ON SCIENTIFIC DISCOVERY

BOSTON STUDIES IN THE PHILOSOPHY OF SCIENCE

EDITED BY ROBERT S. COHEN AND MARX W. WARTOFSKY

VOLUME 34
ON SCIENTIFIC
DISCOVERY
The Erice Lectures 1977

Edited by

MIRKO DRAZEN GRMEK

Ecole des Hautes Etudes, Sorbonne

ROBERT S. COHEN
Boston University

GUIDO CIMINO
Istituto della Enciclopedia /taliana, Rome

D. REIDEL PUBLISHING COMPANY

DORDRECHT: HOLLAND I BOSTON: U.S.A.

LONDON: ENGLAND
Library of Congress Cataloging in Publication Data

International School for the History of Science, Erice, Italy, 1977.

On scientific discovery.

(Boston studies in the philosophy of science ; v. 34)

Revised papers from a course held at Erice, Italy, Feb. 16-22, 1977, and spon-
sored by the Majorana Centre, the Domus Galilaeana, and the Istituto della Enciclopedia
italiana.
Includes bibliographical references and index.
1. Science-Addresses, essays, lectures. 2. Science-Philosophy-Addresses,
essays, lectures. 3. Science-Methodology-Addresses, essays, lectures. I. Grmek,
Mirko Drden, II. Cohen, Robert Sonne. III. Cimino, Guido. IV. Ettore
Majorana International Centre for Scientific Culture. V. Domus Galilaeana, Pisa.
VI. Istituto della Enciclopedia italiana. VII. Title. VIII. Series.
Q174.B67 vol. 34 [QI71] SOls [500] 80-23792
ISBN-13: 978-90-277-1123-6 e-ISBN-I3: 978-94-010-1284-3
DOl: 10.1007/978-94-010-1284-3

Published by D. Reidel Publishing Company,

P.O. Box 17, 3300 AA Dordrecht, Holland.

Sold and distributed in the U.S.A. and Canada

by Kluwer Boston Inc.,
190 Old Derby Street, Hingham, MA 02043, U.S.A.

In all other countries, sold and distributed

by Kluwer Academic Publishers Group,
P.O. Box 322, 3300 AA Dordrecht, Holland.

D. Reidel Publishing Company is a member of the Kluwer Group.

This volume was prepared for publication with the editorial assistance of
Carolyn R. Fawcett.
Support for the translations from Italian was provided by the
Istituto della Enciclopedia Italiana.

All Rights Reserved

Copyright © 1981' by D. Reidel Publishing Company, Dordrecht, Holland
Softcover reprint of the hardcover 15t Edition 198 I
No part of the material protected by this copyright notice may be reproduced. or
utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any informational storage and
retrieval system, without written permission from the copyright owner
 T ABLE OF CONTENTS

PREFACE vii

INTRODUCTION 1

PART I. GENERAL PROBLEMS

MIRKO D. GRMEK / A Plea for Freeing the History of Scientific

Discoveries from Myth 9
GERARD RADNITZKY / Progress and Rationality in Research:
Science from the Viewpoint of Popperian Methodology 43
JOSEPH AGASSI / The Problems of Scientific Validation 103
JOHN D. NORTH / Science and Analogy 115
MARCELLO PERA / Inductive Method and Scientific Discovery 141
VITTORIO SOMENZI / Scientific Discovery from the Viewpoint
of Evolutionary Epistemology 167
DEREK DE SOLLA PRICE / The Analytical (Quantitative) Theory
of Science and Its Implications for the Nature of Scientific Dis-
covery 179
GABRIEL GOHAU / Difficulties Inherent in a Pedagogy of Dis-
covery in the Teaching of the Sciences 191
VINCENZO CAPPELLETTI / Discovery and Vocation 211

PART II. CASE STUDIES

JACQUES ROGER / Two Scientific Discoveries: Their Genesis and

Destiny 229
RICHARD TO ELLNER / Logical and Psychological Aspects of the
Discovery of the Circulation of the Blood 239
vi TABLE OF CONTENTS

LUIGI BELLONI / The Discovery of Duodenal Ancylostoma and

of Its Pathogenic Power 261
SALVO D'AGOSTINO / Weber and Maxwell on the Discovery of
the Velocity of Light in Nineteenth Century Electrodynamics 281
HOWARD E. GRUBER / Cognitive Psychology, Scientific Crea-
tivity, and the Case Study Method 295

BIOGRAPHICAL NOTES 323

NAME INDEX 327

 PREFACE

The 1977 lectures of the International School for the History of Science at
Erice in Sicily were devoted to that vexing but inexorable problem, the
nature of scientific discovery. With all that has been written, by scientists
themselves, by historians and philosophers and social theorists, by psycholo-
gists and psychiatrists, by logicians and novelists, the problem remains elusive.
Happily we are able to bring the penetrating lectures from Erice that summer
to a wider audience in this volume of theoretical investigations and detailed
case studies.
The ancient and lovely town of Erice in Northwest Sicily, 750 m above the
sea, was famous throughout the Mediterranean for its temple of the goddess
of nature, Venus Erycina, said to have been built by Daedalus. As philosophers
and historians of the natural sciences, we hope that the stimulating atmo-
sphere of Erice will to some extent be transmitted by these pages.

We are especially grateful to that generous and humane physician and historian
of science, Dr. Vincenzo Cappelletti, himself a creative scientist, for his
collaboration in bringing this work to completion. We admire his intelligent
devotion to fostering creative interaction between scientists and historians
of science as Director of the School of History of Science within the great
Ettore Majorana Centre for Scientific Culture at Erice, as well as for his
imaginative leadership of the Istituto della Encic10pedia Italiana.
Our thanks, too, to Professor Mirko Grmek for his cooperative work as
Co-Director of the School and in the preparation of this volume; and especially
to Dr. Guido Cimino for his extraordinary and energetic competence at all
stages of the organization and planning of the School and of this book.

Center for the Philosophy and ROBERT S. COHEN

History of Science MARX W. WARTOFSKY
Boston University
August 1980

vii
 INTRODUCTION

Every two years the International School for the History of Science, located
at the Ettore Majorana Center in Erice, Italy, invites historians, epistemologists
and research scientists from all over the world for meetings devoted to discus-
sion of some of the significant problems currently facing the scientific com-
munity and the world of culture. Furthermore, the nearly sixty other schools
of the Majorana Center contribute an indirect but effective indication of the
lines along which scientists and historians can work together.
The purpose of the course of the International School for the History of
Science held at Erice, 16-22 February 1977, sponsored by the Majorana
Center, the Domus Galilaeana and the Istituto della Enciclopedia Italiana,
was to produce a critical examination of the logical, psychological, cultural
and social aspects of scientific discovery. This book is the fIrst publication to
be issued as a result of one of the School's courses, and in presenting it to the
reader, we must emphasize that it is not the 'proceedings' of the course,
namely a faithful record of the lively reports, communications and discus-
sions held at the Ettore Majorana Center. Rather we preferred to ask the
speakers to prepare a new version of their papers which took into account the
subsequent debate and also the requirements of a monographic publication.
Thus, after two years the problems and positions expressed at Erice have
crystallized to produce an image of the course which, though perhaps less
lively than the event itself, is more rigorous in its presentation.
This book sets out to ascertain the contemporary state in a particular
fIeld of the historiography of science and to make an original contribution
to topics bordering on both the philosophy of science and the history of
science. The contributors and the editors are pleased to have the volume
published under the auspices of the Boston Studies in the Philosophy of
Science.
For a long time epistemologists have considered scientific theories to be
parts or examples of a body of knowledge which is free of contradictions
in its structure and based on 'foundations', on propositions which cannot be
reduced to others. A more recent, different and reciprocal way of consider-
ing science is that of an expanding body of knowledge which increases the
domain of the known at the expense of the unknown. The analysis of scientifIc
1
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 1-6.
Copyright © 1980 by D. Reidel Publishing Company.
2 INTRODUCTION

discovery, as the instrument of the growth of knowledge, has taken on ever-

increasing importance. Interest in the genesis and methodology of discovery
by now has taken its place, with the same relevance and value, alongside of
the traditional orientation towards axiomatic reconstructions, in the his-
toriography and philosophy of scientific thought.
The expression 'scientific discovery', however, requires clarification: it is
a broad denotans with an imprecise connotatum. First of all, is discovery a
fact or an idea? This question, of crucial importance, was raised more than a
century ago by Claude Bernard, a great discoverer, in his bookAn Introduc-
tion to the Study of Experimental Medicine:
A fact is nothing in itself, it has value only through the idea connected with it or through
the proof it supplies. We have said elsewhere that, when one calls a new fact a discovery,
the fact itself is not the discovery, but rather the new idea derived from it; in the same
way, when a fact proves anything, the fact does not itself give the proof, but only the
rational relation which it establishes between the phenomenon and its cause. This rela-
tion is the scientific truth which we must now discuss further. 1

In 1877, Hermann von Helmholtz, like Claude Bernard, insisted in his

essay on Das Denken in der Medicin - at a time when the controversy over
J. R. Mayer's priority in discovering the principle of the conservation of energy
was still fresh in everyone's memory - on some particular features of those
'discovered ideas', 'gefundene Ideen' (Bernard's 'new ideas') which, if proven
important, are then called 'discoveries.'
The first discovery of a new law, is the discovery of a similarity which has hitherto been
concealed in the course of natural processes. It is a manifestation of that which our fore-
fathers in a serious sense described as 'wit' ... It is something which cannot be forced,
and which cannot be acquired by any known method . . . . The proof that the ideas
formed do not merely scrape together superficial resemblances, but are produced by a
quick glance into the connection of the whole, can only be acquired when these ideas are
completely developed - that is, for a newly discovered natural law, only by its agree-
ment with facts. 2

Helmholtz went on to say that a printer's case of type contains the entire
wisdom of the world. A discovery must be formulated as a rational tran-
scription, but the transcription must be organic, and it must be in agreement
with empirical data.
Thus the rational character of scientific discovery had already been under-
stood and affirmed during the last century. A new fact, a new 'property of
Nature', as Mach would say, became known to others through the proposi-
tions by which the discoverer describes, dermes and explains the factual. Even
from the discoverer's viewpoint, however, what is discovered is what he him-
 INTRODUCTION 3

self thinks and states about an observation or the result of an experiment.

Sometimes the statement is bare, a simple protocol, but the passage from per-
ceiving to affirming is always fundamental to discovering. This explains how
it can be that there have been discoveries for which the affirmation, in the
form of an hypothesis, clearly preceded observation and which in turn was
not the source of an unforeseen novelty, but the confIrmation of an assumed
one. This would furthermore explain the possibility of purely theoretical and
formal discoveries in that field of logic and mathematics which is rightfully
a part of the realm of scientific knowledge.
A discovery is not a fact, but the statement of a fact. This is, in summary,
one principal conclusion to be drawn from all the contributions gathered in
this volume. A discovery is a new statement, characterized by a novelty,
where novelty is understood as what cannot be deduced, but must be con-
structed within the sphere of logic, or observed in the course of experimental
science, whether physical or biological. 'Knowing' can be 'discovering' if it is
able to unfold itself and grow, and in growing renew its own structure. With
respect to new knowledge, old knowledge may be a cognitive or epistemo-
logical obstacle, almost a meta-obstacle, if it stands in the way of observation
or construction; or in other words, if it prevents the establishment of experi-
mental or constructive concepts, of structures or forms which are different
from those already known. Discovery is knowledge in growth, which thereby
extends its domain, or attributes a different order to itself. Any body of
knowledge which considers itself complete cannot accept the specific moment
of understanding that discovering is: an unfolding, a disclosure, and thereby
a renewing of thought.
Discovery, like intuition - in other words like the constructive act taken
in its presentation of the goal it will eventually reach - and like observation,
is a cognitive tension between participating and affirming. The subject perceiv-
ing a physical event, or representing, constructing an ideal form, is connected
with the object while constituting its very defmition. To express the perceived
event or the constructed form, is to begin to defme them, to insert them into
a universe of premises and relationships. Through the perceptive or construc-
tive unity of the two terms, the object enables the subject to fabricate the
science of a real, existing, or at least coherent world; but the subject, in turn,
enables the object to be placed in that universal framework of which science
is, or can be, the representation. Since object and subject, fact and idea, be-
come united in the act of discovery, history, as well as the philosophy of
discovery, will follow different paths in order to state and expound the terms
of the problem confronting them: on the part of the object, its constitutive
4 INTRODUCTION

structure and the occasion of its offering itself, or at least showing itself, to
the observer; on the part of the subject, personal and social events, psycho-
logical and sociological factors which have been and still are the interpretative
elements for historians and epistemologists. Yet eventually, discoverer and
discovery must feed into a cognitive process, and anything can contribute to
the growth of knowledge, which is all that truly matters. Rational objectivity
is what counts when vicissitudes come to an end, vicissitudes which at times,
have the savor of adventure.
That discovery is the statement of a fact has already been said. We can
now add that this statement must be comprehensible, as expressed in an
accepted and codified language and logic. Discovery therefore becomes a test-
ing ground for scientific knowledge, no matter whether in a positive sense, in
the form of verification, or in a negative sense, in the form of refutation.
However, innovation, the essence of discovery, can never be reduced to more
validation or confutation. Frequently, therefore, existing language does not
suffice. The discoverer thus coins a new terminology, even a new axiomatic
system: he codifies a language that his discovery will justify. Just as it is the
creation of new knowledge, discovery also implies the creation of a new
language, words, definitions and descriptions which had not previously existed.
If science is expanding knowledge, renewing itself on the periphery as well
as at the heart of its own domain, and if innovation is identified with scientific
discovery, then the history of discoveries will be the essential part of the his-
toriography of science and their logical and ontological analysis will be the
core of epistemology. Cognitive growth must be recognized in order to con-
tinue. How far have we come? Only history can provide the answer. How is
our knowledge structured, on which axioms does it depend, what inferences
does it use? The answers lie within the philosophy of science. But there also
exists a pedagogy of science and it too fmds its principal subject-matter and
instruments in scientific discovery in order to advance towards new and open
knowledge. To eliminate the possibility of the above-mentioned meta-obstacles
from the spirit of research is the greatest contribution that can be made to
the progress of knowledge.
In the papers we have brought together in this volume; scientific discovery
is studied not only as an original acquisition, a new body of knowledge
analyzed in its connection with the entire body of constituted knowledge,
but also and above all as a creative act, a mental process by means of which
one goes beyond the known.
If we take scientific discovery into consideration in so complex a manner,
three different approaches arise, overlap, and to a certain extent contradict
 INTRODUCTION 5

one another: logical, psychological, and sociological analysis. The papers

which follow provide evidence of the necessarily incomplete and complemen-
tary nature ofthese three approaches.
Specialists in three great disciplines have been invited to write on thiS
subject, and no attempt has been made to attenuate either their divergencies
or their conflicts. It is precisely this diversity of viewpoints, and not their
homogeneity, which we thought useful for the reader's orientation and for
furthering research in this field.
The first part of the book deals with themes of a general nature. It starts
with an attempt to expose mercilessly some common illusions in the histori-
cal presentation of scientific discoveries. The contributions of two Popperian
philosophers analyze the progress and rationality of scientific thought and
face the problems of validation. In contrast, two other papers, while staying
within the limits of philosophical enquiry and steering clear of psychologism,
set out to demonstrate the pragmatic value of analogic reasoning and to save
induction. The psychological aspect is emphasized in the papers on evolutionary
epistemology and the relationship between discovery and vocation. The
sociological approach is illustrated by a contribution which is outstanding for
the quantitative method it applies and for the significance of its results. Yet
another paper is concerned with the pedagogical application of the historical
and epistemological analyses of scientific discoveries. It recounts a personal
experience that may perhaps open the way to new perspectives in the teach-
ing of science at all levels.
As might have been expected, psychological examination of the scientist
and study of the social implications of his work have been gathered into the
case studies that make up the second part of the volume. Among the cases
presented are the discovery of the spermatozoon and the follicle of viviparous
species mistaken for the egg; the first formulation of the theory of the circu-
lation of the blood; the discovery of the duodenal hookworm; the origin of
some fundamental concepts connected with the velocity of electromagnetic
waves; and the explanation of biological evolution through a process of
natural selection. The essay on Darwin is certainly of considerable theoretical
interest, carrying as it does a number of considerations of a psychological
nature which throw new light on certain aspects of scientific creativity.
Even if well anchored within its specific individual and historical reality,
each of the selected examples goes beyond the single case and serves as a
model and reference point for interpretations of a general nature (which is
why it was included in the volume). Our purpose, however, is not only to
illustrate general theses, but also to enable or even force confrontations
6 INTRODUCTION

between the hypotheses of the epistemologist and the reconstructions of the

historian. The lack of very precise or even genuinely accurate historical
analyses of individual scientific discoveries is one of the reasons why the
history of science has not yet been able to become truly the 'laboratory of
epistemology' .
The chapters of this book are intended to give a clear epistemological
significance to scientific discovery and to identify it as a significant measure
of the growth of knowledge.

M.D.GRMEK
R. S. COHEN
G. CIMINO

NOTES

1 Claude Bernard, An Introduction to the Study of Experimental Medicine, trans. by

H. C. Greene (New York, Macmillan, 1927), p. 53.
2 Hermann von Helmholtz, On Thought in Medicine, trans. by E. Atkinson, Baltimore
(Johns Hopkins Press, 1938), pp. 21-22.
 PART I

GENERAL PROBLEMS
 MIRKO D. GRMEK

A PLEA FOR FREEING THE HISTORY OF SCIENTIFIC

DISCOVERIES FROM MYTH

In the course of the historical reconstruction of a scientific discovery, a

process occurs similar to what Stendhal, who used this image to study the
hazards of love, called 'crystallization'.
In the salt-mines of Hallein, near Salzburg [writes Stendhal (1822)] the miners throw
a bough, stripped of its leaves by winter, into the depths of a disused working of the
mines; two or three months later they find it entirely covered with glittering crystals
by the brine which moistens the bough and then subsides and leaves it dry. The tiniest
twigs, no bigger than a tomtit's claw, are spangled with an inimite number of shim-
mering glistening crystals. The original bough is no longer recognizable .... I call crys-
tallization that process of the mind which discovers fresh perfections in its beloved
at every turn of events. (On Love, trans. H.B.V., New York, Liveright, 1927, p. 359
andp.6.)

Our beloved, for us, historians of science, is the whole of res gestae et
scriptae of past scholars. We embellish these in our fashion and we contrive
to give them a lustre which is in harmony with our general ideas on the nature
of knowledge and on the way it is acquired. So, the dry branch of 'historical
facts' is, by a process of progressive theorization, enveloped little by little in a
thick layer of seductive 'myths'.
Let us take for granted that it is utopian to wish to produce, in accordance
with Ranke's precept (or, to keep within the more limited field of the history
of science, with Sarton's positivist recommendations), a history claiming to be
the absolutely objective mirror ofwhat really happened. Let us take for granted
that historical reconstruction carries out selections, establishes links and pro-
poses 'explanations' which necessitate a philosophical stand, indeed an involve-
ment of an ideological nature. But this recognition of a certain lack of objec-
tivity in historical research should urge us not to relativism and an attitude of
resignation, but to a careful ~nd systematic analysis of the factors which inter-
vene in 'crystallization', to the exposure of the processes which condition our
search for truth. It seems inevitable to us that we demand nowadays that the his-
torianof science submit himself to an epistemological self-examination. In short,
at the moment when he becomes fully conscious of the impossibility of total
objectivity in his discourse, the historian should feel more than ever the necessity
to elucidate, as much as possible, the modalities of his deviations therefrom.
9
M. D. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 9-42.
Copyright © 1980 by D. Reidel Publishing Company.
10 MIRKO D. GRMEK

Certain methodological presuppositions appear so inherent in scientific

thought that they cannot be questioned without jeopardizing the rationality
of discourse. But there do exist key ideas as well that are less sacrosanct.
Historians of science often draw their inspiration from convictions whose
validity is not certain: I propose to call these convictions methodological
myths. These 'myths' have historical roots and fulftll an important psycho-
logical and social role. It is a question, fundamentally, of beliefs and not of
proven assertions, nor of axioms indispensable for all subsequent research.
These 'myths' are rationalized justifications of our desires. I therefore refer
to them also as methodological illusions.
Their great strength is derived from the fact that we are indeed dealing
here with 'truths', but alas, with only partial truths. These myths always hold
an element of truth but never the whole truth. The myth consists most often,
to put it concisely, in the extension of an idea beyond its range of validity
by transforming a partially sound rule of limited application into an illusory
general rule.
The historical reconstruction of scientific discovery is dependent on a
particularly rich and tenacious 'mythology'. Here are a few examples of these
governing ideas which, diverse as they are in character and in their epistemo-
logical level, are all both widely accepted and of doubtful validity. The follow-
ing list is not systematic and has no claim to exhaustiveness. Its sole aim is to
draw the attention of historians of science to the existence of the hidden
traps strewn in the path of the historical rediscovery of scientific discoveries.

1. THE MYTH OF THE EPISTEMOLOGICAL ADEQUACY OF

'ANATOMIZING' ACQUIRED SCIENTIFIC KNOWLEDGE
AND THE PATTERNS OF ITS CONQUEST

Many epistemologists are interested exclusively in the analysis of the struc-

ture of scientific knowledge and put little store by the study of the historical
coming into place of each structural element of present knowledge. Most
neo-positivists belong to this category: they conjure away the problem of the
actual genesis of the network of scientific propositions by deliberately sub-
stituting its formal logical construction for its actual historical becoming.
It is quite significant that the philosopher Scheffler, who teaches the logic
of scientific investigation at Harvard University, entitled his main work
Anatomy of Inquiry. Indeed anatomy, a morphological and static science par
excellence, is concerned with the state of structures and not with the process
of their formation. Undoubtedly, the analogy with the anatomy is inescapable
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 11

to characterize the preoccupations and working methods of certain philo-

sophers of science.
However, nowadays one hears more and more the voice of those who are
interested not only in Constituted Science, nor in what seems immutable in
the processes of scientific research, but also and especially in Science-in-the-
making, in the genesis of so-called scientific structures. By widening the field
of epistemological reflection, they complete 'anatomical' analysis with a kind
of 'embryology' of scientific knowledge.
In this last case it is relatively easy to insist on the necessity of recourse
to historical method. But I feel one cannot altogether do without it even in
an epistemological study which aims at probing into the subtle details of
establiShed science . The analogy aetween the morphology of organized bodies
and the state of scientific knowledge appears unsatisfactory in this connec-
tion: the structural relationships of scientific thought cannot be represented
adequately nor studied in the full richness of their content, if their manifold
historical articulations are not taken into account.
As for those epistemologists concerned with the genesis of scientific
knowledge, one can aim the following criticism at them: they often offer
explanations of the process of scientific discovery when its different stages
are insufficiently known and described. In short - keeping within the limits
of Scheffler's analogy - it appears that a 'causal embryology' is made with-
out a previous grounding in the 'descriptive embryology' of the whole range
of scientific matters.
In another paper I have argued that the philosophical theories concealing
the development of science have a scientific tenor only in so far as they can
be compared with the data of the history of science. This is a fundamental
requirement with which philosophers must comply. I would now like to
emphasize that it puts historians of science in a delicate position, for it assigns
a task to them which is still far from accomplished and raises questions to
which one must beware of replying hastily or in an over-simplified or com-
placent manner. That brings us to the next section.

2. THE ILLUSION CONCERNING THE ADEQUACY OF OUR PRESENT

HISTORICAL KNOWLEDGE FOR RECONSTRUCTING THE PROCESS
OF SCIENTIFIC DISCOVERY OR FOR VALIDATING HYPOTHESES
RELATING TO THIS ENDEAVOR

One is nowadays aware, on the one hand, of an extraordinary proliferation of

publications of a general nature on the creativity and the intrinsic mechanism
12 MIRKO D. GRMEK

of discovery processes and, on the other hand, of an appalling dearth of solidly

documented historical studies of concrete discoveries. This dearth is not due
to a shortage of writings of a biographical nature on the researchers and their
scientific works - in absolute numbers there are actually too many. But they
untiringly repeat the same anecdotes and are satisfied with a superficial docu-
mentation which is not sifted with critical appraisal, and is dressed up to
comply with the demands of a new kind of hagiography. One of the main
weaknesses of present-day publications on scientific discoveries is that the
level of abstraction and generalization is too high, thus creating false problems
and masking certain essential aspects.
The fact is that epistemologists have nowadays at their disposal only a few
precise and detailed descriptions of historical events relating to important
discoveries. That is why general theories on scientific creativity do not inspire
the experienced scholar with confidence and can proliferate with impunity
for the indulgence of philosophical speculation.
L. Pearce Williams notes that epistemologists often base their theories
concerning the genesis of scientific knowledge either - as Kuhn does, for
example - on what scientists do (without, however, providing the irrefutable
historical proof that they really act this way), or - as in Popper's case - on
what they ought to do (with an historical documentation which is too inter-
pretative and restricted). One would rightly wish to base these epistemological
systems on the history of science but 'it cannot bear such a load at this time'
(L. P. Williams, in Lakatos and Musgrave, 1970).
Williams' critical remark has been accepted by Kuhn, but rejected by
Popper, who places logic outside and·above the bounds of history. Certainly,
it is reasonable enough to admit that logic has nothing to learn from history.
But this is not the problem. It lies in the possibility, or rather the great likeli-
hood, that the structure of science and even more the process of its growth
do not belong entirely to the field of traditional lOgic.
Certain attempts at a historical revision of 'classic cases' (let me quote as
an example, in the one field of the life sciences, the work of W. Pagel on
Harvey, that of H. E. Gruber on Darwin and that of F. L. Holmes and my
own on Bernard) show that a new light can be thrown on subjects one might
have thought to be exhausted. In the history of scientific discovery, all or
almost all is yet to be done over again. Historians sensitive to the problems
of modern epistemology must systematically reexamine the past course of
science. Alas, we shall be disappointed too often: historical sources at our
disposal rarely measure up to our wishes. It would be illusory to draw a
supposedly decisive epistemological lesson from what we can know with
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 13

certainty about the discoveries of Aristotle or Galen, even Copernicus or

Vesalius. We must therefore concentrate on a few instances privileged by the
nature and abundance of the surviving documentation. But as we expose the
shortcomings of our historical sources, would it not be possible to improve
our methods of collecting and preserving present-day evidence of scientific
discovery, thus to prepare and render more fruitful the work of future his-
torians?

3. THE MYTH OF A PERFECT AGREEMENT BETWEEN A RATIONAL

RECONSTRUCTION AND THE EXPERIENCE OF DISCOVERY

A. Deformation of Reality in the Researcher's Report

The fundamental sources on which any historical reconstruction of a scien-
tific discovery must draw are documents originating from researchers directly
involved in the different stages of its realization. These sources, whatever
their nature (pUblications, correspondence, note-books, laboratory records,
recorded statements, recollections of conversations bya third party, etc.), can
provide reliable information about the workings of someone's mind and his
conscious motivations at a given moment or, at least, express irrevocably
what that person wished to record for himself or convey to his contempo-
raries or to posterity.
We are here in touch with first-rate 'historical facts' which we must do our
utmost to garner, the contents of which, with no exception, must be taken
into account. But these documents do not in themselves convey the picture
of the process which concerns the historian of science. He has to interpret
them, build an account which via the sum-total of available historical traces,
offers a rational reconstruction of that historical order the existence of which
we take for granted in what I called 'Ie vecu de la decouverte' ('the experience
of discovery'), that is in what really happened in the course of the actualiza-
tion of a significant growth of scientific knowledge (see Grmek, 1973).
In this reconstruction, one runs a great risk of going right off the track in
always accepting literally what the protagonists or witnesses of an event have
to say. Testimony, direct or indirect, is a kind of raw datum which must be
used beyond its immediate and superficial significance. A confrontation of
different sources is called for to avoid the traps of accounts which are too
subjective, made in good faith and yet often false because of their partiality.
An illusion both widespread and deeply rooted for reasons which are
easily understandable, consists in the belief that perfect reconstructions of
14 MIRKO D. GRMEK

the experience of discovery can be found implicitly in the writings of the

researchers concerned. Is not every individual, if not the best judge, at least
the one most acquainted with his personal history, his own life-experience?
This widespread belief and popular opinion appears highly questionable to
me. With regard to events which have a heavy emotional charge and important
social implications for the person in question - which is certainly the case of
most scientific discoveries - an autobiographical account is certainly valuable
but cannot be accepted as definitive.
Scientific works swarm with surveys of the history of the research and
discoveries reported therein. Their authors often present in them the order of
their reasoning and of their experiments, acknowledging the influence exerted
on them by the knowledge of their predecessors' works and opinions, etc.
However, such information has no really historical purpose. Here we find not
history as such, but elements for a future history. Addressing the scientific
community to inform it about the state of their researches, scientists provide
certain details of their personal history so as to explain better the content of
their discoveries, not to analyze the actual process of their attainment.
A few great scholars have published, as an afterthought, autobiographical
texts reconstructing the progress of their reasoning and even containing, as a
by-product, reflections on scientific creativity. To name a few at random (still
taking examples only from the life sciences): Claude Bernard, Hermann von
Helmholtz, Otto Loewi, Charles Nicolle, Karl von Frisch, Walter B. Cannon,
Albert Szent-Gy6rgyi, Hans Selye, John Eccles, James D. Watson, Jacques
Monod, etc. Their contribution to the history of science and epistemology
is invaluable. No one would dare to deny the importance of their testimony
on the 'facts' and their analysis of historical connections regarding their own
self. Nevertheless, historians cannot be content simply with transcribing
these individual presentations and incorporating them in the more general
framework of treatises of the history of science.
In order to be aware of the difference of method and perspective which
separates the autobiographical account from the discourse of the historian of
science, it is useful to compare the writings of Watson and Olby who retrace,
each in his own fashion, the path leading to the discovery of the double helix
of DNA.
With the help of detailed studies of Claude Bernard's research (in particular
that concerning the glycogenetic function of the liver, the so-called piqure
sucree, the vasoconstrictor nerves and the mechanisms of intoxication and
drug-induced reactions) I have tried to show the points of disagreement which
exist between the experience of a discovery and its verbal or written presenta-
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 15

tion. The discoverer's account is an abridgement of the experience he has

lived; yet it involves not only a simplification through omission of the unim-
portant and secondary, but also a rationalizing readjustment, whereby the
actual sequence of events is transmuted in favor of logical rigor and co-
herence. A winding path becomes a straight-lined road, with sudden precise
changes of direction.
Galileo maintains that he invented the telescope as a result of precise
optical reasonings. Now, as Geymonat has clearly shown, in 1609, at the time
of his invention, he did not have the theoretical knowledge necessary for the
logical argument which he claims in his belated reminiscences.
Analyzing Loewi's account of his discovery of 'Vagusstoff', Bacq points
out that this reconstruction of events is too elegant and in some degree in-
evitably false, for Loewi had forgotten that at the moment of his discovery,
his technique of electrical stimulation of isolated frog heart was not yet able
to separate completely the inhibitory cardiac nerves (vagus fibers) from the
sympathetic fibers.
The same internal impossibility can be found in Claude Bernard's famous
account of his discovery of glycosuria provoked by lesion in a well-defined
area of the central nervous system. He quotes as a theoretical starting-point
for his decisive experiment an analogy with certain features of the neural
influence on saliva secretion which were then, in truth, totally unknown to
him.
We should like to stress the interest presented by historical investigations
into the progressive transformations of personal experience into a message
addressed to others, and of the Original information into a fairy-tale intended
to amaze the scientist himself as well as his peers and the general public. A
meticulous comparison of documents bearing different dates but coming
from the same person and recounting the same event, as well as psychophysio-
logical studies on memory and sociopsychological ones on evidence, will
perhaps allow certain constant features to emerge in the distortion of histori-
cal reality made a posteriori by the protagonists themselves.
In my work on Claude Bernard's experimental reasoning, I took a few
tentative steps in this direction. I have been able to establish, for example,
that in the lived experience of his work, the researcher wishes to convince
himself, while in his report, he wishes to convince others. As a result, the real
starting point of the discovery, often too fragile, finds itself reduced more
and more to a secondary argument, to an unimportant and even irksome
entity. Moreover, the most common error of the autobiographical report is
chronological inversion in the sequence of events or the concatenation of ideas.
16 MIRKO D. GRMEK

Be it said in passing that sometimes one comes up against what I have

called the 'illusion of narrative's time dimension'. What is recounted no
longer takes place in real time since the end conditions the beginning. A dis-
covery is achieved in a real area of time in total ignorance of the future, while
the researcher's report and the historian's discourse are made on an artificial
time-fabric which resembles real time in its structuration, but which differs
therefrom in that it has no becoming, that is, it forms a closed whole.

B. The Defonnation of Reality by the Historian or the Philosopher of Science

This deformation may be voluntary or involuntary; in a large number of cases

it is betwixt and between. The most typical is the distortion which takes
place without a conscious decision to cheat, but nevertheless with the intent
of proving, or rather illustrating by an historical example, a preconceived
general idea.
Abel Rey has shown in a polemic against Duhem's conventionalism how
much a particular philosophical attitude leads the latter to separate artificially
the presentation of scientific 'truth' from the real processes of his discovery.
For an observer who places himself, however, a little above the present
epistemological fray, it is evident that the logical reconstruction of the way in
which a science (for example classical mechanics, quantum physics, or mole-
cular biology) is constituted, and further still the way in which a new paradigm
has asserted itself in a defined field of scientific research, does not tally, and
does not in fact aim at tallying, with absolute objectivity, with the actual his-
torical process. On the occasion of another seminar, I have tried to prove that
each historical example can be harnessed to any reasonable epistemological
position whatever, provided selection, with a given slant, of the historical
documentation is made, and certain details are declared unimportant.
The reasons for the distortion with which we deal here, are most often
of an ideological nature. It springs from an idea anticipating the historical
inquiry. This is not at all surprising. The 'patterns of expectation' are aheady
at work at the level of the organization of perception and their interference is
all the more to be feared as one climbs the ladder of theorization.
The advent of awareness, of a prise de conscience vis-a-vis the myth of the
absolute objectivity of historical discourse, should not give rise to a facile
and baneful attitude of historical relativism and the triumph of scepticism.
Criticism is the only healthy reaction. Distortions do exist but they can be
detected by critical analysis. The neutrality of the historian and the philosopher
is never absolute in practice, a fact which does not, however, exclude the
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 17

possibility of acquiring, little by little, an increasingly objective knowledge.

Uke scientific research (in the narrow sense of the term), historical inquiry is
susceptible to progressive objectivation, to an asymptotical approximation to
a reality independent of the researcher's mind.

4. THE MYTH OF AN IMPERSONAL HISTORICAL DEVELOPMENT

OF 'SCIENTIFIC THOUGHT'

Under this heading I wish to denounce both (a) the pseudoproblem of the
conflict between 'internal' history and history referred to as 'external', and
(b) the illusion resulting from the confusion between the personal history of
the scientist and the general history of the progress of scientific knowledge.
Without doubt, a dual aspect exists of the iter mentis in veritatem: on the
one hand the actual itinerary of ideas in the 'discoverer's mind', and on the
other, the historical development of 'scientific thought' considered in abstracto,
that is to say without taking into account the particularities of those individ-
uals who are unavoidably its mainstay.
Experience proves the possibility of a historic discourse on the transforma-
tions of scientific theories which disregards considerations of a biographical
or psychological nature on those who have in effect created that world of
ideas. Is such a discourse legitimate? Although we are convinced that thought
is exclusively the product of concrete and individual nervous systems, we
acknowledge the pragmatic value of studying an 'internal dynamics' of
scientific ideas. However, to overcome the paradox of a thought without a
head to think it, it appears necessary to call upon the help of an intermediary
notion: that of 'scientific community'.
I therefore define 'scientific thought' as a structured aggregate of known
facts and propositions of a certain type which are accepted or debated by a
group restricted in historical time and in social space. In this way, discourse
on the general development of scientific thought relates to a historical reality
and not to an idealistic schema.
Evidently, if the 'internalist' approach is conceived, justified and realized
with the help of the notion of 'scientific community', it concerns, strictly
speaking, not the 'general' development of scientific ideas, but only
numerous sectors of it. The success at the monograph level of such a histori-
cal presentation will provide proof of the reality of a community of scholars
within well-dermed spatio-temporal co-ordinates, with due regard to the
problems they study. 'Internalist' scientific historiography requires a com-
plementary inquiry: the historical examination of the community concerned,
18 MIRKO D. GRMEK

that is to sayan 'externalist' analysis which is essentially of a sociological

nature.
In their common wish to be mutually exclusive, approaches known as
'internal' and 'external' become myths. They are self-limiting, interdependent
and mutually complementary. Their conflict is nothing but a false problem,
born of exaggerated pretensions on both sides.
Even someone interested only in the history of a specific scientific dis-
covery, who would like to isolate it from the texture of general history, must
have an overall picture of the flux of scientific ideas, as well as take into
account economical, political and sociopsychological factors. But we must
never forget that this overall picture refers to a second-degree reality born of
a theoretical construction.
The danger lies in the errors of fact and interpretation which can occur if
this 'history of scientific thought' is projected into the consciousness of a
scientist qua individual. That, alas, happens too often.
To study properly the creative process of new ideas, it is not enough to
know what, at a given moment, was 'known' on the topic under considera-
tion, but one must try hard to state precisely what each scholar involved in
the matter knew or believed he knew. In my work on Claude Bernard's
experimental reasoning, I have wished to show clearly that it is necessary to
distinguish carefully between the two types of historical discourse: in a
synoptic table I have placed side by side a general reconstruction of ideas on
intoxication by carbon monoxide and a reconstruction of the particular
evolution of known facts on this subject in a researcher's consciousness.
These two reconstructions overlap only partially.
An additional task, therefore, devolves on the historian of the process of
scientific discovery: he must establish if, at critical moments, the 'discoverer'
has really known (and, if so, how and in what precise form) certain 'facts'
and certain 'explanations', prior to the discovery and well known to the scien-
tific community. To use the example quoted above, let us remember that
Claude Bernard discovered, on his own, the 'arterial' color of venous blood
poisoned by carbon monoxide. This observation was previously described by
other scientists, but Bernard was not aware of their work; that he achieved it
by his own personal effort greatly influenced his subsequent research.
Let us mention in this context the sometimes positive role of ignorance of
an idea by the discoverer of a better one: in such a case, the former idea was
an 'epistemological obstacle' for other researchers. Some scientific discoveries
and technical inventions have been made because their author did not realize
the 'impossible nature' ofhis endeavor.
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 19

5. THE MYTH OF THE UNITY OF THE DISCOVERER,

OF THE PLACE AND TIME OF THE DISCOVERY

Most important scientific discoveries have not been achieved by a single

person, and even less at a precise moment in such a person's life. However,
in many historical accounts we find in this connection variations of the old
myth of the birth of Pallas Athena, the embodiment of knowledge, springing
fully armed from the head of Zeus. This is a false image: discovery does not
spring in one leap, radiating beauty and strength, from the creator's head.
To defme unequivocally a scientific discovery requires very often a brutal
laceration of a complex spatio-temporal reality, the isolation of certain events
with a near-symbolical value which encompasses not just one point but a
whole field of action. That makes schematization inevitable, but this can
easily become abusive: for one passes almost imperceptibly from simplifica-
tion to historical illusion.
Asking the qu~stion 'Who discovered this (a substance, a relationship,
etc.) and at what time?' assumes an improbably simple train of events, a
particular historical situation the existence of which is sometimes impossible
to fmd at the end of an inquiry, and which one is not allowed to take straight
away for granted. Kuhn [1963] has shown the misleading nature of such a
question with regard to the discovery of oxygen. I reached the same conclu-
sion about the discoverer and the moment of discovery of the mechanism of
curarization. The desire to reply at all costs with a name and an exact date is,
in the cases quoted, tantamount to weakening historical reality in a way
which modifies and betrays its true meaning.
Let us examine, as an example, the following question: 'Who discovered
glycogen and when?' On close inspection we notice that this question, of
the kind often found in scientific text-books (and to which all good encyclo-
pedias give a reply as crisp as it is problematic), is precise only in appearance.
What is the meaning of 'to discover glycogen'? Does it mean to be aware
of its existence, or to give it a name, or to isolate it (and, in this case, how
purely), or to know its chemical composition? Claude Bernard was the first
to have an inkling of this substance; following an original interpretation of
certain experiments, he assumed its existence (1855); he called it matiere
glycogene, then simply glycogene, thus defming its principal functional
property; he also foresaw the possibility ofits extraction from the liver. How-
ever, Claude Bernard was not the first to isolate glycogen. He was preceded,
in July 1856, by a German student, Victor Hensen, who achieved this feat
thanks to the advice of his professor, Scherer, inspired in tum by Bernard's
20 MIRKO D. GRMEK

publications. Was this simply a practical application of Claude Bernard's

'theoretical discovery'? To give credit to Hensen's work, one must acknowl-
edge that Bernard's hypothesis was still quite uncertain, that the procedure
employed by the young German scholar was original and that, in away, he
helped Bernard to perfect his own methods of isolating glycogen. Only in
March 1857 did Claude Bernard obtain a relatively pure glycogen. Pfluger
showed by repeating Hensen's experiments, that the substance obtained by
the latter in 1856 contained more proteins than polymerized carbohydrates.
If it is acceptable to look upon Bernard as 'the' discoverer of glycogen, it is
only on condition that we do not forget that he succeeded in this research
program because he was immersed in the efforts of a scientific community.
But it is illusory to try to assign a precise date to what was, in fact, quite a
long process, made up of many stages.
Replies to questions of the kind described above are 'true' only with
regard to a certain level of historical explanation. Paradoxically, their 'veracity'
can be inverted when this level is changed.
For example, it is said that Claude Bernard discovered hepatic glycogenesis.
Taken literally, such an assertion is false: Bernard was even aggressively
opposed to contemporary scientists who believed in the existence of 'hepatic
glycogenesis' (in the present sense of the term, namely the storing in the
liver, in the form of glycogen, of sugar coming from food-ingestion). His
'glycogenesis' was glucidic synthesis derived from other chemical substances
('glyconeogenesis' for present-day physiologists). Louis Figuier, an opponent
of Bernard's opinions on this topic, was often right in the details, but it is
nevertheless fair to attribute to Bernard almost all the glory of the discovery
of the hepatic metabolism of sugars. The statement quoted at the start of this
paragraph is both false and true: it becomes valid in a meaning which corre-
sponds to a higher level of abstraction in the presentation of the history of
ideas.

