. Schumann und Eichendorff: Studien zum Liederkreis Opus 39. MusikKonzepte, 95. Munich: Edition Text + Kr itik, 1997.
Brown, Roger. How Shall a Thing Be Called? Psychological Review 65 (1958): 14 21.
. Social Psychology. New York: Free Press, 1965.
Burmeister, Joachim. Musical Poetics [1606]. Translated by Benito Rivera. Music Theor y
Translation Ser ies. New Haven, Conn.:Yale University Press, 1993.
Burnham, Scott. The Second Nature of Sonata Form. In Music Theory and Natural Order

337

338

b i bl i og raphy

from the Renaissance to the Early Twentieth Century, edited by Suzannah Clark and Alexander Rehding, 111 41. Cambridge: Cambridge University Press, 2001.
Carroll, Lewis. The Annotated Alice: Alices Adventures in Wonderland and Through the Looking
Glass. Introduction and notes by Martin Gardner, illustrated by John Tenniel. New York:
W.W. Norton, 2000.
Chambers, Jack. Milestones I: The Music and Times of Miles Davis to 1960. Toronto: University
of Toronto Press, 1983.
Christensen, Thomas. Ftis and Emerging Tonal Consciousness. In Music Theory in the Age
of Romanticism, edited by Ian Bent, 37 56. Cambr idge: Cambridge University Press,
1996.
. Rameau and Musical Thought in the Enlightenment. Cambridge Studies in Music Theory and Analysis, 4. Cambr idge: Cambridge University Press, 1993.
Chua, Daniel K. L. The Galitzin Quartets of Beethoven: Opp. 127, 132, 130. Princeton, N.J.:
Princeton University Press, 1995.
Coleman, Linda, and Paul Kay. Prototype Semantics: The English Word Lie. Language 57
(1981): 26 44.
Cone, Edward T. Analysis Today. The Musical Quarterly 46 (1960): 172 88.
. Music, a View from Delft: Selected Essays. Edited by Robert P. Morgan. Chicago: University of Chicago Press, 1989.
. Musical Form and Musical Performance. New York: W.W. Norton, 1968.
. On Der ivation: Syntax and Rhetor ic. Music Analysis 6 (1987): 237 55.
Cook, Nicholas. Analysing Musical Multimedia. Oxford: Clarendon Press, 1998.
. At the Borders of Musical Identity: Schenker, Corelli and the Graces. Music Analysis 18 (1999): 179 233.
Cooper, Grosvenor W., and Leonard B. Meyer. The Rhythmic Structure of Music. Chicago: University of Chicago Press, 1960.
Coulson, Seana. Semantic Leaps: Frame-Shifting and Conceptual Blending in Meaning Construction. Cambridge: Cambridge University Press, 2001.
Cowdery, James R. The Melodic Tradition of Ireland. Kent, Oh.: Kent State University Press, 1990.
Cox, Arnie Walter. The Metaphor ic Logic of Musical Motion and Space. Ph.D. diss., University of Oregon, 1999.
Crawford, Richard. George Gershwins I Got Rhythm (1930). In The American Musical
Landscape, 213 36. Berkeley: University of Califor nia Press, 1993.
Dahlhaus, Carl. Nineteenth-Century Music. Translated by J. Bradford Robinson. Berkeley: University of Califor nia Press, 1989.
. Richard Wagners Music Dramas. Translated by Mary Whitall. Cambridge: Cambr idge
University Press, 1979.
Damasio, Antonio R. Descartes Error: Emotion, Reason, and the Human Brain. New York: Avon,
1994.
. The Feeling of What Happens: Body and Emotion in the Making of Consciousness. New
York: Harcourt Brace, 1999.
DAndrade, Roy G. The Development of Cognitive Anthropology. Cambridge: Cambridge University Press, 1995.
DAndrade, Roy G., and Claudia Strauss, eds. Human Motives and Cultural Models. Publications
of the Society for Psychological Anthropology. Cambridge: Cambridge University
Press, 1992.
Darwin, Charles. The Voyage of the Beagle.With an introduction by Leonard Engel. Garden
City, N.Y.: Doubleday, 1962.
Davis, Ruth. The Effects of Notation on Performance Practice in Tunisian Art Music. The
World of Music 34 (1992): 85 114.

b i bl i og raphy
DeBellis, Mark. Music and Conceptualization. Cambridge: Cambridge University Press, 1995.
Deffaa, Chip. Traditionalists and Revivalists in Jazz. Studies in Jazz, 16. Metuchen: Scarecrow
Press; Institute of Jazz Studies, Rutgers, The State University of New Jersey, 1993.
Deutsch, Otto Er ich. Mit Wrfeln komponieren. Zeitschrift fr Musikwissenschaft 12 (1930):
595.
Duchez, Marie-Elisabeth. La Reprsentation spatio-verticale du caractre musical graveaigu et llaboration de la notion de hauteur de son dans la conscience musicale occidentale. Acta musicologica 51 (1979): 54 73.
Dupr, John. Are Whales Fish? In Folkbiology, edited by Douglas L. Medin and Scott Atran,
461 76. Cambr idge, Mass.: MIT Press, 1999.
Eco, Umberto. A Theory of Semiotics. Advances in Semiotics. Bloomington: Indiana University Press, 1976.
Edelman, Gerald M. Bright Air, Brilliant Fire: On the Matter of Mind. New York: Basic Books,
1992.
. The Remembered Present: A Biological Theory of Consciousness. New York: Basic Books,
1989.
Eichendorff, Joseph Freiherr von. Viel Lrmen um nichts. Neue Gesamtausgabe der Werke und
Schriften, 2. Stuttgart: J. G. Cottasche Buchhandlung Nachf., 1957.
Emge, Charles. Meet Jazzman Jack Webb: Swing Era Helped Make Him Music-Conscious. Downbeat 22 (10 August 1955): 8.
Fahnestock, Jeanne. Rhetorical Figures in Science. New York: Oxford University Press, 1999.
Fauconnier, Gilles. Mappings in Thought and Language. Cambr idge: Cambridge University
Press, 1997.
. Mental Spaces: Aspects of Meaning Construction in Natural Language. 2nd ed. With a
foreword by George Lakoff and Eve Sweetser. Cambr idge: Cambridge University Press,
1994.
Fauconnier, Gilles, and Eve Sweetser, eds. Spaces, Worlds, and Grammar. Chicago: University
of Chicago Press, 1996.
Fauconnier, Gilles, and Mark Turner. Blending as a Central Process of Grammar. In Conceptual Structure, Discourse and Language, edited by Adele E. Goldberg, 113 30. Stanford,
Calif.: Center for the Study of Language and Infor mation, 1996. [Based on papers presented at the First Conceptual Structure, Discourse and Language Conference, San
Diego, 1995.]
. Conceptual Integration Networks. Cognitive Science 22 (1998): 133 87.
. Conceptual Projection and Middle Spaces. UCSD Department of Cognitive Science Technical Report 9401. San Diego: University of California, San Diego, 1994.
. Principles of Conceptual Integration. In Discourse and Cognition: Bridging the Gap:
[Second] Conceptual Structure, Discourse, and Language Conference, Buffalo, 1996, edited by
Jean-Pier re Koenig, 269 83. Stanford, Calif.: Center for the Study of Language and
Information, 1998.
. The Way We Think: Conceptual Blending and the Minds Hidden Complexities. New
York: Basic Books, 2002.
Feld, Steven. Flow Like a Waterfall: The Metaphors of Kaluli Musical Theory. Yearbook for
Traditional Music 13 (1981): 22 47.
Ferris, David. Schumanns Eichendorff Liederkreis and the Genre of the Romantic Cycle. New
York: Oxford University Press, 2000.
Finnegan, Ruth. The Relation between Composition and Perfor mance: Three Alternative
Modes. In The Oral and the Literate in Music, edited byYoshihiko Tokumaru and Osamu
Yamaguti, 73 87. Tokyo: Academia Music, 1986.
Finson, Jon W. The Intentional Tourist: Romantic Irony in the Eichendorff Liederkreis of

339

340

b i bl i og raphy

Robert Schumann. In Schumann and His World, edited by R. Larry Todd, 156 70.
Princeton, N.J.: Princeton University Press, 1994.
Fodor, Jerry A. Concepts: Where Cognitive Science Went Wrong. Oxford: Clarendon Press, 1998.
. The Mind Doesnt Work That Way. Cambridge, Mass.: MIT Press, 2000.
. The Modularity of Mind: An Essay on Faculty Psychology. Cambridge, Mass.: MIT
Press, 1983.
Forte, Allen. The Structure of Atonal Music. New Haven, Conn.:Yale University Press, 1973.
Foucault, Michel. The Order of Things: An Archaeology of the Human Sciences. New York:Vintage Books, 1973.
Freudenthal, Gideon. Atom and Individual in the Age of Newton: On the Genesis of the Mechanistic World View. Translated by Peter McLaughlin. Boston Studies in the Philosophy of
Science, 88. Dordrecht: D. Reidel, 1986.
Grdenfors, Peter. Semantics, Conceptual Spaces and the Dimensions of Music. In Essays
on the Philosophy of Music, edited by Veikko Rantala, Lewis Rowell, and Eero Tarasti,
9 27. Acta Philosophica Fennica, 43. Helsinki: Philosophical Society of Finland, 1988.
Gates, Henry Louis, Jr. The Signifying Monkey: A Theory of African-American Literary Criticism.
New York: Oxford University Press, 1988.
Geertz, Clifford. Religion as a Cultural System. In Anthropological Approaches to the Study of
Religion, edited by Michael Banton, 1 46. New York: Frederick A. Praeger, 1966.
. Thick Descr iption: Toward an Inter pretive Theory of Culture. In The Interpretation of Cultures: Selected Essays, 3 30. New York: Basic Books, 1973.
Gendron, Bernard. Moldy Figs and Moder nists: Jazz at War (1942 1946). In Jazz among
the Discourses, edited by Krin Gabbard, 31 56. Durham, N.C.: Duke University Press,
1995.
Gentner, Dedre, and Albert L. Stevens, eds. Mental Models. Hillsdale, N.J.: Erlbaum Associates,
1983.
Gerigk, Herbert. Wrfelmusik. Zeitschrift fr Musikwissenschaft 16 (1934): 359 63.
Geselle, Cynthia M. The Institutionalization of Music Theor y in France: 1764 1802.
Ph.D. diss., Princeton University, 1989.
Gibbs, Raymond W., Jr. The Poetics of Mind: Figurative Thought, Language, and Understanding.
Cambridge: Cambridge University Press, 1994.
Gibbs, Raymond W., Jr., and Herbert L. Colston. The Cognitive Psychological Reality of
Image Schemas and Their Transfor mations. Cognitive Linguistics 6 (1995): 347 78.
Gjerdingen, Robert O. A Classic Turn of Phrase: Music and the Psychology of Convention. Studies in
the Criticism and Theory of Music. Philadelphia: University of Pennsylvania Press, 1988.
. Courtly Behaviors. Music Perception 13 (1996): 365 82.
. Revisiting Meyers Grammatical Simplicity and Relational Richness. In Cognitive
Bases of Musical Communication, edited by Mari Riess Jones and Susan Holleran, 225 43.
Washington, D.C.: American Psychological Association, 1992.
Gleitman, Lila R., Sharon Lee Armstrong, and Henry Gleitman. On Doubting the Concept
Concept. In New Trends in Conceptual Representation: Challenges to Piagets Theory?
edited by Ellin Kofsky Scholnick, 87 110. Hillsdale, N.J.: Erlbaum Associates, 1983.
Goehr, Lydia. The Imaginary Museum of Musical Works: An Essay in the Philosophy of Music.
Oxford: Clarendon Press, 1992.
. The Quest for Voice: Music, Politics, and the Limits of Philosophy. Berkeley: University
of California Press, 1998.
Goetschius, Percy. Lessons in Music Form: A Manual of Analysis of All the Structural Factors and
Designs Employed in Musical Composition. Bryn Mawr, Penn.: Oliver Ditson, 1904.
Goldman, Herbert G. Banjo Eyes: Eddie Cantor and the Birth of Modern Stardom. New York:
Oxford University Press, 1997.

b i bl i og raphy
Goodman, Nelson. Languages of Art: An Approach to a Theory of Symbols. 2nd ed. Indianapolis: Hackett Publishing, 1976.
Gopnik, Alison, and Andrew N. Meltzoff. Words, Thoughts, and Theories. Cambridge, Mass.:
MIT Press, 1997.
Gopnik, Alison, Andrew N. Meltzoff, and Patricia K. Kuhl. The Scientist in the Crib: Minds,
Brains, and How Children Learn. New York:William Morrow, 1999.
Gosnell, Harold F. Negro Politicians: The Rise of Negro Politics in Chicago.With an introduction
by James Q.Wilson. Chicago: University of Chicago Press, 1966; originally published
with an introduction by Robert Park, Social Science Studies, 35. Chicago, 1935.
Grey, Thomas S. Wagners Musical Prose: Texts and Contexts. Cambridge: Cambridge University Press, 1995.
Grifn, Donald R. Animal Minds. Chicago: University of Chicago Press, 1992.
Groth, Renate. Die franzsische Kompositionlehre des 19. Jahrhunderts.Wiesbaden: Franz Steiner,
1983.
Hampton, James. Prototype Models of Concept Representation. In Categories and Concepts:
Theoretical Views and Inductive Analysis, edited by Iven van Mechelen, James Hampton,
Ryszard S. Michalski, and Peter Theuns, 67 95. Cognitive Science Ser ies. London: Academic Press, 1993.
Hampton, James, and Danile Dubois. Psychological Models of Concepts. In Categories and
Concepts: Theoretical Views and Inductive Analysis, edited by Iven van Mechelen, James
Hampton, Ryszard S. Michalski, and Peter Theuns, 11 33. Cognitive Science Ser ies.
London: Academic Press, 1993.
Harris, Paul L. Thinking by Children and Scientists: False Analogies and Neglected Similarities. In Mapping the Mind: Domain Specicity in Cognition and Culture, edited by
Lawrence A. Hirscheld and Susan A. Gelman, 294 315. Cambridge: Cambridge University Press, 1994.
Harriss, Ernest C. Johann Matthesons Der vollkommene Capellmeister: A Revised Translation
with Critical Commentary. Studies in Musicology, 21. Ann Arbor, Mich.: UMI Research
Press, 1981.
Hauptmann, Moritz. The Nature of Harmony and Metre. 2nd ed. Translated by W. E. Heathcote.
London: Sonnenschein, 1893; reprinted with a new foreword by Sigmund Levarie, New
York: Da Capo Press, 1991.
Hedges, Stephen A. Dice Music in the Eighteenth Centur y. Music and Letters 59 (1978):
180 87.
Hill, John Walter. The Logic of Phrase Structure in Joseph Riepels Anfangsgrnde zur
musikalischen Setzkunst, Part 2 (1755). In Festa musicologica: Essays in Honor of George J.
Buelow, edited by Thomas J. Mathiesen and Benito V. Rivera, 467-87. Festschr ift Ser ies,
14. Stuyvesant, N.Y.: Pendragon Press, 1995.
Hillach, Ansgar, and Klaus-Dieter Krabiel. Eichendorff-Kommentar. 2 vols. Munich: Winkler,
1971.
Hjelmslev, Louis. Prolegomena to a Theory of Language. Rev. ed. Translated by Francis J.
Whiteld. Madison: University of Wisconsin Press, 1963.
Hobbes, Thomas. Leviathan, or the Matter, Forme, and Power of a Common-Wealth Ecclesiasticall
and Civill. London: Andrew Crooke, 1651.
Hockenberry, John. Moving Violations: War Zones, Wheelchairs, and Declarations of Independence.
New York: Hyperion, 1995.
Hoegi, Pierre. A Tabular System Whereby the Art of Composing Minuets Is Made So Easy That
Any Person, without the Least Knowledge of Musick, May Compose Ten Thousand, All Different, and in the Most Pleasing and Correct Manner. London: Welcker, 1770.
Hofstadter, Douglas R., and the Fluid Analogies Research Group. Fluid Concepts and Creative

341

342

b i bl i og raphy

Analogies: Computer Models of the Fundamental Mechanisms of Thought. New York: Basic
Books, 1995.
Holy, Ladislav, and Milan Stuchlik. The Structure of Folk Models. In The Structure of Folk
Models, edited by Ladislav Holy and Milan Stuchlik, 1 34. London: Academic Press,
1981.
Hood, Mantle.The Reliability of Oral Tradition. Journal of the American Musicological Society
12 (1959): 201 09.
Hosler, Bellamy. Changing Aesthetic Views of Instrumental Music in 18th-Century Germany. Studies in Musicology, 42. Ann Arbor, Mich.: UMI Research Press, 1981.
Hoyt, Peter.The Concept of Dveloppement in the Early Nineteenth Century. In Music Theory in the Age of Romanticism, edited by Ian Bent, 141 62. Cambridge: Cambridge University Press, 1996.
. Review of Mark Evan Bonds, Wordless Rhetoric. Journal of Music Theory 38 (1994):
123 43.
Hughes, R. I. G. Reason and Exper iment in Newtons Opticks: Comments on Peter Achinstein. In Philosophical Perspectives on Newtonian Science, edited by Phillip Br icker and
R. I. G. Hughes, 175 84. Cambr idge, Mass.: MIT Press, 1990.
Hurte, Michael. Musik, Bild, Bewegung: Theorie und Praxis auditiv-visueller Konvergenzen. Bonn:
Verlag fr systematische Musikwissenschaft, 1982.
Hutchins, Edwin. Cognition in the Wild. Cambridge, Mass.: MIT Press, 1995.
Irving, John. Mozart: The Haydn Quartets. Cambridge Music Handbooks. Cambridge:
Cambridge University Press, 1998.
Jackendoff, Ray. Consciousness and the Computational Mind. Cambridge, Mass.: MIT Press, 1987.
Johnson, Graham. Die schne Mllerin: The Poems and Their Backg round. In Die schne
Mllerin, D795. Compact disk. Hyperion, [CD] CDJ 33025, 1996.
Johnson, Mark. The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason.
Chicago: University of Chicago Press, 1987.
Johnson-Laird, Philip Nicholas. Mental Models: Towards a Cognitive Science of Language, Inference, and Consciousness. Cambridge, Mass.: Harvard University Press, 1983.
Kargon, Robert H. Atomism in the Seventeenth Centur y. In Dictionary of the History of
Ideas: Studies of Selected Pivotal Ideas, vol. 1, edited by Philip P. Wiener, 132 41. New
York: Charles Scr ibners Sons, 1973.
Keil, Charles. Tiv Song. Chicago: University of Chicago Press, 1979.
Keil, Frank C. Concepts, Kinds, and Cognitive Development. Cambridge, Mass.: MIT Press,
1989.
Keil, Frank C., Daniel T. Levin, Bethany A. Richman, and Grant Gutheil. Mechanism and
Explanation in the Development of Biological Thought: The Case of Disease. In Folkbiology, edited by Douglas L. Medin and Scott Atran, 285 319. Cambr idge, Mass.: MIT
Press, 1999.
Keiler, Allan. Music as Metalanguage: Rameaus Fundamental Bass. In Music Theory: Special
Topics, edited by Richmond Browne, 83 100. New York: Academic Press, 1981.
. The Or igins of Schenkers Thought: How Man Is Musical. Journal of Music Theory 33 (1989): 273 98.
Keller, Hans. Functional Analysis no. 9a: Mozarts Piano Sonata in A Minor, K. 310. In
Essays on Music, edited by Christopher Wintle, 129 38. Cambridge: Cambridge University Press, 1994.
. The Musical Analysis of Music. In Essays on Music, edited by Christopher Wintle,
126 28. Cambridge: Cambridge University Press, 1994.
. Wordless Functional Analysis no. 1: Mozart, K. 421. The Score and I.M.A. Magazine
22 (1958): 56 64.