6. THE POSITIVIST MYTH OF THE STRAIGHT ROAD TO TRUTH

Neither the meanderings of individual thought, nor the advances of scientific

knowledge within a community, proceed by successive approximations,
always in the right direction, towards truth. The path to discovery is a wind-
ing one.
Positivist historiography has accustomed us to representations of the
development of scientific knowledge which are like stairways rising trium-
phantly towards the temple of present-day science, and made up of steps each
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 21

of which represents a 'positive' scientific advance, a partial but definitive

truth. This myth still predominates today in most 'historical introductions'
found at the beginning of chapters in text-books of contemporary science.
We have here a Manichean illusion where good and evil, truth and error, are
clearly separated. This triumphant didactic vision of the history of science
played an important part in the social assertion of scientism in the last
century, but it does not stand up to a thorough examination of historic
reality.
The ascent towards truth is not linear, and error is not necessarily (though
often it is the case) a negative element, that is, an obstacle on the road of
scientific research. Some 'errors' are particularly prolific, more functional,
in a given historical situation, than opinions recognized as more 'true' in the
judgement of posterity.
Thus historical investigations carried out recently or still going on amply
demonstrate tha~ the influence of magic and alchemy on the birth of the new
sciences in the seventeenth century was not altogether a negative one as
positivist historians of the last century would have us believe. Even men like
Francis Bacon and Newton are indebted to the hermetic tradition and to
ways of thinking condemned without appeal by a certain brand of rationalism.
We must beware of confusing the scholar's judgement on an opinion con-
sidered as a scientific image claiming to give the truthful explanation of the
world with the historian's judgement on this same opinion considered as a
master-idea influencing the subsequent development of scientific research.
For a historian blinded by an uncompromising positivism, the fact that
Serveto included in a theological work his description of the passage of blood
through the lungs can only be an accident with no special significance, or at
best, the testimony of that man's anguish and the murky depths of his
chimerical speculations. Yet, looking at it closely, in this case as in so many
others, the light of discovery shone forth owing to a piece of confused and
analogical reasoning wherein mysticism figured as an indispensable com-
ponent.
A particularly striking example of the mixture of 'bad method' and 'good
results' is found in Kepler) work. A modern scholar has gone so far as to say
that this genius would appear much greater to us if three quarters of his
writings could be excised. The 'normative' aspect of this opinion renders it
odious and dangerous. Kepler, Newton and Harvey did not make their dis-
coveries in spite of their astrological, alchemical or Aristotelian prejudices but
through them and in part because of them. About the work of Kepler, C. de
Buzon notes very aptly that
22 MIRKO D. GRMEK

a history of science which limits itself to tracing the skyline of the world of geniuses
prevents itself from understanding the mechanism of scientific research; inversely, the
genuinely philosophical task is not to locate truth and to brand error, but to understand
intellectual processes which are now alien to us.

The discovery of the circulation of the blood has often been cited as a
typical example of the rigorous and successful application either of Baconian
induction, or of the hypothetico-deductive method. These approved interpre-
tations, dear to the particular inclination of every pOSitivist-historian, are not
altogether consonant with what the documents say about the real genesis of
William Harvey's discovery. We may not disregard the influence of the mysti-
cal symbolism of circles, Gnostic philosophy and Peripatetic metaphysics. In
physics even, Harvey drew his inspiration as much from Aristotle as from
GaWeo's new mechanics. When he gives the heart the distinguished title of
sun of the microcosm, Harvey uses something more than a poetic image.
Physiological cardiocentrism is actually historically bound up with astronomi-
cal heliocentrism. The most decisive step in Harvey's reasoning was the
transformation of the Aristotelian analogy between the irrigation of a garden
and the distribution of blood in the body, a transformation carried out with
the help of another 'model' likewise taken from his reading of Aristotle,
namely the place of the sun in the circular motion of terrestrial and atmos-
pheric waters.
I will not dwell on the role of chance, for I have recently published a study
specially devoted to this topic (see Grmek, 1976).
To illustrate how the rational reconstruction of discoveries serves as a
model for a positivist history, let us remember the introduction of vaccina-
tion against smallpox by Jenner (Razzell (1977) has shown certain of its
'mythical' aspects), the first vaccinations against rabies by Pasteur, and
Bernard's experiments on glucosuria. The events leading to these achieve-
ments are not in reality as simply and logically linked up as we are asked to
believe.
A particularly instructive example is provided by the history of the dis-
covery of insulin. It is admirably clear in the justification of the Nobel prize
award by the chairman of the selection committee, and in the autobiographi-
cal accounts of the recipients, but becomes more and more confused as one
proceeds in the critical analysis of the facts. In the first place, it does not
seem permissible to eliminate from the historical report the contribution of
certain 'forerunners' (see Murray's article on experiments by Paulesco, not
unknown to, but wilfully deformed by the winning team). What is even more
interesting, from our point of view, is that one becomes aware (see especially
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 23

the excellent study of Pratt, 1954), that the official version of the way the
discovery was arrived at, does not stand up to an examination of the original
publications nor to a repetition of the experiments. The clue of Banting's
reasoning was simply an illusion: obstruction of the pancreatic ducts does not
give rise to a selective degeneration of the tissues producing trypsin. Contrary
to the working hypothesis he proposes, the extract obtained from a normal
pancreas contains a higher proportion of insulin than that yielded by a
'degenerate' one. The results of Banting and Best's experiments are at vari-
ance with the conclusions that they drew from them. We know today that the
success of the Toronto team is due to the precipitation of proteinic hormone
by 95 percent alcohol, which eliminates toxic ingredients and allows medical
use of the extract. The inventor of this process, the biochemist CoUip, was
put on the side, when the Nobel prize was awarded, to make room for the
head of the laboratory in which the discovery had been made.

7. THE MYTH OF CONTINUOUS EVOLUTION AND

THE MYTH OF PERMANENT REVOLUTION

Continuity and rupture co-exist and any historical venture based on an exclu-
sive mythicizing of one of these two complementary aspects would seem
doomed in advance.
If, in the personal history of a researcher, there are sudden moments of
enlightenment, comparable to Archimedes' Eureka, there is also, inevitably,
prior to them an incubation period, a slow, continuous process of intel-
lectual ripening. In the general history of the sciences, there are likewise
revolutionary reversals - historical falsifications in Popper's sense of the
expression, or paradigmatic reversals in Kuhn's terminology - but as the
latter so well emphasizes, that does not exclude the existence of a slow, not
really discontinuous accumulation of scientific knowledge in the period
separating two crises.
From this standpoint let us look at the history of the discovery of the
circulation of the blood. Baldini has examined Harvey's work in the light of
Popper's epistemology and has severely criticized the idea that his discovery
resulted from a progressive and gradual piecing-together of a mosaic of minor
discoveries, of tiny steps made by his 'forerunners' and by Harvey himself,
necessarily leading by their added momentum, to the new pattern of circu-
latory physiology. Baldini's analysis Seems to us convincing: Harvey's achieve-
ment is revolutionary and, in essence, cannot be explained by the summation
of a series of partial discoveries. It is nonetheless true that such a summation
24 MIRKO D. GRMEK

exists as well, and if it does not necessarily produce nor explain the decisive
discovery, it is nevertheless, for its production, an indispensable precondition.
The advent of Harvey's new physiology is unthinkable without the cumula-
tive contribution of Vesalius, Colombo, Cesalpino, Fabrizio d' Acquapendente
and several other researchers who breached Galen's system and allowed
Harvey to find himself confronted with a particular panorama of known facts.
The 'heroic' view of all scientific research idealizes historic reality. Without
doubt, a 'normal science' (in Kuhn's sense) exists, and we could not partici-
pate, in this connection, in the disillusioned sigh (,yes, alas!') of certain of
Popper's disciples. The existence and proper working of this 'normal science'
is a historical sine qua non precondition of 'extraordinary science'. The alter-
nation of 'puzzle-solving' periods and emergence of new scientific theories
seems to belong to the very nature of the process which produces the growth
of our scientific knowledge.
Certain epistemologists, for example Hanson and Kuhn, compare the act
of scientific discovery to the shift in the interpretation of an ambiguous
drawing (the famous rabbit-duck). According to this analogy, a scientific
theory would explain a group of facts exactly as a perceptual interpretation
organizes lines and spots into a Gestalt. We would thus pass from one inter-
pretation to another, both in the change of paradigm and in the setting-up of
a new apperception, by a total transformation determined by insight and
therefore of an essentially unforeseeable and radically discontinuous nature.
The reversal of the scientific paradigm has been compared to a kind of mystic
crisis.
Without wishing to go as far as the 'anarchical' excesses of Feyerabend,
who speaks of the 'incommensurability of theories' and the poetic determina-
tion of the choices made by scholars, I am ready to accept the existence, in
the progress of knowledge, of qualitative jumps, that is to say, changes which
are irreducible to the accumulated mass of variations of secondary importance.
But we do not wish to fall into the trap of the myth of a permanent revolu-
tion and deny the fundamental importance of slow transformations of scien-
tific theories, which come about by a progressive adaptation, not by a rever~l
of paradigms. The two aspects appear to me complementary.
To go back to the analogy between scientific discovery and the interpreta-
tion of the rabbit-duck drawing, I intend to consider it in a more dynamic
way than that in which it is usually described. Let us imagine that the draw-
ing in question is not a state but a becoming, that is to say that it is not static,
but mobile, in slow, continuous transformations. This can be realized, for
example, on the screen of an analogical computer. Let us imagine further that
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 25

this machine is programmed, so that the drawing presented to the observer

at the outset suggests beyond doubt the outline of a rabbit, and that the con-
tours and spots gradually change to present only at a determinate moment
the ambiguous drawing. It is then and only then that the reorganization of
apperception can take place. It can but is not bound to. The Gestalt switch
depends certainly on the observer's state of mind. To him will be attributed
the revolution in the interpretation of the whole, but it must not be forgotten
all the same that it was impossible without the preceding stage which gradually,
by a cumulative effect, gives the drawing its ambiguity. To bring my analogy
closer to reality, one must also imagine that the new apperception rever-
berates by a feedback on the objective characteristics of the drawing, rein-
forcing the shape of the duck. The Gestalt model of scientific discovery can
be made valid only by supposing that each of its parts, the observer's state
of mind and the drawing observed, are in a condition of flux and that their
link-up works in both directions.
In my opinion, the acquisition of knowledge bears the essential charac-
teristics of biological growth in general. The latter is not simply an increase
by progressive accumulation of elements; it is always a morphogenesis which
is brought about by the setting-up of a structure and by the alternation of
cumulative transformation (minor spells of imbalance with a continuous regu-
lating mechanism) and of restructurations (sudden breaks, accompanied by
the establishment of new levels of equilibrium).
Piaget and Lorenz, the former starting from the observation of children,
and the latter from an analysis of animal behavior, reached fairly similar con-
clusions relating to the existence of isomorphisms between the cognitive
processes and the biological somatic regulating events. According to genetic
epistemology, the intellectual development of the child passes through
progressive restructurations, through conquest of stages (the morphogenesis
of knowledge thus resembles embryogenesis). The psychologist has the im-
pression that progress is made through rapid mutations, but - says Piaget -
the fundamental transformation is slow; what is sudden, is not the process
of building-up but the coming of awareness, the eventual comprehension at
the moment of completion of the structuring of a stage.
Claude Bernard has given a description of the process of discovery: 'illumi-
nation' is only the sudden completion of a slow, 'underground' (the term
'subconscious' was not yet fashionable), maturation process, marked by
assimilation of new data, construction of explanatory patterns.
Modern mathematization, based largely on the notion of function and on
differential and integral calculus, favors an understanding of the continuous,
26 MIRKO D. GRMEK

evolutive aspect of each event. But a new mathematics of discontinuous

entities is in the process of emerging: Rene Thorn's 'theory of catastrophes'.
Born of topological considerations, this outline of a general theory of qualita-
tive models tries to formulate mathematical characterizations of 'catastrophes'
(that is, sudden reorganizations of states of equilibrium). A large number of
biological phenomena belong to this category and I think we must also in-
clude the processes of artistic and scientific creativity.

8. AN ILLUSION WHICH RESULTS FROM THE PROJECTION OF

INITIAL CIRCUMSTANCES IN A DEVELOPED SEQUENCE OF EVENTS

This illusion is well known thanks to the chicken-or-egg priority sophism. It

continues, however, doggedly to live on in a form which is relevant precisely
to experimental reasoning and the genesis of discovery: are 'facts' (or rather
the observation of the facts) anterior to the hypotheses or, on the contrary,
do hypotheses always precede 'facts'? In other words, does experimental
reasoning begin with observation or theory?
As in the case of the chicken-or-egg question, it is not difficult to show
that each of two sequential elements must by turn precede the other.
On the one hand, observation is never, in the actual circumstances of
scientific research, independent of a certain theoretical framework, of a net-
work of hypotheses formulated beforehand. I shall return to this with regard
to induction. Let us simply recall here that, according to Piaget's investiga-
tions, theoretical structuration, already in the child, modulates very strongly
the 'facts' observed (let us quote, as an example, the experiment with the
drawing of the line of level of a liquid in a tilted glass).
On the other hand, all scientific hypotheses are, in one way or another,
induced or influenced by previous empirical knowledge.
In short, we need theories to observe facts and we need observed facts to
produce theories. In actual scientific research, the dialectical interplay be-
tween theory and practice, between epistemic structures and empirical con-
tent, is already so far advanced that the question of the chronological primacy
of observation or hypothesis is improperly formulated and leads to a number
of paradoxes.
To follow another line of argument, it may well be that this interplay
of influences is not only at a very advanced state, but also more subtle and
complex than the simple alternation of the two above-mentioned elements,
an alternation suggested by the generalization of the hypothetico-deductive
method.
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 27

9. THE MYTH OF THE CLEAR-CUT AND ABSOLUTE ALTERNATION

OF THE OBSERVATION OF FACTS AND THE INVENTION OF
HYPOTHESES

According to a common opinion, quite widespread among scientists, experi-

mental reasoning passes through three stages which Claude Bernard had
represented by the following schema: 0 ~ H ~ E (0 = observation;H = hypo-
thesis; E = experiment). Observation gives rise to hypothesis, which is used in
the carrying out of an experiment offering in its turn a new observation
which confirms or invalidates the first hypothesis. Bernard's triad extends and
branches out into several chains of 'experimental reasoning', for each observa-
tion can give rise to hypotheses and each hypothesis can be the starting point
for new experiments. Strictly speaking, these chains are made up of two
constituent elements and not three, for observation and experiment differ
only in their position in the experimental cycle. Experiment is an observation
engineered to test the validity of a hypothesis: it is a terminal point with
regard to the latter, and also, by providing new 'facts' for the researcher's
consideration, a starting-point with regard to subsequent hypotheses. Observa-
tion and experiment are enshrined in the empirical tradition, but between
them hypothesis is thrown as a bridge: hypothesis, that rational element
through which - says Bernard - "the scholar's mind finds itself, as it were,
placed between two observations".
What seems essential in this schema is the on-and-off alternation of theory
and practice, of imagination and 'facts', of the intellectual variation of ideas
and their 'natural selection.'
What mental operations are represented by the two arrows of the triad?
The second signifies that starting from a hypothesis one arrives by logical
deduction at the construction of situations which can be subjected to experi-
mental control. It is the graphic expression of a methodology advocated by
men like Galileo and Newton.
The crucial problem is to explain the birth of hypothesis from observed
facts, that is to say, the significance which is attributed to the first arrow.
The graphic presentation of Bernard's schema could make us believe that
H follows 0 by a constraining local process, especially by the rules of in-
duction. Now; Bernard rightly criticizes the legitimacy of inference by
Baconian induction and affirms that hypothesis is, certainly, founded on a
knowledge of the 'facts' but does not necessarily result from it. Hypothesis is
not inferred from the facts but, according to Bernard, is produced by the
'sentiment' (that is to say by intuition), by a constructive mental activity
28 MIRKO D. GRMEK

which depends both on known facts and on an intuitive forecast of facts

to be known. The researcher is - says Bernard - nature's superintendent.
Similar opinions have been expressed by the chemist Liebig and by the
philosophers Jevons and Meyerson. We shall return to them concerning various
myths about the origin of scientific hypothesis. For the moment what con-
cerns us foremost, is to demonstrate the illusory character of the usual
interpretation of Bernard's schema which makes observation and theorizing
alternate in a clear-cut and absolute way.
In a study of the origins of the notion of living fiber and of the cellular
theory, I paraphrased Virchow's dictum by stating: omnis theoria ex theoria.
In my opinion there is no simple induction from facts to theories and there is
moreover no creation of hypotheses ex nihilo. One always passes not from
observations to theory but from one theory to another. This passage, how-
ever, takes place under the influence of observations, or through indirectly
becoming aware of new 'facts'. Contrary to Popper, I believe that this in-
fluence is not only a falsifying one, that is to sayan eliminator of hypotheses
at variance with reality, but also in some way a positive theorization leading
towards ideas which have a relatively high chance of survival.
My interpretation of the concatenation characteristic of experimental
reasoning differs from Claude Bernard's schema: it is a sequence oftheoreti-
cal views in a state of flux which produce experiments and are modified
following their results. Here is a symbolical representation of it:
Eo E1 E2
.j, t.j, t.j,
-+Ho -+H1 -+H2-+
(H0 = initial theoretical framework; Eo = initial observation, not necessarily
caused by theoretical considerations but acting only through its integration
with Ho; HI initial theory modified by a new hypothesis; E1 experiment
initiated to ascertain the accuracy of the logical consequences of HI).

10. THE MYTH OF BACONIAN INDUCTION

It is rare today to fmd philosophers of science who wish to rank induction

as the principal procedure in the logic of discovery. Although there are
still scientists and logicians who endeavor to save this myth, numerous lucid
critiques, from Hume's to Popper's, allow me to restrict my own apprecia-
tion to a few rather summary considerations. I shall not therefore attend
to the problem of the logical foundation of induction, nor to its proba-
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 29

bilistic aspects, nor to paradoxes of confirmation by the inductive method.

One first difficulty of complete inference from empirical data is already
due to the impossibility of obtaining them in a raw state and of systematizing
them according to their properties alone. What one notices, through observa-
tion, as a 'fact', does not depend on properties of objects (or, in other words,
sense data) and on some mental categories with a general validity alone, but
also on the theoretical framework prior to apperception and to the integra-
tion of the 'fact' in our world-vision. Hanson has brilliantly demonstrated
that 'facts' are in mutual dependence with the formalism which expresses
them and are impregnated with theoretical prejudices. A two-way link exists
between the results of observation and language. The idea we have of an
instrument and the theory which gives birth to an experiment influence the
way facts are noticed and described. Looking down an optical tube one sees
something else, another 'fact', as one takes such an instrument for a micro-
scope, a telescope or a kaleidoscope.
To illustrate the role of prejudice in biological discoveries which, at first
sight, are not far above the level of simply noticing the 'facts', let us recall
the history of Bathybius haeckeli, of which Rupke (1976) has recently
published a good analysis. In the attainment and disclosure of that mistaken
discovery, in 1868, of an extremely primitive animal (it was actually an
inorganic formation), psychological and even epistemological factors inter-
vened which biased the description and interpretation of relatively simple
phenomena. Those factors notably included the enthusiastic acceptance of a
certain form of Darwin's theory and the need to bring new evidence into a
scientific debate with a strong political resonance.
Mendel's discovery is sometimes quoted as a typical example of inductive
empiricism. Yet, thoroughgoing analyses of Mendel's work carry the convic-
tion that the logical unfolding of his demonstration should not be confused
with the actual progress of his discovery. The fundamental hypothesiS on the
segregation and statistical distribution of characteristics has not been drawn,
by simple inductive inference, from experimental results. According to the
biologist and statistician R. A. Fisher, the results obtained by Mendel are 'too
good' (a proportion of 3.01:1 noted on 8000 peas), that is to say rather
improbable without the regulating interference of an underlying hypothesis,
of an anticipatory knowledge of the 3:1 ratio. That is not all: Mendel's
original results in experiments on the hybridization of descendants with
parents are both too good from the point of view of his own fmal theory and
false according to our present knowledge (it would have been necessary to
obtain 37%, instead of 33%, of recessive-character representatives). Therefore
30 MIRKO D. GRMEK

Mendel was expecting certain numerical data; he was testing his hypothesis
and did not infer it from observed facts.
However, certain new historical and mathematical precisions, due especially
to Van der Waerden and Welling, seem to temper Fisher's conclusion accord-
ing to which Mendel first formulated his theory and then performed his
experiments. Theorizing and experimenting went hand in hand: the essence
of Mendel's discovery was acquired during the carrying out of the experi-
ments.
There are certainly no logical rules for acceding with unassailable certainty
to a so-called inductive assertion. But must one go as far as Popper and say
that induction is an illusion in the strongest sense of this term, that is to say,
a mental process not only without logical foundation but also, as it were,
without existence? According to Popper, induction exists only in appearance,
thanks to the selection of hypotheses obtained by a completely different
process. An assertion aspiring to general validity would then be in no way
influenced, in its genesis, by observed facts. It would secondarily be brought
in harmony with these, and not be inferred from them.
For a historian of science, convinced as he is that reality ranges wider than
the excessive rigidity and extreme simplifications of logical patterns, it
appears rash to reject induction in so drastic a fashion. While admitting the
non·validity, in pure logic, of inference from the particular to the general, we
are not obliged to deny the existence of a mental inductive process which
generates hypotheses, without, however, justifying these. If there is no induc-
tion in the logical sense of the word, it may well be that there is a kind of
psychological inference, an orientation of thought via knowledge of empirical
data.
The 'discovery' of the 'cell' by R. Hooke and the progress of micrographi-
cal studies in the seventeenth and eighteenth centuries did not lead directly
to the formulation of the cell theory. This is a historical example of the im-
potence of the inductive method, such as it was taught, precisely at the time
in question, by the brilliant English Chancellor.
Various theories on fibers, globules and 'molecules', understood as ele-
mentary carriers of life, did not give birth to the authentic cell theory before
Schleiden conveyed to Schwann certain data gathered from observation. I
interpret this historic case (for a good exposition onthis, see Florkin's, Klein's
and Canguilhem's (1965) publications) as an instance of subtle dialectic
between attempts at generalizing theorization and concrete observation.
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 31

1l. THE ILLUSION OF THE FUNDAMENTAL EPISTEMOLOGICAL

ROLE OF VERIFICATION

We catch a glimpse of this old illusion in the common parlance of scientists

who, even today, speak much more often of verifying than of falsifying
hypothetical propositions or descriptions of the observable.
The tendency to the preservation of the structures of acquired knowledge,
to the minimal alteration of these in the process of integrating new knowl-
edge therein, appears as a fundamental characteristic of our mentality. This is
an 'epistemological obstacle', in Bachelard's meaning of the term. One's
primary desire is to see knowledge increased, not overturned. Having stressed
that the underlYing theoretical framework influences observation and deter-
mines certain aspects of the 'facts', I should have specified that not only is
one obliged to see in relation to what one knows, but also that one tends to
see in keeping with what one knows.
In my study of Claude Bernard's toxicological experiments, I pointed out
a particular kind of epistemological obstacle which, for a researcher with
original and deep-rooted opinions, makes it very difficult, in a later moment,
to falsify and forsake what was, earlier, the essence of his own discoveries.
There is, without doubt, a preference of a psychological nature for verifi-
cation. However, we must not confuse psychology with logic: preference at
the one level in no way signifies superiority at the other.
The test to which a scientific assertion is submitted, either by examining
its coherence within a system, or by confronting its deductions with observ-
able data, must not prejudge the positive or negative character of the result.
This seems to indicate that, from the point of view of method, there is no
difference between verification and falsification. Indeed, a good many re-
searchers have believed, and some still do, in the myth of epistemological
symmetry between verification and falsification.
Such a myth does not stand up to logical analysis, not even to a few rather
superficial considerations. For assertions with a general validity, falsification
is obviously stronger by its consequences than verification: the first establishes
non-truth, while the second does not establish truth. Falsification appears
as definitive, verification is only provisional. Certain logicians prefer therefore
to speak of confirmation, a concept which allows a probabilistic relativization
of verification.
In short, falsification of a scientific theory seems to be a condemnation
without appeal, while verification would be but a precarious acquittal. By the
logic of falsification, one aims at a defmitive epistemological judgement; by
32 MIRKO D. GRMEK

the logic of confirmation, one tries to secure a process of progressive approxi-

mation to truth.
Karl Popper, who has spotlighted the epistemological superiority of falsi-
fication, upholds the thesis of the completely illusory nature of verification
(even in its refined form of relativized confirmation), which is, according to
him, nothing but a series of aborted attempts at falsification.

12. THE ILLUSION OF A PERFECT ASYMMETRY BETWEEN

FALSIFICATION AND VERIFICATION

Faced with the logicians' method of reasoning, we are forced to admit that
scientific theories are not verifiable in the absolute sense of the word. Histori-
cal examples of theories which have never been falsified evidently prove
nothing: we do not know their future. But a different demonstrative im-
portance would attach to the historical existence of theories which, after
being falsified, have later been recognized as 'true'. Such cases are known;
indeed they are quite numerous. So, to our surprise, historical inquiry seems
to show that, in fact, scientific theories are not falsifiable either in a definitive
way.
The logical explanation for this can be easily found: data provided by
observation can be improperly noted because of technical or psychological
errors and, most important, the reasoning process which leads to falsification
always presupposes tacit acceptance of a conceptual framework which may
be called in question after the event. Such an alteration of certain tacitly
accepted presuppositions leads to a disavowal offalsification: which explains,
on the one hand, the historical possibility of a rebirth of scientific theories
judged at a given moment as certainly false, and on the other, the impossi-
bility of experimentum crucis.
Let us quickly look at a historical example: choosing once again, to make
our analysis easier, the discovery of the circulation ofthe blood. Harvey was
at grips with serious difficulties and could not decide, for many long years,
to publish his new physiological system. Without a knowledge of the capil-
laries and pulmonary alveoli, it was impossible to explain properly the passage
of blood at tissue-level (from the arteries to the veins) and at lung-level (from
the veins to the arteries). Assuming the existence of porosities, invisible to
the naked eye, Harvey acknowledged implicitly the Galenists' right to believe
in the existence of invisible pores in the interventricular partition of the
heart: in other words, he weakened an important criticism of the ancient
system. To explain the venous return to the heart, one had to tum to non-
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 33

mechanical conceptions (the spontaneous return of blood to its natural place),

borrowed from an out-dated physiology. The origin of the motor power of
the heart remained an obscure problem. But what we must remember above
all in the history of Harvey's theory is that, in the context of the positive
knowledge of his time, it was 'falsified' by the attacks of Johann Vesling
and Jean Riolan the younger. For his part, the English physiologist disproved
the Galenic theory (especially by his quantitative considerations); all the
same, he was unable to reply to the serious objections of his opponents. On
a practical level, clinical failures of blood transfusion, that is of the first
therapeutic method based on the new theory, weighed heavy. In the purely
physiological field, the major difficulty lay in the following problem: how to
explain the difference between arterial blood and venous blood, if it is the
same blood which circulates continuously through all the vessels? Replying to
this, Harvey had to maintain that such a difference does not exist in the
body, and that it results from particular conditions in the observation of
blood issuing from injured vessels.
We know today that these refutations of Harvey's theory are illusory, but
we know too that that could not be demonstrated at the time and that
Harvey had to turn to very weak ad hoc hypotheses. In the eyes of many
seventeenth century doctors, his theory had been falsified.
Of course, such historical situations are not unknown to present-day up-
holders of falsificationism (see Antiseri's (1976) lucid and instructive analysis).
'Ingenuous falsificationism' is the term applied now to the belief in the
possibility of definitive refutation of a theory with empirical content. Popper
and his followers take a more subtle position, an attitude of 'methodological
falsification ism' , which recognizes the impossibility of condemnation without
appeal ('disproof'), and which skilfully uses relativized concepts, such as
'illusory refutation' and 'ad hoc hypothesis'. Each theory can be 'immunized'
against formal disproof by auxiliary hypotheses, but, curiously enough, as it
becomes more and more unassailable, it does not improve its chances of being
true. A scientific theory which accumulates saving hypotheses and presents
itself in a more and more impregnable form, becomes more and more suspect
and may even pass out of the field of science altogether.
As there is no sound method of distinguishing, in the heat of a scientific
debate and without historical perspective, an 'ad hoc hypothesis' from a
'justified complementary explanation', the asymmetry between verification
and falsification is, in practice, less important than it seems at first sight.
The historian who studies the process of scientific discovery is often struck
by the precarious character of the first 'good' hypotheses in the solution of a
34 MIRKO D. GRMEK

problem. They are often 'falsified' out of hand, yet their discoverers do not
abandon them, but cling to them and end up falsifying the first-stage 'refuta-
tions'. The best hypotheses often contradict one part of the experiment:
they call for a re-examination of the 'facts' which, at that point in history,
seemed established in definitive fashion.
The strength of a new theory does not therefore lie solely in its resistance
to attempts at refutation. There must be, particularly in the initial phase of a
discovery, factors which incline one in its favor and which are irreducible to
the Single-handed logic of falsification.
Why out of two (or several) non-falsified theories, or two (or several)
theories historically falsified and saved by auxiliary hypotheses, does one
decide in favor of one rather than the other? Neopositivists put forward the
criterion of probability, especially the degree of corroboration. Popper appeals
to the principle of maximization of empiric content. Serious authors have
spoken of factors of an aesthetic nature, of 'poetical', even 'mystical', pre-
ferences. I shall draw attention to a characteristic which could be important
for the decision scholars make (without believing, however, that it is the only
one to turn the scale): the 'perspectivity' of a theory.
To abandon a scientific theory and accept another is more often a matter
of methodological opportunity than of strict refutation. Formal logic inter-
venes generally after the event. It justifies more than it decides the way of
scientific research.

13. THE MYTH OF THE STRICTLY LOGICAL NATURE

OF SCIENTIFIC REASONING

Now, as a result of my last statement, we find ourselves faced with the partic-
ularly delicate problem of the relationship between the logic, psychology
and sociology of scientific discovery. I do not have time to analyze properly,
in this context, the three myths which consist in reducing the total reality of
'science in the making', or at least its essence, into just one of these three
approaches. Let us be satisfied with a quick overall view, hoping to be able to
deal with this subject more fully in the course of another seminar.
My account thus concludes with a summary and curt denunciation of
three illusions which, because of their methodological importance, should
have been placed at the head of a list of errors to be avoided, and which, by
their special epistemological nature, would have deserved the detail and
finesse of a.well-documented monograph.
The three myths in question have a common source: the dissociation of
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 35

the unitary triad made up of the three aspects of the acquisition of knowl-
edge, corresponding to the constitutive triad of the concept of man:

individual
species
/ "'- society
(Concerning the latter, see Edgar Morin's publications.) The distinction be-
tween the logical, the psychological and the social lies not in reality, but in
our means of intellectual analysis. It is a necessary and fruitful process of
dismantling, but it creates difficulties and exposes one to the risk of mythiciz-
ing, in an exclusive, imperialistic fashion, each of the components of the
global system. The three approaches are concurrent, mutually antagonistic
and complementary. They should not be impervious to intercommunication.
In the analysis of scientific discovery, two key-concepts are used: induc-
tion and intuition. These are in my opinion 'amphibian' concepts which,
because of their polyvalence, have a very great operational value and, at the
same time, are the despair of certain philosophers and psychologists who
would like to lock them in the field of a single investigatory discipline. Induc-
tion is a concept invented by logicians, and yet they have never succeeded in
reducing it, in a completely satisfying manner, to formal logic. Factors of a
psychological nature intervene in inductive inference. As for intuition, while
it is a concept belonging to the field of psychology, it cannot be successfully
studied if it does not take into consideration the categories and rules of logic.
Positivists with varying allegiances and neo-Baconians still fight tooth and
nail in favor of the old rationalist myth which reduces scientific investigation
to a practical application of logic. It is an undeniable fact, which we must not
forget in this debate, that a considerable number of scientific discoveries
result simply from the application of an impeccable logical reasoning to con-
crete problems. In the genesis of such discoveries the historian can easily
grasp the 'leading idea' and is bound to admire the precision of observation
(Leeuwenhoek, Redi, Fabre, etc .... ), or the persistence in the application
of certain clear, relatively simple hypotheses (Pasteur, Ehrlich, Ludwig, etc.).
When it is a matter of 'deductions', that is to say, applications of a theory
or an investigatory technique to cases not previously contemplated, rationalis-
tic explanation is triumphant. Ascertaining the origin of a completely new
idea is a more complex matter. It seems that to explain the genesis of certain
discoveries, and precisely the most original and fruitful ones, it is necessary to
refer to irrational processes.
It goes without saying that the term i"ational does not indicate anything
36 MIRKO D. GRMEK

'supernatural', magical or unaccountable in its principle. I only wish to con-

vey by this word that a particular creative element intervenes in the roots of
rationality itself and that scientific thought transcends the limitations of
classical logic.
The strongest condemnation of the myth of the sovereignty of logic has
been uttered by none other than certain logicians themselves, anxious to cir-
cumscribe their territory and abandon litigious fields to feel more comfortable
in their fortress. Does not this withdrawal go too far? It seems dangerous to
me to affIrm that intuition (therefore the actual genesis of discovery) eludes
a really scientific study for the reason that it is outside the framework of
logic in the narrow sense of the word (that is of present-day formalized
logic). Popper considers that "there is no such thing as a logical method of
having new ideas or of a logical reconstruction of this process", and that
every discovery contains 'an irrational element' or 'a creative intuition', in
Bergson's sense of these terms. The act of conceiving or inventing a scientific
theory, just like the birth of a musical theme, is not, by this reasoning,
reducible to logical analysis, but comes within the purview of empirical
psychology .
This opinion seems to me, on the one hand, well entrenched against any
formal attack, and on the other, very dangerous in as much as it seals off
certain paths of enquiry.
The logic of scientific discovery (if there is one, as I believe there is) can-
not be the sort of formal logic that is capable of dealing exclusively with
matters of justification or validation. In this sense, the title of Popper's main
work is a false promise.

14. THE MYTH OF THE STRICTLY IRRATIONAL NATURE

OF THE ORIGIN OF DISCOVERIES

This myth rests, on the one hand, on a strict and rigid rationalism, that is to
say, on the considerations of the logicians we have just mentioned, and, on
the other, on a romantic, irrational, sometimes even surrealistic idealism.
These last tendencies are in fashion again today.
A psychology of scientific discovery, which invokes in a quasi-mystical
manner the 'personal genius' of the researcher, must be judged with the
greatest severity. Even in its variants bearing semblances of scientificity (the
analogy, for instance, between the birth of a new idea and biological muta-
tion, which was proposed by C. Nicolle, or the quantum leap; speculations of
certain psychoanalysts and biographical studies with a psychological veneer),
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 37

such an approach suffers from a fundamental flaw, that of untestability.

Socrates' 'demon' is an amusing hypostasis of the actual experience of inspira-
tion, of sudden illumination, but is also the prototype of 'mythological'
explanations. One must, in this day and age, reject as illusory any explanation
presupposing the existence of 'supermen', of exceptional geniuses thinking
and acting in accordance with modalities differing qualitatively (and not just
quantitatively) from those of the common run of mortals.
Empirical psychology can certainly shed a precious light, which is indis-
pensable for understanding creative thought. Let us recall the results of two
classical methods, that of introspection (see J. Hadamard's (1954) excellent
work on the subject) and that of the observation of a subject under experi-
mental conditions (from the already old and yet still relevant works by
Claparede (1934), by Duncker (1926), by Maier, to more recent research by
Wertheimer, Bruner, Boring, Gruber and so many others). The 'problem-
solving' type of experimental analysis has proved profitable with regard to the
elucidation of certain elementary processes of reasoning, but we are still
unable to draw from it defmite, useful conclusions concerning artistic inven-
tion or scientific discovery. In the laboratory, creativity is limited in time and
applied to simple problems: which does not therefore imitate the actual con-
ditions of scientific research. A psychologist, who is concerned with scientific
creativity at the highest level, must turn towards historical documentation,
for at the present time obstacles of various kinds prevent psychological
experimentation on renowned scholars grappling with important scientific
problems.

15. THE MYTH OF THE SOCIOLOGICAL EXPLANATION

OF SCIENTIFIC DISCOVERIES

Far be it from me to wish to deny the impact of socio-economic factors on

the progress and vicissitudes of scientific thought, as much at the individual
level of each scholar as at that of the community giving rise to, and encourag-
ing, the solutions which we feel are needed; it appears nevertheless illusory to
seek, at all costs, in these factors the complete explanation of creative scien-
tific activity and its results. In the history of science, 'externalism' is a mirage,
and pushed to its limits, an untruth.
There does probably exist a link between Harvey's monarchism and his
ideas on the central role of the sun and the heart, between Magendie's revolu-
tionary education and his iconoclastic pragmatism, between Virchow's middle-
class liberalism and his cellular pathology, but the presence of political
38 MIRKO D. GRMEK

elements in a scientific work, and even in any scientific work, is something

different from the reduction of science to sociology. If it is true that scien-
tific opinions are most often in agreement with the political stands of scholars,
it does not follow that it is possible to deduce such opinions from ideology,
especially when we deal with original ideas. Still more illusory have been
attempts to explain the genesis of scientific theories through the direct in-
fluence of socio-economic factors. The achievements of men like Lomonosov
or Lobachevski are, certainly, influenced by the economy of Tsarist Russia,
but are not epiphenomena of it. It is, without doubt, very illuminating to
detect and follow up as far as possible the connections between the material
determinants of general historical development and the meanderings of scien-
tific thought, but provided one does not wilfully close one's eyes by halting
inquiries of a non-sociological nature.
The difficulty and the appeal of the problem of scientific discovery are
derived to a large extent from the fact that it is placed at the crossroads of
several diSciplines - in at least two ways. First, in spite of the diversification
of present-day branches of sciences and techniques, the fundamental problem
of the creative activity of the human intellect is everywhere the same. The act
of achievement of a scientific discovery or technical invention is analogous, if
not identical, to the act by which a work of art is' created. Secondly, the
position of this problem in the theory of knowledge seems to be so central
that it constitutes the Gordian knot of all possible approaches and that no
unidisciplinary method of inquiry can give, single-handed, results that are
entirely satisfying.

Translated by
MARGARET ROUSSEL

BIBLIOGRAPHY

Agassi, J.: 1973, 'Continuity and Discontinuity in the History of Science', J. Hist. Ideas
34, 609-626.
Antiseri, D.: 1976, 'Prova di una teo ria ed educazione al riconoscimento dell 'errore',
Med. nei secoli 13, 307-364.
Anzieu, D. and coil.: 1974,Psychana/yse du genie createur (Dunod, Paris).
Bachelard, G.: 1938, La formation de ['esprit scientifique (Vrin, Paris).
Bacq, Z. M.: 1974, Les transmissions chimiques de ['influx nerveux (Gauthier-Villars,
Paris).
Baldini, M.: 1974, Epistem%gia e storia della scienza (Citta di Vita [for Univ. di Siena),
Florence).
Bellone, E.: 1976, II mondo di carta (Mondadori, Milan).
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 39

Bernard, C.: 1865, Introduction a ['etude de la medecine experimentale (Bailliere,

Paris).
Beveridge, W. J. B.: 1957, The Art of Scientific Investigation, 3rd ed. (Norton, New
York).
Boirel, R.: 1972, L 'invention. 4th ed. (P.U.F., Paris).
Bruner, J. S. and Postmann, 1.: 1949, 'On the Perception of Incongruity: a Paradigm',
Journal of Personality 18,206-223.
Bunge, M.: 1967, Scientific Research (Springer, Berlin, New York).
Buzon, C. de: 1977, 'Remarques sur !'interpretation de l'oeuvre de Kepler', Arch. Intern.
Hist. Sci. 27, 72-81.
Canguilhem, G.: 1965, La connaissance de la vie. 2nd ed. (Vrin, Paris).
Cannon, W. B.: 1945, The Way of an Investigator (Norton, New York).
Carmichael, R. D.: 1930, The Logic of Discovery (Open Court, Chicago).
Caws, P. J.: 1969, 'The Structure of Discovery', Science 166, 1375-1380.
Claparede, E.: 1934, La genese de l'hypothese (Kundig, Geneva).
Duncker, K.: 1926, 'A Qualitative (Experimental and Theoretical) Study of Productive
Thinking (Solving of Comprehensible Problems)', J. Genetic Psych. 33, 642-708.
Enriques, F.: 1906, Problemi della scienza (Zanichelli, Bologna).
Feyerabend, P. K.: 1975 ,Against Method (NLB, London).
Fisher, R. A.: 1936, 'Has Mendel's Work Been Rediscovered?', Ann. Sci. 1, 115-137.
Florkin, M.: 1960, Naissance et deviation de la theorie cellula ire dans ['oeuvre de Th.
Schwann (Hermann, Paris).
Fourastie, J.: 1966, Les conditions de l'esprit scientifique (Gallimard, Paris).
Geymonat, 1.: 1962, Galileo Galilei. 2nd ed. (Einaudi, Turin).
Geymonat, 1.: 1970-1976, Storia del pensiero filosofico e scientifico. 7 vol. (Garzanti,
Milan).
Giere, R. N. and Westfall, R. S. (eds.): 1973, Foundation of Scientific Method: the
Nineteenth Century (Indiana Univ. Press, Bloomington).
Gingerich, O. (ed.): 1975, The Nature of Scientific Discovery: a Symposium Commemo-
rating . .. Copernicus (Smithsonian Inst. Press, Washington).
Granit, R.: 1972, 'Discovery and Understanding',Ann. Rev. Physiol. 34,1-12.
Grmek, M. D.: 1962, L'introduction de ['experience quantitative dans les sciences bio-
logiques (Palais de la Decouverte, Paris).
Grmek, M. D.: 1966, 'Examen critique de la genese d'une grande decouverte: la piqure
diaMtique de Claude Bernard', Clio Med. 1,341-350.
Grmek, M. D.: 1968, 'First Steps in Claude Bernard's Discovery of the Glycogenic Func-
tion of the Liver', J. Hist. BioI. 1,141-154.
Grmek, M. D.: 1970, 'La notion de fibre vivante chez les medecins de l'ecole iatrophysi-
que', Clio Med. 5,297-318.
Grmek, M. D.: 1973, Raisonnement experimental et recherches toxicologiques
chez Claude Bernard (Droz, Geneva); (ltal. ed.: Psic%gia ed epistemologia della
ricerca scientifica: C. Bernard, Ie sue ricerche tossic%giche. Episteme, Milan,
1976).
Grmek, M. D.: 1976, 'Le role du hasard dans la genese des decouvertes scientifiques',
Med. nei secoli 13,277-305.
Gruber, H. E.: 1974, Darwin on Man: a Psychological Study of Scientific Creativity
(Dutton, New York).
40 MIRKO D. GRMEK

Gruber, H. E., Terrell, G., and Wertheimer, M. (eds.): 1963, Contemporary Approaches
to Creative Thinking (Atherton Press, New York).
Hadamard, J.: 1954, An Essay on the Psychology of Invention in the Mathematical
Field. 2nd ed. (Dover, New York).
Hanson, N. R.: 1965, Patterns of Discovery (Cambridge Univ. Press, London).
Hanson, N. R.: 1965, 'The Idea of a Logic of Discovery', Dialogue 4,48-61.
Harding, S. G. (ed.): 1976, Can Theories be Refuted? (Reidel, Dordrecht).
Hesse, M. B.: 1974, The Structure of Scientific Inference (Macmillan, London).
Holmes, F. 1.: 1974, Claude Bernard and Animal Chemistry; the Emergence of a Scien-
tist (Harvard Univ. Press, Cambridge, Mass.).
Horton, M.: 1973, 'In Defence of Francis Bacon: a Criticism of the Critics of the Induc-
tive Method', Stud. Hist. Phil. Sci. 4,241-278.
Inhelder, B. and coli.: 1976, 'Des structures cognitives aux procedures de decouverte',
Ann. Psycho/' 44,23-60.
Ivanov, G. M.: 1969, 'K voprosu 0 ponyatii fakt v istoricheskoi nauke', Vop. istorii,
No. 2,73-87.
Jewkes, J., Sawers, D., and Stillerman, R.: 1969, The Sources of Invention. 2nd ed.
(Macmillan, Edinburgh).
a
Jung, C. G. and Kertlnyi, C.: 1953, Introduction I'essence de la mythofogie (Payot,
Paris).
Kedrov, B. M.: 1969 'Sur la psychologie de la creation scientifique', Organon, No.6,
49-68.
Koestler, A.: 1967, The Act of Creation (Laurel, New York).
Kordig, C. R.: 1971, The Justification of Scientific Change (Reidel, Dordrecht).
Kuhn, T. S.: 1962, 'Historical Structure of Scientific Discovery', Science 136,760-764.
Kuhn, T. S.:1963, The Structure of Scientific Revolutions (Univ. of Chicago Press,
Chicago).
Lakatos, I.: 1973, The Changing Logic of Scientific Discovery (London).
Lakatos, I. and Musgrave, A. (eds.): 1970, Criticism and the Growth of Knowledge
(Cambridge Uniy. Press, Cambridge).
Largeault, J.: 1976, 'Questions de logique et d'epistemologie', Rev. Synthese 97, 279-
310.
Loewi, 0.: 1965, The Workshop of Discoveries (Univ. of Kansas Press, Lawrence).
Lorenz, K.: 1973, Die Riickseite des Spiegefs (Piper, Munich).
Machamer, P. K.: 1973, 'Feyerabend and Galileo: The Interaction of Theories and the
Reinterpretation of Experience', Stud. Hist. Phil. Sci. 4, 1-46.
Medawar, P. B.: 1969, Induction and Intuition in Scientific Thought (Methuen, London).
Meyerson, E.: 1931, Du cheminement de fa pensee. 3 vols. (Alcan, Paris).
Mikulinski, S. R. and Yaroshevski, M. G. (eds.): 1971, Nauchnoe otkrytie i ego vospri-
yatie (Nauka, Moscow).
Monod, J. et al.: 1975, 'The Analysis of Scientific Method and the Logic of Discovery',
in H. A. Krebs and J. H. Shelley (eds_), The Creative Process in Science and Medicine
(Excerpta Medica, Amsterdam), pp. 3-28.
Morin, E.: 1973, Le paradigme perdu: fa nature humaine (Le Seuil, Paris).
Murray, I.: 1971, 'Paulesco and the Isolation ofInsulin', J. Hist. Med. 26,150-157.
Nagel, E.: 1961, The Structure of Science (Harcourt, Brace and World, New York).
Nicolle, C.: 1932, Biologie de l'invention (Alcan, Paris).
 FREEING SCIENTIFIC DISCOVERIES FROM MYTH 41