b i bl i og raphy
Kerman, Joseph. The Beethoven Quartets. New York:W.W. Norton, 1966.
Klotz, Sebastian. Ars combinatoria oder Musik ohne Kopfzerbrechen: Kalkle des Musikalischen von Kircher bis Kir nberger. Musiktheorie 14 (1999): 231 45.
Knaus, Herwig. Musiksprache und Werkstruktur in Robert Schumanns Liederkreis: Mit dem Faksimile des Autographs. Schriften zur Musik, 27. Munich: Emil Katzbichler, 1974.
Knouse, Nola Reed. Joseph Riepel and the Emerg ing Theory of Form in the Eighteenth
Century. Current Musicology 4 (1986): 46 62.
Koch, Carl. Bernhard Klein (17831832): Sein Leben und seine Werke. Leipzig: Oscar Brandstetter, 1902.
Koch, Heinrich Chr istoph. Introductory Essay on Composition: The Mechanical Rules of Melody,
Sections 3 and 4. Translated by Nancy Kovaleff Baker. Music Theory Translation Ser ies.
New Haven, Conn.:Yale University Press, 1983.
. Versuch einer Anleitung zur Composition. Leipzig: Adam Fr iedrich Bhme; Rudolstadt:
gedruckt mit Schr iften der Lwischen Erben, und Schirach, 1782 (vol. 1); Leipzig:
Adam Fr iedrich Bhme, 1787 93 (vols. 2 3); reprinted, Hildesheim: Georg Olms,
1969.
Kchel, Ludwig Ritter von. Chronologisch-thematisches Verzeichnis smtlicher Tonwerke Wolfgang
Amad Mozarts. 8th ed. Edited by Franz Giegling, Alexander Weinmann, and Gerd Sievers.Wiesbaden: Breitkopf and Hrtel, 1983.
Koffka, Kurt. Principles of Gestalt Psychology. London: Routledge and Kegan Paul, 1935.
Kramer, Lawrence. Music and Poetry: The Nineteenth Century and After. Berkeley: University of
Califor nia Press, 1984.
Krebs, Harald. Fantasy Pieces: Metrical Dissonance in the Music of Robert Schumann. New York:
Oxford University Press, 1999.
. Schenkers Changing View of Rameau: A Comparison of Remarks in Harmony,
Counterpoint, and Rameau or Beethoven? Theoria 3 (1989): 59 72.
Kropnger, Klaus. Wagner and Beethoven: Richard Wagners Reception of Beethoven. Translated
by Peter Palmer. Cambridge: Cambr idge University Press, 1991.
Kuhn, Thomas S. A Function for Thought Exper iments. In The Essential Tension: Selected
Studies in Scientic Tradition and Change, 240 65. Chicago: University of Chicago Press,
1977.
Kuntz, Paul G. A Formal Preface and an Infor mal Conclusion to The Great Chain of Being:
The Universality and Necessity of Hierarchical Thought. In Jacobs Ladder and the Tree
of Life: Concepts of Hierarchy and the Great Chain of Being, edited by Marion Leathers
Kuntz and Paul Grimley Kuntz, 3 14. American University Studies, Series V, Philosophy, 14. New York: Peter Lang, 1987.
Kurth, Ernst. Die Voraussetzungen der theoretischen Harmonik und der tonalen Darstellungssysteme.
Berne: Akademische Buchhandlungen von Max Dreschel, 1913.
Lakoff, George. The Contemporar y Theory of Metaphor. In Metaphor and Thought, 2nd
ed., edited by Andrew Ortony, 202 51. Cambridge: Cambridge University Press, 1993.
. The Invariance Hypothesis: Is Abstract Reason Based on Image-Schemas? Cognitive Linguistics 1 (1990): 39 74.
. Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago:
University of Chicago Press, 1987.
Lakoff, George, and Mark Johnson. Metaphors We Live By. Chicago: University of Chicago
Press, 1980.
. Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought. New
York: Basic Books, 1999.
Lakoff, George, and Mark Turner. More Than Cool Reason: A Field Guide to Poetic Metaphor.
Chicago: University of Chicago Press, 1989.

343

344

b i bl i og raphy

Lang, Paul Henry, ed. Problems of Modern Music: The Princeton Seminar in Advanced Musical
Studies. New York:W.W. Norton, 1960.
Langacker, Ronald W. Conceptualization, Symbolization, and Grammar. In The New Psychology of Language: Cognitive and Functional Approaches to Language Structure, edited by
Michael Tomasello, 1 39. Mahwah, N.J.: Erlbaum Associates, 1998.
. Descriptive Application.Vol. 2 of Foundations of Cognitive Grammar. Stanford, Calif.:
Stanford University Press, 1992.
. Grammar and Conceptualization. Berlin: Mouton de Gruyter, 2000.
. Theoretical Prerequisites.Vol. 1 of Foundations of Cognitive Grammar. Stanford, Calif.,
Stanford University Press, 1987.
Langer, Susanne K. Feeling and Form: A Theory of Art Developed from Philosophy in a New Key.
New York: Charles Scr ibners Sons, 1953.
Lassaline, Mary E., Edward J.Wisniewski, and Douglas L. Medin. Basic Levels in Articial
and Natural Categor ies: Are All Basic Levels Created Equal? In Percepts, Concepts and
Categories: The Representation and Processing of Information, edited by Barbara Bur ns,
327 78. Advances in Psychology, 93. Amsterdam: North-Holland, 1992.
Lee, Peggy. Miss Peggy Lee: An Autobiography. New York: Donald I. Fine, 1989.
Lenat, Douglas B., and R.V. Guha. Building Large Knowledge-Based Systems: Representation and
Inference in the Cyc Project. Reading, Mass.: Addison Wesley, 1989.
Lerdahl, Fred, and Ray Jackendoff. A Generative Theory of Tonal Music. Cambridge, Mass.:
MIT Press, 1983.
Levy, Janet M. Beethovens Compositional Choices: The Two Versions of Opus 18, No. 1, First
Movement. Studies in the Cr iticism and Theor y of Music. Philadelphia: University of
Pennsylvania Press, 1982.
. Gesture, Form, and Syntax in Haydns Music. In Haydn Studies: Proceedings of the
International Haydn Conference, Washington, D.C., 1975, edited by Jens Peter Larsen,
Howard Serwer, and James Webster, 355 62. New York:W.W. Norton, 1981.
Lewin, David. Auf dem Flusse: Image and Background in a Schubert Song. In Schubert: Critical and Analytical Studies, edited by Walter Fr isch, 126 52. Lincoln: University of
Nebraska Press, 1986.
. Generalized Musical Intervals and Transformations. New Haven, Conn.: Yale University
Press, 1987.
. Musical Form and Transformation: Four Analytic Essays. New Haven, Conn.:Yale University Press, 1993.
London, Justin. Riepel and Absatz: Poetic and Prosaic Aspects of Phrase Structure in 18thCentury Theory. Journal of Musicology 8 (1990): 505 19.
Louis, Rudolf, and Ludwig Thuille. Harmonielehre. 7th ed. Stuttgart: Carl Grninger Nachf .
Ernst Klett, 1920.
Lovejoy, Arthur O. The Great Chain of Being: A Study of the History of an Idea. Cambr idge,
Mass.: Harvard University Press, 1936.
Loveland, Franklin O. Snakebite Cure among the Rama Indians of Nicaragua. In Medical
Anthropology, edited by Francis X. Grollig and Harold B. Haley, 81 102. The Hague:
Mouton, 1976.
Lwenstein, Paul. Mozart-Kuriosa. Zeitschrift fr Musikwissenschaft 12 (1930): 342 46.
Lutz, Catherine. Ethnopsychology Compared to What? Explaining Behavior and Consciousness among the Ifaluk. In Person, Self, and Experience: Exploring Pacic Ethnopsychologies, edited by Geoffrey M.White and John Kirkpatr ick, 35 79. Berkeley: University of Califor nia Press, 1985.
Lyle,Wilson. Colour and Music: An Introduction. Music Review 43 (1982): 261 64.
McCauley, Robert N. The Role of Theories in a Theor y of Concepts. In Concepts and

b i bl i og raphy
Conceptual Development: Ecological and Intellectual Factors in Categorization, edited by Ulric
Neisser, 288-309. Cambridge: Cambr idge University Press, 1987.
McClary, Susan. Conventional Wisdom: The Content of Musical Form. Berkeley: University of
Califor nia Press, 2000.
McCreless, Patrick. Song Order in the Song Cycle: Schumanns Liederkreis, Op. 39. Music
Analysis 5 (1986): 5-40.
McNeill, David. Hand and Mind: What Gestures Reveal about Thought. Chicago: University of
Chicago Press, 1992.
Macpherson, Stewart. Form in Music: With Special Reference to the Designs of Instrumental Music.
London: Joseph Williams, 1908.
Macrobius, Ambrosius Aurelius Theodosius. Commentary on the Dream of Scipio. Edited and
translated by William Har ris Stahl. New York: Columbia University Press, 1990.
Macy, Laura. Speaking of Sex: Metaphor and Performance in the Italian Madr igal. Journal
of Musicology 14 (1996): 1 34.
Marks, Lawrence E. The Unity of the Senses: Interrelations among the Modalities. New York: Academic Press, 1978.
Marx, Adolph Ber nhard. Die Lehre von der musikalischen Komposition, praktisch-theoretisch. 4
vols. 4th ed. Leipzig: Breitkopf and Hrtel, 1841 47.
. Ludwig van Beethoven: Leben und Schaffen. 6th ed. Edited by Gustav Behncke. Berlin:
Otto Janke, 1908.
. Musical Form in the Age of Beethoven: Selected Writings on Theory and Method. Edited
and translated by Scott G. Burnham. Cambridge Studies in Music Theory and Analysis,
12. Cambridge: Cambridge University Press, 1997.
Mathiesen, Thomas J. Apollos Lyre: Greek Music and Music Theory in Antiquity and the Middle
Ages. Lincoln: University of Nebraska Press, 2000.
Mattheson, Johann. Kern melodischer Wissenschaft. Hamburg: C. Herold, 1737; reprinted,
Hildesheim: Georg Olms, 1976.
. Das neu-erffnete Orchestre. Hamburg: Printed for the author, 1713; reprinted,
Hildesheim: Georg Olms, 1993.
. Der vollkommene Capellmeister. Hamburg: C. Herold, 1739; reprinted, Dokumenta
musicologica, 1. Reihe, Druckschr iften-Faksimiles, 5, edited by Margarete Reimann,
Kassel: Brenreiter-Verlag, 1954.
Mervis, Carolyn B., and Eleanor Rosch. Categorization of Natural Objects. Annual Review
of Psychology 32 (1981): 89 115.
Meyer, Leonard B. Explaining Music: Essays and Explorations. Berkeley: University of California Press, 1973; reprinted, Chicago: University of Chicago Press, 1978.
Milne, A. A. The World of Pooh: The Complete Winnie-the-Pooh and the House at Pooh Corner.
Illustrated by E. H. Shepard. New York: E. P. Dutton, 1957.
Minsky, Marvin. A Framework for Representing Knowledge. In The Psychology of Computer
Vision, edited by Patrick Henry Winston, 211 77. New York: McGraw-Hill, 1975.
. The Society of Mind. New York: Simon and Schuster, 1985.
Momigny, Jrme-Joseph de. Cours complet dharmonie et de composition, daprs une thorie neuve
et gnrale de la musique. 3 vols. Paris: Chez lauteur, 1806.
. Musique, in Musique, Series 36 of Encyclopdie mthodique, ou par ordre de matires,
edited by Nicolas tienne Framr y, Pierre Louis Ginguen, and Jrme-Joseph de
Momigny, vol. 2, 175 94. Paris: Chez Madame veuve Agasse, 1818.
Monson, Ingrid. Saying Something: Jazz Improvisation and Interaction. Chicago Studies in Ethnomusicology. Chicago: University of Chicago Press, 1996.
Morgan, Robert P. Circular Form in the Tristan Prelude. Journal of the American Musicological Society 53 (2000): 69 103.

345

346

b i bl i og raphy

. Symmetrical Form and Common-Practice Tonality. Music Theory Spectrum 20

(1998): 1 47.
. The Theory and Analysis of Tonal Rhythm. The Musical Quarterly 64 (1978): 435 73.
Mozart,Wolfgang Amadeus. Musikalisches Wrfelspiel: Eine Anleitung Walzer oder Schleifer mit
zwei Wrfeln zu componieren ohne Musikalisch zu seyn, noch von der Composition etwas zu
verstehen. Edited by Karl Heinz Taubert. Mainz: B. Schotts Shne, 1956.
Mller,Wilhelm. Vermischte Schriften. 5 vols. Edited by Gustav Schwab. Leipzig: F. A. Brockhaus, 1830.
Murphy, Gregory L., and Douglas L. Medin. The Role of Theories in Conceptual Coherence. Psychological Review 92 (1985): 289 316.
Musica enchiriadis and Scolica enchiriadis. Edited by Claude V. Palisca. Translated by Raymond
Erickson. Music Theory Translation Series. New Haven, Conn.:Yale University Press, 1995.
Narmour, Eugene. The Analysis and Cognition of Basic Melodic Structures: The Implication-Realization Model. Chicago: University of Chicago Press, 1990.
. The Analysis and Cognition of Melodic Complexity: The Implication-Realization Model.
Chicago: University of Chicago Press, 1992.
. Some Major Theoretical Problems Concer ning the Concept of Hierarchy in the
Analysis of Tonal Music. Music Perception 1/2 (1983 84): 129 99.
Nattiez, Jean-Jacques. Proust as Musician. Translated by Derrick Puffett. Cambridge: Cambridge University Press, 1989.
Newton, Isaac. Opticks, or a Treatise of the Reections, Refractions, Inections and Colours of Light.
4th ed. London: W. Innys, 1730; reprinted, New York: Dover, 1952.
Nicomachus. Enchiridion. In Harmonic and Acoustic Theory.Vol. 2 of Greek Musical Writings.
Edited and translated by Andrew Barker, 247 69. Cambridge Readings in the Literature
of Music. Cambr idge: Cambridge University Press, 1989.
Noguchi, Hideo. Mozart Musical Game in C Major, K.516f. Mitteilungen der Internationalen Stiftung Mozarteum 38 (1990): 89 101.
Oakhill, Jane, and Alan Garnham, eds. Mental Models in Cognitive Science: Essays in Honour of
Phil Johnson-Laird. East Sussex, U.K.: Psychology Press, 1996.
Patrides, C. A. Hierarchy and Order. In Dictionary of the History of Ideas: Studies of Selected
Pivotal Ideas, edited by Philip P. Wiener, vol. 2, 434 49. New York: Charles Scr ibners
Sons, 1973.
Pinker, Steven. How the Mind Works. New York:W.W. Norton, 1997.
Pollard-Gott, Lucy. Emergence of Thematic Concepts in Repeated Listening to Music.
Cognitive Psychology 15 (1983): 66 94.
Porter, James. Context, Epistemics, and Value: A Conceptual Performance Model Reconsidered. In Selected Reports in Ethnomusicology: Issues in the Conceptualization of Music, 7,
edited by James Porter and Ali Jihad Racy, 69 97. Los Angeles: University of California, Department of Ethnomusicology, 1988.
. Jeannie Robertsons My Son David: A Conceptual Performance Model. Journal of
American Folklore 89 (1976): 7 26.
Powers, Richard. Galatea 2.2. New York: Farrar, Straus Giroux, 1995.
Proust, Marcel. Swanns Way.Vol. 1 of Remembrance of Things Past. Translated by C. K. Scott
Moncrieff and Terence Kilmartin. New York:Vintage Books, 1981.
Prout, Ebenezer. Musical Form. London: Augener, 1893.
Ptolemy. Harmonics. In Harmonic and Acoustic Theory.Vol. 2 of Greek Musical Writings. Edited
and translated by Andrew Barker, 275 391. Cambridge Readings in the Literature of
Music. Cambr idge: Cambridge University Press, 1989.
Quinn, Naomi, and Dorothy Holland. Culture and Cognition. In Cultural Models in Language and Thought, edited by Dorothy Holland and Naomi Quinn, 3 40. Cambridge:
Cambridge University Press, 1987.

b i bl i og raphy
Rameau, Jean-Philippe. Gnration harmonique ou trait de musique thorique et pratique. Paris:
Prault ls, 1737. Facsimile edition; Monuments of Music and Music Literature in Facsimile. Second Ser ies, Music Literature, 6. New York: Broude Brothers, 1966.
. Nouveau systme de musique thorique et pratique. Paris: Ballard, 1726.
. Observations sur notre instinct pour la musique, et sur son principe. Paris: Prault ls, 1754.
Facsimile edition; Monuments of Music and Music Literature in Facsimile. Second
Series, Music Literature, 54. New York: Broude Brothers, 1967.
Ratner, Leonard G. Ars combinatoria: Chance and Choice in Eighteenth-Century Music. In
Studies in Eighteenth-Century Music: A Tribute to Karl Geringer on His Seventieth Birthday,
edited by H. C. Robbins Landon, 343 63. New York: Da Capo Press, 1979.
. The Beethoven String Quartets: Compositional Strategies and Rhetoric. Stanford, Calif.:
Stanford Bookstore, 1995.
. Classic Music: Expression, Form, and Style. New York: Schir mer Books, 1980.
. Harmonic Aspects of Classic Form. Journal of the American Musicological Society 2
(1949): 159 68.
Rti, Rudolph. The Thematic Process in Music. New York: Macmillan, 1951.
Riemann, Hugo. Musikalische Logik: Hauptzge der physiologischen und psychologischen Begrndung unseres Musiksystems. Leipzig: C. F. Kahnt, 1874.
. Musik-Lexikon. Leipzig:Verlag des Bibliographischen Instituts, 1882.
. System der musikalischen Rhythmik und Metrik. Leipzig: Breitkopf and Hrtel, 1903.
Riepel, Joseph. Grundregln zur Tonordnung insgemein.Vol. 2 of Anfangsgrnde zur musicalischen
Setzkunst. Regensburg: Johann Leopold Montag, 1755.
. Smtliche Schriften zur Musiktheorie. Edited by Thomas Emmer ig.Wiener musikwissenschaftliche Beitrge, 20.Vienna: Bhlau, 1996.
Robertson-DeCarbo, Carol E. Tayil as Category and Communication among the Argentine
Mapuche: A Methodological Suggestion. Yearbook of the International Folk Music Council 8 (1976): 35 52.
Rosch, Eleanor. Categorization. In Encyclopedia of Human Behavior, vol. 1,V. S. Ramachandran, gen. ed., 513 23. San Diego: Academic Press, 1994.
. Cognitive Representations of Semantic Categor ies. Journal of Experimental Psychology: General 104 (1975): 192 233.
. Human Categor ization. In Studies in Cross-Cultural Psychology, 1, edited by Neil
Warren, 1 49. London: Academic Press, 1977.
. On the Inter nal Structure of Perceptual and Semantic Categor ies. In Cognitive
Development and the Acquisition of Language, edited by Timothy E. Moore, 111 44. New
York: Academic Press, 1973.
. Principles of Categorization. In Cognition and Categorization, edited by Eleanor
Rosch and Barbara B. Lloyd, 27 48. Hillsdale, N.J.: Erlbaum Associates, 1978.
Rosch, Eleanor, and Carolyn B. Mervis. Family Resemblances. Cognitive Psychology 7
(1975): 573 605.
Rosch, Eleanor, Carolyn B. Mervis, Wayne D. Gray, David M. Johnson, and Penny BoyesBraem. Basic Objects in Natural Categor ies. Cognitive Psychology 8 (1976): 382 439.
Rousseau, Jean-Jacques. Essay on the Or igin of Languages (c. 1763). In Essay on the Origin
of Languages and Writings Related to Music, translated and edited by John T. Scott,
289 332. Collected Writings of Rousseau, 7. Hanover, N.H.: University Press of New
England, 1998.
. Letter on French Music. In Essay on the Origin of Languages and Writings Related to
Music, translated and edited by John T. Scott, 141 74. Collected Writings of Rousseau,
7. Hanover, N.H.: University Press of New England, 1998.
Rowell, Lewis. Music and Musical Thought in Early India. Chicago Studies in Ethnomusicology.
Chicago: University of Chicago Press, 1992.

347

348

b i bl i og raphy

Russolo, Luigi. The Art of Noises (1916). Translated by Barclay Brown. New York: Pendragon
Press, 1986.
Rust, Brian. The American Dance Band Discography 19171942. New Rochelle, N.Y.: Arlington
House, 1975.
Saint-Sans, Camille. Le Carnaval des animaux: Grande fantasie zoologique. Edited by Felix
Aprahamian. Zurich: Eulenberg, 1974.
Sams, Eric. The Songs of Johannes Brahms. New Haven, Conn.:Yale University Press, 2000.
Schank, Roger C., and Robert P. Abelson. Scripts, Plans, Goals and Understanding: An Inquiry
into Human Knowledge Structures. Hillsdale, N.J.: Erlbaum Associates, 1977.
Schelling, Friedrich Wilhelm Joseph von. Philosophy of Art. Edited, translated, and introduced
by Douglas W. Stott, with a foreword by David Simpson. Theory and History of Literature, 58. Minneapolis: University of Minnesota Press, 1989.
Schenker, Heinrich. Counterpoint: Book I: Cantus Firmus and Two-Voice Counterpoint. Edited
by John Rothgeb. Translated by John Rothgeb and Jrgen Thym. New Musical Theories and Fantasies, 2. New York: Schir mer Books, 1987.
. Free Composition (Der freie Satz). Edited and translated by Ernst Oster. New Musical
Theories and Fantasies, 3. New York: Schirmer Books, 1979.
. Der freie Satz. Neue musikalische Theor ien und Phantasien, 3.Vienna: UniversalEdition A. G., 1935.
. Harmonielehre. Neue musikalische Theor ien und Phantasien, 1. Stuttgart: J. G. Cottasche, 1906.
. Rameau or Beethoven? Creeping Paralysis or Spir itual Potency in Music? In The
Masterwork in Music: A Yearbook, 1 9. Translated by Ian Bent, edited by William
Drabkin. Cambr idge Studies in Music Theory and Analysis, 5. Cambridge: Cambridge
University Press, 1994.
Schmid, Hans, ed. Musica et Scolica enchiriadis una cum aliquibus tratatulis adiunctis. Verffentlichungen der Musikhistorischen Kommission, Bayerische Akademie der Wissenschaften, 3. Munich:Verlag der Bayerische Akademie der Wissenschaften, 1981.
Schoenberg, Arnold. Brahms the Progressive (1947). In Style and Idea: Selected Writings of
Arnold Schoenberg, edited by Leonard Stein, translated by Leo Black, 398-441. New York:
St. Martins Press, 1975.
. Coherence, Counterpoint, Instrumentation, Instruction in Form. Edited by Severine Neff.
Translated by Severine Neff and Charlotte M. Cross. Lincoln: University of Nebraska
Press, 1994.
. Fundamentals of Musical Composition. Edited by Gerald Strang and Leonard Stein.
London: Faber and Faber, 1967.
. Harmonielehre. Leipzig:Verlagseigentum der Universal-Edition, 1911.
. The Musical Idea and the Logic, Technique, and Art of Its Presentation. Edited and translated by Patricia Car penter and Severine Neff . New York: Columbia University Press,
1995.
. My Evolution (1949). In Style and Idea: Selected Writings of Arnold Schoenberg, edited
by Leonard Stein, translated by Leo Black, 79 92. New York: St. Martins Press, 1975.
. Opinion or Insight? (1926). In Style and Idea: Selected Writings of Arnold Schoenberg,
edited by Leonard Stein, translated by Leo Black, 258 64. New York: St. Martins Press,
1975.
. Theory of Harmony. Translated by Roy E. Carter. Berkeley: University of California
Press, 1978.
Sectio canonis. In Harmonic and Acoustic Theory. Vol. 2 of Greek Musical Writings. Edited by
Andrew Barker, 191 208. Cambridge Readings in the Literature of Music. Cambridge:
Cambridge University Press, 1989.