Olby, R.: 1974, The Path to the Double Helix (Macmillan, London).
Pagel, W.: 1967, William Harvey's Biological Ideas; Selected Aspects and Historical
Background (Karger, Basel).
Pagel, W.: 1976,New Light on William Harvey (Karger, Basel).
Pfliiger, E. F. W.: 1905, Das Glykogen und seine Beziehungen zur Zuckerkrankheit
(Hager, Bonn).
Piaget, J.: 1967, Biologie et connaissance (Gallimard, Paris).
Popper, K.: 1959, The Logic of Scientific Discovery (Basic Books, New York).
Popper, K.: 1963, Conjectures and Refutations (Routledge and Kegan Paul, London).
Popper, K.: 1972, Objective Knowledge; an Evolutionary Approach (Clarendon Press,
Oxford).
Porep, R.: 1971, 'Der Prioritatsstreit urn die Entdeckung des Glykogens zwischen Claude
Bernard und Victor Hensen', Med. Mschr. 25,314-321.
Pratt, J. H.: 1954, 'A Reappraisal of Researches Leading to the Discovery of Insulin',
/. Hist. Med. 9,281-89.
Razzell, P.: 1977, Edward Jenner's Cowpox Vaccine: the History ofa Medical Myth,
(Caliban Books, Newhaven).
Rey, A.: 1907, 'L'energetique et Ie mecanisme au point de vue des conditions de la
connaissance', Rev. phi/os. 46,495-517.
Rossi, P.: 1977, Immagini della scienza (Editori Riuniti, Rome).
Rupke, N. A.: 1976, 'Bathybius haeckelii and the Psychology of Scientific Discovery',
Stud. Hist. Philos. Sci. 7,53-62.
Russo, F.: 1974, 'Typologie du progres des connaissances scientifiques', Rev. Quest. Sci.
145,345-363 and 479-502.
Schaff, A.: 1970, Geschichte und Wahrheit (Europa, Vienna, Frankfurt, ZUrich).
Schaffner, K.: 1974, 'Logic of Discovery and Justification in Regulatory Genetics',
Stud. Hist. Philos. Sci. 4, 349-385.
Scheffler, I.: 1963, The Anatomy of Inquiry (Knopf, New York).
Schilpp, A. (ed.): 1974, The Philosophy of Karl Popper (Open Court, La Salle).
Selye, H.: 1964, From Dream to Discovery (McGraw-Hill, New York).
Stendhal: 1959, De I 'amour, ed. with an introduction and notes by H. Martineau (Garnier,
Paris).
Stent, G.: 1972, 'Prematurity and Uniqueness in Scientific Discovery', Scient. A mer.
227,84-93.
Szent-Gyorgyi, A.: 1962, 'On Scientific Creativity', Persp. Bioi. Med. 5, 173-178.
Taton, R.: 1955, Causalite et accidents de la decouverte scientifique (Masson, Paris).
Thorn, R.; 1974,Modeles mathematiques de la morphogenese (Union generale d'editions,
Paris).
Waerden, B. L. van der: 1968, 'Mendel's Experiments', Centaurus 12, 275-288.
Watson, J. D.: 1968, The Double Helix (Weidenfeld and Nicolson, London).
Weiling, F.: 1971, 'Neue Ergebnisse zur statistischen Vorgeschichte der Mendelschen
Versuche',Biometrics 27, 709-719.
Wertheimer, M.: 1945, Productive Thinking (Harper, New York).
Westcott, M. R.: 1968, Toward a Contemporary Psychology of Invention (Holt, Rine-
hart and Winston, New York).
Worden, F. G. (ed.): 1975, The Neurosciences: Paths of Discovery (MIT Press, Cam-
bridge, Mass.).
42 MIRKO D. GRMEK

Wright, G. H. von: 1957, The Logical Problem of Induction. 2nd ed. (Macmillan, New
York).
Yaroshevski, M. G.: 1971, Problemy nauchnogo tvorchestva v sovremennoi psikhologii
(Nauka, Moscow).
 GERARD RADNITZKY

PROGRESS AND RATIONALITY IN RESEARCH

(Science from the Viewpoint of Popperian Methodology)

PART I. THE IDEA OF A METHODOLOGY AND THE BACKGROUND

OF POPPERIAN METHODOLOGY.
THEORY APPRAISAL, METHODOLOGY APPRAISAL AND
IDEAL OF SCIENCE

o. ON THE CONCEPT OF METHODOLOGY: THEORY APPRAISAL

AND METHODOLOGY APPRAISAL

0.0. The Need for Methodology: Since Decision-making is an Ubiquitous

Moment of Research There Cannot Be a Methodology-free Research
In the process of research the researcher finds himself time and again con-
fronted with problems of decision making: to decide which of two alternative
research programs should be followed, whether it is worthwhile to conduct
a certain experiment, etc., etc. And just as often he is confronted with prob-
lems of appraisal, decision-making ex post so to speak: to decide whether a
certain explanation is adequate, to appraise the comparative achievements of
competing problem-solutions, i.e. to decide which is 'preferable', to decide
above all whether a proposed new theory constitutes progress over its rivals,
and so forth. Ex ante the researcher has to decide which course of action is
'rational', in the sense of purposive rationality (Zweckrationalitiit) - given
his interpretation of his current research situation (which of course may be
mistaken). This part of research activity resembles the Similarly risky business
of financial investment. Ex post he analyzes his past investment decisions,
thereby attempting also to estimate opportunity costs: whether time and
effort if invested in a rival theory program would with a certain likelihood
have yielded better results. Again a delicate problem.
If, in any sort of activity, decision-making is complicated, the need is
felt to systematize it. If this need is taken seriously, a special discipline will
develop to meet it. As regards research, 'methodology' is a suitable label
for such a discipline. Some dispute whether methodology is possible; but a
sceptical view on this issue already reflects a certain image of science, i.e. it
is itself the result of a certain methodology which operates clandestinely. If
43
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 43-102.
Copyright © 1980 by D. Reidel Publishing Company.
44 GERARD RADNITZKY

all research were like the unique acts of creativity in art, all the decision-
making involved might be governed by 'tacit knowledge' in the sense of not
being articulable. If so, a critique of decision-making in research would be like
art criticism, since there would be no general criteria, no statute law. Faute de
mieux, the critic would have to rely on his intuition, sensitivity, Finger-
spitzengejiihl. He would have to bring his personality into play. If on the
other hand all of research were like a routinized procedure of 'problem solv-
ing', the methodology could be an algorithm. Clearly either view is a totaliza-
tion - and patently false. Research obviously contains moments of both
types and above all moments which in various degrees approximate to each of
the above extreme types. Insofar as research or an important section of it is
a rational goal-directed activity, at least some part of the decision-making can
be elucidated by a praxiological study. (,Praxiology' is being used in Kotar-
birlski's sense: roughly as the theory of effective and efficient action.) If so -
and we submit that this is a correct assumption - methodology is a viable
project. Moreover, the researcher cannot avoid it since, whether he likes it or
not, part of his time as researcher is spent in this sort of decision-making:
every researcher is his own part-time methodologist. There cannot be any
methodology-free or methodology-neutral research. (As little as there can be
observation sentences or even communicable, hence formulable, perception
reports free from theoretical ingredients.) The suspicion that methodology as
such, i.e. as a diScipline, might claim to be able to prescribe to the researcher
what he should do is, as already hinted at above, based on a misunderstand-
ing. The existence of certain misguided paternalistic methodologies which
oversell themselves does not warrant such a generalization to the discipline as
a whole. The researcher who does not recognize the interdependence of
research and methodology will be a 'methodologicien malgre lui'. The meth-
odological criteria and gambits he uses in his research activity will remain
latent, and so long as they remain latent they cannot be criticized and hence
there will be little chance of their being improved. We have belabored the
obvious because in contemporary discussion the raison d 'etre and even the
possibility of a methodology have been questioned.
0.1. What Contributions Can Methodology Make?
Most important of all is that it should contribute to the refinement of our
image of science. This image is becoming increasingly important for our
image of man. Questions such as 'Can scientific knowledge be rationally
justified?' 'How does knowledge grow?' etc. have to be attacked if we are to
refme our view about man's capacity of knowing, which is a central part of
 PROGRESS AND RATIONALITY IN RESEARCH 45

our self-conception as species. Moreover the image of science, in particular

the prescriptive part of it, the ideal of science, is essential for the researcher's
understanding of his activity, and hence eventually also for his success. The
ideal of science also serves as a foil for the historiographer of science in his
attempts to develop a descriptive picture of historically given science.
In addition to this general significance, methodology may, indirectly, help
to increase the researcher's efficiency. For instance, through offering means
of conceptualizing research situations and possible alternative developments,
through critically analyzing and appraising the ways in which successful re-
search enterprises have proceeded, through making explicit tacit presupposi-
tions and more or less unnoticed dependencies on certain styles of thought,
etc. All this with the view to increasing his freedom of decision - not with a
view to prescribing to him how he should proceed.
0.2. What Sort of Discipline is Methodology?
If there is to be reasonable hope that it may fulfIl the above-mentioned tasks,
what sort of diScipline would methodology have to be? First let us say what,
in our opinion, it cannot be: let us distance ourselves from two popular
reductivist views. (i) The view that conceives of methodology as an empirical
inquiry, as another scientific discipline and equating it with social science
cum historiography of science. This view is an example of scientism, and it is
based upon an instance of the so-called naturalistic fallacy: the attempt to
base good reasons for following a certain methodological recommendation or
for opting for a certain ideal of science upon what one believes to be the facts,
upon descriptive statements about how certain 'successful' research undertak-
ings did in fact proceed. But to speak at all of 'success' one has to transcend
the realm of description and explanation and enter that of appraisal. (ii) The
view, contrary to the scientistic conception, that methodology should be
applied logic. This view, logicism, totalizes one important aspect, logical
moves in research and the logical aspects of the results of research, by holding
that these aspects are all that matter in methodology. While the scientistic
view rests on a fallacy, this view correctly covers a part of the truth while
concealing others: of course logical aspects are very important, but, just as
evidently, there are many other important features of research.
If methodology is to have a chance of fulfilling the above-mentioned
demands (Section 0.1), it will have to be a discipline that develops a system
of recommendations about how to act in certain types of research situations
in order to facilitate achieving the aim of this activity: scientific progress; a
discipline that articulates and criticizes such recommendations in order to
46 GERARD RADNITZKY

improve them. Thus it cannot be identical with either sociology, psychology

of science cum historiography of science nor with logic applied in the recon-
struction of the results of research, although it will have to cooperate closely
with both history of science and with logic. Methodology, if conceived as
above, will have to identify types of research situations, and formulate re-
commendations stating what it would be rational to do if one were in a certain
type of research situation. A methodological prescription or advice could be
cast, e.g., as follows: "When you have to choose between two competing
theories, it is rational to prefer the one that stands in relation R to its com-
petitor - assuming always that your aim is to achieve cognitive progress".
Then, of course, good reasons have to be brought forward why it would be
rational to follow such and such advice. If two methodologies differ in their
advice, e.g., in a concrete theory appraisal give different verdicts, then we are
faced with the problem of methodology appraisal: appraising the comparative
achievements of rival methodologies. A methodological rule can be defended
or criticized only argumentatively: defended by giving good reasons for accept-
ing the conjecture that it is 'better' than its rival in the sense of having greater
potential for realizing the aim of the activity: scientific progress. Everything
depends upon whether this can be plausibly argued. Strictly speaking two
methodologies can be rivals only insofar as they have the same aim, i.e. to
explicate the aim of research - scientific progress - in roughly the same way.
To this problem we shall return later (Section 1).
In sum, methodology as a discipline is much like philosophical reflection;
it produces prescriptions of the type of the so-called hypothetical imperative
and good reasons to defend them or to criticize them. The system of recom-
mendations articulates an idealized image of research. If research is conceived
with Popper as basically an interplay of conjecture and criticism, of variation
and selective retention, of making and matching, of innovative moments
and moments of quality control, then we will distinguish two broad clusters
of rules. One is rules of quality control, of the ex post appraisal of results
and interim results of all sorts, but also appraisal of procedures, criteria,
arguments, even of problems. To use a convenient pars pro toto label for
this group, we propose to speak of rules of 'theory appraisal'. The other
group are rules referring to the innovative moments. They could be called
'heuristic rules' - heuristic in a narrow sense since all the rules have an advice-
giving function. A convenient umbrella word would be 'rules of theory
formation'.
Popper has focussed on the theory appraisal. Rightly so, because theory
formation always includes an essential element of conjecture, of creativity,
 PROGRESS AND RATIONALITY IN RESEARCH 47

which cannot be accounted for in a structural way. Hence there is no method

of theory formation, and therefore this moment is not of the same interest to
methodology as is quality control. But this does not mean that methodology
may not have something to say about the structural characteristics of hypoth-
esis formation. The often voiced reproach that Popperians have to abandon
the study of hypothesis formation to psychology is unjustified. For instance,
methodology studies the requirements of the output: e.g. the requirements
that the tentative theory, which is the output of the hypothesis-generation
moment, must be such that it is at least in principle capable of solving the
problem at hand, that it must be falsifiable, that it should have as high a
degree of testability as possible (say much), that, in case it is to function as
a revised successor of a falsified theory, it must not be produced by an 'ad
hoc' adjustment of the theory which has met with experimental results
contradicting it.
If the above programmatic defmition is accepted, what sort of activity is
methodology? It has obvious similarities with a technology. However, it
would be too crude to propose that it may be viewed as 'the technology of
scientific progress' because there are striking negative aspects to the analogy
between methodology and technology. Technology is often conceived as law
hypotheses in the context of solving concrete practical problems. We would
prefer to defme it as a system of prescriptions for how to use means to
achieve certain pre-given goals. 1 The most reliable technologies are based on
highly corroborated scientific theories. It need not of course be the best
available theory. It suffices if it is sufficiently reliable since, in the context
of application, the theory is being used only as an instrument of prediction,
and moreover a cost-benefit analysis is always relevant. (Thus, e.g. in space
flight Newton's theory is used rather than Einstein's, although in the context
of basic science, the superseding of Newton's theory by Einstein's is one of
the paradigmatic examples of scientific progress.) However, the theory to be
used must be sufficiently well established. Assessing its degree of 'evidential
support' - however this concept is explicated - is theory appraisal, a task
that ex definittone only methodology can tackle. Hence to attempt to base a
methodology upon empirical science in the same way as a technology can
thus be based would involve a vicious circle. The relevant knowledge would
involve also methodological appraisals, and methodological knowledge (how-
ever explicated) cannot be falsified or supported by empirical evidence in the
same way in which scientific knowledge can. To suppose that it can is an
instance of the scientific fallacy. This is the first important difference between
methodology and technology.
48 GERARD RADNITZKY

Related to it is a second difference. There are methods or rules for accom-

plishing technological tasks such as producing certain sorts of steel. Methodo-
logy, however, can point out means for achieving cognitive progress only in
the sense of facilitating its achievement, of facilitating the growth of knowl-
edge. For instance, it can provide broad rules for theory appraisal and rules
guiding the rational preference of one theory over its rivals. On the other
hand, in connection with problems such as on which premiss to put the blame
for a falsified prediction (Part II, Section 2.3), it can only give very broad
global advice leaving the greatest part of the decision to the researcher's
sensitivity; and in connection with hypothesis formation it can only give very
general guidelines (Part II, Section 1.3), since creativity cannot be planned
nor fully explained.
There is a third difference which is important. The goal of a technology
can be stated independently of that technology. (E.g. if the goal is to produce
steel of a certain specification, this specification can be given in the terms of
physico-chemical properties, and it is not the task of the technology of steel
production to provide these specifications - the goal is given from outside.)
In the case of methodology the situation is different: The specification of the
aim - facilitating scientific progress - is itself one of the major tasks of
methodology. (For many the task of methodology.) Who else could explicate
the idea of cognitive progress? Methodology has not only to explicate this
idea, but also to criticize the explicata that have been proposed, a criticism
that will lead to a comparative appraisal of the ideals of science which under-
lie different explicata. Methodologies will be rivals only insofar as they
attempt to realize roughly the same aim. A methodology, whose proclaimed
aim is to help achieve knowledge that is justified in the sense of having been
shown to be true, and a methodology based on a non-justificationist view of
human knowledge, cannot be appraised with respect to their comparative
achievements without such comparative appraisal leading us to a critique of
the ideal of science, the view of man's capacity of knowing underlying each
of them. (Perhaps, although of course words do not matter, we could say
that methodology is a 'quasi-technology' - a convenient label for epitomizing
the above considerations of the positive and negative aspects of the analogy
between methodology and technology.)
Thus this activity, which is neither empirical investigation nor mere
applied logic, but rather argumentative, has affinity with philosophical reflec-
tion. Insofar as it attempts to find out what course of action is rational given
a certain interpretation of the situation or attempts to develop assessments of
good reasons, i.e. ex post to assess problem solutions, theories, procedures,
 PROGRESS AND RATIONALITY IN RESEARCH 49

decisions and so on, it may be said to develop what Popper calls a 'situational
logic'. When it attempts to assess the efficiency of ways of proceeding, of
actions, of research undertakings, etc., it qualifies as a praxiological study in
Kotarbmski's sense; and insofar as it looks at research as processing a 'research
program', i.e. a system of hypotheses (knowledge) - problems-instruments
(techniques, calculi, etc.)-plans-etc. into a more refined system of that sort,
it might be said to exemplify what nowadays is often labelled 'systems
thinking'. 2 It constitutes a sub-field of philosophy rather than anything else,
a sub-field that is secondary to some other fields as it is secondary to science
itself in the sense that first there must be empirical research before a need for
methodology can arise.
Methodology obviously does not and cannot deal with certain problem
clusters of traditional philosophy. But nonetheless philosophers often re-
proach it for not attacking certain philosophical problems. Hence some proca-
taleptic remarks about its limitations are called for. (i) Some philosophers
(e.g. H. Spinner) accuse Popperian methodology of conventionalism, asserting
that in spite of its being a non-justificationist approach it has not been able to
come to grips with the problem of the so-called 'basic' sentences. So far as
methodology is concerned Popperians hold that no type of sentence is to be
accorded an epistemologically privileged status. In the empirical testing of
hypotheses, data sentences are used, e.g. in physical research statements about
material objects and processes but not about perceptual experiences. Percep-
tual reports form part of the good reasons for the conjecture that a certain
'basic' sentence or data sentence may be accepted pro tempore. When there
appears no reasonable doubt concerning a data sentence (a matter to be
decided by the researcher) it would be pointless for the methodologist to
emphasize that it can always be questioned. For this reason the relationship
between perceptual reports and a data sentence about physical objects is not
analyzed by Popperian methodology. Such an analysis is regarded as a topic
of general epistemology, and ontological analyses of acts of perceiving, etc.
are left to philosophers. (E.g. 'epistemology as the ontology of the knowledge
situation' as it is developed by Gustav Bergmann and his followers is a field
that methodology cannot encompass.) (ii) Similarly Popper's three-world
ontology is intended as means to an end: to provide a suitable ontological
ground-plan for discussing certain methodological problems. It is not in-
tended as an 'Aufbau der Welt', as e.g. 'ontology as an argumentative struc-
ture upon a phenomenological base' by means of which to construct or recon-
struct reality including our experience. This again is a task of 'first philosophy'.
(iii) Popperian methodology presupposes that there exists a language and that
50 GERARD RADNITZKY

researchers are capable of forming what is called 'a communication com-

munity'. For this reason some philosophers (e.g. K.-O. Apel) accuse Popperian
methodology of a sort of abstractive fallacy. But such a reproach would be
justified only if Popperians were to forget that research has such precondi-
tions. This is acknowledged, but the analysis of the conditions of possibility
of science (sinnkonstitutive Bedingungen der Moglichkeit) is left to philoso-
phy proper, because this is a burden methodology cannot bear. All this is
simply a practically necessary division of labor within philosophy. Of course
it would be hubris if methodology were to aspire to encompass these sub-
fields of philosophy. We have emphasized what may seem self-evident because
in the literature there has been some misunderstanding about this situation.

1. METHODOLOGY APPRAISAL LEADING TO THE CRITIQUE OF

SCIENCE; TASKS FOR METHODOLOGY FOLLOWING FROM
THE IDEAL OF SCIENCE

1.0. Appraising methodologies includes consideration of two sorts of ques-

tions: Are the solutions a particular methodology proposes to the problems it
has posed acceptable? Are the problems really those that matter? Will answer-
ing these questions help in facilitating scientific progress? The explication of
the idea of progress will be governed by the ideal of science adopted: progress
will, in general, mean coming closer to that ideal. In this way a comparison of
two methodologies will involve comparing the reasonableness of their ideals.
If two methodologies' fundamental ideals of science, although distinct from
each other, can be shown to be alternative explicata for a common concep-
tion of scientific merit, then a comparative appraisal at least of these explicata
is possible.
In our intuitive ideas about what scientific knowledge should be like and
about the earmarks of progress we seem to have such a common ground:
there exists a common explicandum. Probably - as John W. N. Watkins has
argued 3 - all parties would accept the following naively stated desiderata.
The aim of research is knowledge that is genuine knowledge, as comprehen-
sive and as deep as possible: i.e. it should explain a lot and these explanations
should help us better and better to understand the world and ourselves in
this world. A successor theory's progress over a predecessor consists in its
achieving more in at least one of these three aspects than the predecessor.
The two most prominent methodologies, which have placed the normative
problem in the center, logical empiricism and Popperian methodology, have
proposed differentexplicata for this intuitive ideal (as a common explicandum),
 PROGRESS AND RATIONALITY IN RESEARCH 51

and the Popperian ideal of science has arisen from criticizing the explicatum
proposed by logical empiricism. We propose first to give some hints at the
ideal as explicated by logical empiricism, then to move to the Popperian
criticism of that ideal, and eventually to the alternative offered by Popper.

1.1. The guiding idea of the ideal of science of logical empiricism, of the
foundationalist approach, is to give the desideratum certainty, top priority. As
a result the most important task of methodology becomes that of formulating
and legitimating a role of acceptance in accordance with the basic conviction
that a scientific proposition is ultimately acceptable only if it is true. To
make such a role operational one needs a method of establishing in concrete
cases whether a proposition is true or not, and the method must give us an
infallible criterion of truth. This is the position of verificationism, the justi-
ficationist approach to rationality in science: a sentence is acceptable only if
true and recog~ed as such. It is then weakened in probabilistic verification-
ism: a scientific proposition is acceptable only if it has been probabilified
to a 'sufficient' degree. In this scheme experience plays a positive role (hence
the tag 'positivism'): it serves as the final establishing arbiter. Verificationism,
absolute or probabilistic, seems attractive only if one is willing to countenance
a particular class of empirical propositions whose certainty does not need to
be called in question: a sure source of knowledge about reality as an epistem-
ological fundament on which to erect the edifice of science. Ideally, systems
of propositions would be generated by deductive connections. This is a sort
of 'proof-empiricism'. Again, in view of its obvious unattainability - since
universal propositions infmitely transcend any fmite set of singular sentences
serving as a fundament - one lowers his demands and attempts to construct
propositional systems with partial information-covering from a selection of
'basic' sentences assumed to be certain. Underlying this probabilistic verifica-
tionism seems to be a principle of hope that, as more and more evidence
comes in, it may in the long run be possible at least asymptotically to approach
that ideal state in which the evidence completely 'covers' the information of
the complex sentences. Inductive logic or theory of confirmation is to pro-
vide the connections and to measure the degree of 'coverage'. Cognitive
progress would then be defmed in terms of better and better approximation
to this ideal of science. Logical empiricist philosophy of science may be
viewed as an articulation of various key aspects of this ideal of science. 4

1.2. Popper's critique may, in accordance with what has been said above, be
divided into two parts: the critique of the problem solutions offered, and the
52 GERARD RADNITZKY

critique of the problems. (i) The critique of the solutions offered amounts
roughly to pointing out that the various models (such as the various explicata
offered for the concept of 'empirical significance', the models of explanation
which promise an explication of the idea of causal explanation by stating not
only necessary but also sufficient conditions, etc.) have not been able to help
researchers approach certainty, the ideal's own centerpiece in measuring the
results of research, not even when these results are highly stylized. This must
be so because, firstly, the presupposed certain 'basis' does not exist, and,
secondly, even if for the sake of argument such a 'basis' were conceded to
exist, not only absolute verificationism but also probabilistic verificationism,
inductivism, founders on Hume's criticism, since it follows from the proba-
bility calculus that the logical probability of a universal proposition on the
basis of a finite set of evidential statements is zero. Even a partial retransmis-
sion of truth from verified conclusions to premisses is logically not possible.
In addition, probabilistic verificationism, with its attempt to develop an in-
ductive logic or similar method, will continue to lack the crucial deductive
structure - which is one of the desiderata of the logical empiricists' own ideal
of science. It must ex definitione introduce amplificatory logical moves. (ii)
Popperian criticism of the explicatum of our intuitive ideal of science pro-
posed by logical empiricism asserts that this explicatum is neither fruitful
nor sufficiently similar to the reasonable part of the explicandum to be
acceptable as the result of a successful explication attempt. The desideratum
given top priority, certainty, is unattainable in principle - utopian. Since
certainty and informative content are inversely proportional, and since cer-
tainty is given top priority, the value of 'high content of empirical informa-
tion' must be sacrificed. However explicated, the desideratum of 'depth' -
an important component of the intuitive ideal - is not only lost, but has also
become anathema. (Another reason for the tag 'positivism'.) In sum, the price
to be paid for the search for certainty is a total loss in all dimensions even of
the ideal of science as explicated by the logical empiricists themselves: not
only is certainty unattainable in principle, but in striving for it nonetheless,
other desiderata of the logical empiricist's own ideal become unattainable.
The very idea of knowledge in the sense of certain knowledge, which appears
to be a secularization of the theologican's concept of revealed knowledge, can
have no place in empirical inquiry.
From logical empiricism's explicatum of the intuitive ideal follow certain
problems for methodology. Since the explicatum is mistaken, these problems
are inappropriate. From the ideal it follows that the main task of methodo-
logy is to search for an acceptance rule, to formulate and legitimate such a
 PROGRESS AND RATIONALITY IN RESEARCH 53

rule. It conflicts with that idea of scientific progress which sees novelty as
the essence of progress: insofar as we want theories to go much beyond 'back-
ground knowledge', to lead to new insights and thereby to new and deeper
problems, it is unreasonable to hold that we are ultimately after theories
whose information content will eventually asymptotically approach the state
of being completely covered by the information carried by that 'complete'
evidence. The justificationist, cumulative view of science is mistaken if only
for that reason, and if, interpreted as a descriptive picture of historical
science, it is refuted by the history of science because theories once regarded
as certain, later were falsified and superseded by new theories. In short, the
Popperian side shows that logical empiricism's commitment to foundational-
ism and inductivism has proved untenable - that the foundationalist approach
founders like all Begriindungsphilosophie. Scepticism would be a possible
reaction to this insight, but it is not the only possibility.

1.3. Popper offers an alternative to the foundationalist approach and to scep-

ticism. With his pioneer work of 1934, which has become a classic, he is the
first to work out a deliberate non-foundationalist methodology and also a
non-foundationalist, evolutionist theory of knowledge: a Copernican revolu-
tion in the philosophy of science. In the Popperian explicatum of the intui-
tive ideal the quest for the pivotal desideratum of the explicatum proposed
by logical empiricism, certainty - the Fata Morgana of all foundationalist
philosophy (Begriindungsphilosophie) - is abandoned in favor of a conjec-
turalist-fallibilist view of human know/edge, at least for all knowledge about
the empirical world. Such knowledge is in principle fallible, conjectural. But
fallibilism preserves the idea of absolute truth as a regulative principle,
especially in the comparative notion of a 'more (or less) accurate representa-
tion' (mehr oder weniger zutreffende Darstellung): we know what we mean
by truth or truthlikeness in this sense even if there is no criterion of truth, i.e.,
even if in any concrete case we cannot with certainty tell whether a particular
proposition is true or false.
The explicatum of our intuitive ideal of science must above all meet the
meta-criterion of fruitfulness, implying inter alia such problems for meth-
odology that a methodology which attacks these problems can hope to make
the contributions mentioned above in Section 0.1. It must also meet the
necessary condition of being 'sufficiently similar' to the intuitive ideal in
those respects where the ideal is not utopian. Once it is recognized that
certainty is unattainable in principle, the goal of research can be epitomized
as representing more and more accurately (increasing in truthlikeness about)
54 GERARD RADNITZKY

those aspects of reality whose comprehension (explanation) leads to new,

fruitful perspectives and thus to new and deeper problems, so that we get a
better and better understanding of the world and of mankind (contribution
to the refmement of our world view). Science is concerned with developing
theories further and with replacing theories by better ones, i.e. with cognitive
progress. The question of acceptance has its place primarily when we ask
whether or not we regard it as rational, as justified (given the practical situa-
tion at hand) to use a certain theory as an instrument of prognosis and to
base technologies on it. Basic research is ex definitione concerned with cogni-
tive progress, not with 'acceptance'; and if in the context of a methodological
discussion Popperians speak of 'acceptance', this is short for saying that, since
the theory in question has not (yet) been falsified and indeed has thus far
stood up to all empirical tests, we propose to continue working on it and with
it, i.e. to develop it further by inter alia subjecting it to new sorts of tests.
In sum Popper proposes an alternative to the quest for certainty, the
search for a principle of induction, the 'new philosopher's stone', the search
for growth of knowledge. Hence it is important to explicate the idea of
growth of knowledge, of cognitive progress, and in this context the idea of
one theory's being closer to the truth than its rival is likewise of great impor-
tance.
What tasks for methodology follow from the Popperian explicatum of the
ideal of science? The global tasks will be: making explicit the various com-
ponents of the ideal in more details, suggesting methodological rules supposed
to facilitate the realization of the ideal as explicated, and supporting these
rules by good reasons. If one accepts the Popperian explicatum of the intui-
tive ideal, then in the appraisal of the comparative achievement of rival
methodological rules the key question will be for which of the competing
rules it can plausibly be argued that it is of greater help than the other in
realizing the ideal in the sense of Popper's explicatum.
What specif~c tasks for methodology follow from the above global tasks?
The center of concern will be preference rules, i.e. to formulate such rules
based on rules of appraisal and to fmd out what sort of good reasons might
accompany the conjecture that one of a particular pair of competing problem
solutions, theories, etc., should rationally be preferred over the other. The
rules of comparative appraisal will have to be formulated not only for results
such as theories and explanations, etc., but also for procedures and for past
decisions - all this with a view to improving future deciSion-making. Here we
consider only the issue of theory comparison. It is of course rational to prefer
that theory which is 'better' than its rivals on all counts or on that count
 PROGRESS AND RATIONALITY IN RESEARCH 55

which matters in the particular comparison situation at hand. A theory T' is

'better' than T with respect to z to the extent r. Thus we will have to identify
the relevant respects and explicate 'better' with respect to them. Evaluations
of this kind are internal to science.
Apart from trivial desiderata such as internal consistency and empirical
significance, a commonsensical desideratum is that the theory proposed
should be pertinent to the problem at hand, a potential answer to the ques-
tion raised. (Not even the rare cases of serendipity are without a guiding
question even if it is concealed (M.D. Grmek).) A basic desideratum is that
the successor theory T' should go beyond its rival (predecessor) T. If it con-
tradicts the predecessor (and is successful in predicting), then this is an indica-
tor that it has a greater 'depth' than the predecessor; and if so, it will be more
fruitful, i.e. give rise to still 'deeper' problems. Another basic desideratum is
that T' should say more than T, bring an increase in potential 'explanatory/
predictive power', and that what it says should be correct, in particular its
predictions successful.
Popper and even more so his followers have focussed on this dimension of
'truth likeness', where truthlikeness is conceived as a concept that solders
content and truth. Content is not to be taken in an absolute sense. The
problem of theory appraisal arises, typically, only when theories are com-
petitors, when they attempt to solve the same set of problems. They are
maximally competitive if they give incompatible answers to the same ques-
tion( s). If the task of ascertaining which theory is closer to the truth were not
thus limited, theory appraisal would presuppose a prior appraisal of the
'scientific interest' of the questions. 'Scientific interest' may be explicated
objectively in terms of the contribution an answer to the question at hand is
expected to make to cognitive progress in the discipline; it will be particularly
high if the successor theory contradicts the theory that is the reigning cham-
pion. However, since an element of prognosis is involved, such an appraisal is
very risky. Fortunately at least this problem need not trouble ushere since we
may presuppose that the theories under appraisal, T' and T, are competitors.
Intuitively everything appears fairly clear. The two basic distinguishing
features of progress in the information-theoreticaljepistemic dimension are:
(a) that T' says more than T, i.e., in the area of the two theories, mutual con-
cern; and (b) ideally, that what T' says is true, or, more realistically, what
T' says is closer to the truth than what T says, again, with respect to the
scientific problems at hand. It will be requested (i) that the empirical content
of T' that goes beyond that of T has not been falsified although tested in
tests of a certain severity (Part II, Section 2.1.0); (ii) that T' matches the past
56 GERARD RADNITZKY

explanatory successes of the predecessor theory T, Le. that the corroborated

hypotheses thus far deduced from T (in the presence of auxiliary hypotheses)
must also be deducible from T' (and auxiliary hypotheses) with at least the
same degree of precision (Le. with the same empirical content); and preferably
T' should also refme and correct some of the original predictions. If a pro-
posed successor theory fulfils at least the first two of these three require-
ments, then not only is T' closer to the truth than T but, so far as the scien-
tific problems under consideration are concerned, T' also dominates T in
content.
Of course, degree of corroboration (as the balance sheet of empirical
criticism) together with preservation of explanatory successes provide but a
fallible indicator on which to base the conjecture that T' is closer to the
truth than T, since good performance of a theory to date no more guarantees
high dividends in novel knowledge in the future than the good performance
of a stock guarantees future profits. Nonetheless, the situation of the meth-
odologist is better than that of the fmancial analyst relying on (inductivistic)
extrapolation from charts, because the methodologist can rationally conjec-
ture that the better corroborated theory is closer to the truth than the one
with a lower degree of corroboration thus far. Of course he too may be
proved wrong by future scientific developments.
Since a dramatic increase in 'truthlikeness' in the sense of content-cum-
truth is possible only if the successor theory is 'deeper'than the predecessor,
and makes possible 'deeper' explanations, the 'deeper' theory is to be regarded
as the better one. The idea of depth will, of course, have to be clarified.
From the desideratum of increase in 'truthlikeness' as (fallibly) indicated
by increase in degree of corroboration together with the method of falsifica-
tion (which is thoroughly deductive), it follows that no amplificatory moves
will be permitted in connection with theory testing. Certainty will thus be
retained in the only area where it has a place. But since it follows from other
desiderata, the requirement of deductive procedures need not be mentioned
explicitly.
1.4. What Do the Above Considerations Mean in Terms of Specific Tasks for
Methodology ?
(1) From the desideratum of increase in 'truthlikeness' there follow two
tasks: (la) As much as it is necessary for the methodological problem of
theory comparison, to clarify the concept ' ... being closer to the truth
than ... '; (1 b) to develop indicators by means of which we can produce
argumentatively good grounds for the conjecture that one of an actual
 PROGRESS AND RATIONALITY IN RESEARCH 57

pair of competing theories in fact comes closer to the truth than the other.
The indicators will be fallible but must and can be objective. s (E.g. whether
or not a prognosis about, say, an eclipse is falsified is objective in the sense
that the fate of the prognosis is independent of human influence.) As is well
known, the notion of degree of corroboration is offered as a fallible indicator
that sets out the balance of attempted falsifications to date. The role of
experience here is exclusively that of a critical arbiter. No 'founding' is
sought after. To interpret degree of corroboration as designed to function in
the long run much like degree of inductive support - as some philosophers
have suggested (e.g. G. H. v. Wright) - has no grounds in Popperian method-
ology. (2) Since degree of testability is equivalent with content of empiri-
cal information, the more content a theory has, the greater is the risk of
falsification and thus also its corroboration potential. From the desideratum
of increase in degree of corroboration (as a fallible indicator of increase in
truthlikeness) follows that of content increase (potential explanatory power).
This in turn sets methodology the task of clarifying the concept of empirical
content and of providing instruments for making content comparison. (3)
A dramatic increase in content is possible only when the successor theory
contradicts the predecessor, for only then does it really introduce new
concepts and open new perspectives, thus leading to deeper explanations and
to deeper and deeper problems. This sets methodology the task of explicating
the concept of depth and of providing an indicator of an increase in depth.
One indicator that functions similarly to a sufficient condition 6 is that the
successor theory, in attempting to explain the predecessor theory, corrects
it, i.e. from the successor theory (e.g. Newton's theory) a hypothesis is
deduced (e.g. Newtonian versions of Galileo's law of free fall or of Keplerian
laws of planetary motion) which, although contradicting the original ex-
planandum, may be regarded as an improved successor to the original ex-
planandum (e.g. Galileo's law of fall), or the explanandum mathematically is
an approximation within a limited realm (in the example, when the height
of fall is negligible in relation to the earth's radius) to the improved successor
hypothesis. (The original explanandum could be derived if we make the false
assumption that the earth's radius is infinite or the height of fall zero.) The
successor theory (in our example, Newton's) introduces a new sort of con-
cepts (causal concepts) which are not used in the predecessor theory (Galileo's
and Kepler's law hypotheses do not involve any causal concepts); and it is
plausible that it is these new concepts that enable us to look at the world
in a new way, which in turn makes possible deeper explanations and gives
rise to new problems of a greater level of depth. Therefore the desideratum
58 GERARD RADNITZKY

increase in depth appears to hold a key position. In this way theory appraisal
leads to appraising metaphysics or cosmological hypotheses (as Feyerabend
prefers to call them), to appraising the comparative value, fruitfulness for
research, of competing world-picture hypotheses. Scientists were once de-
servedly termed 'natural philosophers'. The best of them did and still do face
up to the philosophical issues posed by their own work. As Joseph Agassi has
argued, 7 giving top priority in the explicatum of progress to degree of testa-
bility carries an anti-metaphysical flavor; we would add that it is a hangover
from positivism.
Since 1941 Popper has drawn attention to the phenomenon that a deeper
theory corrects the 'observationa1' law-hypothesis - independently of whether
or not the latter has been falsified when the deduction is made - in the very
process of explaining it. 8 It is important to notice the continuity in the
empirical as well as in the mathematical aspects between the hypothesis
corrected and the improved successor hypothesis deduced from the new
theory, although the successor hypothesis contradicts the hypothesis that gets
corrected. Thus, in spite of the break constituted by the new concepts in-
troduced by the successor theory - the new perspectives it opens up and the
new, deeper problems it poses - there still is an element of continuity in
these two aspects.

1.5. In the Popperian ideal of science the essence of progress is seen in moving
from problems to deeper problems. The Kantian-Popperian thesis of the
propagation of problems claims that every solved problem generates new
problems 9 (objectively, i.e. independently of the researcher's wishes, indepen-
dently of whether or not he formulates them or even recognizes them). The
deepening of problems is seen as a measure of progress.! 0 "Science should be
visualized as progressing from problems to problems - to problems of ever
increasing depth."!! When we confront this ideal of science with the picture
of historically given science, we find that the history of science illustrates it
well. Although this fact per se could not be used as a good reason for recom-
mending the ideal of science, it nonetheless shows that the ideal is not utopian
- as is the ideal of science of logical empiricism.
From the point of view of the history of the philosophy of science it is
sweeping, but correct, to say that Popper is the chief critic of the methodology
developed by logical empiricism and of the ideal of science at its root, an
ideal which owes its decisive impulse to the philosophy of the early Wittgen-
stein. The polemic between Wittgenstein and Popper - which has remained
implicit - in fact has a continuation: the most important contemporary critics
 PROGRESS AND RATIONALITY IN RESEARCH 59

of Popper are in turn indebted to the philosophy of the later Wittgenstein.