b i bl i og raphy
Seeger, Anthony. Why Suy Sing: A Musical Anthropology of an Amazonian People. Cambridge:
Cambr idge University Press, 1987.
Seeger, Charles. Oral Tradition in Music. In Funk and Wagnalls Standard Dictionary of Folklore, Mythology, and Legend, edited by Maria Leach and Jerome Fried, vol. 2, 825 29.
New York: Funk and Wagnalls, 1950.
Sessions, Roger. The Musical Experience of Composer, Performer, Listener. Princeton, N.J.:
Princeton University Press, 1950.
Shore, Bradd. Culture in Mind: Cognition, Culture, and the Problem of Meaning. New York:
Oxford University Press, 1996.
Smith, Edward E., and Douglas L. Medin. Categories and Concepts. Cambridge, Mass.: Harvard
University Press, 1981.
Snarrenberg, Robert. Competing Myths: The Amer ican Abandonment of Schenkers
Organicism. In Theory, Analysis and Meaning in Music, edited by Anthony Pople, 29 56.
Cambr idge: Cambridge University Press, 1994.
Spitzer, Mark. Marxs Lehre and the Science of Education: Towards the Recuperation of
Music Pedagogy. Music and Letters 79 (1998): 489 526.
Stark, Lucien. A Guide to the Solo Songs of Johannes Brahms. Bloomington: Indiana University
Press, 1995.
Stone, Ruth M. Let the Inside Be Sweet: The Interpretation of Music Event among the Kpelle of
Liberia. Bloomington: Indiana University Press, 1982.
Sweetser, Eve E. The Denition of Lie: An Examination of the Folk Models Underlying a
Semantic Prototype. In Cultural Models in Language and Thought, edited by Dorothy
Holland and Naomi Quinn, 43 66. Cambridge: Cambridge University Press, 1987.
Talbot, Michael, ed. The Musical Work: Reality or Invention? Liverpool: University of Liverpool, 2000.
Talmy, Leonard. Concept Structuring Systems.Vol. 1 of Toward a Cognitive Semantics. Cambridge,
Mass.: MIT Press, 2000.
Tanaka, James W., and Marjor ie Taylor. Object Categor ies and Expertise: Is the Basic Level
in the Eye of the Beholder? Cognitive Psychology 23 (1991): 457 82.
Taubert, Karl Heinz, ed. Das Menuett: Musikalisches Wrfelspiel, Paris 1786. Zurich: Musikhaus
Pan, 1988.
Temperley, David. Motivic Perception and Modular ity. Music Perception 13 (1995): 141 69.
Thomas, Downing A. Music and the Origins of Language: Theories from the French Enlightenment.
New Perspectives in Music History and Criticism. Cambridge: Cambridge University
Press, 1995.
Tomasello, Michael. Introduction: A Cognitive-Functional Perspective on Language Structure. In The New Psychology of Language: Cognitive and Functional Approaches to Language
Structure, edited by Michael Tomasello, viixxiii. Mahwah, N.J.: Erlbaum Associates,
1998.
Tovey, Donald Francis. Symphonies.Vol. 1 of Essays in Musical Analysis. Oxford: Oxford University Press, 1935.
Treitler, Leo. History and the Ontology of the Musical Work. Journal of Aesthetics and Art
Criticism 51 (1993): 483 97.
. Transmission and the Study of Music Histor y. In Report of the Twelfth Congress,
Berkeley 1977, edited by Daniel Heartz and Bonnie Wade, 202 11. Kassel: BrenreiterVerlag; Philadelphia: The American Musicological Society, 1981.
Turner, Mark. Aspects of the Invariance Hypothesis. Cognitive Linguistics 1 (1990): 247 55.
. Design for a Theory of Meaning. In The Nature and Ontogenesis of Meaning, edited
by Willis F. Overton and David S. Palermo, 91 108. Hillsdale, N.J.: Erlbaum Associates,
1994.

349

350

b i bl i og raphy

. Figure. In Figurative Language and Thought, edited by Albert N. Katz, Cristina Cacciari, Raymond W. Gibbs, Jr., and Mark Turner, 44 87. New York: Oxford University
Press, 1998.
. An Image-Schematic Constraint on Metaphor. In Conceptualizations and Mental
Processing in Language, edited by Richard A. Geiger and Br ygida Rudzka-Ostyn,
291 306. Cognitive Linguistics Research, 3. Berlin: Mouton de Gruyter, 1993.
. The Literary Mind. New York: Oxford University Press, 1996.
Turner, Mark, and Gilles Fauconnier. Conceptual Integ ration and Formal Expression.
Metaphor and Symbolic Activity 10 (1995): 183 204.
. Conceptual Integration in Counterfactuals. In Discourse and Cognition: Bridging the
Gap: [Second] Conceptual Structure, Discourse, and Language Conference, Buffalo, 1996,
edited by Jean-Pier re Koenig, 285 96. Stanford, Calif.: Center for the Study of Language and Infor mation, 1998.
Tversky, Barbara, and Kathleen Hemenway.Categories of Environmental Scenes. Cognitive
Psychology 15 (1983): 121 49.
. Objects, Parts, and Categor ies. Journal of Experimental Psychology: General 113
(1984): 169 93.
Verba, E. Cynthia. The Development of Rameaus Thoughts on Modulation and Chromatics. Journal of the American Musicological Society 26 (1973): 69 91.
Waidelich, Till Ger rit, ed. Franz Schubert: Dokumente, 18171830. Verffentlichungen des
Internationalen Franz Schubert Instituts, 10. Tutzing, Germany: Hans Schneider, 1993.
Walser, Robert. Out of Notes: Signication, Interpretation, and the Problem of Miles
Davis. In Jazz among the Discourses, edited by Krin Gabbard, 165 88. Durham, N.C.:
Duke University Press, 1995.
Weber, Gottfried. Theory of Musical Composition, Treated with a View to a Naturally Consecutive Arrangement of Topics. Edited by John Bishop. Translated by James F.Warner. London:
Robert Cocks, 1851.
Welker, Robert L. Abstraction of Themes from Melodic Variations. Journal of Experimental
Psychology: Human Perception and Performance 8 (1982): 435 47.
Wendt, Lloyd, and Her man Kogan. Big Bill of Chicago. Indianapolis: Bobbs-Mer rill, 1953.
Will, Richard. Time, Morality, and Humanity in Beethovens Pastoral Symphony. Journal of
the American Musicological Society 50 (1997): 271 329.
Wittgenstein, Ludwig. Philosophical Investigations. Translated by G. E. M. Anscombe. New
York: Macmillan, 1953.
Wolff,Viktor Er nst. Lieder Robert Schumanns in ersten und spteren Fassungen. Berlin: H. S. Hermann, 1913.
Wolzogen, Hans von. Guide through the Musical Motives of Richard Wagners Tristan and Isolde:
With a Preface on the Legend and the Poem of Wagners Drama. Leipzig: Feodor Reinboth,
1889.
. Thematischer Leitfaden durch die Musik zu Richard Wagners Tristan und Isolde: Nebste
einem Vorworte ber den Sagenstoff des Wagnerschen Dramas. Leipzig: Feodor Reinboth,
1890.
Youens, Susan. Behind the Scenes: Die schne Mllerin before Schubert. 19th Century Music
15 (1991): 3 22.
. Retracing a Winters Journey: Schuberts Winterreise. Ithaca, N.Y.: Cornell University
Press, 1991.
. Schubert: Die schne Mllerin. Cambr idge Music Handbooks. Cambridge: Cambridge
University Press, 1992.
. Schubert, Mller, and Die schne Mller in. Cambridge: Cambridge University Press,
1997.

b i bl i og raphy
Zanten, Wim van. The Tone Mater ial of the Kacapi in Tembang Sunda in West Java. Ethnomusicology 30 (1986): 84 112.
Zbikowski, Lawrence. The Blossoms of Trockne Blumen: Music and Text in the Early
Nineteenth Century. Music Analysis 18 (1999): 307 45.
. Large-Scale Rhythm and Systems of Grouping. Ph.D. diss.,Yale University, 1991.
Ziolkowski, Theodore. German Romanticism and Its Institutions. Princeton, N.J.: Princeton
University Press, 1990.
Recordings
This reference list makes use of the following abbreviations: CD = compact disc; LP = 33
rpm record; 78 = 78 r pm record.
Bennie Kruegers Orchestra. Bye Bye Blackbird. Brunswick, [78] no. 3186, 1926.
Benny Goodman Sextet. I Got Rhythm. In Benny Goodman. CBS Records, [LP] P5 15536
CP, 1981.
Bobby Short. Bye Bye Blackbird. In Bobby Short. Atlantic, [LP] 1230, 1956.
Charlie and His Orchestra. Bye Bye Blackbird. In Charlie and His Orchestra: German Propaganda Swing 19411942. Harlequin, [CD] HQ CD 03, 1990.
Charlie Parker and others. I Got Rhythm. In Bird: The Complete Charlie Parker on Verve.
Verve/Polygram Classics, [CD] 837 142-2, 1988.
Chick Webb and His Little Chicks. I Got Rhythm. In Spinnin the Webb: Chick Webb and
His Orchestra. GRP Records, [CD] 513678L, 1994.
Derek Bailey, Pat Metheny, Gregg Bendian, and Paul Wertico. The Sign of 4. Knitting Factory
Works, [CD] KFW 197, 1997.
Don Redman and His Orchestra. I Got Rhythm. In Don Redman and His Orchestra,
19311933. Classics, [CD] 543 Classics, 1990.
Eddie Cantor. Bye Bye Blackbird. In Songs by Eddie Cantor from The Eddie Cantor Story.
Capitol, [LP] L-46, 1954.
Fletcher Henderson and His Orchestra. Yeah Man. In Fletcher Henderson, Developing an
American Orchestra, 19231937. Smithsonian Collection, [LP] R 006, 1977.
Fred Rich and His Orchestra. I Got Rhythm. In 1930Youre Driving Me Crazy: Portrait
of a Year in Music. Phontastic, [CD] PHONT CD 7618, 1992.
Freddie Schnickelfritz Fisher and His Orchestra. Bye Bye Blackbird. Decca, [78] no.
25357/3788, 1941.
Gene Austin. Bye Bye Blackbird.Victor, [78] no. 20044-B, 1926.
George Gershwin. I Got Rhythm. In Gershwin Performs Gershwin: Rare Recordings
19311935. Musical Her itage Society, [CD] 512923A, 1991.
George Gershwin and Ira Gershwin. Girl Crazy. Elektra Nonesuch, [CD] 79250-2, 1990.
Jimmy Dorsey and His Orchestra. I Got Rhythm. In Jimmy Dorsey and His Orchestra, Vol.
2: 194244. Hindsight Records, [LP] HSR-153, 1980.
Johnny Maddox. Bye Bye Blackbird. In Tap Dance Rhythms. Dot, [LP] DLP 3008, 1955.
Lester Young with the Kansas City Six. I Got Rhythm. In The Tenor Sax: Lester Young, Chu
Berry and Ben Webster: The Commodore Years. Atlantic, [LP] SD 2-307, 1973.
Marlene Fingerle and Arthur Schutt. Medley:I Know That You Know, Baby Face, Bye
Bye Blackbird. In Song Hits of 1926. Songs of Our Times. Decca, [LP] DL 5170, 1950.
Miles Davis. Round about Midnight. Columbia, [CD] CK 40610, 1987.
Peggy Lee.Bye Bye Blackbird. In Songs from Pete Kellys Blues. Decca, [LP] DL 8166, 1955.
Pete Kelly and His Big Seven (Matty Matlock and others). Pete Kellys Blues. BMG Special Products; [CD] DRC12081, 1998.

351

352

b i bl i og raphy
Red Nichols and His Five Pennies. I Got Rhythm. In Rarest Brunswick Masters, 19261931:
Red Nichols and His Five Pennies. MCA Records, [LP] MCA 1518, 1982.
Red Norvo and His Swing Sextette. I Got Rhythm. In Red Norvo. Time-Life Records,
[LP] STL-J14, 1980.
Sidney Bechet and the New Orleans Feetwarmers. Shag. In Sidney Bechet. Time-Life
Records, [LP] STL-J09, 1980.

index

Page numbers in italics indicate that a musical example appears on the given page. Page numbers in
boldface indicate substantial analytical discussions with numerous musical examples and gures.
Abelson, Robert P., 112
Absatz, 293, 299
Acadmie Royale de Musique. See Opra
Agawu,Victor Ko, 51n45, 244 45, 286
Aldwell, Edward, viii
analysis, musical
and music theory, 19, 134 35, 333
analysis of song, 18, 244 45
and conceptual blending, 254 55, 264 65,
285 86, 332
Anlage, 296
Anstieg, 130, 261
anthropomor phism, 79 81, 91 95, 113, 127n59,
248, 315, 327
Antrag (Propositio), 295
Aristotle, 10, 311
Aristoxenian music theory, 7, 12 14, 16
Aristoxenus, 10 12, 14, 63
Arom, Simha, 217
Auffhrung (Executio), 295
Ausarbeitung (Elaboratio), 295 96
Ausfhrung (Executio), 295 96
Austin, Gene, 231 32
Baby Face (by Harry Akst and Benny Davis),
234
Bach, Johann Sebastian, 126
Bakhtin, Mikhail Mikhailovich, 18, 204, 224 25,
22729
Bali. See pitch relations, descriptions of
Bamberger, Jeanne
research on musical understanding, 96 99
work with Jeff, 17, 99 101, 103 7
Barker, Andrew, 63

Barsalou, Lawrence W., 41, 69

Bechet, Sidney, 210
Beethoven, Ludwig van, 18, 27, 48, 59, 169,
201 3, 242
Fifth Symphony (Symphony No. 5 in C
minor, Op. 67), rst mvmt., 34, 35 40,
42 44, 46 48
Ninth Symphony (Symphony No. 9 in D
minor, Op. 125), 120
piano sonatas, 120
Sixth Symphony (Symphony No. 6 in F
major, Op. 68; Pastorale), 89 94, 327,
331
String Quartet, Op. 18 No. 1, rst mvmt.,
177 91, 191 94, 196, 198, 200, 329
String Quartet, Op. 18 No. 1 [rst version],
rst mvmt., 177, 191 97
String Quartet, Op. 18 No. 6, rst mvmt.,
169 77, 199, 329
Bekrfftigung (Conrmatio), 295
Bericht (Narratio), 295
Berlin, Brent, 32, 115
Berliner, Paul F., 220, 331
Boethius, Ancius Manlius Severinus, 15
Brahms, Johannes, 126, 279
compositional technique, 25, 57 58, 284
In der Fremde (Op. 3 No. 5), 27984, 286, 331
Sextet in Bf major, rst mvmt., Op. 18, 26,
26 27, 28
Brailoiu, Constantin, 216 17
Brechung, 127
Brown, Roger, 31, 33
Bruckner, Anton
Sixth Symphony, 57 58

353

354

i nde x
Bruno, Giordano, 312
Bye Bye Blackbird (by Mort Dixon and Ray
Henderson), 18, 203 4, 219, 223 25,
229 39, 240
racial associations, 223 24, 239 40
recording history, 230 39
cadence, 121, 151 52, 284
complete (vollstndiger Schlu ), 299
garbled (verstmmelter Schlu ), 299
perfect authentic, 61
Cage, John
4' 33", 242
Cantor, Eddie (Edward Israel Iskowitz), 235
Carroll, Lewis (Charles Dodgson)
Alices Adventures in Wonderland, 287 88, 305,
324
Through the Looking Glass, 169
categor ies, 12, 16, 326, 332 33
articial (see categor ies: Type 2)
basic-level, 31 33
basic-level, and musical motives, 34, 36
classical (see categories: Type 2)
natural (see categor ies: Type 1)
Type 1, 39 42, 44, 48 49, 61n59, 149, 326
Type 2, 40 41, 48, 61n59
typicality effects in, 36 40
typicality effects in, and musical motive,
43 44, 46 49
categor ies, musical, 59 62, 168, 176, 242, 326
and compositional strategy, 139 40, 150 52,
159, 167, 191, 199 200
and musical ontology, 203, 213, 222 23
typicality effects in, and compositional strategy,
47, 50 51, 149, 153, 200
categor ies, musical motives as, 34, 43 44, 46 52,
58 59
in Beethovens String Quartet Op. 18 No. 1,
177, 179 80, 182 84, 186, 188 89, 191,
193 94, 196, 198
in Beethovens String Quartet Op. 18 No. 6,
169, 171 74, 176
in Mozarts String Quartet K. 465, 155,
158 60, 164, 166 67
categor ization, 12 17, 24, 30 33, 37, 43 46,
48 49, 51, 58, 60 61, 77, 102, 108,
203 4, 222, 327 28, 333
and compositional strategy, 139, 199 200
research on, 30 31, 108
Catel, Charles Simon, 301n39
Chambers, Jack, 224
Chambers, Paul, 223, 239
changes (i.e., harmonic progressions; in jazz
performance), 211n21
Charlie and His Orchestra, 234

Chomsky, Noam, 4
chorus (in jazz perfor mance), 210
CINs. See conceptual integration networks
cognitive capacities, 328, 331, 334
cognitive domain, 110n20
cognitive linguistics, 5, 17, 64, 66, 138 39
cognitive structure, 4 6, 17, 19, 108, 111, 119,
133 34, 330, 333
cognitive theory, 19
Coleman, Linda, 226, 242
Colston, Herbert L., 69
Coltrane, John, 223 25, 239 40
competence, linguistic, 4
competence, musical, 4
compositional strategy, 47, 54, 138 41, 148 49,
152 54, 159, 167 68, 177, 180, 184, 191,
198 200, 264, 329
and musical syntax, 138 40, 148, 153 54, 167,
198 200, 264
concepts, 108 9
categories and, 60 61
musical, 17, 64, 102, 326 27 (see also categories, musical)
conceptual blending, 17 18, 288, 323, 331
in Brahmss In der Fremde, 284
in Kleins Trockne Blumen, 254, 256,
26465
in Mllers Trockne Blumen, 248 49
and program music, 89 94
role in text painting, 65, 82 84, 87 89
in Schuberts Trockne Blumen, 263 65, 286
in Schumanns Im Rhein, 270 72
in Schumanns In der Fremde, 278 79
theory of, 77 82
See also analysis of song
conceptual domains, 103, 108 9
basic features of, 112 13
and mental spaces, 113
and theor ies, 113
conceptual integration networks (CINs), 78 80,
82, 331
for anthropomorphism, 91 92
for Brahmss In der Fremde, 282
for Kleins Trockne Blumen, 253 56
for Palestr inas text painting, 82 83
for Schuberts Trockne Blumen, 262 63
for Schumanns Im Rhein, 270
for Schumanns In der Fremde, 277
topography of, 79 80, 254, 265, 272, 283,
28586
for Toveys analysis of Beethovens Sixth
Symphony, 92 93
for Werts Tirsi mor ir volea, 87 88
conceptual models, 14 17, 102, 108 9, 324, 327,
33233

i nde x
basic features of, 109 12
cross-domain mapping and, 71 72
global, 46 48, 50, 59, 62, 72, 74, 76, 112, 167,
218, 221 22
idealization of, 111
Jeff s development and use of, 101 2, 104 7,
109, 218
local, 48, 50, 59, 62, 112, 218, 221 22
and music theory, 332
musical motives and, 46 48, 50 52, 57 58
and musical notation, 221 22
and musical ontology, 203, 213 14, 242, 330
and musical perfor mance, 216 23
negotiation of, 219 22
representing, 112
role in categor ization, 44 48
and theor ies, 15 17, 62, 98 99, 102 3, 107 9,
111, 125, 131, 133, 317, 323 24
conceptual models for musical categor ies
for Beethovens Fifth Symphony, 46 48
for Beethovens String Quartet Op. 18 No. 1,
179 80, 184, 187, 198
for Beethovens String Quartet Op. 18 No. 6,
171 72
for Bye Bye Blackbird, 233, 235, 238 39
for I Got Rhythm, 212 16
for Mozarts String Quartet K. 465, 155, 160,
167
conceptual slippage, 105 6, 112, 159, 183 84
conceptualization of music, 12, 14, 59 61, 67,
334
concords, 8 9, 11, 13, 15
Cone, Edward T., 333
analysis of Mozart Piano Sonata K. 331,
307 10, 323 24
on motivic der ivation, 27 28, 54n48, 59
Conr matio (Bekrfftigung), 295
Confutatio (Wiederlegung), 295
Conservatoire (Conservatoire National de
Musique et de Dclamation), 300 301
consonant intervals. See concords
Cook, Nicholas, 245, 265, 332
Cooper, Grosvenor W., 307
corps sonore, 123, 127n59
Cowdery, James R., 217
Crawford, Richard, 204n7, 210, 214, 222
cross-domain mapping, 13 17, 19, 71, 73,
287 89, 291, 295 97, 326 27, 332 33
accounts of music and, 64, 76
between grammar and music, 291 92,
294 96, 299, 301
between language and music, 291 92, 299
between rhetor ic and music, 295, 297 99
image-schema theory and, 68 69
metaphor theory and, 65 66