This is true - in spite of all the differences between them - of Thomas Kuhn,
Stephen Toulmin and Paul Feyerabend, to mention only the most well
known. But in the long run they lose their grasp of the normative problem,
and must lose it; and so Kuhn and Feyerabend see theoretical developments
between which a 'scientific revolution' lies as incommensurable entities,
similar to Wittgensteinian monadic forms of life which can only be evaluated
from within and cannot be rank-ordered. 12
1.6. Brief Comment on the Problem Situation as Reflected in the Literature
1.6.0. The problem situation arises from the attempts to clarify the ideal of
science outlined above and to deal with the resulting methodological prob-
lems. There is a cluster of problems centering around explicating the concepts
of content and truthlikeness. The approach generally adopted is to define
logical content as the set of non-tautological consequences of a theory. But
how could this be measured? Cardinal numbers obviously provide no viable
measure. A measure of content has been dermed only for certain very simple
formalized language systems, using the concept of absolute logical probability
as primitive. Aside from the fact that such model 'languages' have their value
only as instruments for 'logical underpinning', this approach presupposes that
one regards the concepts of absolute logical probability as at least as clear as
that of content. (After all, one is the converse or complement of the other.)
Since in the most interesting cases of scientific progress the successor
theory revises the predecessor theory, content comparisons have to be made
between incompatible theories. Popper has pointed out that the successor
theory has a greater content if the questions answered by the predecessor
theory are a proper subset of those answered by the successor theory. (Of
course, the questions that are of interest here are scientific questions, not just
any questions.) D. Miller (1975) has questioned whether it is at all possible
to ascertain that more accurate predictions are derivable from one theory
as a whole than from its rival; Griinbaum has questioned whether it can be
ascertained that one theory answers more questions than another. 13
If one attempts to give an adequate explication for the idea of one theory's
being closer to the truth than its rival, being a more accurate representation
of the aspects of reality that interest us at the moment, the following approach
seems natural: either the amount of information conveyed by the true con-
sequences of the successor T' is larger than that conveyed by the true con-
sequences of the predecessor T and its amount of false information not larger,
or the amount of information conveyed by the false consequences of T' is
60 GERARD RADNITZKY

smaller than that of T and its amount of true information not smaller. A set-
theoretical interpretation of the comparative concept, called 'verisimilitude',
takes the statement "the amount of information conveyed by the true con-
sequences of T' is larger than that conveyed by the true consequences of T"
to mean that the true consequences of T are a proper subset of the true
consequences of T', and similarly, mutatis mutandis for the other parts of
the definition. This interpretation proves inadequate if taken as an expli-
catum of our intuitive idea of 'more accurate representation than'. Miller
(1974) showed that if one uses this set-theoretical interpretation of the
definition of verisimilitude, the required subset relations can only obtain
between axiomatizable theories if both are true. The next explicatum proposed
hinges on the idea that the information conveyed by the true consequences of
one theory minus the information conveyed by its false consequences must
be larger than the information conveyed by the true consequences of the rival
theory minus the information conveyed by its false consequences. This
approach too has proved to be unfeasible. Andersson 14 has shown that some
points of the criticism advanced by Miller and Tichy can be met, but con-
cedes that there remains a fundamental difficulty which cannot be overcome
even with fmer measures of content and verisimilitude: it turns out that all
false theories with the same measured content have the same degree of veri-
similitude if this explicatum is used. This is an absurd consequence since,
according to this explicatum, the verisimilitude of a false theory depends only
on how much it says, not on what it says. Hence this explicatum does not
meet the metacriterion of 'sufficient similarity' between explicatum and
explicandum, a necessary condition for fruitfulness. Moreover the comparison
of false theories is very important, since cognitive progress often consists in
one falsified theory's being replaced by another which, although likewise
falsified, is regarded as closer to the truth. (We need not to go to the history
of science; a primitive example can illustrate this: the hypothesis 'The planets
move in triangular orbits', although false, contains a kernel of truth (e.g.
that the planets have closed orbits); the hypothesis 'The planets move in
circles' is likewise false, but intuitively is closer to the truth than the first one.)
In the literature various positions can be discerned. Some writers go so far
as to deny that our intuitive idea of one hypothesis being closer to the truth
than its rival is fruitful. (For example, K. Hiibner, A. J. Ayer and G. S. Robin-
son hold this view.) The rest would certainly agree that it is fruitful. Some of
them have shown that the problem of explicating our intuitive idea via a con-
cept introduced by a formal defmition (Le., a defmition formulated in an
idealized language schema, IL, based on standard logic) remains unsolved
 PROGRESS AND RATIONALITY IN RESEARCH 61

(D. Miller, P. Tichy, G. Andersson). Undoubtedly this is true. If it should turn

out to be unfeasible to make the intuitive idea more precise with this method,
the absence of a formal definition will induce some to become sceptical about
the value of the intutive idea for methodology (e.g. Miller), while others
would question whether this sort of precision is necessary 'for the concept
to be fruitful, and would still regard the intuitive idea as indispensable (e.g.
H. Albert, G. Andersson).

1.6.1. In this situation it seems appropriate to recall the real issue here: our
aim is to legitimate (in the sense of giving good reasons) a preference rule,
which is based on a rule of theory appraisal- at the moment in the informa-
tion-theoretical and epistemic dimensions. Hence the adequacy of the expli-
cata proposed must be judged in terms of their fruitfulness for this meth-
odological task. Throughout, of course, we must clearly distinguish between
two sorts of tasks, that of explicating concepts and that of developing indica-
tors.

1.6.2. Some unorthodox reflections.

1.6.2.0. Intuitively we distinguish between the information conveyed by a
theory, i.e., the explicitly formulated theses, which constitute what the
theory says, and the set of all the theory's consequences, which is indepen-
dent of whether or not these consequences have been or ever will be formu-
lated (or 'discovered'). In any case, the information of the theory, in the
above sense, is so condensed that in practice it is difficult to test it directly.
For this reason, in order to criticize a theory empirically we must derive
testable consequences; we must extract from the information contained in
the theory empirical information in small enough doses that it is technically
possible to test it.
As is well known, in his classic exposition Popper dermes the 'logical
content of a theory T' as the set of non-tautological consequences of T.
Consequences whose truth rests on their logical form or on definitional con-
ventions are, of course, irrelevant to our present task, since in this respect
all theories are on a par. (For analogous reasons it would be pointless to
compare the sets of consequences of inconsistent theories, since any sentence
is deducible from an inconsistent theory. This demand for consistency thus
remains, is necessary in principle, although P. Feyerabend is certainly right
when he emphasizes that in research one deduces not x number of arbitrary
consequences, but only those which represent a potential answer to a scien-
tific problem, and that in the attempt to find an answer to a particular
62 GERARD RADNITZKY

scientific problem an inconsistency between two of a theory's components

need not always be relevant. On the other hand, when Feyerabend says that
the demand for consistency would be unrealistic, since establishing consis-
tency first requires axiomatization of the theory, which would take too long
because in practice the theory would have already been superseded by then,
this has to do only with methods of ascertainment, and not with the explica-
tion of concepts, which is what is here at stake.) At any rate, those con-
sequences of a theory, which are generated by exploiting the peculiarities of
the V-connective (as defined in standard logic), are quite irrelevant here, since
they do not constitute possible answers to our scientific problems. In real
science researchers do not make trivial deductions by joining to a deduced
predictive hypothesis another hypothesis, connecting them with an 'or' -
this would indeed be a 'philosophical joke'.
1.6.2.1. But, even admitting such qualifications, a subdivision is still needed
within the set of consequences thus restricted. For in the context of theory
comparison only a part of the synthetic consequences is relevant. These non-
logical consequences are either metaphysical or empirical. The metaphysical
portion of a scientific theory has metaphysical implications, implications for
philosophical cosmology (Part II, Section 1.5). However, a theory will have
such repercussions, such 'philosophical implications', only if it constitutes a
major breakthrough, and a necessary condition for this is that the theory be
regarded as closer to the truth than its predecessor. Hence it appears per-
missible to bracket the issue of a theory's philosophical implications. Given
our aim (Section 1.5 .), the class of relevant consequences can be limited even
further. Since in the final analysis the issue is which of the competing theories
is closer to the truth, what matters are the testable consequences. In some
cases it will be possible to deduce testable consequences from the theory
alone. Given our aim, what matters are only such consequences as constitute
potential answers to scientific problems - either potential answers to our
pressing scientific problems or potential answers to questions whose answers
would eventually be of importance for improving our world-picture (cf. Part
II, Section 1.5). Such consequences are as a rule not derivable from the
theory alone. (For example, in order to be able to test Newton's theory of
gravitation empirically, at least in astronomy, we need in addition, as auxiliary
hypotheses, theories of optics.) As is well known, ,Popper defines the 'empiri-
cal content of a theory T' as the set of potential falsifiers of T. In the context
of actual empirical criticism of a theory, what is relevant is the empirical
content in the sense of the informative content of the set of potential falsifiers
of T plus auxiliary hypotheses A (because normally such additional premisses
 PROGRESS AND RATIONALITY IN RESEARCH 63

are needed) i.e., the set of all conjunctions of appropriate statements of initial
conditions and the negation of a hypothesis deducible from T in the presence
of auxiliary hypotheses A, so that this conjunction contradicts the conjunc-
tion of Tand A.
Thus it appears advisable to stipulate that the assertion that T' is better
than T in the information-theoretical dimension is to mean that the informa-
tion conveyed by the set of potential falsifiers for T' with the auxiliary
hypotheses A necessary for deducing them is greater than the information in
the potential falsifiers deducible from T and A. It is presupposed here that
the hypotheses deduced concern scientific problems and that the theories are
competitors, i.e., that they attempt to answer the same questions.
This yields a preference rule: before testing, prefer that theory for which
there are good reasons to conjecture that its empirical content (in the pre-
sence of the necessary auxiliary hypotheses (part II, Section 2.1.1) is greater
than that of its competitor, that it is more falsifiable than its competitor.
1.6.2.2. All this has to do only with the issue of explicating concepts. The
good reasons for such a conjecture hinge upon the use of some indicator
(fallible but objective) of relative falsifiability. Such an indicator is, in prin-
ciple, provided by the information conveyed by each of the sets of potential
falsifiers thus far deduced as potential answers to our scientific problems.
Thereby the proposed explicatum has plainly guided the production of
indicators - as should be the case. In actual practice, of course, there is no
usable measure of a unit of infOImation, and comparing two theories here
seems feasible only if one theory entails the other. On the other hand, there is
no point in content comparison anyway unless the theories are competitors,
and ultimately the decisive question is whether what T' says in answer to
the problems common to T and T' is a more accurate representation than
what T says; what counts is the situation after testing. In brief, the proble-
matic of content comparison seems, at least to this writer, to have received
undue attention considering that the global aim is to legitimate a preference
rule for the situation after testing.
In conclusion we can now return to the distinction initially mentioned.
The information in a theory, the group of formulated theses constituting it,
or, more accurately, constituting a particular version of it, is so condensed
that in practice it is difficult to test it directly. The more general and the
deeper a theory is, the more highly condensed the information will be. This is
why the whole business of deducing testable consequences has to be gone
through to make empirical criticism possible at all. Since every sentence
(trivially) entails itself logically, the theses themselves are of course included
64 GERARD RADNITZKY

in the set of the theory's consequences though not in the set of its testable
consequences. However, although for our present aim only the content of the
potential falsifier (or, in certain contexts, of the potentially falsifying hypoth-
eses (Part II, Section 2.2.2» matters, the two intuitive ideas - the content
of a theory and the set of its consequences - are not identical. And it seems
inadvisable to make them identical by defmition. The hypothesis that a
particular formulation of a theory has such-and-such content is again a con-
jecture for which good reasons can be provided, good reasons to be based
upon our interpretation of the theory's theses. Making such a conjecture is a
hermeneutic task. The conjecture may be criticized, inter alia, by confronting
it with the information in the consequences including that in the 'metaphysi-
cal' component of the theory.
1.6.2.3. The above preference rule is applicable in the situation before test-
ing. The simplest case would be that in which the two competitors give incom-
patible answers to the same scientific question. In this case, the theory that
gives the correct answer or a more accurate answer than the rival does would
be preferred. This consideration goes beyond the information-theoretical
dimension. In the epistemic dimension, the assertion that T' is better than T
is to mean (Le., it is so explicated) that T' is closer to the truth than T. But
this is just another way of saying: by 'truthlikeness' we mean that the theory
is nearer to the truth whose answers to our scientific questions represent the
relevant aspects of reality more accurately than those given by the competitor
theory. The preference rule is: after testing, prefer the theory for which the
conjecture that it is closer to the truth than its competitor is supported by
good reasons. These good reasons will make reference to an indicator (fallible
but objective): whether the potential falsifiers (or the potentially falsifying
hypotheses (Part II, Section 2.2.2» thus far deduced have or have not stood
up to the empirical tests thus far carried out.
This is all that is needed to reach our aim. It is presupposed that in our
language the idea of truth and the concomitant idea of truthlikeness (mehr
zutreffende Darstellung als) function successfully. This may seem problematic.
But it cannot be stressed too strongly: without the descriptive function
(Darstellungsfunktion) there is no language in the full sense, no human
language. Karl Buhler's work (1934) is highly pertinent here; Tarski's famous
semantic definition of truth, pace Popper, is not. 15 It is, of course, a task of
philosophy to clarify and explicate the idea of truth and its derivative con-
cept of one hypothesis being closer to the truth than another. This is indeed a
perennial task of philosophia prima. It cannot be a task for methodology
(cf. Section 0.2, s.f. on the division of labor). Nor can methodology wait until
 PROGRESS AND RATIONALITY IN RESEARCH 65

'fust philosophy' has produced answers that are deeper and more relevant
than those it has produced during the last two and a half thousand years.
Hence to reject these concepts because we have not been able to provide (and
perhaps never will be able to provide) an explication of them such that the
explicatum is introduced via a formal defmition couched in an idealized lan-
guage schemaIL based on standard logic would be pathetic, ineffectual,
indeed even quixotic. Moreover it would be self-stultifying: the very argu-
ment for rejecting them would use the concepts of truth and truthlikeness
and would presuppose their functioning - surely such an argument would
claim to be true, correct, if it were to be taken seriously.
1.6.2.4. Thus we can conclude this section by returning to our starting
point, the typology of positions on the issue of truthlikeness. I would join
H. Albert and G. Andersson, but would also conjecture that the attempts to
explicate the idea ' ... is closer to the truth than .. .' in terms of an IL will
carry with them a repetition of the degenerating problem shifts we have
witnessed in the logical positivists' attempts to explicate the concept of
'empirical significance' by means of the IL : a host of problems will be induced
by the very instruments introduced in order to solve the originall~xplicatory
problem. 16

2. CRITICAL RATIONALISM: THE PHILOSOPHICAL FRAMEWORK

OF POPPERIAN METHODOLOGY AND IDEAL OF SCIENCE

2.0. Methodology is not self-sufficient. Just a few examples. The material of

its Gedankenexperimente it gets from the history of science. It needs the
history of science in many ways. Even in theory appraisal it cannot do with-
out it if only to clarify what exactly is to be appraised. For instance 'the
Newtonian theory' refers ambiguously to a historical succession of various
formulations and to different versions having been developed in one and the
same period. Hence historical studies are an indispensable preparation for
getting started. Methodology has to import some of its tools from studies
about formalized language and other studies using formalized languages. It
interacts with, e.g., the studies of creativity, which are an interface between
history and psychology. And so forth.
Methodology is (as mentioned in Section 0.2) not independent of other
sub fields of philosophy either: every methodology is embedded in a philo-
sophical framework. On the other hand methodology interacts with some of
these other subfields: e.g., while it presupposes work in ontology, it makes
contributions to the philosophical anthropology of knowledge.
66 GERARD RADNITZKY

2.1. Critical Rationalism Rests on Three Pillars: Realism, Fallibilism and

Meliorism

2.1.1. Realism. We do not use the customary label 'critical realism' because
naive realism, the view that the world is what it appears to be, has been dis-
carded since antiquity. Realism is a necessary presupposition of methodology.
Falsification presupposes the idea of error, hence that of truth and of the idea
that one hypothesis may be a more or less accurate representation/description
of certain aspects of reality. Whether or not a hypothesis is falsified depends
on reality: the idea of an experiment is that reality 'gives an answer' which is
independent of human influenceP (To take a trivial example, whether or not
gold is heavier than iron is something that cannot be influenced by human
beings.) Ontological realism as the thesis that material entities 'exist' in the
full sense (are, inter alia because of their independence, given full ontolOgical
status) or as the thesis of the existence of the external world and of other
minds is scarcely contested. Its main support is the unattractiveness of its
denial since ontological idealism relentlessly leads to solipsism, a position,
which, even if it may be consistent, is patently absurd. Also its corollary, the
thesis that only what (physically, materially) exists can be the object of cogni-
tion for the natural sciences, appears scarcely controversial. Epistemological
realism is the thesis that at least the properties of physical entities, physical pro-
cesses and the 'reality' of a physical event are independent of any process of
cognition, in particular of observation. This sort of independence is a precon-
dition for the possibility of objective indicators of comparative truthlikeness.
One can combine epistemological idealism (the denial of epistemological
realism) with ontological realism. This position implies an instrumentalistic
view of scientific theories. It has become increasingly popular among quantum
physicists. Popperians would not distinguish 'theory realism' as a special
sort of realism since such a distinction is the result of the artificial distinction
between 'theoretical language' and 'observation languages', of the two-
language approach characteristic of positivistic-foundationalist philosophers
who wish to give 'observation sentences' or 'observation predicates' a pri-
vileged epistemological status. For Popperians a theory - this holds good for
theories of quantum mechanics no less than for Newton's gravitational
hypothesis - talks about the world and makes truth claims which are in
principle the same as those made on behalf of a data sentence deduced from
the theory in presence of suitable additional premisses. The epistemological
idealists hold that the theories of microphysics do not represent/describe
aspects of the micro-world, but are, rather, nothing but instruments for
 PROGRESS AND RATIONALITY IN RESEARCH 67

deducing testable consequences; hence they regard the theoretical system

consisting of the theory to be tested plus auxiliary hypotheses as an opaque
instrument for transforming input information into (new) output informa-
tion. They claim that it is impossible to draw a sharp line between observer
and object: micro-object, apparatus, and the observer constitute a black box
that cannot be analyzed. 18 The scatters are attributable to the entire box
rather than to the microo{)bject - the more so if the micro-world is allotted a
sort of 'reality' which is being 'created' by the observation itself.19 The con-
temporary tendency towards epistemological idealism and the instrumentalist
view of physical theory connected with it is exemplified by many famous
scientists and philosophers of physics. Recently Paul Feyerabend has also
joined the club. Thus some hold that quantum theory does not deal with
(say) elementary particles and their properties as 'existents', but only with
'experimental arrangements' (Philipp Frank); C. F. von Weizsiicker speaks of
the 'unobjectifiability' of microphysical attributes; Heisenberg writes, "The
conception of the objective reality of elementary particles has ... evaporated
. . . into the transparent clarity of a mathematics that represents no longer
the behavior of the elementary particles but rather our knowledge of this
behavior".20 If 'represents' were taken in the sense of expressing (ausdriicken) ,
the point of the passage would evaporate and the passage would be trivialized.
'Represents' must therefore be construed in the sense of describing. The
passage says that the mathematicized theories of quantum physics con-
stitute knowledge not about the behavior of elementary particles but knowl-
edge about 'our knowledge of this behavior'. This is its point. But then
according to this dictum physics or at least that part of physics has become
one of the Geisteswissenschaften. Its idealistic thrust (anti-realism) has be-
come unmistakably clear. These philosopher-physicists claim that episte-
mological idealism follows from quantum mechanics. Mario Bunge has shown
that this is not so, but that, e.g., for the empirical indeterminacy interpreta-
tion of Heisenberg's inequalities, its "only support is a positivist-philosophy
popillar in the 20's and 30's". 21
Popperians point out that an instrumentalistic view of theories totalizes
one ingredient in the testing of theories: deriving a potential falsifier from the
theory, in the presence of suitable additional premisses,22 and that if this
totalization is made, the success of a prediction becomes totally mystical:
why can it be that we deduce from the theory successful predictions if the
theory does not more or less correctly represent some of the aspects of the
reality about which the predictions are made?
I would surmise that, while epistemological idealism and the concomitant
68 GERARD RADNITZKY

instrumentalist view does not follow from quantum theory, the intellectual
motive underlying it stems from the historical situation of quantum theory.
Elementary particles have properties which do not fit in with the world-
picture, neither with that of common sense nor with that built out of the
contributions of classical theories. Hence the question 'What sort of entities
are they?' becomes disturbing. All difficulties are avoided if one holds that
the statements about the behavior of elementary particles are not descriptive,
that they are nothing but fictions to be used as instruments for predicting
what will happen in the laboratory when a certain experiment is carried out;
in the last resort they are predictions about the perceptual experiences of
the experimenters. If this gambit is adopted, the question of the ontological
status of these entities does not arise - they are but fictions and the theories
nothing but black boxes used as instruments. If so, then the results of quantum-
physical research have no repercussions on the level of world-picture. The
task to examine the mteraction between the results of quantum-physical
research and our world view is eschewed. Hence instrumentalism is a lazy
philosophy. (It may be convenient for experimental researchers if it is used
only as a short-term moratorium on metaphysical questions in the hope later
on to be better equipped to deal with them.)
For Critical Rationalism the ontological pillar, realism, is a philosophical
presupposition of methodology. The defence of the realist posit cannot be
the task of methodology because a methodologist embarking on this enter-
prise has eo ipso turned ontologist. Nor can methodology aspire to develop
ontological analyses of, say, acts of perceiving, thinking, etc. The division of
labor within philosophy requested in Section 0.2 and Section 1.5 is indeed
indispensable. But methodology gives problems to ontology: e.g. Popper's
precious insight that certain problems are literally discovered poses the
problem, for ontology, of accounting for their partial independence (of
Popper's world-3 entities), which problem is at the same time a test case for
any ontological groundplan.

2.1.2. Fallibilism. While realism remains a posit and a philosophical (input)

presupposition of Critical Rationalism, the conjecturalist approach, fallibilism,
is the result of the thorough criticism of foundationalist philosophies. Hans
Albert 23 has convincingly argued that the justificationist approach (in-
ductivism, which has led to the problem of the justification of an Inductive
Principle) leads to a trilemma: infinite regress, vicious circle or stopping the
justification procedure at some epistemologically privileged sentences (such
as the empiricists' sense data statements or observation sentences or the
 PROGRESS AND RATIONALITY IN RESEARCH 69

rationalists' various apriorisms, or the 'transcendental pragmatists' (K.-O.

Ape!) quasi-transcendental conditions of possibilities).24 But this means a
breaking off of the justification procedure at some juncture, regarded as
absolutely certain, which runs counter to the very demand of 'founding' in
a non-dogmatic manner. Fallibilism is moreover a contribution which Pop-
perian methodology makes to the refinement of our image of man (man as a
researcher) and hence to philosophical anthropology. The later Popper con-
tributes also his evolutionary theory of knowledge, which is both a generaliza-
tion of and a broader frame for his methodology. 25

2.1.3. Meliorism or cautious optimism, the means for achieving progress:

critical methodology. While realism is a metaphysical posit and fallibilism
the result ofthe criticism of the foundationalist approach to science, the label
'meliorism' could be used to express the flavor of Popperian methodology.
When the demand for absolute justification is consistently upheld, then -
this is the lesson to be learned from Albert's trilemma - prima facie the only
available position appears to be scepticism. Popper is the first to have worked
out an alternative to the pendulous movement in the history of philosophy
between unfulfillable demands such as the foundationalist demands and the
reaction to them, wholesale scepticism. The Popperian alternative holds that
empirical knowledge cannot be proved to be true, but it can be improved. We
know at least roughly what we mean by 'scientific progress'; such progress
is possible not only in principle, but is also exemplified in the history of
science; although there is no guarantee that we will be successful in the future,
there is a chance of it. With the help of Popperian methodology the chances
of realizing progress in the sense of the ideal of science outlined above (Sec-
tion 1) are better than with any other presently available methodology. This
is the bold promise of Popper and his followers. Everything depends on
whether or not we can plausibly argue for this conjecture. While realism is a
presupposition and fallibilism a result, the critical methodology is the answer
to the question 'Given realism and fallibilism, what is it rational to do in
research?'; and meliorism epitomizes the recommended attitude towards
research and methodology. The criticistic methodology is a general theory
of rational (purposive-rational) action. It has been generalized from meth-
odology in the narrow sense and it has research as its paradigmatic field of
application. But it is claimed to be applicable in principle at least to all sorts
of problems, not only to problems associated with knowledge production.
The core of the critical approach may best be expressed in Popper's own
words: It is "the general idea of intersubjective criticism, or in other words,
70 GERARD RADNITZKY

of the idea of mutual rational control by critical discussion."26 This general

idea can be explicated (in the etymological sense of that word) by several
interrelated principles. In Figure 1 they have been sketched by a few major
rules: the 'master rule' bans all immunization strategies; it functions much
like a meta-rule stipulating that no rules may be used that would prevent
discorroboration, falsification. It is the core of the so-called demarcation
criterion. From it follows the 'basic operational rule', which prescribes
'severe' testing not only for conjectures, but also for falsifiers, for proce-
dures, etc. Derived from it is a 'preference rule'. It has two aspects: Prefer
theories with higher content to those with less content, and, after testing,
prefer, ceteris paribus, that theory which has the higher degree of corrobora-
tion. From fallibilism follows moreover the rule indicated in Figure 1 as
'revision clause'. Since there is no epistemological rock bottom, it applies to
all components of the scientific enterprise including data sentences, and as,
therefore, falsification cannot be conclusive either (since one of the premisses
in the argument is not conclusive), falsification must not be exempted from
possible revision. These global rules are to give some guidance to the research
process. In Part II we will attempt to spell out in some detail how, according
to Popperian methodology, research should proceed in order to facilitate
achieving cognitive progress in the sense of Popper's explicatum of our
intuitive ideas of cognitive progress.
Before embarking upon this outline a final remark about the realm of
application of the critical method: it is far wider than science. The critical
method is used to distinguish rational from non-rational procedures: a pro-
cedure is rational if and only if it adopts the critical policy. If from a particular
statement a consequence has been deduced that is 'unacceptable', then this
statement has been criticized to that extent. This concept of criticism is
applicable also to normative-evaluative issues (except 'ultimate values', which
for the believer, are by definition exempted from criticism so that statements
about them do not form part of purposive-rational activity).
Within the rational manners of proceeding we find a descriptive distinction
between science and non-science. To draw this distinction is a problem of
explication: for certain goals - and we shall shortly return to what these are
- the intuitive idea of science is to be replaced by a concept of science that
is a better instrument for these particular goals. The intuitive idea of science,
our concept to be explicated, is partially defined by the goal of the activity,
cognitive progress: Cognitive progress is the goal of empirical enquiry in
general; scientific research is that empirical enquiry that can demonstrate at
least a minimum amount of method. Since the goal of research is cognitive
 PROGRESS AND RATIONALITY IN RESEARCH 71

progress, i.e., improvement, expansion and deepening of our knowledge about

empirical reality, within scientific research the critical method or policy must
include empirical criticism as an essential component.
This has two consequences. (1) The demand that theories must be tested,
are to be subjected to empirical criticism, makes sense only if the theories are
falsifiable. That falsifiability thus constitutes a component of the explicated
concept of science is therefore a corollary of the insight that in empirical
research the critical method must essentially contain empirical criticism.
(2) Insofar as science is prinlarily seen as an activity, as research, methods
and strategies are more important than theories. Now it is possible to interpret
most theories so that under this interpretation they are falsifiable. But a
falsifiable theory can always be rescued from a falsification by adding ad hoc
hypotheses. From this it follows that a general method, a policy, is scientific
if and only if auxiliary hypotheses are not introduced ad hoc or, if such an
introduction is expressly declared to be a temporary, purely heuristic measure,
then the method is scientific if and only if these hypotheses are retained only
if they lose their ad hoc character. For this reason the question, "When is
introducing an auxiliary hypothesis ad hoc allowed, and when is it illegitimate
to retain an additional hypothesis, which was originally introduced ad hoc as
a temporary heuristic expedient?" is a topical problem for every meth-
odology. Popper's answer can be summarized as follows: (i) Introducing an
additional hypothesis ad hoc is illegitimate if this is done to preserve the
theory from falsification and if the price to be paid for this is a decrease of
the theory's empirical content, that is, of the information contained in the
class of potential falsifiers. For this reason the scientific method dictates that
a potential falsifier must be specified in advance: that one must be able to say
what kind of experimental result or observation one would recognize as
falsifying the theory. The as yet unsolved difficulty consists in defining 'ad
hoc' objectively - to speak of the intention of the researcher would be to
lapse back into psychologism. (ii) But the salient point is whether an auxiliary
hypothesis which was originally introduced ad hoc as a heuristic expedient
but without reducing the empirical content of the theory is retained even if
there is no reason to assume that it will be testable independently of the
theory and will stand up to such a test. The circumstance alone that such an
ad hoc auxiliary hypothesis is falsifiable is insufficient. And so the core of the
scientific method as critical method ala Popper is the prohibition of immuni-
zation strategies. The demand for the falsifiability of scientific theories is
only a corollary of the requirement of the method of empirical criticism. The
critical method is for still another reason more important than falsifiability. A
72 GERARD RADNITZKY

non-falsifiable hypothesis belongs to the realms of non-science. This assertion

is descriptive, not evaluative. But if a hypothesis claims scientificality and is
simultaneously immunized against falsification, then it is not only non-science
but also pseudo-science. This is a negative evaluation, and this version of the
solution of the demarcation problem has an important function in political
debate and critique of ideologies.
This answer of Popperian methodology to the question of what scientific
method is relieves us of the task of first having to indicate what is meant by
'science' before being able to reflect on 'scientific method'. The critical
method including empirical criticism is the distinguishing feature of science,
since it is the core of the 'scientific method'. The prohibition of immuniza-
tion methods plays an important role in research. The demand that theories
be falsifiable results as a precondition for the realizability of the method of
empirical criticism. In the context of comparing theories, on the other hand,
falsifiability hardly plays a role, since the researcher is almost never faced
with the task of choosing one theory from a pair, one of which is unfalsifiable,
that is, has no empirical content at all.
Outside methodology, it is very important within non-science to separate
out pseudo-science, because this is indispensable in combatting the pollution
of the intellectual environment by theories masquerading as science although
not falsifiable, and it is still more important to unmask a policy that im-
munizes falsified theories against acknowledging their falsification. The
importance of the demarcation criterion in the political context can scarcely
be overrated, considering how rewarding it is for a propagandist if he succeeds
in having an ideological, non-scientific doctrine illegitimately profit from the
prestige of science. No wonder communists insist on the title of 'scientific
socialism'. At the University of Moscow, there even exists a chair for 'scien-
tific atheism', although, of course the theme of atheism forms part of theo-
logical inquiry no less than that of theism, and thus is in principle outside
the realm of empirical inquiry. That such inquiry is non-science does not
speak against it; only when it pretends to be empirical science - as 'scientific
atheism' or 'scientific theism' - does it turn into pseudo-science.

PART II. OUTLINE OF POPPERIAN METHODOLOGY

o. Among other things, I wish to make good the claim that Popperian meth-
odology implies pluralism rather than pure falsificationism, hence that it is
not a form of 'logicism'. Two diagrams will be used as a means of exposition.
 PROGRESS AND RATIONALITY IN RESEARCH 73

Their function is twofold: first, to portray basic features of the methodology,

and also to identify problems for the methodologist. The figures are, of
course, dispensable, but are a convenient, space-saving device. Methodology,
as conceived here (part I), consists largely of hypothetical imperatives.
According to the Popperian methodology, a research enterprise's chances of
success are increased if the project is governed by these rules. The idea of
success, cognitive progress, in the 'game of science' is explicated by filling out
the ideal of science outlined in Part I. The 'cybernetic' models of research
pictured by the two figures represent the gambits and moves recommended
by the general rules (such as the master-rule of anti-conventionalism, the
falsification rule and the preference rule) that follow from the explicatum of
cognitive progress obtained from the ideal of science (N and NW parts of
Figure 1). Since these methodological gambits and moves are intended to
apply to any research enterprise worthy of the name, they must necessarily
be schematic. 27 Applying them in a concrete research enterprise will involve
taking the substantive preconceptions and presuppositions behind that enter-
prise into account. As in everything else, in any concrete research enterprise
there are certain substantive presumptions. Of primary importance here are
presuppositions about the general nature of the object(s) under investigation
(cosmological hypotheses) and a programmatic definition stipulating how the
discipline ought to look ideally, a program that arises from applying the
general ideal of science to this particular discipline. (In Figure 1 the discipline
investigating a realm of objects X has been labelled 'X', the 'X-ology' in ques-
tion.) For this ensemble of cognitive (in part 'metaphysical') and prescriptive
components we have proposed the umbrella title of 'internal steering factors'
of the research enterprise. 28 (NW part of Figure 1, abbreviated as 'ISF'.) If
only by supplying criteria for appraising products,29 the internal steering
factors (which are close to a Kuhnian paradigm in one of its senses or the
'hard core' of a Lakatosian research program) give a group of various research
enterprises a certain unity and direction, gather them together into a research
direction, tradition, school, style of thought.

1. CONJECTURES AND CRITICISM ('ENTWURFE UND

UBERPRUFUNGEN'), VARIATION AND SELECTIVE RETENTION

1.0. According to Popper "science begins with problems".30 Of course every

question has its presuppositions, is askable, formulable only if some particular
knowledge is presumed. Which comes first? This is the old question of the
temporal priority of the chicken or the egg, or in philosophical parlance 'the
74 GERARD RADNITZKY

hermeneutics of the question'. With respect to science, Popper sees successful

movement as a progression from problems to deeper problems (cf. above Part
I, Section 1.S). This process does not have any defInite beginning nor end. It
is an unending quest: there cannot be any internal limits to science since
every solved problem creates new problems (the Kant-Popper thesis of prob-
lem propagation (Part I, Section 1.5)). Yet a problem may be suitably used to
mark the beginning of a concrete research enterprise or even of a research
tradition. In the literature Popperians are often accused of not considering
the problem of where the problems come from because they concentrate only
on criticism. We will argue that this reproach is unjustifIed. However, we want
to point out immediately that the setting of priorities for sorts of problems
(e.g. whether efforts should be concentrated on space research or molecular
biology - just to take an example), on ranking disciplines for funding, etc.,
is a question of global research policy and not a problem of methodology.
Methodological considerations presuppose that the task is cognitive progress
in the discipline concerned.
1.1. Where Does the Problem Come From?
From where does one get the problem that sets a research enterprise in motion?
There always arises need for an explanation: our 'pre-existing knowledge', no
matter whether it consists of commonsensical background knowledge, which
contains the precursors of scientific theories, or consists of scientific theories,
is bound, sooner or later, to run into an observation it cannot explain. (In
Fig. 1 the arrow from Ti to Pi.) This situation can become critical, namely
when the observation contradicts the pre-existing knowledge or the pertinent
theory. (In Fig. 1 this is represented by the arrow 'falsification' pointing at
Pj.) The situation may be dramatic when the reigning theory encounters
'anomalies', when it either clashes manifestly with some experimental result
or 'runs out of steam', as would be the case if, e.g., a set of 'observational' law
hypotheses grows, which resists all attempts at explanation by means of the
dominant theory, while an approach to explaining them by means of a newly
conceived theory contradicts the reigning theory. An example might be
spectroscopic laws before 1913. 31
But our theories do not even have to have run into diffIculties. Even if
they are highly successful, there automatically arises a problem of explana-
tion: to explain the key components in the explanans of the successful
explanation, the theory by means of which we could achieve these explana-
tions. This problem too is an objective problem: the researcher frods himself
confronted with it or discovers it independently of his wishes and plans; it
 PROGRESS AND RATIONALITY IN RESEARCH 75

MODEL OF RESEARCH STRATEGY OF SCIENTIFIC PROGRESS

f
IDEAL OF SCIENCE
/_.---..... EXPLICATION OF
/' CRITICAL,
I SCIENTIFIC PROGRESS
( RULE SYSTEM /" RATIONALISM
meth~ology ) /'."" REAL'SM I
(
/·~nlj..................... FALLIBILISMI MORE AND MORE ACCUAATE
< conventionalism
.ppriliUI

:,;- REPRESENTATION
. / .......................... ,/"/

/ .. -. . . . . . . <: / ' ......... .......... INCREASE IN

t.tsifiCition '> EMPIRICAL CONTENT
/' ', ......... rule /'"
............... /
<,
/'

,
'''''''''''''>
lUll /'/
y
I
......., . / ' I
Y i
world picture hVPQlhne'S ("metil phYI'CS'"

BACKGRO"ND
KNOWLEDGE
scienullc thrones about rewarch irei X:X

~.-1,-, L_____ , ____ _

i_ i+l i i CONJECTURES AND FALSIFICATION ATTEMPS
RESEARCH
PROCESS

'l
~
1-+-1 -+-"'- ---~
I 1
, <p
I I ··1 P il
I 1
r-
r
-----n:iEccoC-Rl--A,."','-.A-IS-A-L----++-.

I 1

1
....,...
phil_ion

I
I
IL..-~)""
I
I
-'.........
cl'it:icitm

I
B i_i+1

I
I
I
I FALSIFICATION """.. 'oFH

~----------
CORROBORATION ..'...... ------- 0 FH

x x x x
Fig. 1.
(This figure copyright © 1978, G. Radnitzky.)
76 GERARD RADNITZKY

arises as an unintended consequence of his research activity. This independ-

ence of the majority of problems is itself a compelling argument for attribut-
ing ontological status to them, as entities of Popper's world-3. 32
In particular, one will attempt to extend a successful theory to new realms
of application. And methodology also demands this. The 'preference rule'
whose first part advises us before testing to prefer a theory with more content
over rivals with less, urges us to generalize the successful theory. Sooner or
later this will very likely have the result that the theory thus generalized runs
into difficulties; it may get falsified, something that may eventually lead to an
improvement of it, e.g. by making possible a more precise delineation of its
range of application. (In Figure I this is represented by the arrow from the
'preference rule' to Pd Of course, the 'falsification rule' also demands that
we continue to devise severe tests for the theory that hitherto has been
successful - last but not least so that we can improve it, e.g. in the way just
mentioned. These two 'rules' are intertwined since severe testing presupposes
a high degree of testability, Le. of empirical content. In sum, in Figure 1 the
two arrows converging on Pi depict the two main sources of problems. Once a
research enterprise has been embarked on in order to solve such a problem, a
host of other problems arises on the way to its solution. Some of them are
portrayed in Figures 1 and 2. These ensuing problems are obviously objective
in that they come as unintended consequences of the decision to embark on
that research enterprise.

1.2. Where Does the 'Predecessor Theory' Come From?

In Figure 1 only the two limiting ideal-typical cases have been portrayed:
either the theory is one of the fairly well-developed scientific theories, at best
the reigning champion theory of the discipline concerned (in Figure I, i); or
it is embedded in the intuitive background knowledge, Le., on the basis of
preconceptions and presuppositions about the subject under investigation
(in Figure I, X), an attempt has been made to formulate an approach towards
a theory, which in the beginning is naturally rudimentary. Thus in those cases
where the field of study has just been opened up, in a hitherto unexplored
realm of phenomena, 'metaphysics', cosmological hypotheses (in Figure I,
ISF) are structurally indispensable, and not merely indispensable as a part
of psychological heuristics (world-2).33

1.3. Where Does the Successor Theory Come From?

In one sense, everything is permitted in hypothesis generation since the critical
 PROGRESS AND RATIONALITY IN RESEARCH 77

appraisal of the output comes afterwards, i.e. intermittently during the various
stages of interim products. Only the structural characteristics of hypothesis
generation are the concern of the methodologist. In Part I (Section 0.2) we
mentioned that Popperians by no means treat the hypothesis generation as a
black box whose study is to be left to the psychologist. Rather, psychological
investigation or hypothesis formation needs the methodological study of
structural characteristics (world-3) before getting off the ground. 34
In Fig. 2 (p. 82) hypothesis generation is represented by a box marked
'HG'. This box has inputs from all levels of background knowledge. Hence its
operation must include a selector. In Figure 2 this is symbolized by a fre-
quency fIlter. The selection will be governed by the prior assumptions about
the general nature of the phenomena studied. In Figure 2 this is indicated by
the 'internal steering- factors' JSF (whose operation has been mentioned in
Section 1.0). When working within a research tradition, a selection criterion
for both input assumptions and output hypotheses will be their coherence with
the cosmological hypotheses of the tradition in question. When a dramatic
shift in perspective is in the offmg, the assumptions about the phenomena
under investigation may conflict with the cosmological assumptions of the
reigning tradition since an output theory (Ti+ 1) may conflict with the pre-
decessor theory.
A pOSSIble source of input into the 'hypothesis-formation box' are theories
about other sorts of phenomena, which may belong to the same or to another
discipline. Hence in Figure 1 this input is shown as coming from Y. (i = abbre-
viation of the diScipline studying the realm X, the X-ology in question, Y=
abbreviation for studies other than the one in which the theory under consi-
deration is developed.) Example: Bohr conjectures a possible analogy between
atoms (a realm about which no scientific knowledge existed at the time) and
the planetary system (a realm about which full-fledged scientific theories
were extant). Such a guess would belong to what is here labelled 'JSF'. It
governs input from the extant theories about planetary motion into HG to
generate hypotheses about electrons, which are then subjected to falsification
attempts. One expects that if there is an analogy there will also be negative
aspects to it. Whether there are, and, if so, how far the analogy can be carried
must be found out by empirical testing. In this way a basic conjecture of
analogy provides a structural heuristic governing selection of input from
neighboring scientific theories as well as experiments.
The overall operational characteristics of hypothesis formation (HG box)
will be governed by methodological rules; in particular by the 'preference
rule' (cf. Figure 1), because in order to make severe testing possible the out-
78 GERARD RADNITZKY

put theory should have as much empirical content as possible: preferred are
'daring conjectures'.
We said above that in hypothesis formation 'everything goes' since criticism
follows upon each interim product. (In actual research there is a continuous
interplay between hypothesis formation and criticism - which cannot be
mirrored by Figure 1.) However, inductive procedures will not be used - not
even on the level of mental processes (world-2). This of course is a question
for the psychologist to investigate, an interface between psychology of re-
search and history of science. The methodological considerations strongly
suggest that in mental processes a number of observed cases will only be
recognizable as similar if at least an implicit, perhaps often subconscious
conjecture has been made that they are entities that have something in com-
mon. That means that hypotheSis formation as a mental process will never
proceed in a way that would correspond to the algorithmic functioning of a
Baconian inductive machine. (That is why Popper puts 'observational' law
between shudder quotes; as there are no theory-independent descriptive state-
ments, so there are no inductively, 'purely observationally-experimentally'
produced law hypotheses.)
1.4. Appraisal of the Successor Theory
How is the output of the hypothesis-generation-station to be subjected to
criticism, appraised? Most commentators regard this as the core of Popperian
methodology. We hope to have made it apparent that the critical element is,
however, comprehensible only as an integral part of the prescriptive picture
of research. Obviously appraisal makes sense only in the light of a guiding
ideal of science, an explicatum of progress. This ground has been covered in
Part I. Figure 1 gives a skeleton diagram of research, Figure 2 a blow-up of the
component theory appraisal.
The output of the hypothesis-generator automatically creates an objective
problem: to examine the output critically, to appraise the successor theory.
This criticism is basically a falsification attempt. It has two logical steps. To
answer the question, 'What exactly does the theory say?', we have to deduce
testable consequences, and in particular, consequences which enable us to test
the theory severely. This stage is represented in Figure I by the left box within
the box marked 'theory appraisal'. Then we have to answer the question, 'Is
what T says true?', by ascertaining whether or not it stands up to the empiri-
cal tests made possible by the consequences deduced in step one. In Figure 1
the second stage is represented by the box marked 'HC' (as short for 'hypoth-
esis-checking/control'). The HC-box is provided on the output side with a
 PROGRESS AND RATIONALITY IN RESEARCH 79

decision triangle. The plus and the minus signs indicate the answer decided
upon to the question, 'Did the attempts to falsify the potential falsifier (de-
scribing an individuated event) succeed or not?' If it has not succeeded then
we have a falsifying hypothesis (describing a reproducible effect, which has a
higher level of generality than the potential falsifier, which is a singular state-
ment). In both figures the falsifying hypothesis is symbolized as 'FH'. If we
accept pro tempore the falsifying hypothesis, then there must be at least one
false premiss in the argument for the consequences whose negation constituted
the potential falsifier. This again automatically creates an objective problem;
the so-called Duhem-problem: on which of the premisses to put the blame.
In Figure 1 this is represented by a decision triangle at the output side of
the theory-appraisal box. The minus and the plus signs indicate the possible
answers to the question, 'Should the blame for the falsification be put on
premisses other than the theory under test?' If this question is answered in
the negative, the theory is falsified - relative to the falsifying hypothesis,
that is, provided FH is not problematized. Since all empirical hypotheses are
fallible, revision of FH is also in principle possible. Falsification cannot be
conclusive since at least some of the premisses are not conclusive. Figure 1
renders this circumstance by means of an arrow from the 'revision clause'.
In Figure 1 two clear-cut cases are portrayed. In one, the falsification
attempt has succeeded - which automatically produces a new objective
problem: Pi has been transformed into Pi+ 1. With the new problem a new
turn in the research enterprise begins, an iteration of the feed-back loop that
has just been completed. But the new problem is more advanced, if only
because one of the inputs into the HG-station is now the knowledge gained
by the preceding investigation, including the falsification itself. This novel
knowledge can help to refine the theory falsified, or, if you please, help to
produce a refmed version of it. In the second case, the falsification attempt
has miscarried. Here the methodology advises us to regard Ti+1 as corrobo-
rated to some degree - the degree of corroboration depending upon the
severity of the testing - and to retain it until further notice, i.e., to continue
to test it, to generalize it - but above all to regard it as preferable to the pre-
decessor theory, at least in this respect.
1.5. Possible Repercussions on the 'Cosmological Hypotheses' on the World-
Picture Hypotheses
Figure 1 indicates also that the replacement of the predecessor by the suc-
cessor theory may have repercussions on our 'background knowledge.' It
may have philosophical implications which feed back to the 'cosmological
80 GERARD RADNITZKY

hypotheses'. In connection with a great scientific advance, where according to

Popper the successor theory contradicts the predecessor, changes will have
to be made in the assumptions at the level of the world-picture. Since the
theory contradicts its predecessor it may tum out that in its construction
the researchers presupposed (perhaps without being clearly aware of it)
cosmological hypotheses that are incompatible with those underlying the
predecessor theory, with parts of its [SF. In Part I it was suggested therefore
that the 'scientific importance' of a question was maximal if successfully
answering it would lead to a substantial change in our cosmological assump-
tions, in what hitherto we had taken for granted. If you prefer to use terms
of the philosophic tradition, you may, instead of speaking of a feedback
circle, say that this illustrates an aspect of the so-called hermeneutic circle
or spiral,35 a (non-vicious) 'circle' in the sense that the existence of certain
presuppositions in its original, less adequate state was a prerequisite for
developing the successor theory and therefore for refining, improving the
preconceptions of the cosmological hypotheses, in particular by eliminating
false items from them. These improved cosmological hypotheses will then
form the 'internal steering factors' for the next tum in the scientific enter-
prise, in the unending quest. Thus even the 'metaphysical' components among
the consequences of the successor theory or the cosmological presumptions
underlying it are not immune to criticism emanating from empirieal investiga-
tions, not totally immune to scientific developments, even if they are more
resistant to change than the other components of the 'body' of knowledge at
a given point in time. If this were not so, one would risk abandoning a set of
'internal steering factors' long before it had given a substantial part of the
dividends in novel knowledge derivable from it.

2. THE COMPONENT OF THEORY APPRAISAL MORE CLOSELY

EXAMINED
2. O. On the Logical Structure of the Move of Falsification
In the literature it is usually regarded as simply modus tollens. The sentence
'All swans are white' is indeed falsified by 'a is a swan (statement of 'initial
conditions J) and a is not-white'; 'J and not-P' is a potential falsifier for 'All
J are P'. (Incidentally, as Mario Bunge often has pointed out, of interest for
methodology are scientific statements (e.g. Mendel's laws) but not 'Swans are
white'. To use such sentences even for purposes of illustrating logical moves,
in my opinion, carries with it a certain danger of 'model Platonism'.) For the
hypothesis, 'The orbit of planet M is a circle', a potential falsifier may be a
 PROGRESS AND RATIONALITY IN RESEARCH 81

conjunction of four data sentences each describing an individuated event and

based upon an observation of the planet's position. In this case the statement
of initial conditions would be trivial, e.g. 'The heavenly body observed ... is
planet M'. From a logical point of view, to deduce not-T or not-A from 1- (T
& A) -+ (J -+ P) we need only J, P*, and 1- p* -+ not-Po In actual research
practice, however, the logical structure of the move of falsification often
consists of two steps, a modus ponens part followed by a modus tollens part.
(Its conclusion - this cannot be stressed too much - can never be the falsifi-
cation of the theory tested, but only a disjunction.) The modus ponens part
is portrayed (in Figure 2 in the left box up to but excluding the box 'negat-
ing'). Its structure is basically this. From the theory T and auxiliary hypo-
theses follows a testable consequence (in Figure 2, Hj).36 The following
empirical premisses are posited hypothetically: T, the theory to be tested, A,
the conjunction of the auxiliary hypotheses we assume pro tempore, and J,
the initial conditjons we intend to produce in the experimental situation. The
conclusion is P, a hypothesis describing a reproducible effect (Vorgang); it
may then be particularized so as to describe a single event (in Figure 2, dj).
In historical studies the empirical testing often concerns a singular sen-
tence, and the use of the modus ponens part is explicit. Consider the follow-
ing simple example. An archaeologist makes the conjecture that the house he
is about to excavate was the house of a physician (J). How can he test J
empirically? He may assume a theory about the way physicians' houses were
built during the period concerned (T). T is not tested in the argument at
hand, but used. From T and J in the presence of auxiliary hypotheses, a pre-
diction, e.g. that the distance between two particular walls will be such-and-
such (P). From P we derive, in the presence of singular statements about
initial conditions (i), e.g. that we proceed with the excavation in a certain
way and, eventually we derive a data sentence, e.g. that the archaeologist
will find a particular str-ucture of stones in a particular location (d). Then he
looks at the fact, i.e. d is empirically tested.
The ensuing modus tollens part (right box in Figure 2) has basically the
following structure. We cannot directly get the negation of the deduced
consequences (not-P) as an experimental result; what we get is something
positive, a description of a certain type of initial condition J together with a
deSCription of a certain type of reproducible effect P *. The conjunction J
and p* has been experimentally established, with due regard for the principle
of the fallibility of all empirical knowledge, of course. If P and p* are incom-
patible, then from p* not-P is deducible, hence not-P follows. From J and
not-P not-(T and A) can be deduced; hence not-T or not-A, or both. This
 82 GERARD RADNITZKY

THEORY APPRAISAL
,/
"- '-

r ---.----- --<./'
;:.....
preference -........ '>
rule ,/
<Jl '- ,/
I V

.i
"BACKGROUND KNOWLEDGE"

cp-

x x x y
Fig. 2.
(This figure copyright © 1978, G. Radnitzky.)
 PROGRESS AND RATIONALITY IN RESEARCH 83

leaves us with the options already mentioned above (in Section 1.4) as the
so-called Duhem-problem. We will postpone discussing it until Section 2.3.

2.1. aose-up on the Information-Theoretical Part of Theory Appraisal

2.1.0. Deducing testable consequences: to get a potential falsifier which will

open the way for 'severe' testing. The operational characteristics of the stage
of obtaining a potential falsifier (Falsifikationsm6glichkeit) - in Figure 2,
the left box - are governed by the 'principle of falsification or of severe test-
ing'. In Figure 2 this is indicated by 'PST' as short for 'the problem of severe
testing', the problem which regulates the selector at the output side of the
theoretical component of appraisal. This oversimplifies the role of PST' but
the diagram would be illegible otherwise. Of course, this problem already
guides the deduction of testable consequences from the very start (in Figure
2, of the predictions HI, ... ,Hn). The point is to deduce consequences that
constitute 'severe' testing, and for the sake of simplicity a triangle is used;
there should really be a frequency fIlter ('Schmalbandfilter') here, since a
test is a test only if its outcome is not known in advance. Hence deducing a
hypothesis which already has evidential support independently of the argu-
ment in question, in a pragmatic way of speaking, 'an explanation', cannot
constitute a test; nor can the repetition of the same test constitute a test in
the full sense. In short, for an experiment to be a test, it must be about a
deduced hypothesis, which contains information that goes beyond that
carried by the 'background knowledge' or the predecessor theory. In Part I,
Section 1.4 it was mentioned that, intuitively, a test is maximally severe if the
potentially falsifying hypothesis, by contradicting one of the consequences
deduced from the predecessor theory, conflicts with that theory. To clarify
the idea of 'severe testing', the notion on which it hinges, the concept of
'background knowledge', has to be explicated.

2.1.1. When can a successful prediction playa role in evaluating a theory? 37

On the problem of explicating 'background knowledge'. We here disregard
the problem of determining when a prediction is successful and attempt to
concentrate on the information-theoretical aspects of appraisal. What we
want is a rule that tells us when it is rational to credit a theory with a success-
ful prediction, and when not. 38 The rule must be such that it prevents the
theory from being immunized against possible failure by the simple device of
making deducible from it only such hypotheses as are already 'known'. i.e.,
which already have evidential support of their own before the deduction is
84 GERARD RADNITZKY

made. Successful explanation is not a test of the theory. The idea of 'severity'
is built into the idea of a test: to constitute a test of the theory, the informa-
tion of the deduced hypothesis must go beyond what we claim to know on
the basis of background knowledge.
The intuitive idea of background knowledge is roughly that it is all puta-
tive knowledge available at a certain time. In Figure 1 this is represented by
the horizontal beam which is labelled in the right margin 'background knowl-
edge'. In Figure 1 various inputs from this background knowledge are por-
trayed, with the corresponding selection mechanisms - for in a particular
research enterprise only a small segment of 'what is known or assumed to be
known' will be relevant. Moreover what is judged relevant will vary with the
research tradition and of course with time, with the state of the discipline.
This maximal interpretation of 'background knowledge' is seen to be proble-
matic when we remember that one desideratum is 'boldness' of theories, i.e.,
that a theory should make predictions which are 'unlikely' in the light of our
expectations based on what we think we know. If it is successful, we may say
'based on what we thought we knew'. Perhaps no more should be required
than the following: background knowledge is the set of those theses which
(are relevant for the scientific problem at hand and which) there is at present
no reason to doubt. If more is required, viz. that we have reason not to doubt
them, we seem to get entangled in circularity: if we were to require that they
be theses that so far have a sufficiently high degree of corroboration, an expli-
cation of 'degree of corroboration' and hence of 'severity of testing' and
hence of 'background knowledge' would already have to be at our disposal.
While the 'maximal' interpretation of 'background knowledge' appears
problematic, the minimal interpretation is unproblematic. Testable conse-
quences are deduced from the theory in the presence of auxiliary hypotheses
and statements of initial conditions. The minimal construal of our intuitive
idea of background knowledge explicates it as the information of the addi-
tional premisses, which at present we have no reason to doubt and are neces-
sary for deducing testable consequences from the theory, i.e., for deducing
consequences that constitute at least potentially an answer to our current
scientific problems. 39 In Figures 1 and 2 background knowledge in this
minimal interpretation is depicted as the output from the various selectors
whose input comes from the beam labelled 'background knowledge'. The
requirements that the consequences not be deducible from the additional
premisses alone, Le. that the theory's empirical information go beyond that
of the additional premisses, is obviously needed since we want to test the
theory, to extract information from it and not from the other premisses,
 PROGRESS AND RATIONALITY IN RESEARCH 85

which function in the argument somewhat like a catalyzer. However, it would

be best to avoid the term 'background knowledge' here and to speak explicitly
of the information conveyed by the additional premisses.
A hypothesis we have no reason to doubt at present (a 'fact') cannot count
as a test for the theory if its information is necessary for constructing the
theory, necessary in order to formulate T. (In Figure 1 this information is
portrayed by the input into the box HG, hypothesis-generation.) Again the
rationale behind this rule is that the outcome of such a 'test' would be known
in advance. J. Worrall and E. Zahar have elaborated this 'heuristic' construal
of background knowledge as the idea that 'a fact must not be used twice'. On
the other hand, the genesis of the theory is irrelevant to its appraisal. If the
theory contains more empirical information than that of the total input into
the theory-generator (HG-box in Figure 1), then this new, additional informa-
tion can certainly, if extracted in a suitable form, be the stuff for a severe
test. In my opinion the above rule is justifiable only if the restriction is
judiciously used: to prevent an allegedly new theory from merely summariz-
ing the observations we have made, the 'data base' on hand.
However, since we are primarily interested in ascertaining whether or not
the successor theory constitutes cognitive progress, it appears best to focus on
the comparison of a successor and a predecessor theory, i.e. on the situation
where there already exists a scientific predecessor theory and not just vague
commonsensical knowledge. (In Figure 1 portrayed by the arrow from i, the
discipline in question, to Ti, the predecessor theory.) The foil against which
the new theory, together with the additional premisses necessary for deducing
a potential answer to one of our current scientific problems, operates is con-
stituted by the system consisting of the rival theory and its auxiliary premisses.
If the answers to this problem given by the two competing systems are incom-
patible, then the corresponding empirical test will be an experimentum crncis.
In this situation we should be able to appraise the comparative achievement
of the two competing systems, although of course not of the two competing
theories alone. I. Lakatos and A. Musgrave have focussed on this situation.
Uke them I would place it in the center, but abstain from using the expres-
sion 'background knowledge' in this context.

2.l.2. Deducing testable consequences: explanation vs. prediction. The first

step in the logical part of theory appraisal is deducing testable consequences,
i.e., extracting information from the theory in chunks suitable for actual
empirical testing. This is necessary because the theory itself condenses so
much empirical information that it is not feasible to test it directly. In Figure
86 GERARD RADNITZKY

2 this stage is represented by the box marked 'deducing'. Its output is a grow-
ing system of deductive arguments such that the premisses of each argument
have a common component: the theory (Ti+l in Figure 2). The conclusions
are hypotheses that describe a reproducible effect (Vorgang). The arguments
may be either explanatory or predictive. In Figure 2 one pattern T and A -+
Ho ,40 portrays an explanation; the other patterns predictions (HI," . ,
Hn)·
In an explanatory argument the conclusion is 'known', i.e. it has evidential
support independent of this argument. An interesting case is that in which,
although the pattern is still an explanatory one, some new knowledge is gen-
erated because the original explanandum gets corrected by the explanation.
Popper has emphasized this most significant phenomenon of the improve-
ment of an 'observational' law in explaining it with a more universal law or
theory since at least 1941,41 and has also pointed out its close link to the
idea of depth: " ... whenever ... a new theory of a higher level of universality
successfully explains some older theory by correcting it, then this is a sure
sign that the new theory has penetrated deeper than the old ones".42
This improvement by explanation is best dealt with by means of a simple
example. Let the explanandum E by Galileo's law of free fall, i.e. his conjec-
ture based on experiments and on many thought experiments that the accele-
ration of a freely falling body in a vacuum on earth is constant. Let Ho in
Figure 2 portray the Newtonian version of Galileo's law. Ho contradicts E.
But within a certain realm, for falls where the height above the surface of the
earth is negligible in comparison with the earth's radius, the numerical values
of E are so close to those of Ho that this constitutes a good reason for con-
sidering E mathematically (and empirically) as an approximation of H o , the
hypothesis deduced by means of Newton's theroy, or Ho as an improved
successor of E. Although we say that Newton's theory explains Galileo's,
what is 'assimilated' into Newton's theory as one of its consequences can only
be the improved successor hypothesis of E. This is still more obvious if we
consider Einstein's theory and Newton's. For a certain realm the consequences
deduced by means of Newton's theory may be considered mathematically
(and empirically) as approximations of the consequences deduced by means
of Einstein's. But this relation of approximation holds only for certain
deduced consequences of both theories. It would be inappropriate to take
this as a good reason for considering Newton's theory, its basic theses, as an
approximation to Einstein's since the latter introduces new concepts, new
ways of looking at the world that are incompatible with the Newtonian
perspective (Part I, Section 1.4).
 PROGRESS AND RA TIONALITY IN RESEARCH 87

The hypotheses HI, ... ,Hn differ from Ho only in that when the deduc-
tions are made they do not possess any evidential support. They constitute
potential knowledge, so to speak. The respective deductive arguments yield
virtual predictions: if Ti+ 1 is a conjecture that has not yet been tested, the
consequences deduced by means of it cannot profit from the theory's degree
of corroboration. But in the present context the predictions function primarily
to test the theory and we are only secondarily interested in the pieces of
novel knowledge they may convey. Assuming that we do not have reasons now
to problematize the additional premisses, the negation of Hi, viz. J and not
P constitutes a potentially falsifying hypothesis for Ti+l.
How can we find out whether the negation of H can be generalized to a
falsifying hypothesis describing a reproducible effect? Even HI , describing
a reproducible effect, condenses too much empirical information to permit
testing it directly. Thus again from HI in the presence of statements of initial
conditions (singular sentences) consequences are deduced, hypotheses of such
a low degree of generality that we can directly confront them with the 'facts'.
We can call them 'data' sentences; they have singular form and describe an
individuated event. Such a data sentence is not a report of a perceptual
experience. Reports about perceptual experience (e.g. about my own or
NN's perceiving a certain pointer reading) do, as Popper long ago pointed
out, form part of our reasons for accepting a certain data sentence, i.e. for
seeing no point in questioning its correctness for the time being. Other parts
needed in these good reasons will involve a theory of the hardware instru-
ments used, of their functioning and the causal link between object of study
and measuring apparatus, a theory of perception, and a theory of communica-
tion, since more than one observer is involved and the findings have to be
formulated in language to be communicable at all. We shall return to this
presently.
At this point the process of deduction ends. By negating the deduced data
sentence and combining it with the statement of initial conditions (J & -,P)
we get a potential falsifier (Falsijikationsmoglichkeit) for H. The potential
falsifier describes an individuated instance of a kind of state of affairs
(Zustand) or process outlawed by the law hypothesis H. If even a single
individual belonging to its realm of application exemplifies that state of
affairs or effect, then the law hypothesis is falsified unless the blame is put on
a different premiss (or premisses). Although this is so in logic, in actual re-
search more is needed. Before returning to this we wish to make two com-
ments.
We speak of a theory, e.g. Ti+ 1 in Figure 2, when we assert that the
88 GERARD RADNITZKY

successor theory is a common component in all the explanantia of the set of

arguments which are outputs of the deductive operation. However, the
successor theory is conceived as a theory with a career dimension, or if you
please, a sequence of successive versions of a theory. In the course of testing
the theory and of producing new knowledge by means of it, the theory itself
is often processed, refined. Thus in terms of Figure 2 the theory in the
pattern (T and A -+ Ho) is really To: in the history of science it is typically
a step towards a new theory, e.g. a central component of the later theory.
Tk could stand for a full-fledged theory, which has grown out of To or for a
well-developed version of the original thoery. Tn may well stand for a so-
called incorrigible theory (Heisenberg's term is 'abgeschlossene Theorie'),
i.e., a theory that cannot be substantially improved, processed into a 'better
version', but can only be superseded by another theory in the sense in which
Newton's theory was 'finished' quite early in its career but was eventually
superseded by Einstein's. Of course the degree of corroboration of the theory
also changes over its history: T has to be evaluated in terms of its career.
Imre Lakatos has placed this in the center and Ernan McMullin is very likely
justified when he views a Lakatosian research program as a theory over its
history rather than a series of theories; but already in Popper's classic of
1934 he clearly emphasizes this career dimension - with the idea of apprais-
ing a theory in terms of past performance and the focus on the idea of improv-
ing a theory through testing it.
The predicted consequences (in Figure 2 Hl to Hn) can function as
potential corroborators of T; the explained hypothesis (Ho) cannot. This is
clear from what has been said in Section 2.1.1 in connection with 'background
knowledge'. In the measure in which the predicted hypotheses constitute
severe tests and prove to be correct, the theory gets credit in terms of degree
of corroboration - although always revocable credit. In this process the
pattern (T & A -+ H 0), which was but a potential explanation as long as Thad
not been tested at all, turns into an explanation proper. Although neither H 0
nor the original explanandum E in our example can function as corroborators
of T, T should get some credit for this achievement, i.e. for the explanation
itself - apart from whether or not the hypothesis explained gets corrected in
attempting to explain it. In Part I we mentioned that one requirement on the
successor theory was that it should match the past explanatory successes of
the predecessor theory; it should explain everything the predecessor theory
explains with at least the same precision. This is usually regarded as a necessary
condition for considering the change to be cognitive progress and nothing
more is said about it. But should not the successor theory get some explicit
 PROGRESS AND RATIONALITY IN RESEARCH 89

credit also for this (explanatory power), since it is an achievement in itself?

Having been able to match the past explanatory successes of its rival could
serve as an indicator for 'conservation of corroborable content of conse-
quences deducible from the predecessor theory at least in the mathematical
and empirical aspects of that content'. For instance, Einstein's theory has but
few corroborators in relation to Newton's but also matches all the latter's
explanatory successes. We think that, apart from the fact that it may explain
these hypotheses with greater precision, his achievement should also be taken
into account in addition to the degree of corroboration, when good reasons
are produced for the assertion that Einstein's theory comes 'closer to the
truth' than Newton's. In any case, we would like to bring this matter up for
discussion.

2.2. The Empirical Component in Theory Appraisal (right box in Figure 2)

2.2.0. Comparing the negation of the deduced data sentence with the experi-
mentally obtained data sentence; the role of technical data-generating systems.
The logical structure of this component has been dealt with in Section 2.0.
From the logical point of view a singular data sentence, incompatible with
a data sentence deduced from Hl in the presence of statements of initial
conditions, J, if assumed true, permits the deduction of either not-H 1 or
not-J. However, in actual research more must be requested. First, of course,
the potential falsifier, the conjunction of the negation of the deduced data
sentence and the statement of initial conditions, has to be empirically tested.
As mentioned in Section 2.0, an experiment will yield a positive result,
e.g., J and P*, a statement of initial conditions together with a statement
describing the observed event; and if p* logically entails not-P (if I- p* ~ ,P),
P is falsified unless we go back to question the correctness of the statement of
initial conditions. Thus basically the test of the potential falsifier consists
in an attempt to falSify it by comparing it with an observed, an experimentally
established data sentence about the type of event in question. In Figure 2 the
SW box within the right box (empirical testing) depicts this comparison of
the negation of the deduced data sentence with the experimentally obtained
data sentence (where J is not problematized for the moment). The input into
this box is the result of our experiment (in Figure 2 d*). To get an experi-
mental result one needs hardware instruments, technical data-generating
systems. In Figure 2 these are represented by the box DGS. A technical data-
generating system is based upon some physical theory (Ty in Figure 2).
Hence in an obvious sense all data are 'theory-Iaden'to some extent. 43 But the
90 GERARD RADNITZKY

data are thereby made dependent not upon the theory under test, but on the
various theories used (not in this context tested) in this theory appraisal. The
theories used are theories underlying the construction and manipulation of
the hardware instruments as well as theories used as software, e.g., logico-
mathematical techniques (used mainly in the information-theoretical part of
the appraisal) or statistical theories used in connection with data-generation. 44
If only because the theories underlying the data-generation are in principle
fallible, data sentences that could not be produced without these theories
must themselves be fallible in principle. There can be no epistemically privileged
'basic' sentences.

2.2.1. At the output side of the 'comparison station' there is a decision point:
the decision concerns the answer to be given to the question, "Has the attempt
to falsify the potential falsifier (J & -P) succeeded?" If the answer is positive
(in Figure 2 indicated by a plUS-Sign at one end of the decision-triangle),
i.e., if the negation of the deduced data sentence has been falsified (J being
considered to be fulfIlled), this potential falsifier is to be rejected and the
theory has received corroboration. If, on the other hand, the potential falsifier
(in Figure 2, not-d 1) has withstood testing, then the researcher automatically
has the new problem of finding out whether or not the potential falsifier,
thus established, can be processed into a falSifying hypothesis (in Figure 2,
PPH)·
Before continuing, a remark may be in place about the accusation of
conventionalism often voiced in the literature with respect to the procedures
just outlined. The decisions involved are neither based merely on conventions
nor do they give rise to conventions. They are based on good reasons. In
principle any of the data sentences may be questioned any time. But it would
be stultifying to do so without good reason. For instance, one may decide to
stop re-checking a particular statement about initial conditions or re-checking
the experimental apparatus simply because there are at the moment no
reasons for supposing the statement to be incorrect or the apparatus to be
working improperly. This may turn out to be wrong - and may get corrected.
But here a convention has been established only in a Pickwickian sense. The
decisions are based on good reasons; they are controllable and corrigible and
if necessary often get corrected by the objective method of falsification. (As
was pointed out in Part I, this correction is objective since the outcome of a
properly conducted experiment cannot be changed by human intervention.)
Decisions are unavoidable in research. But if the moments of decision were
considered to be all that mattered in research, then this methodological view
 PROGRESS AND RATIONALITY IN RESEARCH 91

would indeed merit the label of 'conventionalism'. It is as much a totalization

as the opposite pole, 'logicism', the view that the logical moves are all that
really matter in research. Conventionalism would imply an instrumentalist
view of theories. In the Popperian schema the moment of decision is kept in
balance by theory realism: by the view that whether or not a hypothesis gets
corroborated in an empirical test is 'decided upon' by reality itself so to
speak, ascertainable in an perfectly objective way. Fallibilism by no means
excludes the objectivity of indicators and methods.

2.2.2. Let us return to the problem we left: Can the potential falsifier be pro-
cessed into a falSifying hypothesis? (PFH in Figure 2.) First it is conjectured
that J together with the experimentally obtained data sentence, which con-
flicted with the data sentence deduced from HI, describes an instance of a
reproducible effect. This is again a move that can be labelled 'hypothesis
generation' (in Figure 2, the boxHFHG) whose output is more general than a
data sentence but less general than the hypothesis HI . It describes a repro-
ducible effect; it claims that for all occasions the conditional (J ~ P*) obtains.
It is a potentially falsifying hypothesis for HI and A. (In Figure 2, hi in the
upper middle of the right box.) In order to get it corroborated, if possible,
we have to test the conjecture, attempt to falsify it. In Figure 2 the testing
process is portrayed in the box HFHCE. For the purpose of testing, further,
additional data are requested from the data-generating system (arrow from
HFHCE to DGS). The data obtained constitute one of the inputs into the
testing station. At the output side is again a decision point. The minus sign at
one end of the triangle represents the case where the attempt to falsify h I has
miscarried. We have now a falsifying hypothesis proper, not only a potentially
falsifying hypothesis. This process can also be described in sociological terms
or, better, it has its counterpart in Popper's world-2: To secure the objective
status of the falsifying experiment, to secure intersubjectivity, other re-
searchers must be able to repeat the experiment with the same result. Other-
wise the scientific community will not accept the hypothesis as a falsifYing
hypothesis. But when a hypothesis has been accepted (pro tempore) as a
falsifying hypothesis, the severity of any future repetition of the same
empirical test will be zero.
A law hypothesis may get corrected in attempting to explain it if the con-
clusion of the explanatory argument can be considered an improved successor
of the original explanandum. Analogously for the falsification of a predicted
hypothesis: it may give rise to a refmement or, if you please, to an improved
successor of the predicted bypothesis, which was falsified. Again a very simple
92 GERARD RADNITZKY

example may save words. If HI stands for a rudimentary so-called gas-law

(whose core is the formula pv = constant, at constant temperature), a certain
experiment will falsify it, and the falsification leads directly to a more refmed
version, a somewhat more precise successor hypothesis H; stating that the
relationship holds only for temperatures within certain specified limits.
Successful attempts to falsify H; in tum to give rise to further refinements,
further specification of the realm of validity of the law-hypothesis.
2.3. Deciding Which Premiss is to Blame for the Prediction Failure: The So-
called Duhem-Problem
We ended the outline of the logical structure of the falsification move (Sec-
tion 2.0) by drawing attention to the options left by the conclusion either
not-T or not-A or both. Let us have a closer look at the tw0 45 options, to
save the theory or to retain the auxiliary hypothesis. (Assuming that there are
good reasons for not questioning the reliability of the experiment or the
correctness of the deduction of P, for not deciding to ignore the contradic-
tion or for postponing judgment etc.)
Option A: change the auxiliary hypotheSis, i.e. replace A by A I in order to
save T. A I must meet certain requirements: it must be incompatible with A
and the experimentally established hypothesiS p* must be deducible from T
and A I, i.e. T and A I must provide a potential explanation of P *.46 Potential
because to begin with the explanatory pattern is ad hoc and will remain ad
hoc until A I has been corroborated, until there is independent evidence for
A I, i.e. evidence apart from the experimental result (J and P*). Hence this
potential explanation of the experimental results does not by itself provide a
good reason for deciding to replace A by A I and to retain T.
Option (B): Replace T by T* while retaining A. T* must meet the follow-
ing requirements: T* must contradict T since T* must enable us to deduce a
sentence that contradicts P, and T* in the presence of the old additional
premisses A must provide a potential explanation of our experimental or
observed result (J & P*). However, even if a theory becomes available that
fUlfils both requirements, this does not by itself entitle us to regard option
(B) as obligatory, since A does not follow from the theory's fulflIling both
requirements, and hence it does not follow that T is assuredly false. These
two negative rules are already some contribution the methodologist can make
to the Duhem-problem.
Examples will substantiate the above considerations. Let T stand for the
Newtonian theory of gravitation, A for auxiliary hypotheses including a
description of the planetary system without Neptune, J for statements of
 PROGRESS AND RATIONALITY IN RESEARCH 93

initial conditions such as some initial position and momentum of Uranus at

some specific point of time; let P stand for the false description of the orbit
of Uranus deducible from these premisses, and p* for the hypothesis that
correctly describes the observed orbit of Uranus. If one uses option (A) in
this example, then A I differs from A in that it contains a different description
of the planetary system, viz. it includes an existential hypothesis about the
planet 'Neptune' (not contained in A).
Example No.2 (used by Elie Zahar): T stands again for Newton's theory,
A for a conjunction of ordinary auxiliary hypotheses including a description
of the solar system which includes a singular statement assuming that the sun
can be treated as a point-mass, J for the statements such as some initial posi-
tion and momentum of Mercury at some specified point of time, and P the
false hypothesis that the orbit of Mercury is a stable ellipse. p* stands for a
hypothesis describing (correctly) Mercury's perihelion's precession. (In
Figure 2 it would be portrayed by the falsifying hypothesis hi which appears
in the middle between the two large boxes.) T* stands for General Relativity
Theory. A I differs from A in that in it the above assumption about the sun is
replaced by the hypothesis that the sun's density is not evenly distributed (a
singular statement, although of a higher generality than one describing an
individuated event).
Since in none of these situations can logic alone tell us which option is
rational, what good reasons can be adduced for opting for one rather than
for the other? Everything hinges upon whether or not the potential explana-
tion can be processed into an authentic explanation, i.e., whether or not the
new component originally introduced ad hoc into the explanans REMAINS
ad hoc. Only if it gets corroborated, gets evidential support independent of
this explanation, is option (A) the rational choice. A methodology, whose
rules must ex definitione remain schematic, since they are to cover all kinds
of research enterprises, cannot give the researcher more advice - he must
decide everything else for himself, case by case. 47
In the Uranus example the singular statement about the existence of an
additional planet gets straightforward independent support when the
postulated planet has actually been observed. Holding that the blame for the
falsification of the hypothesis deduced from Newton's theory can be put on
one of the Singular premisses of the argument enables the alleged falsification
to be turned into a triumph for the theory. In the Mercury example there is
good reason, sufficient for rational action, for rejecting the option to change
the additional premisses while retaining T if it appears impossible to provide
independent evidence for the assumption about the uneven distribution of
94 GERARD RADNITZKY

mass in the sun. This good reason becomes almost compelling if, in addition
to this, the successor T* has been corroborated, has gotten independent
evidential support by making possible the deduction of novel knowledge. But
this development is not a necessary condition for having good reasons to
abandon T, i.e., for rejecting the rescue operation, which introduces A '. If A '
does remain ad hoc, then, even without any available successor theory, T
would have to be considered falsified, Le., to have some falsity content. But
since it has so many other successes to its credit it would still not be irrational
to continue to work with T or on refming T.
In Figure 2 the large triangle between the two boxes represents the deci-
sion point described by the Duhem-problem. Where to put the blame for a
falsification of a testable consequence is something only the researcher him-
self can decide. But methodology does have an uncompromising prohibition:
"If you decide to put the blame on the theory, you must never attempt to
repair the situation by reducing the empirical content of the theory. This is
forbidden by the master rule, the 'anti-conventionalist' rule which outlaws
any strategies that would immunize a theory against criticism." In Figure 2
the arrow from this rule to the 'Duhem-triangle' marks this guiding precept.
In addition, methodology tells the researcher that in constructing good reasons
for his decision everything depends upon whether or not the component
introduced in response to the challenge of the falsification - be it A' or T* -
remains ad hoc. If A' or T* has just been invented and lacks independent
evidential support, the ex ante hypothesis that it will or will not remain ad
hoc is a risky conjecture. The researcher himself is in the best position to
make such a conjecture.
2.4. The Situation after It Has Been Decided Whether or Not the Successor
Theory is Falsified, at Least in Its Present Form
If the successor theory is falsified, then this situation automatically creates
a new problem, a return to the problem that was the starting point of the
research enterprise in Figure 1, but now on a higher level. In Figure 2 this is
portrayed by the arrow which runs in the left margin from Tio+1 to Pi+ 1.
This turn brings us back to Figure 1, although Pi is now replaced by Pi+ 1. It
is needless to point out once again that falsification cannot be conclusive
since some of the premisses in the argument cannot be conclusive. It is fallible,
like all our empirical knowledge, yet it is perfectly objective.
If the falsification attempt has failed, the successor theory has thereby
been corroborated to the extent corresponding to the severity of the test
involved. Then the problem automatically arises of appraising the comparative
 PROGRESS AND RATIONALITY IN RESEARCH 95

achievements of the predecessor and the successor theory in other areas. In

Figure 2 the arrow that runs downward in the left margin depicts this move
and this brings us back to the themes dealt with in Part I.
If T has been replaced by T* or A by A I, this move automatically creates
the urgent problem of testing the new component in order to find out whether
our provisional faith in it is indeed justified. Truly an unending quest.

EPILOGUE
From the Quest for Certainty to the Search for Cognitive Progress:
The Non-Foundationalist, Meliorist View of Human Knowledge
With a brief glance backwards into the history of ideas, one can see in Popper
a Copernican revolution not only in methodology but also in our view of
human knowledge: the change from a justificationist to a non-justificationist,
fallibilist view of human knowledge. Impressed by the advent of the scienza
nuova and by the triumph of Newtonian physics, philosophers committed to
a foundationalist view of human knowledge saw their task in proving scientific
knowledge to be true. This quest was basically a theologoumenon: the idea of
scientific knowledge was modelled upon the idea of revealed knowledge,
which for the believer is infallible. This is reflected in the ordinary-language
use of 'to know' and is particularly clear in scientism. If a foundationalist
program is to be attractive, one needs an infallible source of knowledge. The
philosophers sought this either in an empiricist basis (e.g. Bacon) or in an
intellectualist foundation (e.g. Descartes).
Hume's insight that, even if a secure foundation were granted, since the
evidential basis is always finite, the justification of a general statement would
need an amplificatory move discredited the whole approach. A vain search
ensued for a Principle of Induction that might bridge the gap, and the strict
foundationalist was left with either infinite regress or resort to apriorism. This
intellectual experience opened the gates to irrationalism on a large scale, not
only in epistemology but also in political thinking - e.g. the disastrous
influence of Rousseau.
Kant attempted to rescue a place for rationality by means of apriorism, an
apriorism which in a way combined elements of both the above-mentioned
foundationalist scheme, with his synthetic a priori truth. But his examples
lost their status, some of them in consequence of scientific developments.
The logical empiricists a limine rejected the idea of a synthetic a priori.
They wanted to construct an inductive logic that could answer Hume's objec-
tions by yielding a measure of degree of inductive support applicable also to
96 GERARD RADNITZKY

general statements from a fmite evidential base. They were more impressed
by the way the crisis in mathematics had been resolved than by the crisis in
physics; and they took metamathematics as their model for philosophy of
science. They even flirted with the idea of finding a secure foundation in
observation sentences or observational predicates; and even if this founda-
tionalism has by and large been abandoned nowadays, their ideal of science
is clearly within the foundationalist line (cf. Part I, Section 1.1, 1.2).
Popper's work signified a Copernican revolution both in methodology and
in epistemology: he produced the first worked-out non-foundationalist view
of human knowledge. His criticism of both the solutions offered and the prob-
lems raised by logical empiricism has been dealt with in Part I, Section 1.3. A
critique of Kant's rescue attempt on Popperian lines would run: that the
synthetic a priori truths which for Kant were paradigmatic examples have
dissolved;48 that the question of whether there is a synthetic a priori needs
first a more precise distinction between a priori and a posteriori, which must
partly take recourse to psychology, and second a precise interpretation of an
alleged synthetic a priori sentence so that we could examine its status. But
the most decisive question is: is the conceptual structure one gets by crossing
'analytic/synthetic' with 'a priori/a posterion"' fruitful for methodology? That
does not appear to be the case - at least to this writer. Are the sentences
which Kant viewed as synthetic a priori truth 'conditions of the possibility' of
'experience'? To answer this query we would also have to clarify Kant's con-
cept of experience. However, we would transform the question into: Are
there conditions of the possibility of cognitive progress, and if so, which? A
Popperian would answer this last question in the affirmative, and would in
this sense be a Kantian. Our capacity for conjecturing creatively, for prolife-
rating proposals, as well as our capacity to learn from our errors- in sum the
critical attitude and the criticist method - he would see as conditions of the
possibility of cognitive progress, although not a guarantee of such progress.
After the foundering in principle of all foundationalist philosophizing, 49
how can rationality be restored to its rightful place? By the critical method.
Rationality can also be given a place in political activity by the method of
criticizing and of testing proposals. 50 This is of course a rejection of utopian
schemes referring to 'the totality' and ending in totalitarianism.

NOTES
Cf. Radnitzky (1978b), Section 1.2, and Radnitzky (1979b).
2 Radnitzky (1972) and Radnitzky (1974b) attempt to spell out this way oflooking at
research.
 PROGRESS AND RATIONALITY IN RESEARCH 97

3 Cf. Watkins (1978) in Radnitzky and Andersson (1978); see also Radnitzky (1980).
4 This is the main thesis of Radnitzky (1968/1970, Vol. I); (see Radnitzky, 1972).
5 Since normative-evaluative utterances can be questioned in a way in which descriptive
statements cannot, this alone is a sufficient reason for the distinction between facts and
values, between 'Is' and 'Ought'. This is spelled out in Radnitzky (1979c).
6 Cf., e.g., Popper (1972), pp. 196-204, esp. p. 197.
7 Cf. Agassi (1975).
8 Cf., e.g., Popper (1972), pp. 204f; Popper (1963), p. 222; Popper (1975), p. 75.
9 Cf. Kant (1783), Section 57, in the edition Kant (1911), p. 352.
10 Cf., e.g., Popper (1963), p. 222; Popper (1972), esp. p. 118; Popper (1975), p. 75.
11 Cf. Popper (1963), p. 222.
12 This is spelled out in Radnitzky (1977), Section 6.
13 Cf., e.g., Griinbaum (1976) and Griinbaum (1978).
14 Cf. Andersson (1978) and comments on it in Radnitzky (1977), Section 1.34.
15 Cf. Radnitzky (1977), Section 1.34, on Popper's use of Tarski and also Section 1.2
on the good reasons for working with formalized languages.
16 Cf., e.g., Radnitzky (1968/1970), the chapter on Empirical Significance.
17 It cannot be influenced in the same way in which, say, a norm expressed by a state-
ment such as, e.g., 'It is forbidden to ... ' can be revised by human action. The connec-
tion with the distinction between 'Is' and 'Ought' has already been mentioned. The
importance of this distinction for liberal democracy, its role in the political philosophy
of Critical Rationalism, is examined in Radnitzky (1979c).
18 In Figure 2, X, DGS, together with the observer, all merge into a black box. For a
penetrating examination of this idea of the 'participation' of the observer in the 'crea-
tion' of reality cf. Jammer (1977).
19 For a brief critical examination of this trend the reader is referred to Jammer
(1977).
20 Cf. Heisenberg (1958).
21 Cf. Bunge (1977a), p. 151; see also Bunge (1973) and Bunge (1977b).
22 The left part of the box 'falsification attempts' in Figure 2.
23 Cr., e.g., Albert (1968/1975), pp. 11, 13, 15,24,27,31,35,56,61,72,170.
24 Cf. Hans Albert's critique: Albert (1975).
25 Cf., e.g., Popper (1972), and Popper and Eccles (1977).
26 Cf. Popper (1934-1959), p. 44 n. 1, and Popper (1934), p. 18.
27 Cf., e.g., Radnitzky (1974b), pp. 7f. The claim that the rules apply to all research
should not be taken to imply any form of a 'unity of science' thesis. We only wish to
imply that the sciences humaines - a very heterogeneous group - also contain certain
facets (explanation and criticism of explanations and descriptions) to which the model
of Figure 1 and 2 may be adapted. Again, to claim this is not to deny that the typical
humanities, the Geisteswissenschaften, e.g. philology, may have additional methodological
problems of their own not covered by these models: there is a great difference between
the task of 'giving an explanation of understanding' and the task of 'understanding expla-
nation'. One such peculiarity is the phenomenon of self-fulfilling (self-stultifying) proph-
ecy. Whenever a prediction refers to a process which may be influenced by the action of
those whose future behavior is predicted, communicating the prediction to these people
may influence the outcome. Hence, if it cannot be excluded that the predicted effect
may be due to a self-fulfilling prophecy, the prediction cannot constitute a test of the
98 GERARD RADNITZKY

theory by means of which the prediction was made, and hence a successful prediction
cannot corroborate the theory. The difference between, e.g., whether forecasting and
economic forecasting (if people believe the prediction of a bullish tendency on the stock
market, there will be a bullish tendency!) is a facet of the distinction between 'Is' and
'Ought' (cf. here too Radnitzky, 1979c. Section 2.1).
28 E.g. in Radnitzky (1972), (1974a), (1974b).
29 For example, if you have made the assumption that the real system studied suffi-
ciently approximates an isolated system, then you will suppose that a deterministic
system can be used to model it and it will be reasonable to require of your law hypo-
theses that they be univer8.l1 in form; if you regard it as an 'open system', you may
be willing to accept statistical laws. Such a 'metaphysical' posit will influence your
programmatic defmition of your discipline. E.g., your conception of what a human
being is will influence your view on whether the discipline of psychology should be
much like a natural science or should be a Geisteswissenschaft or a Handlungswissenschaft.
30 E.g., Popper (1972), p. 144.
31 Cf., e.g., Radnitzky (1972), Section 223, and Radnitzky (1974a), p. 86.
32 Cf., e.g., Popper (1972), pp. 160, 118f, 147.
33 Cf., e.g., Popper (1972), pp. 160, 118f, 147.
34 Popper's 'transference law', e.g., Popper (1972), p. 114.
35 Popper would be able to agree with this; e.g. (1972), p. 259 "growth of our knowl-
edge ... as consisting throughout of corrections al.J modifications of previous knOWl-
edge", or (the growth of knowledge) "is largely dominated by a tendency towards
increasing integration towards unified theories" (1972, p. 262), or p. 71 "the growth of
all knowledge consists in the modification of previous knowledge"; cf. also (1959,
~. 276). .
6 Usually rendered by the meta-sentence: I- IT &A -+ (I -+ P)).
37 This problem is dealt with in details in, e.g., Radnitzky (1979a), Section 2.2.
38 The inductivist's answer is 'always'. Because he wants to be able to stylize con-
ditionals with the connective for material implication, for' 'All swans are white' every-
thing that is not a non-white swan functions as a potential satisfier although intuitively
cases of non-swans are completely irrelevant for the 'credibility' or estimated truthlike-
ness of the conditional. Among other things, the inductivist's answer leads him to make
the paradoxical claim that the richer theory is easier to probabilify than the less rich
theory. Cf. Musgrave (1978).
39 Of course it would be pointless trivially to proliferate consequences by making use of
the peculiarities of the v-connective.
40 Expressed in the symbolism used in Section 2.0. Ho might be (/0 ""* Po).
41 Cf., e.g., Popper (1972), pp. 204f (first published in 1957).
42 Cf., e.g., Popper (1972), p. 202 or Popper (1975), p. 97.
43 Popper (1934/1959), p. 107, n. 2; Radnitzky (1974a), Section 2212 (p. 80) and n.
75 and Radnitzky (1974b), p. 24.
44 These 'auxiliary' theories or even the hardware itself can take the lead and research
can become 'governed' by the instruments instead of being oriented toward the original
scientific question. Such a problem shift, if temporary, may constitute an interlude with
a high growth rate of new knowledge, in which new hardware techniques and instru-
ments make possible the production of new knowledge, or new software, e.g. mathe-
matical techniques, make the deduction of novel conclusions possible. However, if it is
 PROGRESS AND RATIONALITY IN RESEARCH 99

'totalized', i.e. if the instruments, hardware or software, take the lead for some time, the
ensuing problem shift deteriorates into a 'pathology' of the discipline concerned.
45 From a logical point of view there are three options; but in praxis the third option is
not interesting.
46 I-A'->-,A, I-[(T&A')->(J->P*)).
47 It can take quite a long time before such an issue can be decided. E.g., in the case of
the clash between Miller's experimental results about an alleged 'ether-effect' and a
system whose key component was Einstein's theory, it took about 25 years to find out
that one of the additional premisses used in Miller's falsification argument was falsified
in empirical testing (temperature variations in the apparatus).
4 8 As regards arithmetic, both logicist and formalist conceptions contradict Kant's
view; in physical geometry Euclid's system has lost its monopoly position; as regards
the propositions of 'reine Naturwissenschaft' (e.g. the principle of causality) the only
reasonable construal appears to be to interpret them as methodological advice.
49 In today's German philosophical scene the foundationalist approach takes various
forms. The so-called Erlangen School manifests the whole syndrome: a constructivistic
approach leading to the 'ortholanguage' (Orthosprache), 'protophysics' and a 'pragmatic
grounding' (pragmatische Begriindung) of sentences. According to K.-O. Apel's own
interpretation of his philosophy, it is a continuation of Kant in the pragmatics of lan-
guage and it attempts to provide a 'transcendental-pragmatic foundation' from the con-
ditions of the possibility of our ability to speak with each other. J. Habermas attempts
to reach the same goal with a 'quasi-transcendental/quasi-empirical' foundation, with
'compelling arguments' which could guarantee the sameness of opinions in the long run.
Thereby the idea of truth is played down and it is replaced by a consensus conception
of 'truth'. They all appear to commit the same basic error: discourses are taken to be the
forum for establishing truth and value (wertsch6pfende Instanz) - but discourses can
only be the forum for examining claims to truth or value (wertpriifende Instanz). This
would be the gist of a Popperian criticism of these contemporary trends in German
philosophy, which are again steeped in the foundationalist approach.
50 Cf., e.g., Radnitzky (1977).