music theory and, 72 76

and musical categor ies, 330
role in conceptual models, 109 10
See also hierarchy; metaphor: contemporary
theory of
cultural knowledge, 109 10, 115 16, 202 4,
219 20, 222 23, 240 42
cultural model, 110n20
cultural theory of information exchange, 225 29
Czerny, Carl, 305
Dahlhaus, Carl, 23 24, 49, 57 58
Damasio, Antonio R., 70, 108n16, 331
Davis, Miles, 223 25, 239 40
Davis, Ruth, 222
Debussy, Claude
Reets dans leau, ix
Decoratio (Schmckung or Zierde), 295
Dever,William E., 223
diezeugmenon, 6
discords, 8 9, 11, 13, 15
Dispositio (Einrichtung), 295 96
dissonance, 25, 121
dissonant intervals. See discords
Dixon, Mort, 229
Dorsey, Jimmy (James), 210, 213 16
double-voiced discourse, 204, 224 25, 227 29,
23941
Dylan, Bob (Robert Allen Zimmer man), 139n73
Eco, Umberto, 137 38, 154, 168 69, 199
Edelman, Gerald M., 60, 69, 108
Eichendorff, Joseph Freiherr von
In der Fremde, 272 73, 276 77, 279,
28284
Viel Lrmen um nichts, 272 73, 276
Eingang (Exordium), 295
Einrichtung (Dispositio), 295 96
Einschnitt, 293, 299
Elaboratio (Ausarbeitung), 295 96
embodied exper ience. See image schemata
enthymeme. See cadence: garbled
Erndung (Inventio), 295
ethnobotanical classication, 32, 115 16
Executio (Auffhrung or Ausfhrung), 295 96
Exordium (Eingang), 295
fake books, jazz, 209
Fantasia (Walt Disney movie), 94n47
Fauconnier, Gilles, 78, 82
Feld, Steven, 67, 116
Ferris, David, 276n28
Ftis, Franois-Joseph, 317
Fichte, Johann Gottlieb, 302 3
Fingerle, Marlene, 234

355

356

i nde x
Finnegan, Ruth, 220
Fisher, Freddie Schnickelfritz, 234
Fitzgerald, Ella, 236
Fodor, Jerry A., 29, 60n53
form, musical, 18, 288. See also grammar, and
musical for m; sonata for m; theories of
musical for m
Frster, Friedrich, 246
Forte, Allen, 132
Foucault, Michel, 295n23, 298n36
frames, 41 44, 110n20
Freudenthal, Gideon, 312 13
Friedrich Wilhelm III (king of Prussia), 301
games, musical, 139 40
Garland, Judy (Frances Gumm), 212
Garland, Red (William M.), 223, 239
Gates, Henry Louis, Jr., 18, 204, 224, 228
Geertz, Clifford, 111, 201 2, 241 42
German universities, 301 2. See also music
education: German
Gershwin, George, 204, 209, 213 16. See also I
Got Rhythm
Gershwin, Ira, 204
Gibbs, Raymond W., Jr., 69
Girl Crazy (George and Ira Gershwin musical),
204 6, 212, 213n25
Gjerdingen, Robert O., 43n33, 168
Goetschius, Percy, 290 91, 305
Goodman, Benny (Benjamin), 212 15
Goodman, Nelson, 203n5
Gopnik, Alison, 117 18
grammar and music, 290, 292 95, 297 99.
See also cross-domain mapping: between
grammar and music
grammar, and musical for m, 298 99
Graun, Carl Heinr ich
Der Tod Jesu, 297
Greater Perfect System, 6 7, 10 13
Greek
music, 6, 12, 330
theories of music, 6 16, 330
Grifn, Donald R., 60
Groth, Renate, 317
Guarini, Giovanni Battista, 84
Hammerstein, Oscar, 224
Happy Birthday (Good Mor ning to All,
by Mildred J. and Patty Smith Hall),
97 98
Harris, Paul L., 118n39
Hauptmann, Moritz, 320
Haydn, Joseph, 154, 174
head (in jazz perfor mance), 210n15
Hegel, Georg Wilhelm Friedrich, 302

Heine, Heinrich
Im Rhein, 265 66, 272
Lyriches Intermezzo (1823), 265, 272
Henderson, Fletcher, 210
Henderson, Ray, 229
Hensel, Luise, 246
Hensel, Wilhelm, 246
hierarchy
atomistic, 310, 313, 320
mapping on to music, 320 23
structure of, 314 16
chain-of-being, 310 13, 320
mapping on to music, 316 18, 321 23
structure of, 314 15
models of, 289, 310, 314, 316, 324
musical, 18 19, 59 60, 120, 128, 131, 306 8,
32324
Hobbes, Thomas, 312
Hockenberry, John, 73n21
Hofstadter, Douglas R., 105
Holland, Dorothy, 109 10
Hood, Mantle, 220
Hot Cross Buns (folk song), 99
Hutchins, Edwin, 9n9, 115
Hypate hypaton, 6, 10
Hypate meson, 6, 10
hypaton, 6
hyperbolaion, 6
I Got Rhythm (by George and Ira Gershwin)
18, 203, 204 12, 213 16, 218 22, 231,
233, 238
recording history, 205 6, 208 12
Rhythm changes, 211, 331
tunes der ived from, 210
I Know That You Know (by Vincent
Youmans), 234
Im a Little Blackbird Looking for a Bluebird
(by Grant Clarke, Arthur Johnston,
George W. Meyer, and Roy Turk), 237
idealized cognitive model, 110n20
Im Rhein. See Heine, Heinrich; Schumann,
Robert
image schemata, 68 71, 314 15, 322 23,
327
and rhetor ical gures, 138 39
theory of, 68 70
In der Fremde. See Brahms, Johannes; Eichendorff, Joseph Freiherr von; Schumann,
Robert
India, music theory of, 13
intervals. See pitch relations, descriptions of
Invariance Principle, 70 71, 322 23
Inventio (Erndung ), 295
invention (process of ), 295 97, 302 3

i nde x
Jackendoff, Ray, 59
Java. See pitch relations, descriptions of
Jeff (experimental subject). See Bamberger,
Jeanne: work with Jeff
Johnson, Mark, 65 66, 68
Johnson, Robert
Cross Road Blues, xxi, 74n22
Jones, Philly Joe (Joseph Rudolph Jones), 223,
239
Kaluli (Papua New Guinea). See pitch relations,
descriptions of
Kay, Paul, 226, 242
Keats, John
Ode to a Nightingale, 66
Keller, Hans, 61n57
Kerman, Joseph, 169, 176
Klein, Bernhard, 243
Trockne Blumen, 249 57, 264 65, 286
knowledge structures, 110n20
Koch, Heinrich Chr istoph, 148, 152, 293 94,
297 98, 301
Kopfton, 127, 129 30n64, 261, 321
Kramer, Lawrence, 244
Krueger, Bennie, 230 32
Kuhn, Thomas S., 118
Kurth, Ernst
character ization of tonality, 76
Lakoff, George, 65 66, 70, 111
Langer, Suzanne, 244
language and music, 5, 17, 61, 64, 84, 89, 220, 245,
285, 297, 326, 331. See also cross-domain
mapping: between language and music
language, origins of, 284 85, 291
Lee, Peggy (Nor ma Delor is Egstrom), 224,
236 40
Leibniz, Gottfried Wilhelm, 313
Leidensmotiv. See Wagner, Richard: Leidensmotiv
from Tristan und Isolde
Lerdahl, Fred, 59
Levy, Janet, 174, 177, 191
Lewin, David, ix, 98, 326n2
Lichanos hypaton, 6
Longinus, 138 39
Louis, Rudolf, and Ludwig Thuille
character ization of tonality, 74 76, 83
Lully, Jean-Baptiste, 120
recitative from Armide, 120, 123
McClary, Susan, xxi, 59, 74n22
McNeill, David, 69
Macpherson, Stewart, 290n9
Macrobius, Ambrosius Aurelius Theodosius, 311
Maddox, Johnny, 236

Mahler, Gustav, 242

Marcello, Benedetto, 295
Marx, Adolph Bernhard (A. B.), 48, 177, 302 5
denition of musical motive, 304
Die Lehre von der musikalischen Komposition,
3025
Mattheson, Johann, 292, 294 97, 333
analysis of anonymous minuet, 292, 293
Meltzoff, Andrew, 117 18
Melvin and Howard (Jonathan Demme movie),
239n78
mental models, 110n20
mental spaces, 78, 80, 82, 89, 331. See also conceptual domains: and mental spaces
Merman, Ethel (Ethel Agnes Zimmer man),
204 5, 213
Mese, 6
meson, 6
metaphor
conceptual, 66 70, 300
contemporary theory of, 14, 17, 65 70, 108
linguistic, 66, 300
Meyer, Leonard B., 307 10, 321 23, 333
Milne, A. A. (Alan Alexander), 77
Minsky, Marvin, 112
modular ity, theory of, 29 30
Momigny, Jrme-Joseph de
theory of tonality, 316 17
Monson, Ingrid, 220, 224 25, 240
Montessor i bells, 96
Morgan, Robert P., 50n43, 52n47, 54nn48 49, 333
analysis of Mozart K. 331 theme, 308 10, 321,
323
motif. See motive, musical
motive, musical, 3, 16, 18, 325
as category, 16
as dynamic construct, 289 90, 305
Goetschiuss denition of, 290
Marxs denition of, 304
and musical for m, 289 91
and process of derivation, 27 29
Prouts denition of, 289 90
Riemanns denition of, 305
as static building block, 290
See also theories of musical motive
Mozart,Wolfgang Amadeus, 18
Marriage of Figaro, Overture, 301
musical game (see Musikalisches Wrfelspiel )
Piano Sonata in A major, K. 331, rst mvmt.,
306 10, 321 24
String Quartet K. 465 (Dissonance), rst
mvmt., 15468, 199, 330
Mller,Wilhelm, 243, 285
Die schne Mllerin, 246 47, 272
Trockne Blumen, 247 49