BIBLIOGRAPHY

Aggasi, 1.: 1975, 'Between Metaphysics and Methodology', Poznan Studies in the Philos-
ophy of Science and the Humanities 1, 2-8.
Albert, H.: 1968, Traktat iiber kritische Vernunft (J. C. B. Mohr, Tiibingen, 1968; 3rd.
enl. ed. 1975).
Albert, H.: 1975, Transzendentale Traumereien. Karl-Otto Apels Sprachspiele und sein
hermeneutischer Gatt (Hoffmann & Campe, Hamburg).
Albert, H.: 1978, 'Science and the Search for Truth. Critical Rationalism and the
Methodology of Science', in Radnitzky and Andersson (1978).
Andersson, G.: 1978, 'Truth and Scepticism: The Problem of Verisimilitude', in
Radnitzkyand Andersson (1978).
Antiseri, D.: 1977 a, Epistem%gia e didattica delle scienze (Armando Armando, Rome).
Antiseri, D.: 1977b, 'Prova di una teoria ed educazione al risconoscimento dell'errore',
in Epistem%gia, metod%gia clinica e stona della scienza medica (Arti Grafiche
E. Cossidente, Rome), pp. 111-168.
100 GERARD RADNITZKY

Biihler, K.: 1978, Sprachtheorie (Gustav Fischer, Jena, 1934; repro UJlstein, Frankfurt,
1978).
Bunge, M.: 1972, 'On Method in the Philosophy of Science', in Dockx (1972), pp. 97-
120.
Bunge, M.: 1973,Philosophy of Physics (Reidel, Dordrecht).
Bunge, M.: 1977a, Foundations of Physics (Springer Verlag, Berlin).
Bunge, M.: 1977b, 'The Interpretation of Heisenberg's Inequalities', in Pfeifer (1977),
pp. 146-155.
Cohen, R., Feyerabend, P., and Wartofsky, M. (eds.): 1976, Essays in Memory of Imre
Lakatos. Boston Studies in the Philosophy of Science, Vo!. 39 (Reidel, Dordrecht).
Dockx, J. et al. 1972,De Iii methode (Office International de Librairie, BruxeJles).
Grmek, M.: 1976, 'Le role du hasard dans la genese des decouvertes scientifiques', Medi-
cina nei Secoli 13, 277-306.
Griinbaum, A.: 1976, 'Can a Theory Answer More Questions than One of its Rivals?',
British Journal for the Philosophy of Science 27, 1-23.
Griinbaum, A.: 1978, 'Popper versus Inductivism', in Radnitzky and Andersson (1978).
Harre, R. (ed.): 1975, Problems of Scientific Revolution. Progress and Obstacles to
Progress in the Sciences (Clarendon Press, Oxford).
Heisenberg, W.: 1958, 'The Representation of Nature in Contemporary Physics', Daedalus
87,95-108.
Jammer, M.: 1977, 'Physics and the Search for the Absolute', in The Search for Absolute
Values in a Changing World. Proceedings of the Sixth International Conference on
the Unity of the Sciences. San Francisco, Nov. 1977 (International Cultural Foun-
dation Press, New York, 1978).
Jammer, M.: 1979, 'A Consideration of the Philosophical Implications of the New
Physics', in Radnitzky and Andersson (1979).
Kant, I.: 1783, Prolegomena zu einer jeden kiinftigen Metaphysik, die als Wissenschaft
wird auftreten k6nnen; reprinted in Kants Gesammelte Schriften (Reimer, Berlin,
1911, Bd. 4).
Lakatos, I.: 1970, 'Falsification and the Methodology of Scientific Research Program-
mes', in Lakatos and Musgrave (1970), pp. 91-195.
Lakatos, I.: 1974, 'Popper on Demarcation and Induction', in Schilpp (1974), pp. 241-
273.
Lakatos, I. and Musgrave, A. (eds.): 1970, Criticism and the Growth of Knowledge
(Cambridge University Press, Cambridge).
Miller, D.: 1974, 'Popper's Qualitative Theory of Verisimilitude' , British Journal for the
Philosophy of Science 25, 166-177.
Miller, D.: 1977, 'Verisimilitude Redeflated', British Journal for the Philosophy of
Science 27, 378-381.
Musgrave, A.: 1974, 'The Objectivism of Popper's Epistemology', in Schilpp (1974), 1
560-596.
Musgrave, A.: 1978, 'Evidential Support, Falsification, Heuristics, and Anarchism', in
Radnitzky and Andersson (1978).
Pfeifer, H.: 1977, Denken und Umdenken. Zu Werk und Wirkung von Werner Heisenberg
(Piper, Munich).
Popper, K.: 1934-1959, The Logic of Scientific Discovery (Hutchinson, London);
revised trans!. of Logik der Forschung (Springer, Vienna, 1934).
 PROGRESS AND RATIONALITY IN RESEARCH 101

Popper, K.: 1963, Conjectures and Refutations (Routledge & Kegan Paul, London).
Popper, K.: 1972, Objective Knowledge. An Evolutionary Approach (Oxford University
Press, London).
Popper, K.: 1975, 'The Rationality of Scientific Revolutions', in Harre (1975), pp. 72-
101.
Popper, K.: 1976, 'A Note on Verisimilitude', British Journal for the Philosophy of
Science 27, 147-160.
Popper, K. and Eccles, J.: 1977, The Self and its Brain. An Argument for Interactionism
(Springer International, Berlin).
Radnitzky, G.: 1968, Contemporary Schools of Metascience (Humanities, New York;
Esselte Studium Giiteborg, 1968; 2nd ed. 1970; enl. paper ed. Chicago, Gateway,
1973).
Radnitzky, G.: 1972, 'Towards a "Praxiological" Theory of Research', Systematics 10,
129-185 (repr. in enlarged 1973 ed. of Radnitzky, 1.968/70).
Radnitzky, G.: 1974a, 'From Logic of Science to Theory of Research', Communication
and Cognition 7, 61-124.
Radnitzky, G.: 1974b, Preconceptions in Research (Literary Services & Production,
London).
Radnitzky, G.: 1976, 'Popperian Philosophy of Science as an Antidote against Rela-
tivism', in Cohen et al. (1976), pp. 505-546.
Radnitzky, G.: 1977, 'Philo sophie und Wissenschaftstheorie zwischen Wittgenstein und
Popper' in ZeIger (1977) pp. 249-281.
Radnitzky, G.: 1978, 'The Boundaries of Science and Technology', in The Search for
Absolute Values in a Changing World. Proceedings of the Sixth International Con-
ference on the Unity of the Sciences (San Francisco, Nov. 1977) (International
Cultural Foundation Press, New York), (1978) Vol. 2 pp. 1007 -1036.
Radnitzky, G.: 1979a, 'Justifying a Theory vs Giving Good Reasons for Preferring a
Theory', in Radnitzky and Andersson (1979), pp. 213-256.
Radnitzky, G.: 1979b, 'Metodo', in Encyclopaedia of the Twentieth Century (Istituto
della Enciclopedia Italiana, Rome, 1979).
Radnitzky, G.: 1979c, 'Die Sein-5011en-Unterscheidung als Voraussetzung der liberalen
Demokratie', in Salamun (1979), pp. 459-493.
Radnitzky, G.: 1980, 'From Justifying a Theory to Comparing Theories and Selecting
Questions', Revue Intemationale de Philosophie 34, 179-228.
Radnitzky, G. and Andersson, G. (eds.): 1978, Progress and Rationality in Science.
Boston Studies in the Philosophy of Science, Vol. 58 (Reidel, Dordrecht).
Radnitzky, G. and Andersson, G. (eds.): 1979, The Structure and Development of
Science. Boston Studies in the Philosophy of Science, Vol. 59 (Reidel, Dordrecht).
Russo, E.: 1974, 'Typologie du progres des connaissances scientifiques', Revue des
questions scientifiques 145, 345-363,479-502.
Salamun, K. (ed.): 1979, Sozialphilosophie als Aufkliirung. Festschrift for Ernst Topitsch
(J. C. B. Mohr (Paul Siebeck), Ttibingen), 1979.
Schilpp, P. (ed.): 1974, The Philosophy of Karl Popper (The Library of Living Philoso-
phers. 2 vols.) (Open Court, LaSalle, Ill.).
Toulmin, S.: 1974, Human Understanding. Vol. 1: General Introduction and Part I
(Princeton University Press, Princeton).
102 GERARD RADNITZKY

Watkins, J.: 1970, 'Against "Normal Science" " in Lakatos and Musgrave (1970), pp.
25-38.
Watkins, J.: 1978, 'Corroboration and the Problem of Content-comparison', in Radnitzky
and Andersson (1978).
Zeiger, J. and Marek (eds.): 1978, Osterreichische Philosophie unil ihr EinflufJ auf das
analytische Denken der Gegenwart, Vol. I (Sonderband Conceptus, nos. 28-30).
 JOSEPH AGASSI

THE PROBLEMS 0 F SCIENTIFIC V ALIDA TION

I. DISTORTIONS IN THE PHILOSOPHY AND IN THE

HISTORY OF SCIENCE

The purpose of the present study is to illustrate the need to remodel our
studies of science, both philosophical and scientific, and together. I wish to
illustrate a point by an examination of a case history. Let me state, first, that
my history is going to be crude: the history I wish to oppose is crude because
it is a popular prejudice; the history I wish to present is crude because it is
only in outline there - I am no historian and I can do no better. Of course,
my view of the history of science may be a prejudice too: I leave it for others
to decide matters.
The thrust of my argument is this: the history of science as we are taught
it, is largely modelled after the history of physics. The history of physics as
we are taught it is the history of the validation of physics, and by way of
the validation of both Copernicanism and Newtonianism, more specifically
of Newtonianism with Copernicanism as a corollary to it. But what is New-
tonianism and how was it validated?
I shall come to all this soon. Let me now give a slight historical point that
I personally fmd intriguing. It is this. Today no physicist will include under
the heading of Newtonian physics the topic of the stability of the solar
system. Even historians of science hardly notice it. Yet it is at least a signifi-
cant link between the history of physics and the history of biology - via the
history of geology. It was the need to study the stability of the system that
has led to the need to study its origins, and it is the need to decide empirically
between competing hypotheses that led to the upheavals in geology that
preceded the Darwinian revolution. But let us leave the origins. It was one
of the revolutionary ideas of Alexandre Koyre that the stability of the solar
system was an essential topic of debate between the Newtonians and their
detractors, and it was therefore only after Laplace oiled the celestial clock-
work and put it on a firm foundation that he could tell Napoleon he could do
without its maker.
But was it so? Historians of science do not report the Laplacean proof of
the stability of the solar system. This may be because the proof is complex.
103
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 103-114.
Copyright © 1980 by D. Reidel Publishing Company.
104 JOSEPH AGASSI

But then they could do with an outline perhaps. Except that this might be
dangerous: after all, Jacques Hadamard did prove that Laplace's proof is
faulty in a sense. But I think things go deeper. Laplace's proof of the stability
of the solar system is very famous though hardly known. Yet Kant's proof
of the instability of the solar system is the other way round: it is far from
famous yet very well-known. That is to say, hardly anyone will be puzzled
if I told him that since gravitation causes tides and tides cause heat and heat
dissipates, and since energy conserves, gravitation dissipates. If anyone is told
that this proof is Kant's, he usually shows no interest. Not even in the face of
the fact that the law of conservation of energy is a mid-nineteenth century
discovery, namely, long after Kant's death. But what kind of history of science
do we have, where an erroneous proof is presented as correct, a seemingly
correct one suppressed and its foundations ascribed to later writers? It is, I
submit, a history of science devoid of problems. The reason so many histo-
rians of science suppress problems, I think, is that they are still doing mis-
sionary work, propaganda science. The result is a mess not only in history but
also in philosophy: science does not need propaganda, validation, excuses. At
least it should not repress the interesting problem: why was stability part of
Newtonianism that was later lost?
So much for the example. Let me repeat my thesis: the problem of valida-
tion of scienc.e or of a scientific hypothesis seems to me to be an archaism
sustained by the division of labor between philosophers and historians of
science: each of them relies on the other to validate their concern with valida-
tion. Let me now backtrack and start allover again, and slowly.

II. BELLARMINO SETS THE TONE

Before explaining why I think St. Robert, Cardinal Bellarrnino, is the man
who raised the problem of validation which I want to eliminate, I feel I must
introduce him again, and in a way shared by no historian of science known
to me. I am a profound admirer of the Cardinal, but I feel I cannot do him
justice unless I say at once I think that the legal murder of Giordano Bruno in
1600 is his almost exclusive guilt: he was ready to burn at the stake both
Bruno and Galileo; but he was not an evil man. As George Bernard Shaw said
of the legal murder of Joan of Are, tragedy is only possible when terrible
things are done by men of good intentions and hard thinking: a person shot
by a stupid criminal dies by a mishap akin to a natural disaster. Unlike a
stupid murderer, Bellarrnino cared for neither fame nor fortune: he cared, we
know full well, for the stability of the social order. And we all now agree that
 THE PROBLEMS OF SCIENTIFIC VALIDATION 105

the facts validate - notice the word! - his thought that if enquiry is allowed
to roam free and discovery left unchecked, then soon the established social
order will be inquired into and questioned too, not only the established
physical doctrine; and soon after, he realized, discovery will bring forth social
ideas as well, not only new ideas about physical matters. Nor did he wish to
repress discovery and innovation. He only felt that Church hegemony about
such matters must remain quite unquestioned, and that the way to do so is by
guarding very jealously the need for each discovery to be validated by the
Church: no idea, he felt, should be allowed all freedom, unless it had the
Church's nihil obstat and imprimatur.
We know that this idea is a new idea, and we may safely ascribe it to the
not-so-saintly Saint Robert Cardinal Bellarmino. That it is a new idea is
declared by Galileo Galilei himself, in his celebrated samizdat essay, Letter
to Castelli, and its expansion, Letter to the Grand Duchess. Of course, we can
question Galileo's judgment. We can even say he is unreliable since he was an
interested party, especially since he accused the innovator - whom he leaves
nameless - of an ulterior motive, namely of the love of power. But I will
not go into that now: I assume that Galileo needed - and received - the
imprimatur only because of Bellarmino's say-so. I wish to stress that at least
Galileo saw here a major point of philosophic principle. He said he was pro-
posing to the Church of Rome the policy that it be quite indifferent to what
ideas a thinker endorses on matters of nature. The Lord wrote two books -
the Book of Nature and the Holy Writ - and so the two cannot be in any real
conflict. Hence, the guardians of the Holy Writ need not trouble themselves
with the Book of Nature.
Not only Galileo but also Bellarrnino saw this as the essential point, and
Bellarmino was willing to concede: unless we realize that Bellarmino was a
reasonable man we miss the point. Reasonably enough, Bellarmino felt that
things are less simple from the viewpoint of the Church than Galileo was
pretending (and others too, for example Kepler). Whereas Galileo, following
Pico della Mirandola's classic Oratio on the Dignity of Man, I suppose, quoted
Psalm 19 in defense of the study of the Book of Nature, Bellarmino too
quoted the same psalm to say that the sun revolves around the earth. Of
course, this can be reintegrated; but what should be reintegrated and what is
up to the official interpreter, namely the religious authority? Again, we must
admit, Bellarmino's idea was validated by facts: geology, biology, archeology,
all came in series and forced the scientific believers to decide between the
Book of Nature and the Holy Writ. Bellarmino was right in fearing that sci-
ence threatened his religion.
106 JOSEPH AGASSI

Contrary to popular opinion, Bellarmino was more than willing to side

with the Book of Nature and reinterpret Holy Writ. But he insisted that
everyone had to wait till the nihil obstat be given. And, he felt, the nihil
obstat is a tool for preserving stability and so it should be no matter of mere
fashion: it cannot be given today and withdrawn tomorrow. Hence the dis-
coverer who calls on the Church and requests from it a reinterpretation of the
Holy Writ so as to make room for his discovery must show that his is not a
whim, not a mere fashion: he must validate it. All that Bellarmino really
wanted, in short, is an instrument to preserve social stability in the form of a
patent office to validate any new discovery that goes against official readings
of any Sacred Text.
I do not mean to side with Bellarmino. I do think that all conservatives
cheat in the same way: they discourage debate of established ideas but re-
quire more debates about new ideas in the hope of at best destroying them
and at worst winning precious time. And this fits Bellarmino too. He believed
people's thoughts should be guided. Again Galileo agreed that this was the
point. At a crucial point in his Dialogue on the Two World Systems, perhaps
the most crucial point, where Aristotle's ailthority is finally shaken to every-
one's satisfaction, one argument in his favor remains: whom shall we trust if
we cannot trust the "maestro di color che sanno" (as Dante called Aristotle)
or the 'prince of philosophers' (as GaIileo called Aristotle)? Trust no one, said
Galileo; trust only your own intellect.
With this the problem of validation of both science in general and of scien-
tific hypothesis x or y or z should have totally disappeared: let anyone accept
anything that takes his fancy, be it scientific hypothesis x or y, be it science
or astrology. This is, I contend, how things should have happened. But things
very seldom happen just the way they should. GaIileo was ignored and science
followed Bellarmino.

III. BELLARMINO'S CHALLENGE ANSWERED

What was wrong with the development of the scientific movement is that it
developed its hagiography and had its saints and martyrs. And martyrs need
enemies. And so BelIarmino entered the history of science and his challenge
was taken up and answered. The answer is very very simple: Copemicanism
is indeed proven and GaIileo was right.
The answer, as I have just now stated it, is just terrible. It incorporates
all that is most erroneous and evil within science at its very best. First and
foremost, who is Bellarmino and why on earth should he be answered? To
 THE PROBLEMS OF SCIENTIFIC VALIDATION 107

defend Galileo against Bellarmino may be as evil as to defend Einstein,

whether by saying that though he was a Jew he was all right, or by saying
that though he was a member of the bourgeois society he was all right,
or even by saying, what can we do: he was right. We still do not know
whether Einstein was right and on what (like the best of us and like the
worst of us, he was right on some matters and wrong on other matters). But
perhaps this is not fair of me: it is one thing to discuss Einstein and another
to discuss Galileo. Galileo and Bellarmino, one might object, should be
discussed in a historical setting. This is terrible too. It leads to the question,
assuming we now know enough to validate Copernicus, did Galileo know
enough for that? If not, then even if Copernicanism is validated today, then
it was not. Now suppose it was not. Was Galileo right to fight? The level-
headed will say, no: he should have gathered more evidence and waited till he
had enough to hang the man, just like Inspector Clouzot. The Romantic will
say, yes: he had to fight for the right cause, and the proof is that history
vindicated him.
This is all very silly, especially since Copernicanism is just false, and a false-
hood cannot be validated. It is an empirical fact that ever so many scientists,
philosophers of science, historians of science, radical politicians, and other
friends of science, all get very angry when someone says, Copernicanism is
false, becausl; this is exactly what Bellarmino said. But I really cannot be so
bothered by what Bellarmino said and what he did not say. When I say that
Copernicanism is false and my interlocutors get angry, I usually ask them
if they know what Copernicanism is. This makes them more angry, either
because they do not know or because they know it says that the sun is the
center of the universe and they themselves are convinced - because they have
read something about Einstein usually - that the sun is defmitely not in the
center of anything. It is, by the way, not even the center of the solar system.
Newton knew it, and said so; but he added, the sun is almost the center of
gravity of the solar system and so Copernicus was right after all. That is
simply not good enough, in Newton's own book of rules. For, to be right,
Newton thought, was to be absolutely right. Why? Do we have to agree? Does
validation have to be so very strong and water-tight? This depends on what
we need validation for. What do we need validation for? Stability, says
Bellarmino; and so we need proofs!
Here history goes doggedly the wrong way. Rather than ask, can we do
without validation, history went on foolishly quarreling with Bellarmino.
Validation is not the nihil obstat or the imprimatur; validation is a matter
between each person and his own innermost conscience. There is a rational
108 JOSEPH AGASSI

criterion of validation, not the whim of the scientist and less so the whim of
the Cardinal. What, then, is this criterion?

IV. THE RADICALIST THEORY OF VALIDATION

Radicalism is the strongest answer to conservatism. It concedes to the conser-

vative demands and meets them fully. That radicalism is dangerous is obvious:
conceding all the demands of the conservative is dangerous even when it is
possible to meet them. If the conservative's demands are not reasonable, then
it is always dangerous to meet them - just as it is always dangerous to pay a
blackmailer - even when he demands only a few pennies and when resisting
him is unpleasant, etc.
The radicalist philosophy of science is a philosophy of validation. It con-
cedes the demands which Bellarmino put to men of science and it claims that
they can meet them. Nay, more. The radicalist theory cleverly puts the same
demands of Bellarmino: can you validate your theori~s? Bellarmino was for
the status quo: as long as a question is in doubt the received opinion has the
day. The naive, even the incidental and unnoticed opinions of the Church
Fathers regarding the meaning of biblical texts, he considered binding until
and unless overthrown by men of science with a successfully validated alter-
native. Why the status quo? Is it validated? If not, let it go too! Radicalism,
that is, calls the bluff of the conservatives.
More. Sir Francis Bacon had a smashing radicalist argument against the
status quo: as long as you accept, endorse, teach, current doctrine, you take
it as valid, and though it is not correct, as you think it is correct you will
never seek an alternative; and what you do not seek you will not find. More
than that: accepting a view as correct we force the facts to fit it and so refuse
to ever allow the facts to overthrow the current doctrine. Hence the stagna-
tion that lay heavily over the world between ancient and modern times! The
Middle Ages were blessed with schools, and schools are blessed with conflict-
ing doctrines that become school dogmas. Ask for full validation and banish
all unvalidated doctrine, and you will have no school doctrines and hence no
schools, only the validated truth or nothing at all.
This is why Newton wanted nothing but the absolute truth. He knew the
risk of spreading a school doctrine, however good. But as we saw, he himself
realized that ifhis theory was absolutely true, then Copernicus's doctrine was
not. Yet, at the time it was still felt that science has enemies and so needs
defense and so its weaknesses need not be stressed. And so I cannot complain
about Newton. But it is time to be less defensive about science.
 THE PROBLEMS OF SCIENTIFIC V ALIDA TION 109

Radicalism, I say, is an excessive, a somewhat exaggerated, response to

conservatism. But it has an inner logic of its own: once started it does not
go on until victory alone, but it becomes the new doctrine that generates
its own conservatism, i.e. the new defense of the new status quo, and so
we have radicalist conservative attitudes, as we see everywhere after a suc-
cessful radicalist revolution. And so the radicalist philosophy of science of
Bacon and of Descartes that demanded of every adult individual to believe
nothing but fully validated theories developed with its inner logic. It spilled
over to all fields of thought, politics, religion, the arts, and made revolutions
everywhere. But while it was extrovertly so very successful, militant and
triumphant, while it extrovertly became the new status quo, it was introvertly
going bankrupt because it could not possibly meet its own demands of
validation.
I shall not discuss the theories of validation and their criticisms. I shall
only say as a matter of historical record not questioned by anyone, that the
history of theories of knowledge is the history of such theories and their
criticisms. And the fact is that we still have no theory of validation that is
recognized by all. Hence something is fundamentally at fault, something is
not just in bad shape but in an impossible situation: the problem of validation
itself is, in Kant's term, a scandal in philosophy. On this everyone is agreed,
and I wish to explain this fact.
Things always fall short of the expected, yet life goes on. But imperfection
raises problems, allows for difficulties, and permits, however remotely, all
sorts of unpleasant possibilities. Normally, we know all that, and learn to live
with it, and take up the problems as they impose themselves on us: we cross
the bridges when we come to them, as the saying goes. Except that this is the
defense of the status quo, and since the conservative delays reforms in pre-
ference to the status quo and the radicalist gives no weight to the status quo,
it is the radicalist who cannot hide behind the status quo in lieu of the prob-
lem of validation. He fights the status quo because it is no utopia-here-and-
now; and so he must give up his utopia-here-and-now.
And so the logic of the situation seems amply clear. We have the conserva-
tive defense of the status quo as long as it is not specifically challenged and
successfully overthrown. The conservative does not have to like the status
quo, and he need not defend it. He can say, we have to accept it and let go
only when it breaks down. In the seventeenth century the conservatives were
hostile to the new philosophy so-called, namely to scientifically oriented
philosophy. Today when science is established there is the possibility of being
for the scientific status quo, and of defending a conservative philosophy of
110 JOSEPH AGASSI

science. This was done by Michael Polanyi, who amongst scientists is better
known through the writings of Thomas S. Kuhn.
The alternatives to conservatism are two. First, radicalism, which can only
rest on a proper, fully attested criterion of validation existing and validated
already; second, reformism, which can rest on a proper and balanced attitude
towards the status quo. The radicalist should offer a validated criterion to
tell us when a theory or a hypothesis or an institution is properly validated,
and validation means one-hundred-per-cent-validation. Reformism may allow
for so-so validation, which means at times tolerating the status quo, but with
the help of a criterion telling us when. It is very important to notice that
Bellarmino's ploy was to demand total validation, and as long as his demand
is answered we are either conservative or radicalist, but any compromise
requires us to refuse to meet his demand.

v. THE DIFFICULTIES OF COMPROMISE

We have no current theory of compromise yet. Any attempt to temper the

theory which radicalists had concerning validation turns out to be not a com-
promise at all but a weaker radicalist criterion for validation. This is a point
that I think is obvious yet I find that people find it difficult to comprehend.
Bacon and Descartes demanded absolute proof. And most people soon found
absolute proof impossible. And so they demanded less than absolute proof:
probability. But probability is as absolute as absolute proof. For, what does it
matter whether science, or scientific theory, is absolutely proven or the only
probable or the most probable or most verisimilar one: as long as it is proven
or most. probable it is the only one to endorse. But what if there is no theory
proven or probable or simple or veri similar? In that case, says the radicalist,
there should be no accepted theory, no mere status quo. And so not only for
proof or probability, but for any other criterion, the question is of the status
quo: this is a central point. There is, no doubt, the possibility that no theory
answers a criterion, or that it is hard to say which of the competing theories
answers the given criterion; in such cases the radicalist must tell us to endorse
no theory until the difficulty is resolved. And so, radicalism, the theory of
validation, remains an all-or-nothing theory, a properly radicalist theory
that makes no allowance for any status quo, which allows for no schools of
thought and for no dissent.
It is here that the historians of science commit the biggest blunder. Ironing
out the historical situation, they allow for no legitimate dissent, for no con-
troversy where both sides made sense. Just as with Galileo and Bellarmino the
 THE PROBLEMS OF SCIENTIFIC VALIDATION 111

majority say Galileo was absolutely right and the minority say the opposite,
so the majority say the phlogistonists were absolutely wrong, and the minor-
ity say they were absolutely right until one day the anti-phlogistonists were
absolutely right. But there is a difference here. The defenders of Bellarmino
say he was right because he was defending the status quo. They are conserva-
tives proper. The defenders of the phlogistonists pro-tern do not defend them
in the name of the status quo. Their argument is still radicalist, as I shall now
explain.
This is a strange case, a very subtle one, and one which is not easy to ana-
lyze - at least I have made mistakes about it. It is the case of the philosophy
known as instrumentalism and conventionalism - a philosophy shared by
some radicalists, some conservatives, and some reformers. Unlike the tradi-
tional radicalists, the instrumentalists allow validation to shift. Yet they do
require validation and do not rely on the status quo. The question when valid-
ity is withdrawn depends on the question, when is validity granted. Usually
validity is said to go with simplicity and so when a theory loses its simplicity
it may lose its validity to its simpler successor. There still is no room here for
a compromise.
What is a compromise, then? How can there be a principle of compromise
at all? Is compromise not between principle and reality?
Indeed, compromise is between principle and reality. Suppose validation
is only of principle. Then, in the life of compromise we do not act validly
but otherwise. How? What is this otherwise? When is this otherwise inevitable
realism and when is it the cowardice, the sluggishness, the ineffectiveness that
we all deplore? Suppose science is principle and the otherwise of the com-
promise something else. It follows that a man of science as a man of science
cannot compromise. Empirical experience shows this an error.
But what are the facts about validation? Do scientists in fact validate? If
so, all we need is a social scientist to make an empirical study and find how
scientists validate, and he should validate his discovery and close matters. This
may be difficult because we may need a historical perspective to differentiate
the work of a scientist as a scientist from his work as a person living in the
real world. Also, there are differences of opinion that have to be settled, and
this also requires time for a historical perspective. And then the historian of
science becomes our empirical social scientist who irons out the history of
science and who shows how scientists did validate their discoveries.
Thus, the historian of science eliminates the differences and the com-
promises - between schools, between the status quo and the innovation,
between principle and reality, between theory and observation.
112 JOSEPH AGASSI

All this must change, because such history of science is only post mortem
- on principle, since in life there is always dissent, doubt, and compromise.
When this kind of post mortem is presented as analysis in vivo it is sheer
poison.

VI. THE NEW PROBLEMS OF VALIDATION

There is, none the less, something missing in all that I have said thus far.
There is no doubt that we all take seriously some ideas and dismiss other
ideas off-hand. On what ground? There is no doubt that some debates go on
for millennia, some get settled. How? On what ground? Are the endless
debates not a waste of time?
Are these not problems of validation?
The question is ill put. There is no doubt that the Copernican views were
serious and could not be dismissed. Yet Bellarmino asked for a validation and
historians of science still debate, these days, the question, was Bellarmino
right. Hence these problems are not problems of validation in the traditional
sense. We can call them the new problems of validation. These are soluble,
and even with ease!
The first point to notice is that there is no fmal authority, neither Church
nor Reason: we are the sole judges, and we are fallible. But we have a limited
ability to say what we want and whether what we have is satisfactory, and to
what extent.
If we specify what we want, we can put this to critical debate. If we agree
about what we want, we may seek whatever satisfies these wants; for exam-
ple, we want a simple explanation of a given body of facts and we look for a
hypothesis that may serve the purpose. The more desiderata we have, the
more likely it may be that one hypothesis answers some, another answers
some, and we may have to make do.
In other words, we have to start not with validity but with minimum
necessary criteria and allow for diversity both of criteria and of possible solu-
tions by these criteria, in the light of these criteria. In other words, not what
view is reasonable, but what disagreement is reasonable, and what criteria for
these are reasonable. Obviously, the Copernicians disagreed with the Ptole-
means in a very reasonable way!
Karl Popper's philosophy goes a long way towards this solution, since he
stresses rejection, not acceptance. But he offers criteria that are obligatory and
which make the choice of the winning hypothesis unique; and so his highly
original view is still too much in line with traditional radicalist philosophy.
 THE PROBLEMS OF SCIENTIFIC VALIDATION 113

Applying his own idea of rationality in science as the avoidance of dogmatism,

we can say rationality is the avoidance of irrationalism, the allowance that we
may be mistaken and the desire to correct ourselves.
To conclude, I propose validation to mean not correctness but reasonable-
ness in the sense of a reasonable possibility. Can we check these views empiri-
cally? I do not think so. Men of science do try to validate views, and they
often enough demand validation. Often enough they mean this to be the
demand for correctness. And philosophers and historians of science take this
demand for correctness as correct and as a matter of course. I, on the con-
trary, claim that, empirically, when scientists are on the defense, they shift
from validation as proof of correctness to validation as argument for possible
seriousness. Therefore I suggest the following. And I suggest it to scientists,
to philosophers and analysts of science, and to historians of science above
all. A scientist validating or demanding a validation has an idea of the canons
of justification. This idea may but need not be the same as the one he says
when he recites'the catechism. This idea may be better than the one practiced
in his scientific community in his own day and age, or not so good. And the
validation offered may, but need not, live up to the standard.
Thus, when a man of science today validates or requires validation, we
may find out that he looks for explanatory power, for testability, for test,
for the rigor of test, etc. Or he may be looking for, say, operational defin-
ability. Thus, when P. W. Bridgman rejected general relativity on the ground
that it is not given to operational defmition, Einstein accepted his reasoning
but rejected his criterion.
When we come to philosophy of science, the situation gets a bit messier,
since often philosophers do not explain what criteria of validation they are
after and why. This is no criticism, and is not true of everyone anyway. When
we come to the history of science things get even messier: for example in the
eighteenth century men of science were extremely radicalist and so they all
shared the view that there is no room in science for schools and for contro-
versies, and so they had no choice but to call their opponents prejudiced and
unscientific. Some historians of science accept this judgment and declare one
school - the phlogistonists - prejudiced; others do not and they iron out
history and overlook the fact. But history of living people is full of strife, of
falling short of ideal, of controversy, and of compromise. And the eighteenth
century scientists argued while disavowing argument. Such an intriguing
historical fact should not remain so neglected as it still is.
I am practically fmished. I must mention, again, the status quo. Bellannino
gave too much weight to the status quo; the radicalists made it ever impossible
114 JOSEPH AGASSI

to pay any attention to the status quo qua status quo. Here in the present
suggestion the status quo is left as utterly marginal. How we should approach
it depends on us and we may agree to handle it one way or another. For
example, clearly, when big science is involved, i.e. much money invested in
research, we may expect more attention to the status quo. This, said Imre
Lakatos, makes big science quite necessarily conservative. But in fact it is
not necessarily conservative: the criteria big science may have to meet - and
does at times - may force it to radicalism.
And with this I must conclude: the status quo, the politics of science -
external or internal - are not always important; and the problem of valida-
tion as classically understood - of criteria that forbid all concession to the
status quo - is therefore a great exaggeration. We all lead our lives in the way
we think best, yet we still differ from each other: some of us, however, are
open to others' opinions and to suggestions that may be improvements. More
abstractly, we have some disagreements that are silly, some where all parties
are intelligent and listen to each other. These facts must be taken care of by
any theory of science and by any progressive history of science.
 JOHN D. NORTH

SCIENCE AND ANALOGY

I have chosen to speak about the use of analogy in scientific argument, and I
ought to begin by clearing away a number of potential misconceptions: I
cannot completely ignore the historical origins of the word 'analogy', but the
time at my disposal is much too valuable to be spent in tracing its Aristotelian
pedigree in detail. On the other hand, I have no wish to do what so many
contemporary philosophers do, that is to say, treat the word as synonymous
with the word 'model.' What I shall say has much to do with the notion of
scientific model, but my perspective will be rather different from that of a
historian searching for the various uses of models in science. I repeat that my
concern is with analogical argument. As you will see in due course, the sorts
of argument that count as analogical are usually reckoned to be rather weak,
and even rather dangerous. There is no point in my trying to persuade you
that things- are otherwise. Mine will not be a history of 'positive science', in
Comte's sense, but a history of tentative science, and of certain methods of
conjecture. Analogy is the basis for much scientific conjecture, but even con-
jecture is an art, which can be done well, done rationally, that is, even though
it might prove in the end to have yielded a false conclusion.
This last remark might well seem highly paradoxical, but it is one that I
wish to emphasize, both because I believe it to be true, and because one's
attitude to it affects one's whole approach to the history of science. Many
historians - more in the past than in the present, I should say - are interested
only in scientific success, and in the gradual progress of mankind towards
the truth. Other historians, aware of the distorted image created by history
practised in that style, boast that they are equally concerned with scientific
error, falsehood, and misunderstanding. Some historians - even historians of
ideas - profess not to be interested in the quality of the arguments they
chronicle - 'That's not a historian's job', they say - but anyone who is so
interested must take into account that no-man's-land between truth and false-
hood, Le. those arguments that were reasonably based in the light of the
knowledge of the time, but in our own time are judged to have been mis-
taken. The past, after all, like the present, framed irrational as well as rational
truths, rational as well as irrational falsehoods. Not all nonsense is equally
foolish. To show what I mean, I shall first take a number of examples from
115
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 115-140.
Copyright © 1980 by D. Reidel Publishing Company.
116 JOHN D. NORTH

the work of Isaac Newton - a respectable enough scientist, you will agree. I
shall then look at the way in which the concept of scientific analogy devel-
oped in the middle of the nineteenth century, in the hands of Thomson and
Maxwell. I shall not have time to say much about the history of what one
might call the logic of analogy, but I will include a brief sketch of Mill's ideas
on the subject, if only to show that this was not abreast of the best developed
scientific uses of analogy at the time.
In June 1672, Newton wrote a long letter 1 to Henry Oldenburg, Secretary
of the Royal Society, concerning some objections raised by Robert Hooke 2
against Newton's work on colours and the refraction of light. I don't want to
be sidetracked into a discussion as to which of the two men was first to main-
tain that light is a periodic phenomenon: following E. T. Whittaker,3 it is
commonly said that Newton's letter of 1672 contains the first statement
''that homogeneous light is essentially periodic in its nature, and that differ-
ences of period correspond to differences of colour"; but this is somewhat
too generous. On the evidence, one could as easily ascribe the discovery of
periodicity to Hooke - as does Richard Westfall 4 - although for my own
part I think neither claim is particularly illuminating. I am content to frame
the discussion in the words used at the time. As Newton said, Hooke's hy-
pothesis was that the parts of bodies, when briskly agitated, excite vibrations
in the aether, and that these in due course, acting on the eye, cause us to have
the sensation of light, in much the same way as vibrations in the air cause a
sensation of sound, by acting on the organs of hearing. s The analogy is all I
want to consider, namely the analogy between sound and light. Newton does
not actually say he is arguing by analogy when he goes on to say that the
largest vibrations in the aether give rise to a sensation of red, and the shortest
a sensation of deep violet; but it is clear that he was indeed using an analogical
mode of reasoning. He goes on to draw the parallel: variation in the size of
the vibration of the air, he says, is responsible for variation in the tones in the
associated sound. There is no doubt that by 'size' ('depth' or 'bigness') he
means not our 'amplitude', but something like 'wavelength.' He certainly
thought of the vibrations as being longitudinal, rather than transverse. He
knew from his experiments that aether vibrations of various 'sizes', that is,
light of various colours, could be separated and recombined by refraction,
and he tried his hand - not very successfully in this letter of 1672 - at ex-
plaining the colours of thin plates and bubbles. His greatest concern with the
Hooke-Newton hypothesis (if I may call it thiS)6 was that waves or vibrations
in a fluid would be expected to spread out into the adjacent medium, rather
than be confmed to straight lines. For his own part, Newton believed he could
 SCIENCE AND ANALOGY 117

put forward his 'doctrine' without reference to any hypothesis at all. He

could, he thought, 'consider light abstractly'; and when, by his example, he
shows us what he means by this, we find him falling back on the analogy with
sound.
In his new use of the analogy, however, there is a slight change of style. He
finds it easy to conceive, he says,? that the different parts of a body may
emit rays of different colours "and other qualities" in the same way as the
several parts of an uneven string [on a musical instrument] , or of water in a
stream or waterfall, or the pipes of an organ all sounding together, or "all the
variety of sounding bodies in the world together", should send a confused
mixture of sounds through the air. He can, he says, conceive of bodies capable
of reflecting one tone and stifling or transmitting another, and just as easily
can he conceive a body reflecting only that one of a mixture of colours which
it is disposed to reflect.
Consider now. his next use of the analogy with sound. Hooke had con-
sidered the possibility that there are only two fundamental colours. Newton's
discussion of the experimental evidence is much clearer than Hooke's, and
quite equal to a total demolition of Hooke's conjecture. Nevertheless, Newton
throws in a personal aside: he would as soon admit that reds and yellows (or
blues and indigos) are merely different dilutions of the same colour as that
two thirds or sixths in music are different degrees of the same sound, rather
than different sounds. Not much of an argument, you may think; but I will
come back to this question later.
The next use I want to consider, by Newton, of the analogy between light
and sound occurs in yet another letter to Oldenburg, written in December
1675.8 Newton is still not prepared to admit to any hypothesis about the
nature of light, but he reserves the right to explore the consequences of
different hypotheses; and he now confesses that If he were to assume an hy-
pothesis it would be that light is something capable of exciting vibrations in
the aether. 9 You may well be wondering how this differs from the hypothesis
of the earlier letter. It is simply that Newton now seems to be thinking of
light as a series of small bodies (corpuscles) emitted from luminous bodies,
and of the corpuscles (rather than the bright bodies) as stimulants of the
vibrations of the aether. Newton does not wish to say categorically, however,
that light really is a series of corpuscles. Those who wish, he says ironically,
may suppose it to be an aggregate of Aristotelian qualities. 10 How does light
move? He is inclined to believe in a mechanical principle of motion, but
others, he says, may look for a spiritual principle, if they so wish. To avoid
dispute, and to make the hypothesis general, he adds, "let every man here take
118 JOHN D. NORTH

his fancy." Newton believes that he can abstract from the alternative theories
a hard core of indisputable truth. He insists, in fact, that however we may
think of light, we are at least to think of it as a succession of rays, differing
from one another in such contingent circumstances as size, shape, and strength
- as do almost all things in nature. We are to think of light, moreover, as dis-
tinct from the vibrations of the aether, which - as we saw worrying Newton
earlier - have the unfortunate property of going round corners. We are to
think of light as being alternately reflected and transmitted by thin plates,
according to their thickness; and we are to suppose that just as light stimu-
lates the aether, so the interaction is mutual, and the aether refracts the light.
(The greater the density of the aether, the greater the refraction.) Reflection
is to be considered the result of secondary vibrations in the surface of the
reflecting body, some going into the aether within the body, and some being
returned to the aether outside it. 11 I will not elaborate further on the precise
mechanism suggested by Newton, or on the difficulties he encounters, or even
on the fact that he continues to draw analogies with sound vibrations. I wish
merely to point out that the model has changed in a rather subtle way. Des-
pite the change, the model is still compatible with Newton's experimental
findings, and although the old analogy between light and sound is no longer
as clear as it was, he continues to develop it.
Newton wants, he says, "to explain colours." The emphasis is mine, but the
phrase is Newton's. He supposes, he says, that just as bodies of various sizes,
densities and tensions 12 "by percussion or other action" excite sounds of
different tone, that is, vibrations of different wavelengths ('bignesses'), so
rays of light, by impinging on the aether both inside and outside bodies, 13
excite vibrations of different wavelengths in the aether. I leave aside the
physiological part of the explanation. The first point I want to make is that
the analogy is no longer quite as good as it was, for there is nothing carried
through the air, in the case of sound, analogous to the rays of light which
stimulate the aether on passing through it. My second observation is that
Newton has begun to take his analogy very seriously, talking as he does of
the 'Analogy of Nature' in the style of his later 'Third Rule of Philosophizing'
in the Principia. 14 Newton conjectures that just as harmony and discord of
sounds proceed from the ratios between their vibrations in the air, so may
the harmony and discord of colours proceed from the ratios of corresponding
vibrations in the aether.
Here is one argument by analogy, and the analogy is continued imme-
diately after. He describes how he and a friend independently divided the
spectrum of light from a prism into its seven component colours. (I assume
 SCIENCE AND ANALOGY 119

that the friend was told to distinguish the seven colours named by Newton.
This, at least, is what his phraseology suggests.) He and the friend arrived at
much the same division. Newton took the mean of each pair of alternative
divisions, and found on measuring the length of the fmal divisions of the
image of the spectrum that they were in approximately the same ratios as the
divisions of a string capable of sounding the notes in an octave. 15
Newton was clearly very fond of this analogy, in its newly extended form,
for when he published his treatise Optieks in 1704 he repeated the material
on the octave of colour in much the same style. 16 What is more, when he
came to summarize his measurements of the diameters of what we now call
'Newton's rings',17 he again used the musical scale to do so. Remember that
at first he did not produce the rings by monochromatic light, and that the
rings were therefore coloured. 18 It was natural enough, under these circum-
stances, that he should extend his musical comparison. What he does is
calculate the thi~knesses of the wedge of air between the glasses at those
points where the rings are made by his seven spectral colours. He fmds that
these thicknesses are in the ratio of the lengths of a string yielding the notes
of the octave, raised to the power 2/3. 19 These thicknesses he subsequently
equated with what he calls "the Intervals of the following Fits of easy re-
flexion and easy Transmission."2o The explanation of the rings offered by
Newton on the basis of his theory of fits of easy reflexion and transmission
is remarkable and interesting, but does not concern the analogy with sound
I am now discussing. As far as I know, Newton does not develop the analogy
any further.
I have mentioned so many details in the course of my account of Newton's
analogy between light and sound that the shifting character of the analogy
has probably been lost to view. I will summarize the six examples I have now
given, three from 1672 and three from 1675:

N(I) Correspondences (some would call them analogies) are set up (or im-
plied) between the following concepts:
air (a 1 ); aether (a 2 ); vibration in a sounding body (b 1 ); vibration
in a luminous body (b 2 ); the tone of sound (e 1 ); the colour of
light (e 2 ); the sensation of sound (Sl); the sensation of light (S2);
vibration in the air (V1); vibration in the aether (v 2 );21
and also between the following:
causation of v 1 by b 1 , causation of v2 by b 2 ; causation of Sl by
v 1 , and ofs 2 byv 2 •
120 JOHN D. NORTH

As a premiss, we have it clearly stated that c l is a function of the length of

vibrations in air.
As the conclusion, we have

c 2 is a function of the length of vibrations in aether.