357

358

i nde x
music education
demand for instruction books, 291
French, 300 301, 305
German, 289, 302, 304 5
institutionalization of, 324
Musica enchiriadis (anonymous treatise), 291
musical notation, 67, 72, 330
Jeff s development of, 100 101, 103 4, 106 7
and the negotiation of conceptual models,
221 22
musical ontology, 18, 202 4, 214, 222, 241 42,
330
Musikalisches Wrfelspiel (attributed to W. A.
Mozart), 140 54
organization of game, 141 43, 145, 148
template for compositional strategy, 149 54
My Favorite Things (by Oscar Hammerstein
and Richard Rodgers), 224 25, 240
Napoleon I (Napoleon Bonaparte), 301
Narmour, Eugene, 62n60, 322n93
Narratio (Bericht), 295
Naturklang, 127, 131, 318
navigation in Polynesia and Micronesia, 115 16
Neoplatonic cosmology, 311 12, 318
Neoplatonism, 311
Newton, Isaac, 313
Nichols, Red (Ernest Lor ing Nichols), 206,
213 15
Nicomachus of Gerasa, 8 9
Norvo, Red (Kenneth Norville), 213 15
notation, musical. See musical notation
obligate Lage, 130 31
Odette (from Prousts A la recherche), 19, 325 27
Opra (Acadmie Royale de Musique), 300
organicism, 113, 127, 130 31, 300, 302 5,
314n71, 317 18
Palestr ina, Giovanni Perluigi da
Pope Marcellus Mass, Credo, 6364, 73, 82 83
use of text painting, 63 64, 73 74, 77, 82 84
Parhypate hypaton, 6
Parker, Charlie (Charles), 211, 213 15
Pastoral Symphony. See Beethoven, Ludwig van:
Sixth Symphony
perceptual symbols, 70, 111n21
Peroratio (Schluss), 295
Pete Kellys Blues (Jack Webb movie), 236
pitch relations, descriptions of
Aristoxenus, 11, 14, 63
Bali, 67 68, 71 72
basis in image schemata, 69
Greek, 14, 63, 67
Java, 67 68, 71 72

Kaluli (Papua New Guinea), 67 68, 116 17

Suy (of the Amazon basin), 68
pitch relations, musical notation and, 67, 72
pitch-event, 101 5, 107, 110, 112
Plato, 10, 311
Plotinus, 311
Pollard-Gott, Lucy, 43n33
Porter, James, 217
Powers, Richard, 218n34
Propositio (Antrag), 295
Proslambanomenos, 6, 10
Proust, Marcel, 16
A la recherche du temps perdu, 34, 81 82,
325 27, 334
aesthetic theory, 328
Prout, Ebenezer, 289 91, 305
Ptolemy, 15
Pythagoras of Samos, 7 9, 12, 311n62
Pythagorean music theory, 7 8, 10, 12 14, 16
Quinault, Philippe, 120
Quinn, Naomi, 109 10
Rama Indians (Nicaragua)
snakebite cure, 115 16
Rameau, Jean-Philippe, 17, 119, 285, 330, 333
analysis of recitative by Lully, 120 24
theory of music, 121 24, 132, 148
theory of music as a cognitive theory, 124 26
ratio (applied to music), 8 10, 13, 15
epimeric, 13, 15
epimoric, 13, 15
multiple, 13, 15
Ratner, Leonard G., 168 69, 179 80, 200
Redman, Don, 208, 213 15
Reicha, Antoine, 301 2, 305
Rti, Rudolph
theory of motive, 26 27
rhetor ic, 139n74, 225, 228, 292 93, 303
and music, 296 98
musical, 288 89, 324
and musical for m, 298 99
See also cross-domain mapping: between
rhetor ic and music
rhetor ical gures, 138 39, 295 96, 299
Rich, Fred, 208, 213 15
Richardson, Samuel, 291
Richter, Jean Paul, 303
Riemann, Hugo, 305
character ization of tonality, 76
denition of musical motive, 305
theory of meter and rhythm, 320
Riepel, Joseph, 148, 152, 292 94, 299
Robertson, Jeanne, 217
Rodgers, Richard, 224

i nde x
Rooney, Mickey (Joe Yule, Jr.), 212
Rosch, Eleanor, 16, 32 33, 41
Rousseau, Jean-Jacques
analysis of recitative by Lully, 122
on the or igins of language, 284 85
Saint-Sans, Camille
Le Carnaval des animaux, 71
Satz, 48n40, 293
Schachter, Carl, viii
Schank, Roger S., 112
Schelling, Friedrich von, 288, 302 3
Schenker, Heinrich, 17, 119 20, 331, 333
analysis of Schuberts waltz Op. 77 No. 1,
128 30
theory of hierarchy, 317 18
theory of music, 127 30
theory of music as a cognitive theory, 130 32
theory of musical motive, 26 27
Schiller, Friedrich, 302
Schlegel, August Wilhelm, 303
Schleier macher, Friedrich, 302
Schluss (Peroratio), 295
Schmckung (Decoratio), 295
Schoenberg, Arnold, 242, 333
character ization of tonality, 76
theory of musical motive, 25 27, 30, 34, 42,
57 59, 139, 154
Schll, Adolf, 273
Schopenhauer, Arthur, 328
Schubert, Franz, 243, 332
Der Wanderer (Op. 4/1, D. 489), 28, 29
Trockne Blumen, from Die schne Mllerin
(D. 795), 257 64, 265, 332
Valses nobles, Op. 77 No. 1, 128 29, 130
Wanderer Fantasie (Op. 15), 27 28, 29
Schumann, Robert, 265, 272 73, 279
Der frohe Wandersmann (Op. 77 No. 1), 276
Dichterliebe (Op. 48), 265
Im Rhein from Dichterliebe, 26672, 278, 286
In der Fremde (Op. 39 No. 1), 273 79, 281,
286
Liederjahr of 1840, 265, 284
Schutt, Arthur, 234
Sectio canonis (anonymous treatise), 13, 15
Seeger, Charles, 221
semantics, musical, 17 18, 138, 199 200, 329
musical syntax and, 54, 199 200, 299, 329
semiotics, musical. See semantics, musical
Short, Bobby, 237 39
Signifyin(g), 18, 224, 229, 240 41
Skolsky, Sidney, 235
Sleepless in Seattle (Nora Ephron movie), 239n78
sonata for m, 152 53
Spaun, Joseph von, 243, 285

Stgemann, Friedrich August von, 246

Stgemann, Hedwig von, 246
Sulzer, Johann Georg, 296 98, 302 4, 313n67
Suy (of the Amazon basin). See pitch relations,
descriptions of
Swann, Charles (from Prousts A la recherche),
3 5, 16, 18 19, 61, 81, 325 28, 333
theory of music, 332
Sweetser, Eve E., 225 27, 242
syllogismus. See cadence: complete
symbols, 72, 103, 106. See also perceptual symbols
synaesthesia, 89n38
syntax, musical, 17 18, 52, 54, 75, 95, 132 33,
137 38, 140, 199 200
in Beethovens String Quartet Op. 18 No. 1,
189, 191, 196, 198
in Beethovens String Quartet Op. 18 No. 6,
172 73, 176
dynamic approach to, 329
formal, 143, 167, 199, 329
harmonic, 124 25, 130, 146, 153 54,
198 200, 329, 331
melodic, 140, 154, 159, 167 68, 331
in Mozarts String Quartet K. 465, 154, 159,
16768
multiple levels of, 198
musical categor ies and, 51 52, 54, 61 62, 329
and the Musikalisches Wrfelspiel, 151 54
Rameaus theory of music and, 121 22,
12425
and rhetor ic, 54n48
Schenkers theory of music and, 127 28,
130 31
and song texts, 254 55, 284 86
thematic, 198, 200
Talmy, Leonard, 69
tetrachords, 6
tetraktys of the decad, 9
text painting, 17, 63 65, 73 74, 77, 82 84,
87 88, 95
theme, musical
conceptual models and, 47 48
theories, 5, 15 17, 62, 98 99, 102 3, 108 9,
119, 125, 131, 133, 317, 332 33
ad hoc, 132
cultural, 114 17
extended, 132
function and structure, 106 7, 111, 113 14
implicit, 116 17
scientic, 113 14, 117
of very young children, 117 19
See also conceptual models: and theor ies
theories of meter and rhythm, 307 10,
31923

359

360

i nde x
theories of music, 5, 15, 121 24, 127 30, 132,
148, 323 24, 332, 334
as cognitive theory, 119, 124 26, 130 34
of early India, 13
Greek, 6 16
Jeff s development of, 101 2, 104 7
and syntax, 121 22, 124 25, 127 28, 130 31
See also Aristoxenian music theory;
Pythagorean music theory
theories of musical for m, 300 306
theories of musical motive, 25 30, 34, 42, 57 59,
139, 154, 289 90, 304 5
theories of tonality, 74 76, 83, 316 18
Thompson, William (Big Bill), 223
Thuille, Ludwig. See Louis, Rudolf, and Ludwig
Thuille
Ton, 316 17. See also Momigny, Jrme-Joseph
de: theory of tonality
tonality
character izations of, 76
as metaphysical, 317
See also theories of tonality
topics, musical, 168, 174, 176 77, 184, 198, 200,
329
topography. See conceptual integration networks
(CINs)
Tovey, Donald Francis, 90 94, 327, 331, 333
Trockne Blumen. See Klein, Bernhard; Mller,
Wilhelm; Schubert, Franz
Tunisia, music of, 222
Turner, Mark, 70, 78
on rhetor ical gures, 138 39
Twinkle, Twinkle Little Star (folk song), 100
Jeff s work with, 99 108

University of Berlin, 302

University of Halle, 301
University of Jena, 302
Urlinie, 127, 130 31
Ursatz, 127 31, 318, 321
Verwandlung, 128
Vinteuils sonata (from Prousts A la recherche),
3 4, 16, 18 19, 325, 327 28, 332 33
Wagner, Richard
compositional strategy, 49 51, 58
Leidensmotiv from Tristan und Isolde, 49 52, 54,
57 59, 62, 138
Sehnsuchtsmotiv from Tristan und Isolde, 49
Tristan und Isolde, 23 24, 24, 4958
Webb, Chick (William Henry Webb), 209,
21315
Webb, Jack, 236
Weber, Gottfr ied, 319 20
Welker, Robert L., 43n33
Wert, Giaches de
Tirsi mor ir volea, 84, 8586, 87 89
Wieck, Clara, 273
Wiederlegung (Confutatio), 295
Will, Richard, 90
Wittgenstein, Ludwig, 30
Wolzogen, Hans von, 49
work of music. See musical ontology
Young, Lester, 211, 213 15
Zemlinsky, Alexander von, 25
Zierde (Decoratio), 295