N(2) The correspondences in N(1) are obviously still meant to hold, in addi-
tion to a correspondence between the two types of reflection, of sound and
of light. Newton now offers, not an argument, in the usual sense, but a state-
ment of a conceptual possibility. He can, he says, conceive a mixture of Vi,
selectively reflected, and he can (therefore) just as easily conceive a mixture
of v2 , selectively reflected.
Some would label N(2) a 'heuristic analogy.' It seems to me to be better
described as a new correspondence relation. It has illustrative value, but is
not an analogy in any of the senses I shall define later.

N(3) The same correspondences apply as before. Newton now says, in effect,

What I regard as two different sounds (e l , 8 1 , or vi?) I cannot regard as different degrees
of one fundamental sound. Therefore, what I regard as two different colours (e 2 , 8 2 , or
v2 ?) I cannot regard as different degrees of one fundamental colour.

Insofar as Newton here gives a reason for anything, the reason is a psycholog-
ical one. In fact the 'argument' could be looked upon as a statement of intent,
rather in the style of N(2), as to what theoretical concepts are to be utilized.
It is no doubt supposed that by showing them to be translatable, their plausi-
bility is increased. What is well worth noticing here is that Newton spent
much of his time in the documents of 1672 and 1675 - and of course else-
where - denying that he made any use of hypotheses. The correspondence
relations, as a whole, as well as the statements of what is to be taken as con-
ceivable, are, however, good examples of fallible hypotheses.

N(4) In the 1675 document, the correspondence relations are different from
those of 1672. As I have already explained, light rays are brought into the
aether side of the analogy, having no obvious counterpart in air. (In fact,
there is a difference between the functioning of the systems at this point, for,
as already noted, sound does not travel along straight 'rays' whereas light was
thought to do so.) The change does not affect the argument offered. There
are some new correspondences, namely between harmony of sounds (h 1) and
harmony of colours (h2), and also perhaps between ratios of properties of
 SCIENCE AND ANALOGY 121

sounds (rl) and ratios of corresponding properties of colours (r 2 ). I think one

should spell out these correspondences explicitly, even though Newton did
not do so, using as he did words (,harmony' and 'ratios') which were the same
on both sides of the analogy.
The first 'argument' is now, in effect:

hi is a function of rl . Therefore h 2 is a function of r2 .

N(S) Here Newton attempts to determine what to him must have seemed a
significant property of colour. How is one colour separated from the rest? It
is almost always assumed by commentators on Newton that he took colour to
be subject to infmite gradation, but there is clearly a sense in which he wished
to preserve the traditional division of the spectrum into a limited number of
named colours.22 His reasoning is not altogether clear, but the four premises
seem to be these:

There are seven points of division of the octave Qf musical tones.

The tones are reproduced by the division of a string's length in
ratios kl ... k 7 •
There are seven divisions of the spectrum of colour.
The divisions are (in ratios of the length of a spectrum) approxi-
mately kl ... k 7 •

The conclusion is then that the divisions are at points dividing the spectrum
in exactly those ratios.

N(6) The reasoning follows closely that of N(S). The rings are coloured, and
by a suitable stretch of the imagination, take the place of the seven (much
purer) colours of the spectrum. Newton seems to have thought that in the
musical scale he had found an item of conceptual apparatus suitable for
investigating the phenomena of light and colour.

It is not, I suppose, really surprising that commentators on Newton's analogies

have been unsympathetic towards them. In the words ofW. S. Jevons, "even
the loftiest intellects have occasionally yielded, as when Newton was misled
by the analogy between the seven tones of music and the seven colours of his
spectrum."23 Jevons was discussing analogies where the resemblance is only
a numerical one - and in particular, one involving the number 7. As he went
on to say, "Even the genius of Huygens did not prevent him from inferring
that but one satellite could belong to Saturn, because, with those of Jupiter
122 JOHN D. NORTH

and the earth, it completed the perfect number of six."24 I am not in a

position to comment on the importance of numerology to Huygens - who
certainly, in his study of Saturn's ring, made far more profound uses of anal-
ogy than Jevons might lead one to suppose. 2S In Newton's case, however,
there was clearly much more than numerology in his arguments - and indeed,
it would be difficult to show that he was here influenced in any way by the
mystical associations of the number 7 - even though his religious writings
show distinctly numerological tendencies. This is not the place for a discus-
sion of Newton's mysticism, although I would like to emphasize the very
great importance of theological debate in the history of analogical thought. I
do not think it too strong a thesis, that "The entire vocabulary of religion is
based upon the perception of analogies between the material and the spiritual
worlds." 26 Both the vocabulary and the arguments of natural religion are
heavily dependent on analogy, and the justification of the analogy between
human nature and the nature of God has always been at the centre of Chris-
tian theology.
I mention these things because any historical study of analogy - even as it
has been used within science - will be deficient if the philosophico-theological
discussions of such writers as Aquinas, Cajetan and Suarez are ignored. There
are differences, of course, between scientific and theological analogy. (In
particular, in theological argument only one side of the analogy is known.)
Philosophical usage of the words 'univocal', 'analogical' and 'equivocal' was
nevertheless ftxed by the scholastics, as were the closely related expressions
'literal', ('analogical'), and 'metaphorical'. I think there is something to be
learned from these writers about the functioning of analogy and metaphor,
even in science; but there is still another reason why they are of relevance to
scientific history.
As is well known, there was a great efflorescence of natural religion at the
end of the seventeenth century and throughout the eighteenth, and the result-
ing interplay of science and religion is an important part of the intellectual
history of the time. Historians of philosophy are obliged to take this intellec-
tual movement into account, if only because (to take a rather parochial
English perspective) it was an important influence on John Locke and his
philosophical contemporaries. Two theologians who, like Locke, made con-
tributions to the understanding of analogy were Archbishop King, of Dublin,
and Bishop Browne, of Cork (and earlier of Trinity College, Dublin), who in
turn entered into controversy with the well-known philosopher Berkeley -
who was also to become an Irish bishop. There are strong links between the
theological discussion of the early eighteenth century and the logical writings
 SCIENCE AND ANALOGY 123

of Richard Whately (another Archbishop of Dublin!) and thence John Stuart

Mill in the nineteenth century .27
Now one of Mill's principal aims in writing his A System of Logic 28 was to
put inductive reasoning on a secure footing, or, as he explained in his preface,
"the task ... of generalising the modes of investigating truth and estimating
evidence, by which so many important and recondite laws of nature have, in
the various sciences, been aggregated to the stock of human knowledge". Mill
was to the nineteenth century what Bacon had been to the seventeenth,
namely empiricist philosopher and self-appointed arbiter of scientific method.
Neither worked in a historical vacuum, and in fact Mill in his Preface acknowl-
edges a debt to Whewell's History of the Inductive Sciences, which he uses, for
example, as a source of 'false analogies'! Both Bacon and Mill paid attention
to analogy, in their account of induction, and Mill's passage on the subject
was especially influential in philosophical circles.29 I am sure, nevertheless,
that neither Mill nor Bacon had very much influence on the analogical tech-
niques of scientists. They offered insights into the character of scientific
argument, but even then the examples they gave were very remote. indeed
from the best science of their times. Mill, for example, had probably never
even heard of a brilliant physical analogy newly developed by the young
William Thomson, who had still not reached the age of twenty when Mill's
Logic was published. Over the follOWing decade Thomson developed numer-
ous large-scale analogical arguments which greatly affected the course of the
history of phYSics. He wrote little about the technique of analogy, but an
even younger scientific contemporary, James Clerk Maxwell, made numerous
historically interesting asides on the subject, quite apart from developing
many important examples of his own. It will be instructive to set Mill and
Maxwell side by side, to contrast the logical aspects of their utterances on
analogy, and at the same time to compare Maxwell's chosen examples of
analogical argument with those I have already spelled out at length from
Newton, and with others I shall outline from Thomson. 30 I shall then try to
decide whether there was any greater subtlety in the use of analogy in the
nineteenth century than in Newton's time. I will begin with Thomson.
To Thomson belongs the credit for drawing the attention of physicists to
the power of analogy. He did so, not by writing a logic of analogy, but by
developing a notable example. He showed, in fact, that the equipotential sur-
faces in a space occupied by electrostatically charged conductors may be
made to correspond to isothermal surfaces in an infmite solid in which heat
is flOwing. An electric charge corresponds to a heat source, and so on. Maxwell
later listed the details of the analogy as follows:
124 JOHN D. NORTH

Electrostatics.
The electric field.
A dielectric medium.
The electric potential at different
points of the field.
The electromotive force which
tends to move positively elec-
trified bodies from places of
higher to places of lower po-
tential.
A conducting body.
The positively electrified surface
of a conductor.
The negatively electrified surface
of a conductor.
A positively electrified body.
A negatively electrified body.
An equipotential surface.
A line or tube of induction.
Heat.
An unequally heated body.
A body which conducts heat.
The temperature at different
points in the body.
The flow of heat by conduction
from places of higher to places
of lower temperature.
A perfect conductor of heat.
A surface through which heat
flows into the body.
A surface through which heat
escapes from the body.
A source of heat.
A sink of heat, that is, a place at
which heat disappears from
the body.
An isothermal surface.
A line or tube of flow of heat.

Thomson's analogy with Fourier's theory of heat was only the first of several
he developed. They may be summarized in brief as follows:
T(1) 1841. The foregoing analogy with heat (this being the same as M(4)
below). Faraday does not seem to have known of the analogy until 1845. By
1850 at the latest, he had begun to make conceptual use of it in his formula-
tion of the notion of a field, with lines of force in empty space independent
of conductors, dielectrics, or magnets.
T(2) 1845. 31 Analogy between Coulomb's theory of electrical action at a
distance and Faraday's theory of action by contiguous particles in a con-
tinuous medium. The common formal element was a mathematical frame-
work in which Green's potential function played an important role.
T(3) 1845 (British Association meeting, Cambridge). Sketch of possible
analogies between optics, electricity and magnetism. This strongly influenced
the direction of Faraday's research. One consequence of Faraday'S exchanges
with Thomson was the discovery of the rotation of the plane of polarization
of light by magnetism. 32 Within a year or so Faraday had been led to formu-
late a number of new concepts, including that of diamagnetism, and that of
continuous and polarized lines of force capable of vibration, and thus of
transmitting optical 'forces'. He no longer needed his aether particles. 33
T(4) 1846. 34 Analogy between elastic solids and magnetic and electrical
 SCIENCE AND ANALOGY 125

phenomena. (The magnetic induction was made to correspond to the curl of

a vector-potential representing the elastic displacement. An electric force
corresponded to an absolute displacement of a particle of the solid.) Thom-
son, perhaps influenced by Faraday's rejection of his old view of the aether
as a particulate elastic solid, soon lost interest in his new analogy, although
he took it up again in 1889.
T(5) 1849. 35 Analogy between magnetism and a system of 'imaginary
matter', capable of attraction and repulSion. (Cf. Maxwell's 'imaginary prop-
erties', p. 129 and n. 54 below.) Thomson made 'solenoidal' magnetism
equivalent, for example, to equal concentrations of opposite polarity at its
two ends. He noted the analogy between the familiar 'equation of continuity'
(for an incompressible fluid) and the mathematical condition for a solenoidal
distribution of magnetism. He subsequently made use of the 'close analogy
which exists between solenoidal and lamellar distributions of magnetism' to
lead him to several new formulae.
Thomson's analogy led him to conclude that magnetic action, determined
as it is by the imaginary matter, is not located in ordinary matter. One might
look upon this conclusion as the thin end of some ontological wedge, but
Faraday was converted, and henceforth argued that lines of magnetic force
belong to space rather than to matter. 36
T(6) 1856. 37 Multiple analogy involving electro-magnetism, a luminiferous
aether with elastic properties, microscopical vortices within it, and the dy-
namical theory of heat. (The spiral structures are not to be interpreted as
vortex atoms, at this stage.) The aim was to explain magneto-optic rotation,
and Thomson seems to have believed that he had found the only possible
explanation for it. Knudsen (see Note 37) argues that Maxwell inherited this
conviction, and that he retained the vortex theory through all his revisions of
the electromagnetic theory. Heimann suggests that Maxwell began by taking
lines offorce as the fundamental physical entities, but later (1861?) sought to
explain them on the basis of molecular vortices. 38
In 1855, fourteen years after the first of Thomson's analogies, Maxwell
expressed himself very strongly in favour of what he called 'physical anal-
ogies'.39 Maxwell was then only 25. He was at Cambridge, but had already
spent three years at the University of Edinburgh, where he came under the
influence of the physicist J. D. Forbes and the philosopher Sir William
Hamilton. To them he undoubtedly owed much of the style of the 1855
paper - 'On Faraday's lines of force.' That he owed methodological as
well as scientific ideas to Thomson is also clear, and in fact several letters
from Maxwell to Thomson are extant, in which the evolution of Maxwell's
126 JOHN D. NORTH

ideas is well illustrated. 40 Thus in a letter of 15 May 1855, Maxwell writes:

I am trying to construct two theories, mathematically identical, in one of which the

elementary conceptions shall be about fluid particles attracting at a distance while in the
other nothing (mathematical) is considered but various states of polarization tension etc.
existing at various parts of space. The result will resemble your analogy of the steady
motion of heat. Have you patented that notion with all its applications? for I intend to
borrow it for a season, without mentioning anything about heat (except of course his-
torically) but applying it in a somewhat different way to a more general case to which
the laws of heat will not apply.41

On 13 September 1855 he refers to Thomson's 'allegories', showing that he

knew of all Thomson's work to date:

In searching for these notions I have come upon some ready made, which I have appro-
priated. Of these are Faraday's theory of polarity ... also his general notions about 'lines
of force' with the 'conducting power' of different media for them. Then comes your
allegorical representation of the case of electrified bodies by means of conductors of
heat ... Then Ampere's theory of closed galvanic circuits, then part of your allegory
about incompressible elastic solids and lastly the method of the last demonstration in
your R.S. paper on Magnetism. I have also been working at Weber's theory of Electro
Magnetism as a mathematical speculation which I do not believe but which ought to be
compared with others and certainly gives many true results at the expense of several
startling assumptions.
Now I have been planning and partly executing a system of propositions about lines
of force etc. which may be afterwards applied to Electricity, Heat or Magnetism or
Galvanism, but which is in itself a collection of purely geometrical truths embodied in
geometrical conceptions of lines, surfaces etc.
The first part of my design is to prove by popular, that is not professedly symbolic
reasoning, the most important propositions about V and about the solution of the
equation in the last page ... and to trace the lines of force and surfaces of equal V.42

On 14 February 1856, three days after Maxwell had read the second part
of his paper 'On Faraday's lines of force' to the Cambridge Philosophical
SOciety,43 he notes that he left the paper with Thomson, whom he asks to
return it, because he wishes to write up the second part 'On Faraday's elec-
trotonic state.' "I think I left an abstract too", he adds. 44 On an unspecified
date in the same month, Maxwell read an essay on analogy to the Apostle's
Club in Cambridge. 45 This light-hearted essay, in a flippant style characteristic
of university societies, adds nothing to the argume.nts offered in the scientific
paper, although it might well be used to settle a number of disputes over
Maxwell's early Weltanschauung. I am less concerned with this than with his
rather specific claims on behalf of analOgical arguments. I will begin with some
remarks made at the beginning of the paper 'On Faraday's lines of force.'
 SCIENCE AND ANALOGY 127

In simplifying previous investigations one may choose to express the

results, Maxwell said, either in a purely mathematical formula or in a physical
hypothesis. In the first case, he thought, we lose sight of the phenomenon to
be explained, and fail to obtain an extended view of the connections of the
subject. In the second case "we see the phenomena only through a medium,
and are liable to that blindness to facts and rashness in assumption which a
partial explanation encourages". The middle way, the way of analogy, allows
us to grasp "a clear physical conception, without being committed to any
theory founded on the physical science from which that conception is bor-
rowed ... ". When he goes on to speak of our thus avoiding being "carried
beyond the truth by a favourite hypothesis", he reminds us very much, not
only of Mill, but of a very powerful 'Newtonian' tradition of methodological
comment, particularly strong in Scottish philosophy, and in time influencing
Maxwell's mentor, William Thomson. 46
Maxwell now says what he means by 'physical analogy', namely
that partial similarity between the laws of one science and those of another [the em-
phasis is mine I which makes each of them illustrate the other. Thus all the mathematical
sciences are founded on relations between physical laws and laws of numbers, so that the
aim of exact science is to reduce the problems of nature to the determination of quan-
tities by operations with numbers. Passing from the most universal of all analogies to a
very partial one, we fmd the same resemblance in mathematical form between two
different phenomena giving rise to a physical theory of light. 47

Maxwell goes on to outline the following analogies (I will refer to that be-
tween the laws of science and the laws of number as M(l»:
M(2) That between light undergoing refraction and a particle moving in an
intense force-field.
M(3) That between light, the vibrations of an elastic medium (elasticity
being a sort of midwife); and electricity.
M(4) That between attraction at a distance (according to an inverse square
law) and the conduction of heat in uniform media. This is Thomson's first
analogy.48
M(5) That (which it is the purpose of the paper to explore) between a
system of electrical and magnetic poles, acting under an inverse square law,
and a field of incompressible fluid, moving within tubes directed along Fara-
day's lines of force. The lines are analogous to the streamlines in the fluid.
Maxwell's comments on these analogies are of some interest. It is said that
M(2) was "long believed to be the true explanation of the refraction of light",
and that ''we still fmd it useful in the solution of certain problems, in which we
employ it without danger, as an artificial method". I will anticipate a passage
128 JOHN D. NORTH

in which Maxwell says, in effect, how analogical argument does not oblige us
to accept the prior theory. He is now saying that we may continue to use an
analogy of highly restricted validity. We may use it for 'certain problems'; but
how the argument is to be kept under control, Maxwell does not say.
Of analogy M(3), Maxwell says that it extends further, and yet "is founded
only on a resemblance in form between the laws of light and those of vibra-
tions". (He here adds a sentence that I find confusing, which can be ig-
nored.)49 The drift of his meaning is plain as soon as he discusses analogy
M(4), which we might refer to as 'Thomson's first analogy':
The laws of the conduction of heat in uniform media appear at first sight among the
most different in their physical relations from those relating to attractions. The quan-
tities which enter into them are temperature, flow of heat, conductivity. The word force
is foreign to the subject. Yet we find that the mathematical laws of the uniform motion
of heat in homogeneous media are identical in form with those of attractions varying
inversely as the square of the distance. We have only to substitute source of heat for
centre of attraction, flow of heat for accelerating effect of attraction at any point, and
temperature for potential, and the solution of a problem in attractions is transformed
into that of a problem in heat. 50

When he said of M(3) that the resemblance was 'only one of form, Maxwell
was, even if only half consciously, making a distinction between this sort of
analogy and analogies in which the 'objects' in the two related domains so
closely resemble each other that the same word may even be used for both.
(An example is N(1), where vibrations (in air) correspond to vibrations (in
aether).) One may well ask about the danger that two domains will be con-
fused in a carefully prescribed scientific analogy. This is a genuine problem,
but it does not seem to have been what was most worrying to Maxwell, who
went on to say, in connection with M(4), that
the conduction of heat is supposed to proceed by an action between contiguous parts
of a medium, while the force of attraction is a relation between distant bodies, and yet,
if we knew nothing more than is expressed in the mathematical formulae, there would be
nothing to distinguish between the one set of phenomena and the other. 51

In other words, the phenomena are very different, and the formulae fail to
reveal the difference; and yet part of the value of the best analogical argument
is that it allows the mind "clear physical conceptions". 52 The stimulation of
mathematical ideas is only a part of what one should hope for:
It is true, that if we introduce other considerations and observe additional facts, the two
subjects will assume very different aspects, but the mathematical resemblance of some of
their laws will remain, and may still be made useful in exciting appropriate mathematical
ideas. 53
 SCIENCE AND ANALOGY 129

In introducing his own analogy, here numbered M(S), Maxwell made no

more than the modest claim that he would show, by the applications of
Faraday's methods, the mathematical ideas underlying the phenomena of
electricity. If this were all he had done, it would be stretching a point to say
that Maxwell had offered a truly analogical argument. What of his hopes for
a 'clear physical conception'? At its most physical, his 'conception' is one
grounded in what one may call a 'geometry of fluids'. S4 He establishes what
he calls a "geometrical model of the physical phenomena", by which he hopes
"to attain generality and precision, and to avoid the dangers arising from a
premature theory professing to explain the cause of the phenomena". ss
Thomson, I shall later suggest, gave Maxwell a healthy reluctance to lay claim
to causal explanations. His geometrical model is one side of an analogy, to
be sure, but it is artificially created for the purpose of 'arranging and inter-
preting' experimental findings. The principles governing the motion of the
incompressible fluid in his so-called 'geometrical model' are the principles
of classical fluid mechanics, but the model itself is a particular system, one
out of many possible systems. S6 The function of even the most 'formal' of
analogies is obviously not to transfer mathematics as a whole from one do-
main to another. Maxwell's analogy M(1), between the laws of science and
the laws of number is of such generality that it can hardly count as a partic-
ular example. The transfer of mathematics in an analogical argument is one
of particular mathematical results, derived in the prior system after the im-
position of specific mathematical conditions. The conditions may be of a very
general sort - as, for example, the laws of mechanics in a Langrangean form.
And, historically speaking, these conditions might have been imposed long
before - as in Newton's musical-harmony analogue - or they might be newly
imposed in a very contrived way - as in Maxwell's case M(S) - by the man
who is creating a new conceptual analogue, or model. If an analogy is an
explanation of the unfamiliar by the more or less completely familiar, then
this is not a case of analogy. Perhaps we should distinguish between estab-
lished analogues and newly contrived analogues.
Putting Maxwell's own views aside, for the moment, we can now see that
analogies between two different domains are likely to be of value in argument
only if the prior domain (and therefore also the other) is structured by re-
stricting conditions, that is by laws or rules of some sort. This is what such
traditional lOgicians as Whately meant when they said analogy was a 'resem-
blance of relations' (a statement repeated in effect by Maxwell), and it is a
point obscured by Mill when he rejected the traditional view. S7 Of course, if
one forms analogies between simple and familiar situations, as did Mill for his
130 JOHN D. NORTH

examples, the rules governing terms in the prior domain will be a matter of
common sense and - one hopes - of common agreement. But the rules are
nevertheless there, to be taken into account in any fonnal rendering of a
logic of analogy .
In discussing Newton, I spelt out in some detail six analogical 'arguments',
whereas in Maxwell's case I have only hinted at the basis for such arguments.
Analogy M(l), as I have said, lends itself to no argument in particular. Anal-
ogies M(2) and M(3), providing the familiar corpuscle and wave explanations
of light phenomena, are no doubt so familiar that I need not describe them in
detail. I would like to point out that both M(2) and M(3) were, in a sense I
have explained, based on newly contrived analogues. 58 This applies also to
M(4) and M(5), the systems which are considerably more sophisticated than
M(2) and M(3). (Even so, I think it could be argued that M(3), namely
Huygens' wave analogy, contained what to fellow physicists was the least
familiar analogue of the four, at the time of framing.)
I should like to make a distinction here between two sorts of analogy.
Some of Maxwell's examples were, in a sense I have explained, based on
analogues specially contrived for the occasion. Huygens' wave analogy is
an example. How different are analogies with an artificial basis (models, in
one sense of that word) from those with a pre-established analogue - as in
Newton's case, where he did not have to invent a theory of sound? A pre-
liminary and obvious answer is that the first may be modified again and
again until it satisfies its creator, whereas with the second - something con-
ceived of as given - we are obliged to distinguish between so-called 'positive'
and 'negative' analogies, i.e. respects in which the analogues agree and dis-
agree. But matters are rarely so simple. We are reminded of the analogies
under headings N(2), N(3), and N(6). There we found Newton establishing
conceptual possibilities, rejecting and refming a concept, and confirming the
value of a concept. Newton was there, in fact, establishing in this way restric-
tive conditions of a sort which I mention later as having been obscured by
Mill in his analysis of the subject, conditions limiting the functioning of key
concepts (selective reflection, dilution of sound, octave division, and so on).
There is another side to this question of the difference between pre-
established and artificial analogues. I refer to the ontological problem. When
Maxwell said of his incompressible fluid that it was "merely a collection of
imaginary properties", 59 he was not saying anything likely to colour our
views of the real nature of the space occupied by electrical charges. This is
less true of Thomson's analogy (viz. M(4», although the influence of this was
oblique. Here is what Maxwell wrote in his Elementary Treatise on Electricity:
 SCIENCE AND ANALOGY 131

We must bear in mind that at the time when Sir W. Thomson pointed out the analogy
between electrostatic and thermal phenomena men of science were as firmly convinced
that electric attraction was a direct action between distant bodies as that the conduction
of heat was the continuous flow of a material fluid through a solid body. The dissimilar-
ity, therefore, between the things themselves appeared far greater to the men of that
time than to the readers of this book, who, unless they have been previously instructed,
have not yet learned either that heat is a fluid or that electricity acts at a distance. 60

No one reading his work was likely to come away with the idea that heat is
really electricity, or that electricity is really heat ;61 and yet Thomson's paper
did persuade many that Faraday had been right to suggest that electrical
action was effected through a continuous medium. This was the outcome of
an interesting clash of paradigms. Throughout the eighteenth century (and up
to about 1820) Newtonian dynamics had been considered as an almost essen-
tial mode for physical science, and the successes of molecular physics, as
practised by such men as Laplace, Navier, Cauchy and Poisson, confirmed a
majority of the scientific community in their belief - a belief which lingered
on in England rather longer than it did on the continent of Europe. There it
had been challenged indirectly by Fourier's theory of heat. In commenting in
his second paper (first printed 1845)62 on the analogy between Fourier's
theory and Faraday'S theory of electrical action in a medium, Thomson hints
at the uneasy compatibility of action at a distance and contiguous action.
Since his style was mirrored in some degree by that of Maxwell's paper of
1855-6,63 I will quote a more extended passage than is necessary to illustrate
the new ontological situation:
Now the laws of motion for heat which Fourier lays down in his Theorie analytique de
la chaleur, are of that simple elementary kind which constitute a mathematical theory
properly so-called; and therefore, when we find corresponding laws to be true for the
phenomena presented by electrified bodies, we may make them the foundation of the
mathematical theory of electricity: and this may be done if we consider them merely as
actual truths, without adopting any physical hypothesis, although the idea they naturally
suggest is that of the propagation of some effect by means of the mutual action of con-
tiguous particles; just as Coulomb, although his laws naturally suggest the idea of mate-
rial particles attracting or repelling one another at a distance, most carefully avoids
making this a physical hypothesis, and confines himself to the consideration of the
mechanical effects which he observes and their necessary consequences.
All the views which Faraday has brought forward, and illustrated or demonstrated by
experiment, lead to this method of establishing the mathematical theory, and as far as
the analysis is concerned, it would, in most general propositions, be even more simple, if
possible, than that of Coulomb. (Of course, the analysis of particular problems would be
identical in the two methods.) It is thus that Faraday arrives at a knowledge of some of
the most important of the general theorems, which, from their nature, seemed destined
never to be perceived except as mathematical truths. 64
132 JOHN D. NORTH

If Fourier had not displaced mechanics from its position as unchallenged

king of the physical sciences, the analogies of Thomson and Maxwell had
certainly done so by the late 1850's.
I want now to leave Thomson and Maxwell for a time, to consider some
passages in John Stuart Mill's famous System 01 Logic, first published in
1843, that is, at much the same time as Thomson's first analogy. I shall later
make some comparisons between Mill's ideas and those of the philosophically
unsophisticated Maxwell.
Mill begins his account of analogy by trying to put the subject on a very
general footing. He rejects tradition in rejecting Whately'S equation of analogy
with a resemblance 01 relations, and in its place explains analogical reasoning
by the following formula:
Two things resemble each other in one or more respects; a certain
proposition is true of the one, therefore it is true of the other. 65
This formula covers, as he points out, induction as well as analogy. I will not
say anything about Mill's views on induction - a very thorny subject - except
that for him, in what he calls a 'complete induction', the properties shared by
the two things are invariably conjoined. In an analogy, on the other hand,
"no such conjunction has been made out."
In order to simplify Mill's rather lengthy explanations of his meaning, I
shall introduce a notation which extends the one used in connexion with
Newton. The two 'things' resembling each other in the properties I will be
denoted by a l and a 2 , and the proposition by p(a) (a = ai, a 2 , etc.). Mill's
formula then becomes simply:
f(a l ), f(a 2 ), p(a l ) 1- p(a 2 ).
He insists that the fact 66 stated in proposition p must be dependent on some
property of a l (and presumably a 2 ), but we simply do not know on which. If
a l and a 2 resembled each other in all their ultimate properties, he claims, the
truth of p would be guaranteed, and the greater the number of resemblances,
the greater the probability of the truth of p.67 (The statement is highly con-
troversial but I cannot discuss it here.) 68 Mill's remarks on ultimate properties
and derivative properties are of dubious value. Let us denote the ultimate
properties by F, G, etc., and derivative by II, gl, ... ,[2,g2 ••• etc. By de-
fmition, then,

F(a) ~ II (a) ,
F(a) ~ 12 (a), etc.
 SCIENCE AND ANALOGY 133

Mill claims that if "resemblance be in a derivative property, there is reason to

expect resemblance in the ultimate property on which it depends, and in the
other derivative properties dependent on the same ultimate property". 69 In
my notation,
1- F(a l ) & F(a 2 ) 'is reasonable',
whence, by definition,

and so on for h .14 , and all other derivative properties.

It would be a mistake to suppose that Mill set great store by arguments of
this sort. Analogy, for him, was "a mere guidepost, pointing out the direction
in which more rigorous investigations should be prosecuted". It must be said
that his attempt to make analogical reasoning look respectable has a modestly
convincing appearance, until he provides an example. Here my f2 may be
taken as 'has inhabitants', while a l and a 2 are the Earth and the Moon. We
may take fl to be a conjunction of such properties as follows: spherical form,
solid, opaque, volcanic, receiving light from the Sun. Mill considers two
counter-arguments: first, there may be factors, dissimilarities, working against
the inference to life on the Moon. It is in connexion with these that his vague
references to probability are introduced. Second, different ultimate proper-
ties may give rise to the same derivative properties (and so the inference from
f to F will be invalid). This 'giving rise to properties' was meant to be a matter
of causation, and to pass from f to F would thus be to commit the 'fallacy
of many causes', so called. 70 But the greatest of all the shortcomings of the
analysis outlined by Mill is that he entirely evades the enormous problem of
what it means to be an 'ultimate property'. Even in his example, he talks
vaguely of the property of having inhabitants as "depending ... on some
of its properties as a portion of the universe, but on which of those proper-
ties we know not". If his view of the overall structure of science had been
clearer, he might have been able to explain his meaning better; but his ac-
count of scientific procedure was very rudimentary, and even in his discussion
of analogy it is noticeable that the foreground is constantly occupied by what
was for him the greatest problem of all, namely the problem of induction. 71
He began, bravely enough, speaking in very general terms of resemblance in
certain 'respects', rather than merely resemblance of relations; but still he
spoke of resemblance of 'things'. He gave no clear sign that he had considered
one of the most important uses of analogy in scientific argument, namely
analogy between entirely different scientific domains. In short, he threw no
134 JOHN D. NORTH

light whatever on the sort of arguments I gave from Newton, Thomson, and
Maxwell.
Perhaps Mill would have claimed that his analogies extended to the case of
different domains, but I think not. At all events, this case is only found in
Mill, so far as I can see, under the rubric 'false analogy'.72 The examples given
are not such as to have gained much sympathy from his readers, involving as
they do numerology and Pythagorean harmony.

I have now considered excerpts from the writings of four men - Newton,
Thomson, Maxwell, and Mill - three of whom made important use of anal-
ogical argument in a scientific context, and two of whom wrote about the
theory of analogy. I have tried to avoid imposing my own logical views on the
historical material, and I hope the result was not too loosely shaped. The
subject of analogy is a large and difficult one, extending as it does into every
region of human activity. Analogies have two sides to their nature: they are
instruments of argument, prediction, and validation, and they are instruments
of cognitive meaning, understanding, formalization and classification. The
problem of meaning and categories is not easily disentangled from the prob-
lem of argument and law. Are the planets the same sort of thing as bodies in
free space? Newton said Yes, Descartes thought No. Are terrestrial motions
governed by the same laws as celestial? Aristotle said No, Galileo said Yes.
But what a thing is is obviously to a large extent decided by how a thing
behaves; and this is what scientific laws inform us about.
This contrast between problems of meaning and of argument is closely re-
lated to one of these dogmas of the logic of analogy which has been so often
repeated that it is frequently taken for granted. I am referring to the idea that
analogies can be easily divided into two classes - namely of so-called substan-
tial and formal sorts. A substantial analogy is supposedly one where there
is a correspondence of simple properties, while a formal analogy is taken to
involve a correspondence between relations, or, in a more sophisticated ver-
sion of the idea, between relations among constituents somehow stripped of
all their properties. Now it is very difficult to comment on this view, unless
we are clear as to the epistemology of the person who is proposing it, but
the view usually goes with the doctrine that linguistic conventions (and the
formulae of scientific theories) are somehow models of complex facts. The
'formal' relations are, it seems, regarded as though they were fixed for all
eternity - for otherwise, how can we be sure that phenomena which are now
(to take a crude case) explained predominantly in terms of properties will
not in future be better explained in terms mainly of formal relations? If I am
 SCIENCE AND ANALOGY 135

being much too vague, it is in response to what seems to me to be a very

vague thesis, namely that analogies are either substantial or formal. On close
inspection, those two sorts of analogy turn out to be only special cases of
what I earlier called co"espondence rules, rules linking air with aether, sound
with light, and so on, in my earlier examples. This is only the beginning of an
argument by analogy. But wait! There are some who would like to get a valid
argument out of the correspondences by the following procedure: if there is
a total substantial analogy between two things (they say), then all properties
are shared. From this it follows that the things are identical; and therefore the
constituent parts, stripped of their properties, will be identical, and thus there
will be a formal analogy.
This argument is taken from the work of a reputable modern philosopher
(Mario Bunge), who follows very closely a lead given by an excellent his-
torian of science (Helene Metzger). 73 One is also reminded of what Mill said
about 'ultimate and derivative properties'. It seems that those who write on
the structure of analogical argument cannot relinquish the idea that there is
a formula to be found which will guarantee the outcome of at least some
arguments from analogy. Most would want to deny the charge. ("Arguments
from analogy may be fertile but they are all invalid" - Bunge.) But why,
then, arguments of the sort I have just given?
The sharp division of analogies into formal and substantial is often asso-
ciated with another misconception. There is a comparative concept of sub-
stantial analogy, it is sometimes said. The degree of similarity is supposed to
be determined by counting attributes (cf. Mill, Note 68 above). This seems to
me to be very naIve. The properties of an object are not countable. What is
countable are the properties human beings agree to count as properties; and
they are far from being absolute, or stable enough for an argument to be
based on a count of them.
The views I have been referring to here are, it seems to me, a by-product
of a traditional phase of analysis of the logic of analogy which should have
ended at the time of Thomson and Maxwell, if not before. Roughly speaking,
one may say that this phase was marked by undue attention to analogies
between things (or between corresponding terms, in more cautious accounts),
and the subsequent categorization of those things. Perhaps the biggest im-
petus ever received by the theory of analogy of this sort was at the hands of
Aquinas, who wanted a theory in order to justify the meaning of predicates
which were to be applied to things in different Aristotelian categories. There
might well be parallels between the epistemological needs of Aquinas and
those of the modern contemplative scientist. But these common problems
136 JOHN D. NORTH

have little or nothing to do with the problem set by those who, in the natural
sciences, have advanced their knowledge by analogical techniques that they
have seldom tried to justify. One often reads the platitude that analogical
arguments are inevitably limited in their scope, because what is radically new
is precisely that which cannot be accounted for in familiar terms. The whole
purpose of analogies, however, is to explore, and to explore in the hope that
what seems to be radically new will have unsuspected elements in common
with what is familiar.
Perhaps the sociologist will categorize this as reactionary thinking. Perhaps
the logician will dismiss it as invalid or illogical. It is the philosopher's job,
nevertheless, to offer an analysis of this very common mode of thought, and
if he can offer a satisfactory analysis of real historical examples, so much the
better. I hope that I have shown something of the way in which the use of
analogy matured between the seventeenth century and the nineteenth, and
how in Thomson and Maxwell there is a conscious awareness of the function
of a mathematical calculus as an intermediary between analogues - between,
that is to say, a theory and a model for that theory. And theirs was more than
an idle philosophical observation, for it suggested ways of applying a powerful
tool for conjecture and for the unification of the physical sciences.

NOTES

1 Newton to Oldenburg, 11 June 1672 in Isaac Newton, Correspondence, ed. by H. W.

Turnbull. 7 vols. Cambridge, Cambridge University Press, 1959-1977. Vol. I: 1661-
1675,1959, pp. 171-193.
2 Ibid., pp. 110-16.
3 Opticks, repr. New York, Dover, 1952, Introduction, p. Ix viii. (The edition is an
enlarged reprint of that published by G. Bell, 1931, and is based on the 4th ed. of
1730.)
4 "Uneasily Fitful Reflections of Easy Transmission", The Texas Quarterly 10 (1967),
86-102. See especially p. 88 and n. 31. Westfall speaks of 'periodicity' as being 'strongly
implied' by the pattern of (interference) rings observed by Hooke, but it is clear that the
strength of the implication depends on the ambiguity of the word 'periodic'.
5 Newton to Oldenburg, op. cit., p. 174.
6 Ibid., p. 175. Newton speaks of this hypothesis, which was one of several espoused by
Hooke, as being in conformity with his own theories. He does not seem to have wanted
his name associated with it at the time, as he showed by the sentence: "But how [Hooke]
will defend it from other difficulties [viz. those I mention below] I know not ... ". Even
so, where he now admits (p. 175) that the "fundamental [part of Hooke's] supposition"
seems impossible, he was earlier (p. 174) prepared to say that "the fundamental part of
it is not against me".
7 Ibid., p. 177.
 SCIENCE AND ANALOGY 137

8 /bid., pp. 362-89.

9 "Were I to assume an hypothesis it should be this if propounded more generally, so as
not to determine what Light is, farther than that it is something or other capable of
exciting vibrations in the aether . "." (/bid., p. 363).
10 Ibid., p. 370.
11 Ibid., p. 374. Cf. Huygens' secondary wavelets.
12 He is thinking of strings.
13 That this is what is meant by "refracting superficies of bodies" is evident from this
same letter at p. 371, op. cit.
14 The stimulation of sensations of colour is effected, he says in his letter, "much after
the manner, that in the sense of Hearing Nature makes use of aerial vibrations of several
bignesses to generate Sounds of divers tones, for the Analogy of Nature is to be ob-
served". Ibid., p. 376.
15 Ibid., p. 377.
16 Opticks, ed. cit., pp. 126-8. (I. ii. Prop. III. Prob.i.)
17 Hooke was the first to publish experiments on the rings (Micrographia, Observation
9).
18 He did realize that by illuminating his glasses with light of fewer colours than were
found 'in the open air' he could cut down the coloration and greatly increase both the
number of the rings and their sharpness.
19 Newton, Opticks, p. 212 (II, i. Obs. 14). He uses the Dorian mode. The thicknesses
are then proportional to the following numbers:

20 Ibid., p. 284 (III. iii, Prop. XVI).

21 I could add: sound, light. But I take it that these were for Newton synonymous with
vi and v2 , respectively, at least under the hypothesis he was exploring in 1672.
22 I know of no study of the various conventions of division of the spectrum into six,
seven, or other numbers of colours.
23 Jevons, W. S., The Principles of Science. 2 vols. London, Macmillan, 1874.
24 /bid.
25 De Satumi Luna observatio nova, The Hague, 1656 (Oeuvres, Vol. 15). Huygens
made much use of the analogy between the innermost satellite of Jupiter and the Moon
(of the Earth), each of which has a period much longer than the period of rotation of the
parent body. The same is true of the period of Mercury in relation to the Sun's period
of rotation.
Cf. Galileo's argument - satellites: Jupiter:: Moon: Earth.
26 Joyce, G. C., art. 'Analogy', Hastings Enc., p. 416.
27 The connections are several. Note, for example, the Copleston-Grinfield controversy
of 1821, which revived a certain interest in the earlier writers. Whately was a friend and
follower of Copleston, and reprinted King's Discourse on Predestination, with additional
notes. Mill quotes Hooker, Copleston, and Whately, on analogy. (System of Logic, V.v.6.)
28 The full title speaks for itself: A System of Logic, Ratiocinative and Inductive, being
a Connective View of the Principles of Evidence and the Methods of Scientific Investiga-
tion, London, Parker, 1843 (etc.).
138 JOHN D. NORTH

29 Most modern discussions of analogy show signs of its influence.

30 Maxwell, 1. C., An Elementary Treatise on Electricity, Oxford, Clarendon Press,
1881 (posthumous ed. W. Garnett), p. 51. The heat flux through a spherical surface due
to a point source at its centre is inversely proportional to the area, and hence varies
inversely as the square of the radius. The analogy with Coulomb's inverse square law
should be obvious.
Thomson's paper, written in 1841, when he was only 17, was fIrst printed in Cam-
bridge Mathematical Journal 3 (1843), 71-84. The paper was reprinted in the Philosoph-
ical Magazine for 1854, and (as paper l) in Reprint of Papers on Electrostatics and
Magnetism by William Thomson, London, Macmillan 1872. It contains no methodolog-
ical asides of the sort given by Maxwell. E. T. Whittaker was apparently following
Maxwell in mentioning an analogy somewhat similar to Thomson's, arrived at by Ohm
in 1827. (See A History of the Theories of A ether and Electricity. Rev. and enl. ed., 2
vols, London, Nelson, 1951. Vol. 1, p. 241.)
31 See Thomson, Reprint of Papers, p. 29.
32 For the circumstances of the new experiments, see Williams, L. P.,Michael Faraday,
New York, Basic Books, 1965, pp. 383-94.
33 'Thoughts on Ray-Vibrations', Phil. Mag. 28 (1846).
34 Published 1847;repr. inMath. and Phys. Papers, 1, pp. 76-110.
35 'Mathematical Theory of Magnetism', Phil. Trans. Roy. Soc., June 1849 and June
1850; paper XXIV, Reprint of Papers ... pp. 340-405 (see especially pp. 381-5;
398-401).
In addition to this early work, note the important series of articles (dated 1870 and
1872) in the same volume, at Sections 573-583 and 733-763, on the hydro-kinetic
analogy for magnetism, an analogy Thomson holds to have been fust appreciated by
Euler.
36 By 1852 he would argue that magnetism in matter is wholly dependent on the
surrounding medium (which, he said, was 'perhaps the aether').
37 'Dynamical Illustrations of the Magnetic and Helicoidal Rotatory Effects of Trans-
parent Bodies on Polarized Light', Proc. Roy. Soc. 8 (1856), 150-8; repro Phil. Mag.
25 (1857), 198-204; reprinted in Baltimore Lectures on Molecular Dynamics and the
Wave Theory of Light, London, 1904, pp. 569-77. For a close study of this work, see
Knudsen, 0., 'The Faraday Effect and Physical Theory, 1845-73', Archive for Hist. of
Exact Sciences 15 (1976), 235-81.
38 Heimann, P. M., 'Maxwell and the Modes of Consistent Representation', Archive for
the History of the Exact Sciences 6 (1970), 171-213. See esp. p. 189.
39 'On Faraday's Lines of Force', repro in Niven, W. D. (ed.), The Scientific Papers of
James Clerk Maxwell, Cambridge, Cambridge University Press, 1890, 1, pp. 155-229.
See esp. pp. 155-9. The report of the original, as printed in Proc. Cambridge Phil. Soc. 1
(1843-65), 160-6 (read 10 December 1855 and 11 February 1856) does not contain
the preamble on analogy, but this is found in the version in Trans. Camb. Phil. Soc. 10
(1856),27-83.
40 Published by Sir Joseph Larmor, in 'The Origins of Clerk Maxwell's Electric Ideas, as
Described in Familiar Letters to W. Thomson', Proc. Camb. Phil. Soc. 32 (1936), 695-
750.
41 Ibid., p. 705.
42 Ibid., p. 711.
 SCIENCE AND ANALOGY 139

43 See Note 39 above.

44 Maxwell to Thomson, op. cit., p. 714.
45 Printed in full by his biographers, Campbell, L. and Garnett, W., The Life of James
Derk Maxwell, London, 1882, pp. 235-44. Some of the main points made are: that
analogies do not exist without a mind to recognize them; that "causes ... are reasons,
analogically referred to objects instead of thoughts"; that, from a scientific point of
view, relations are of paramount importance; that there is a remarkable analogy between
the intention of a man making a machine which will work, and the principle according
to which it is made. Number, space, and time are discussed, but nothing more scientific.
46 A recent work, which I have not yet seen, but which presumably deals with this
subject, is Richard Olson's Scottish Philosophy and British Physics, Princeton University
Press, 1975.
47 Maxwell, 'Faraday's Lines of Force', p. 156. The point had been made by others
before Maxwell, of mathematics in general rather than simply of numbers. Thus Joseph
Fourier: "Mathematical analysis ... brings together phenomena the most diverse, and
discovers the hidden analogies which unite them". The Analytical Theory of Heat, tr. by
A. Freeman, Cambridge, Cambridge University Press, 1878 (from the 1824-6 edition),
pp. 7-8. Thomson often remarked on the strong influence Fourier had on his own ideas
(see S. P. Thompson's Life, 1910, I, pp. 13-20), and it is significant that Thomson's
demonstration that the formulae derived for electricity from Coulomb's law are identical
with those for heat flow (Le. T(1), based on an apparently different conceptual basis,
viz. contiguity rather than action at a distance) was a demonstration of one of Fourier's
'hidden analogies'.
48 See the notes to T(1) above.
49 "By stripping [the analogy) of its physical dress and reducing it to a theory of 'trans-
verse alternations', we might obtain a system of truth strictly founded on observation,
but probably deficient both in the vividness of its conceptions and the fertility of its
method". Maxwell, 'Faraday's Lines of Force', p. 156.
50 Ibid., p. 157.
51 Ibid.
52 See above, p. 127.
53 Maxwell, 'Faraday's Lines of Force', p. 157.
54 "The substance here treated ... is not even a hypothetical fluid which is introduced
to explain actual phenomena. It is merely a collection of imaginary properties which
may be employed for establishing certain theories in pure mathematics in a way more
intelligible to many minds and more applicable to physical problems than that in which
algebraic symbols alone are used." Ibid., p. 160.
55 /bid., pp. 158-9.
56 The beginning of the explanation as to how the particular system is selected is ex-
plained at pp. 158-9.
57 See p. 9 above.
58 Unlike sound (analogue for light) in Newton's analogy. There was a pre-existing
theory of sound, however weak, not artificially set up for the purpose of the analogy.
59 See Note 47 above.
60 Maxwell, Elementary Treatise on Electricity, p. 52.
61 Maxwell went very far in the direction of caution, in his Treatise on Electricity and
Magnetism, lst ed., Oxford, Clarendon Press, 1873, para. 72. After drawing -parallels
140 JOHN D. NORTH

between fluid pressure, electrical potential, and temperature, he went on: "A fluid is
certainly a substance, heat is as certainly not a substance, so that though we may fmd
assistance from analogies of this kind in forming clear ideas of formal relations of elec-
trical quantities, we must be careful not to let the one or the other analogy suggest to us
that electricity is either a substance like water, or a state of agitation like heat". (My
italics.)
62 Paper II in Thomson, Reprint of Papers ... ; paper first appeared in Cambridge and
Dublin Math. Journal, November 1845, and then in an extended version in Phil. Mag.
1854.
63 That is, the paper from which I have quoted much already (,On Faraday's Lines of
Force'; see Note 39). It was read in two parts, in December 1855 and February 1856.
64 Thomson, Reprint of Papers . .. , p. 29.
65 Mill, System of Logic, III. xx. 2 (see Note 28 above).
66 He calls this 'the fact m'.
67 Loc. cit.
68 Cf. Ibid., III. xx. 3: "If, after much observation of B, we find that it agrees with
A in p out of 10 of its known properties, we may conclude with a probability of 9 to
1 that it will possess any given derivative property of A". Dissimilarities are said to
furnish counter-probabilities. Mill makes no attempt to decide whether some properties
may not be more fundamental than others; or whether there is any limit to the known
properties of a thing; or to their triviality.
69 Ibid., II. xx. 2. Further references to Mill are to this section, unless said to be other-
wise.
70 Cf. Ibid., V. v. 6: "It has to be shown that in the two cases asserted to be analogous,
the same law is really operating; that between the known resemblance and the inferred
one there is some connection by means of causation".
71 Mill wished to settle for nothing less than absolute truth in science. He was deeply
suspicious of hypotheses, which he admitted might be fruitful; but for this very reason
they might - if fruitful yet false - be an "impediment to the progress of real knowledge
by leading inquirers to restrict themselves arbitrarily to the particular hypothesis which
is most accredited at the time". The wave and emission theories of light are instanced as
"un susceptible of being ultimately brought to the test of actual induction", even though
they are not "worthy of entire disregard"!
72 Ibid., V. v. 6.
73 Bunge, M., Scientific Research, 2 vols., New York, Springer, 1967, Vol. 2: The
Search for Truth, esp. ch. 15; Metzger, H., Les concepts scientifiques, Paris, Alcan, 1926,
passim.
 MARCELLO PERA

INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY

How was it that the man was ever led to

entertain that true theory? You cannot say
that it happened by chance. - C S. Peirce

This essay intends to argue that induction is the method of scientific discovery
and that the current objections against the inductive method are not correct.
To this purpose I shall first specify the various meanings of 'method' and I
shall use these meanings as an Ariadne's thread in order to draw a map of the
problems of methodology and a model of inductive procedure (Section I).
Then I shall go on to show that the main arguments against induction put
forward by modern hypothetico-deductivists and by Popper are fallacious
(Sections II-III). Lastly, after rejecting the view of those who maintain that
a distinction should be made between discovery (by induction) of laws and
discovery (via hypotheses) of theories (Section IV), r shall try to prove my
thesis by showing that the act of conceiving or inventing a hypothesis is an
inductive inference from observational premises (Section V). I shall outline
also the advantages of such a view with respect to the hypothetico-deductivist
(or 'trial and error' or 'conjectures and refutations') approach.

I. MEANINGS OF 'METHOD' AND MODELS OF PROCEDURE

It is not always recognized that the widely used expression 'scientific method
is a pollakos legomenon. In actual fact, it contains at least three different
explicanda which it is indispensable to distinguish.
In the first place, the scientific method is a procedure, a general strategy
that indicates an ordered sequence of moves (or steps) which the scientist has
to make (or go through) in order to reach the goal of his research. This is the
sense the word carries in the phrases 'deductive method', 'inductive method',
'hypothetico-deductive method', etc. Thus, to give an example, when Bernard
writes that "the experimental method is based on feeling, reason and experi-
ment, in that order" (1865, p. 57), he is using the term 'method' in this sense
and he therefore specifies, together with the ordered sequence of moves he
141
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 141-165.
Copyright © 1980 by D. Reidel Publishing Company.
142 MARCELLO PERA

deems typical of all scientific research (observation - hypothesis - experiment)

the nature of the relationship which connects them (intuition in order to
invent the hypothesis, logical deduction in order to test it).
In the second place, the scientific method is a set of rules of conduct or
recommendations for each of the moves of which the procedure is composed.
Thus, if the procedure - as in the instance of Bernard - contemplates that
one of the moves should be the invention of hypotheses, and another move
that of subjecting them to experimental tests by deriving observational con-
sequences, the scientific method in this second sense is the set of rules pre-
scribing admissible hypotheses and evaluating acceptable test arguments.
Examples of these rules would be such statements as: 'Put forward simple
hypotheses'; 'Make hypotheses highly falsifiable'; 'Do not introduce ad hoc
hypotheses'; 'Test hypotheses with repeatable observational evidence', etc.
The celebrated definition of 'method' given by Bacon in his Novum Organum 1
or by Descartes in his Regulae ad directionem ingenii,2 like Popper's 'logic of
scientific discovery' or Lakatos' 'methodology of scientific research pro-
grammes', all hark back to this meaning of the term, even though in the
former the rules are seen as 'rules of discovery' but in the latter only as 'rules
of appraisal'. 3
Finally, in the third place, the scientific method is a technique, conceptual
and operational, by which one of the moves contemplated by the procedure
and regulated by the rules is actually carried out. It is in this sense that one
speaks, for example, of methods (or techniques) of observation, of classifica-
tion, of computation, of execution of experiments, etc. Still in the same sense
it is said that sociology uses the sampling 'method', psychology the 'method'
of thinking out loud, psychoanalysis the 'method' of free associations, and so
on. 4
It is the task of the methodology of science to explicate each of these
three meanings of method. In particular, in order to afford a satisfactory
criterion of demarcation of science, methodology is committed to:
(1) identifying the procedure of science, establishing whether or not it is
invariant with respect to the multiplicity of scientific disciplines, giving it a
justification that is both philosophical (in terms of an epistemological theory)
and factual (in terms of its correspondence to the actual practice of research);
(2) derming and justifying, in the same twofold way, a system of rules to
regulate each of the moves of the selected procedure;
(3) determining the admissible techniques or the criteria which the tech-
niques should satisfy in order to make their application admissible.
Since it is concerned with the explication of the procedure, the present
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 143

essay deals with the frrst problem of methodology. The problem is currently
given little attention, as can easily be seen by comparing the amount of
discussion dedicated to it with that reserved for the second problem, which is
at present the focus of interest in the philosophy of science. 5 The explana-
tion for this is doubtless to be found in the persistent and almost uncontested
hegemony exercised by the current hypothetico-deductivist paradigm.
This paradigm is characterized by two theses: the anti-inductivist thesis,
according to which scientific discovery does not come about by induction
from facts, and the anti-continuist thesis, which holds that the process of
discovery is not unitary but rather is divided into two discontinuous thought
episodes, one being a-logical or pre-logical and belonging to the 'context of
discovery', and the other being, instead, logically or rationally reconstructible
and belonging to the successive 'context of justification'.
There are two versions of this paradigm, which we shall call the weak and
the strong versions. The weak version is the most widespread; it offers as the
explicatum of scientific procedure a model that can be represented by the
following schema:
(1) P ... H~Oc-+Hc·
This means that an investigation starts off with a problem P, introduces a
conjecture or hypothesis H through a mental jump (' .. .'), from this it deduces
('~') certain observational test statements Oc and then, if these statements
prove to be true, induces ('-+') the truth or probability ofthe hypothesis. 6
The strong version of the paradigm is Popper's. It offers a variant of
schema (1) where the last step has been cut off, i.e. the schema
(2) P ... H~Oc.
According to this variant 7 (which Popper expresses by the better-known
schema PI - IT - EE - P2, i.e. problem PI, tentative theory, elimination of
the errors, and problem P 2 ) the very argument by which the hypothesis is
tested or 'corroborated' is held to be completely devoid of induction. The
whole procedure is thought to consist of two types of attempts, an attempt
to guess H and an attempt to falsify it by Oc: the former being an intuitive
jump, the latter a deductive argument. Therefore, as Popper writes, "the
method of falsification presupposes no inductive inference, but only the
tautological transformations of deductive logic" (1959, p. 42).
As an alternative to the hypothetico-deductivist paradigm I will propose
as explicatum of scientific procedure the following schema:
(3) 0; -+ Hp ~ Oc -+ He.
144 MARCELLO PERA

According to this model a complete scientific investigation starts off with

initial observations (or facts), from these it proceeds to induce a plausible
hypothesis, and then, after deriving further observational consequences, it
once again goes on to induce the probability of the hypothesis. The model is
inductivist because it represents the outcome of research (i.e. the scientific
discovery) as the conclusion of a sequence of inductions from facts related to
one another by deductive arguments; it follows from this that the model is
also continuist because, by virtue of its representation of research as a con-
tinuum of arguments, it rules out the possibility of distinguishing within it
a non-logical phase from a phase which is instead logical or rational.
The most delicate problem raised by schema (3) is doubtless how the
hypothesis can be inductively derived from the initial observations or facts.
This is the problem on which we will chiefly concentrate in what follows. We
shall see that, in spite of the logical objections raised by the hypothetico-
deductivists and by Popper, this problem can be solved. Moreover, we will
show that it is precisely the opposite conception - namely the hypothetico-
deductivist methodology represented by schemas (1) and (2) - which turns
out to be untenable for logical reasons.
Before beginning the discussion, I would like to make the following point
clear. When I talk of 'induction from facts', I mean to refer to inductive'
arguments the premises of which are observational or factual reports, irrespec-
tive of the logical form of these arguments. In other words, I do not consider
induction only in the form of generalization or simple enumeration. Any
ampliative argument whose conclusion does not follow logically from the
premises will be called an induction: if an argument of this type starts from
observational premises and if these describe facts, situations or data of a
problem, then the conclusion that it can be advanced as a proposed solution
to the problem is a hypothesis and the induction is an 'induction from facts'.

II. HYPOTHESES AND THEORIES.

AN ANTI-INDUCTIVIST PARALOGISM

According to the upholders of the hypothetico-deductivist approach, scien-

tific discovery cannot be inductive because hypotheses (laws or theories)
cannot be derived from facts. The reason for this is held to be purely logical:
facts and observations - it is claimed - presuppose hypotheses, for without
a specific viewpoint or a clearly defmed project, observations would be mean-
ingless and in any case it would be hard to find one's way in the infinite
domain of observables.
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 145

This conception is quite widespread and just a few quotations will suffice
to document it.
According to Cohen and Nagel (1934, pp. 200-201):
We cannot take a single step forward in any inquiry unless we begin with a suggested
explanation or solution of the difficulty which originated it. Such tentative explana-
tions are suggested to us by something in the subject matter and by our previous knowl-
edge. When they are formulated as propositions, they are called hypotheses. The func-
tion of a hypothesis is to direct our search for the order among the facts.

Even more explicitly Hempel maintains that the viewpoint that sparks off
an investigation must be a specific hypothesis and not merely a problem. As
he puts it (1966, pp. 12-13):
Perhaps all that should be required in the fIrst phase is that all the relevant facts be
collected. But relevant to what? ... Let us suppose that the inquiry is concerned with a
specific problem. Should we not then begin by collecting all the facts - or better, all
available data - re~evant to that problem? This notion still makes no clear sense .... And
rightly so; for what particular sorts of data it is reasonable to collect is not determined
by the problem under study, but by a tentative answer to it that the investigator enter-
tains in the form of a conjecture or hypothesis .... Empirical 'facts' or fmdings, there-
fore, can be qualifIed as logically relevant or irrelevant only in reference to a given hypo-
thesis, but not in reference to a given problem .... Tentative hypotheses are needed to
give direction to a scientifIc investigation. Such hypotheses determine, among other
things, what data should be collected at a given point in a scientifIc investigation.

But undoubtedly the main champion of the thesis that a hypothesis or

theory must, for logical reasons, precede observations and facts is K. Popper.
Consider, for example, the following two passages:
I believe that theories are prior to observations as well as to experiments, in the sense
that the latter are signifIcant only in relation to theoretical problems .... I do not
believe, therefore, in the 'method of generalization', that is to say, in the view that science
begins with observations from which it derives its theories by some process of generaliza-
tion or induction (1957, p. 98).
Every observation is preceded by a problem, a hypothesis (or whatever we may call it);
at any rate by something that interests us, by something theoretical or speCUlative. This
is why observations are always selective, and why they presuppose something like a
principle of selection (1972, p. 343; italics mine).

Finally, among the Popperians one may cite Medawar, who also holds that
hypotheses are indispensable to make observations meaningful:
We cannot browse over the fIeld of nature like cows at pasture .... Our observations no
longer range over the universe of observables: they are confmed to those that have a
bearing on the hypothesis under investigation (1969, p. 29 and 51).
146 MARCELLO PERA

Despite its apparent plausibility, I believe that it can be shown that this con-
ception - the conception according to which without hypotheses there can
be no relevant observations - is incorrect. To this end I will first avail myself
of a celebrated example of scientific research.
From September 1854 onwards - [Darwin writes in his Autobiography I - I devoted all
my time to arranging my huge pile of notes, to observing, and experimenting, in relation
to the transmutation of species .... After my return to England it appeared to me that
by following the example of Lyell in Geology, and by collecting all facts which bore in
any way on the variation of animals and plants under domestication and nature, some
light might perhaps be thrown on the whole subject .... I worked on true Baconian prin-
ciples, and without any theory collected facts on a wholesale scale, more especially with
respect to domesticated productions, by printed enquiries, by conversation with skilful
breeders and gardeners, and by extensive reading (1958, pp. 118-19).
Darwin's fame among the hypothetico-deductivists is above all linked to the
well-known slogan: "How odd it is that anyone should not see that all obser-
vation must be for or against some view if it is to be of any service!" 8 In the
light of this, the reference made in the above-quoted passage of the Auto-
biography to 'true Baconian principles' ought then to be explained as a
mishap or seen in the psychoanalytic terms of the unconscious persistence of
a "complex of the Lord Chancellor". 9
In actual fact what we are faced with here is neither a mishap nor a com-
plex. Admittedly, when Darwin set out on the Beagle journey gathering
together his 'huge pile of notes' and when he collected 'all facts' he thought
to be relevant to the 'transmutation of species' he was dealing with scientific
problems; but it is equally certain that he did not have the hypothesis of
natural selection in mind. If we use the term hypothesis in the same meaning
as the hypothetico-deductivists - namely as meaning a "suggested solution
of the difficulty which originated an inquiry" (Cohen-Nagel) or a "conjecture
for the solution of a problem" (Popper) or again as an "imaginative pre-
conception of what might be true" (Medawar) - then what can be said is
that Darwin's observations during the Beagle journey were not preceded by
any hypothesis and that he was observing in order to find a hypothesis rather
than to test a hypothesis.
Must we therefore conclude - as the hypothetico-deductivist would object
- that Darwin's mind was a tabula rasa or, to use Popperian imagery, an
empty bucket and that Darwin - as Medawar puts it - was browsing like cows
at pasture? Of course not! Darwin was extremely knowledgeable, he had vast
scientific learning, he had general ideas about how to solve his problems (one
need only think of his reference to the "example of Lyell in geology"). In a
word, he had a theoretical framework.
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 147

What is a theoretical framework? This expression should be understood as

meaning the store of fundamental scientific knowledge that dominates a
given discipline in a given era; therefore a theoretical framework is that part
of background knowledge that is represented by the theory or theories within
which researchers set up their problems and to which they make reference,
positively or critically, in their attempt to solve them.
It must therefore be conceded to the hypothetico-deductivist point of
view that observations are never 'pure' and that the mind is given the incen-
tive to observe by problems and by theories. But theories are not hypotheses.
Theories are knowledge accepted as true before embarking on an investiga-
tion, while hypotheses are tentative explanations put forward after an
examination of the facts to which the investigation refers. There is thus no
logical reason for claiming that hypotheses cannot be derived from facts.
When they uphold this claim, asserting that hypotheses logically precede
observations, the hypothetico-deductivists assimilate theories to hypotheses
and thereby commit a paralogism.
The image of the mind as a searchlight forever projecting hypotheses in
order to test them is as misleading as that of the mind conceived as an empty
bucket to be filled with empirical content.

III. THE 'TRANSCENDENTAL FALLACY'

OF POPPER'S MAIN ARGUMENT AGAINST INDUCTIVISM

It may seem that the distinction between theoretical framework and hypo-
thesis has merely circumvented the obstacle, transposing into psychological
and historical terms the problem raised by the hypothetico-deductivists,
which is really of a logical kind. For indeed, especially when one takes into
account that the theoretical framework of one era is constituted by the
theories that were hypotheses or tentative research programs in the preceding
era, one may legitimately ask: what precedes the theoretical framework?
Now if the answer is to be that the theoretical framework is derived from
previous observations, one falls back, by regression, into the empiricist pitfall
of pure observations; on the other hand, if this solution is rejected (as it must
be), then apparently one has to fall back onto the conception of the logical
priority of hypotheses over observations.
This is exactly what Popper claims. He maintains in the first place that a
theoretical viewpoint must always be presupposed in order for our observa-
tions to be meaningful.
148 MARCELLO PERA

Each observation is preceded by expectations or hypotheses; by those expectations,

more especially, which make up the horizon of expectations that lends those observa-
tions their significance... . In this way science appears clearly as a straightforward
continuation of the pre-scientific repair work on our horizons of expectations. Science
never starts from scratch; it can never be described as free from assumptions; for at every
instant it presupposes a horizon of expectation - yesterday's horizon of expectations, as
it were (1972, p. 346).
From these considerations Popper derives his chief argument against induc-
tivism. If hypotheses and scientific theories are the continuation of our
expectations, then the thesis that they can be derived by induction from
observations is illusory, is a 'myth'. In actual fact, Popper argues, we do not
acquire our knowledge according to Lamarck's model of instruction, but
according to Darwin's model of selection, by anticipating conjectures and
eliminating errors.
All this led me to the view that conjecture or hypothesis must come before observation
or perception: we have inborn expectations; we have latent inborn knowledge, in the
form of latent expectations, to be activated by stimuli to which we react as a rule while
engaged in active exploration. All learning is a modification (it may be a refutation) of
some prior knowledge and thus, in the last analysis, of some inborn knowledge .... I
solved this problem of induction by the simple discovery that induction by repetition
did not exist (any more than did learning something new by repetition) (1974a, p. 40).
This argument of Popper's - of unequivocal Kantian ancestry - is an
excellent confutation of that extreme form of empiricist inductivism (that
may, however, be difficult to identify historically) which holds that our
knowledge is completely reducible to experience and that laws or theories
are wholly derived from repetition or generalization. But the argument does
not imply the conclusion Popper draws from it, namely that "induction by
repetition does not exist" or that, as Popper also puts it, "there simply is no
such logical entity as an inductive inference" (1974b, p. 1015). In particular,
the argument does not imply that hypotheses cannot be derived inductively
from observations that have been carried out and from facts that have been
ascertained previously and independently of the hypotheses themselves. I
will try to show here that this conclusion derives from a failure to distin-
guish between expectations seen as inborn or a priori assumptions or struc-
tures, and expectations seen as conjectures or hypotheses. This assimilation
of assumptions to hypotheses will be termed - with Kantian terminology and
intention - the 'transcendental fallacy'.
We have already hinted that the theoretical framework does not com-
pletely span the background knowledge a researcher possesses. In actual fact,
above and beyond its idola theatri, the mind of the researcher is endowed
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 149

with knowledge and convictions whose most immediately obvious charac-

teristics are a strong measure of stability, the property of being tacitly widely
shared and that of being obstinately protected against contrasting experiences.
We are dealing here, to use Baconian terminology again, with the idola tribus,
more precisely, with assumptions.
A systematic study of the nature of assumptions as also of the role they
play in scientific research has yet to be carried out, although we already have
a goodly amount of material on individual episodes, obtained above all as a
result of the stimulus exerted in our day by the methodologies of 'research
programmes' (Lakatos) or of 'paradigms' (Kuhn) or of 'images of science'
(Elkana).10 Fairly synthetically and without going into details, we may say
here that in relation to a scientific investigation an assumption is: (1) in the
widest sense, the expectation of regularities in facts or of the existence of laws
in nature; (2) in a still fairly wide though yet more specific sense, a particular
conception of the type of regularity or lawfulness presupposed, such as, for
example, simplicity, causality, determinism; (3) in a considerably more
specific sense, a theory or part of a theory accepted and established as a
certain foundation, a true axiom, for an investigation; fmally, (4) a general
belief concerning the nature and aims of science or a specific belief concern-
ing what is to be understood by explanation, law, confirmation and so on. 11
If one keeps this inventory of meanings in mind and takes note of the
'family resemblance' that for all their heterogeneity still appears, it seems
legitimate to affirm that assumptions are beliefs, at times explicit but more
often unconsciously accepted and unformulated, which have the value of
an orientating presupposition for the gathering of facts as well as being
normative or regulative for the proposal and evaluation of hypotheses and
research programs. This amounts to recognizing that assumptions operate
on the plane of transcendental functions (albeit not in a strictly Kantian
sense): they are the a priori conditions for the possibility of scientific research,
a categorial framework to be filled with contents, a system of general expecta-
tions to be satisfied by specific answers.
What, on the contrary, is a hypothesis? One may take this term to mean:
(1) in a pejorative sense, an untestable assertion, being (a) absolutely untest-
able ('hypotheses non fingo') or (b) untestable independently of the facts
relative to which it has been proposed (ad hoc hypothesis); (2) in a positive
sense, a testable but for some reason as yet untested conjecture; (3) in an
extended sense, any theory or scientific law that has been successfully sub-
jected to tests but is still considered to be provisional, being (a) absolutely
proviSional, should the scientific truths all be deemed to be revisable, or
150 MARCELLO PERA

(b) contingently provisional, should the tests be estimated to be uncertain or

not sufficiently strict.
Leaving aside the evaluative connotations and once again bearing in mind
the 'family resemblance', we may then state that a hypothesis is an explana-
tory project deliberately formulated in order to give an account of certain
facts, a specific answer to a specific problem. A hypothesis thus comes into
play on the empirical plane, and is a posteriori with respect to the facts, com-
ing into effect (if and when it does) after these have been ascertained.
Thus assumptions are not hypotheses. Distinguishing the former from the
latter is a matter of the utmost importance. It is not a question of linguistic
admonishments, for after all everyone should be allowed the freedom to
baptize things with the words he likes best and to make use of the conven-
tions he thinks most befitting. Rather, it is a question of approaching state-
ments like 'observations presuppose hypotheses' in such a way as to prevent a
clumsy use of language from leading to misinterpretations. Max Planck put it
very aptly when he referred to an assumption in the second meaning on our
list: "Let us indeed define the causal nexus as a hypothesis; it is not the
terminology that counts. In any case we are not dealing with just an ordinary
hypothesis, but with the chief fundamental hypothesis, with the premise
necessary for hypothesis making to be meaningful" (1923).
Kant considered all cases of mistaking the subjective rules and maxims of
reason for objectively valid principles to be a 'transcendental illusion'; using
similar terminology I define as a 'transcendental fallacy' the misguided step
of confusing the plane of a priori conditions (the assumptions) with that of
the empirical contents made possible by the former (the hypotheses). It is
my contention that when Popper says that observations only exert a selective
function or when he claims, by virtue of the argument that hypotheses
precede observations, to have (negatively) resolved the question of whether
hypotheses (or theories) can be obtained inductively, he actually commits
this fallacy. This on account of an excess of Kantianism which leads him to
multiply unnecessarily the a priori elements of knowledge: only assumptions,
and not hypotheses too, logically precede observations.

IV. DISCOVERY OF LAWS AND DISCOVERY OF THEORIES.

AN UNTENABLE DUALISM

The principal epistemological obstacles that have usually hindered an induc-

tivist theory of scientific method can at this point be considered to have been
overcome. If hypotheses do not logically precede observations, then there is
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 151

no principled reason for rejecting the thesis that the former can be derived
from the latter through an inductive inference.
Before upholding this thesis, it is desirable, however, to take a look at a
more flexible conception than the hypothetico-deductivist position as regards
the logical structure of scientific discovery. I will call this conception 'qualified
inductivism'; it maintains that a distinction should be made between the
discovery of empirical laws and the discovery of theoretical laws.
J. S. Mill may be said to be the first of the qualified inductivists. As is well
known, according to Mill, wherever the plurality of causes and the intermixture
of effects make it impossible to investigate the coming into being of some
phenomenon by the direct method (that is, simply by applying the canons),
then this method must be replaced by the deductive or a priori method or by
the hypothetical method. The deductive method is a three-stage procedure:
the first stage is direct induction from observed phenomena in order to ascer-
tain the laws of causes, the second is calculation or ratiocination, to deter-
mine the effect produced by the combination of the causes, and the third is
empirical verification to establish that the conclusions calculated accord with
experience. The hypothetical method is similarly in three stages but with the
hypothetical formulation of the law instead of its inductive derivation.
According to Mill, therefore, the choice of the deductive or hypothetical
method imposes itself upon the scientist as a result of the degree of complex-
ity of the phenomena awaiting an explanation or, as he puts it"of "the very
nature of the case". This complexity has a nomologic concomitant: on
account of it, the laws of phenomena can no longer be formulated in terms of
observational predicates on a par with the laws obtainable by the canons, but
rather must be expressed in theoretical terms. Mill recognizes this explicitly
when, after having said that it is to the deductive method (and with all the
more reason, one may say to the hypothetical method) that "the human
mind is indebted for its most conspicuous triumphs in the investigation of
nature", he goes on to add: "to it we owe all the theories by which vast and
complicated phenomena are embraced under a few simple laws, which, con-
sidered as the laws of those great phenomena, could never have been detected
by their direct study" (1843, p. 304). Mill takes Newton's theory of gravity
to be one example of such theories.
This attitude, which holds that empirical laws are discoverable by induc-
tion while theoretical laws, on the contrary, are discoverable by hypotheses,
has been taken over, in a certainly much more elaborate and conscious
version than Mill's, by modern qualified inductivists, among whose ranks one
may place G. H. von Wright, J. P. Day, M. Bunge and R. Carnap.12
152 MARCELLO PER A

For his exemplary clarity on this question, Carnap himself can be picked
out as spokesman. In the Philosophical Foundations of Physics he writes:
Theoretical laws are, of course, more general than empirical laws. It is important to
understand, however, that theoretical laws cannot be arrived at simply by taking the
empirical laws, then generalizing a few steps further. How does a physicist arrive at an
empirical law? He observes certain events in nature. He notices a certain regularity. He
describes this regularity by making an inductive generalization (1966, p. 228).
On the other hand, Carnap enquires,
How can theoretical laws be discovered? We cannot say: 'Let's just collect more and
more data, then generalize beyond the empirical laws until we reach theoretical ones'.
No theoretical law was ever found that way. We observe stones and trees and flowers,
noting various regularities and describing them by empirical laws. But no matter how
long or how carefully we observe such things, we never reach a point at which we observe
a molecule. The term 'molecule' never arises as a result of observations. For this reason,
no amount of generalization from observations will ever produce a theory of molecular
processes. Such a theory must arise in another way. It is stated not as a generalization
of facts but as a hypothesis (1966, p. 230).

l. P. Day, upholding the same conception, had previously adduced the

further reason that whereas inductive generalizations are tested by means of
instantial evidence, hypotheses are tested by means of consequential evidence.
The evidence for primitive inductions or generalizations is always instantial evidence,
whereas the evidence for hypotheses is always consequential evidence .... The fact that
Tom's fmgerprints are on the knife and that the depth of this shadow is so-and-so are
consequential but not instantial evidence for the respective hypotheses that Tom did
it and that light travels in straight lines (1961, p. 67).

Making use of these argumentations, qualified inductivists agree with the old
inductivists in concluding that induction is operative in the discovery of
scientific laws, but agree with the anti-inductivists in maintaining that it is
operative only in the simplest, most elementary laws, those with the lowest
empirical level. From a historical point of view, induction is thus seen as
the childhood method of science. In principio erat inductio, deinde fit hy-
pothesis.
In order to evaluate this point of view, let us consider the following law-
like statements:
HI: 'Lions are carnivorous'.
H 2 : 'Magnets are composed of electric currents'.
As we have said, according to the qualified inductivist, HI is discoverable
by induction while H2 is not. And this is because the qualified inductivist
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 153

would argue that (1) H 1 is a generalization of facts that can be described

individually by means of singular statements containing only observational
predicates, while H2 contains a theoretical predicate or in any case refers to a
property not directly observed but surmised by an act of originality (in this
specific case the originality of Ampere); (2) H 1 is established or confirmed by
means ofinstantial evidence while for H2 the evidence is consequential.
And instead, H l' which can be expressed by means of the universal con-
ditional
(x) (Lx :J Cx)
is confirmed by any conjunction of singular statements such as
La . Ca, Lb . Cb, Le . Ce, ... ,
which are instances of it (L = lion, C = carnivorous), whereas H 2 , although
also having the form of the universal conditional
(x )(Mx :J ex)
(M = magnet, C = composed of electric currents), is confirmed by conjunc-
tions of singular statements of the type
Wa . Oa, Wa· Aa, ...
that is, by statements such as 'a is a metal wire through which passes electric
current (= W) and a orients a magnetic needle (= 0)" or 'a is a metal wire
through which passes electric current and a attracts iron filings (= A)" which
are manifestly not instances of it (of H 2 ) but rather are logical consequences
obtained with the aid of correspondence rules.
Now, it is certainly plausible to maintain that (1) and (2) constitute
evidence or criteria for two types of universal statements, H land H 2, the
first describing a mere accidental regularity and the second a genuine law of
nature. Although there are diverging opinions on the subject of the criteria of
lawlikeness,13 there can be no doubt that semantic and pragmatic differences
between the two do exist. However, differences of this kind have no bearing
whatsoever on the question of whether the regularities denoted by statements
of the first type are describable by means of induction or those denoted by
statements of the second type by means of the invention of hypotheses.
Indeed, these very differences show that such a distinction is untenable, since
on closer inspection they either disappear altogether or else remain, but only
in terms of degree.
Consider the first difference mentioned, namely that H 2 , unlike H l ,
154 MARCELLO PERA

introduces a predicate ('composed of electric currents') denoting a non-

observational property. From the theoretical point of view, this difference
does not subsist, because it is false that H 2 contains a new theoretical concept
imposed on the observed facts and HI does not. For insofar as HI is a hy-
pothesis that genuinely seeks to explain phenomena, it too contains a new
concept ('carnivorous') which is not passively registered or gleaned from the
facts, but instead is imposed on them. Thus, to take an example, if I am
walking through the savannah and at a certain point I ask myself 'Why do
lions attack antelopes?' and introduce the hypothesis HI, 'lions are carni-
vorous', then the predicate 'carnivorous', although observational, is new and
not contained within the domain of the observed facts. Pursuing further the
theoretical point of view, my hypothetical situation in the savannah is identi-
cal to the real situation facing Ampere in 1820 when he was confronted with
the experiments of Oersted and Arago: to the questions, 'Why does a conduc-
tor through which current is passing attract iron filings?', 'Why does a wire
through which current is passing deflect a magnetic needle?' Ampere produced
a reply that introduced the hypothesis H 2, 'magnets are composed of electric
currents'. That the predicate 'carnivorous' is supposed to be observational
because that is what the property it denotes is, while the predicate 'composed
of electric currents' is supposed to be theoretical because its denotation is not
accessible to direct observation, is of very little importance as regards the
inventiveness required to introduce them. Exactly the same act of conceptual
colligation, to use an expression of Whewell's, is required in both cases.
Admittedly, from a practical point of view few are the minds that know how
to formulate hypotheses in terms of non-observables like that of Ampere
about magnets, while practically all tourists on excursions through the
savannah reach the same hypothesis as I did about lions. But this is not
enough to warrant making some kind of split between the two kinds of hy-
pothesis regarding the modality of discovery. Anyone who wants to maintain
that all laws are discovered by induction is not obliged to also maintain that
all inductions are equally easy.
Let us now examine the second difference between HI and H2 men-
tioned above, namely that the former is tested only by instantial evidence and
the latter only by consequential evidence. Like the other, this is a difference
which can be invoked to distinguish mere regularities with a low nomological
content from true scientific laws; but, again like the other, this difference is
not specifically relevant to the question of the discovery of the two types of
law. Moreover, careful consideration shows that it is false that all hypotheses
of type HI are tested differently from those of type H 2 • Against J. P. Day'sl4
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 155

contention one can in the first place invoke those existential hypotheses that
refer to the existence of observable entities. Bouvard, Bessel, Adams and Le-
verrier's hypothesis 'there exists a transuranic planet' was certainly put to the
test by means of implications and predictions. But there are also many hy-
potheses that, having the form of universal statements like those of the type
HI, are nevertheless tested in the same way as those of the type H 2 : thus
the hypothesis, 'migratory birds orientate themselves on the basis of the
celestial signs', is confirmed by logically deriving predictions of certain kinds
of behavior under certain kinds of situations.
By denying that laws and theories can be sharply differentiated according
to the method of testing them and above all according to how they are dis-
covered, we may seem to be upholding the anti-inductivist thesis: if, even
where it would appear to be legitimate to speak of mere generalization
processes, it turns out that the discovery of laws originates from an inven-
tion of hypotheses in the form of a conceptual colligation, one might well
conclude that discovery is not a matter of induction or reasoning. It appears,
however, that one can equally well arrive at the opposite conclusion; it is to
the legitimacy as well as to the advantages of this alternative point of view
that I will now turn.

V. HYPOTHESIS AND INDUCTION.

THE LOGICAL POVERTY OF HYPOTHETICO-DEDUCTIVISM

The arguments on the basis of which the hypothetico-deductivists deny that

the method of scientific discovery is inductive basically amount to the
following two: (1) there is no 'discovery machine' or set method or logic
for arriving at new ideas; (2) the discovery of ideas is a creative act requiring
intuition, imagination and individual talentY BaSically, hypotheses (or
theories) are regarded as being, in Popper's words, ''free creations of our own
minds, the result of an almost poetic intuition" (1963, p. 192).
But even though both these arguments are correct, they do not, on the
other hand, support the thesis they are meant to prove. Popper and the hypo-
thetico-deductivists maintain that the procedure of collecting data in order to
derive conclusions inductively from them is impossible for logical reasons; on
the contrary, in this fmal section I will attempt to show that it is precisely the
opposite procedure - the procedure of conjectures and refutations (or con-
firmations) - which is crippled by two logical paralyses. We shall see that the
only effective therapy consists in recognizing that hypotheses are derived by
156 MARCELLO PERA

induction; likewise, the selfsame manner of argument will go to show the

continuous nature of the process of scientific discovery.
As is well known, according to the hypothetico-deductivist conception
a hypothesis is a conjecture or guess which, once conceived, is put to test by
deriving deductively certain observational consequences 010 O 2 , ... , On, and
then comparing them with some facts: if the outcome of the comparison is
negative, the hypothesis is falsified, while if it is positive, the hypothesis is
confirmed or (according to Popper's anti-inductivist exorcism) 'corroborated'.
The usual logical schema for this segment of the scientific procedure is as
follows (with .~, indicating logical deduction):
(1)
H~Ol' O 2 , ... , On
01,02,···,On
H is probable
This schema, in addition to being deliberately simplified by omission of
reference to the initial conditions under which 0 1, O 2 , ... , On derive from
H and to the auxiliary hypothesis which, taken in conjunction with H, allows
the same derivation, is imprecise on one important point. No mention is made
in it of the initial probability of H, and this makes the conclusion logically
incorrect. For, as can be grasped intuitively and as can easily be proved with
the aid of Bayes' theorem,16 no hypothesis can be considered probable no
matter what the amount and quality of its observed consequences may be if
it does not possess a certain finite degree of probability to start off with.
The corrected, if still simplified, version of the confirmation schema is
the following:
(2) H is probable
H~01o O 2 , ... , On
01> O 2 , ... , On
H is more probable
The initial probability of a hypothesis is its plausibility. Thus, schema
(2) tells us that if a hypothesis is plausible then it can become probable but
not otherwise: without initial plausibility, the verification of the consequences
of a hypothesis does not carry any weight, does not afford any 'corrobora-
tion', i.e. as the very word suggests, any increase of strength.
This is a well-known circumstance, but it raises a question to which not
enough thought has been devoted. The question is: to which phase of the
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 157

research does the argument by which a hypothesis is judged to be plausible

belong? Does it belong to the 'context of discovery' or to the 'context of
justification'?
The suggestion that the plausibility issue pertains to the context of dis-
covery was first put forward - with a reference to Peirce - by N. R. Hanson
and then later further elaborated by P. Achinstein. 1 7 W. Salmon objects,
however, that three, not two, thought episodes should be distinguished,
namely: (1) the invention of a hypothesis, (2) plausibility considerations and
(3) the testing arguments. Salmon's contention is that (2) is an essential part
of (3), but that (2) and (3), which belong to the context of justification,
should be logically differentiated from (1), which does not have the nature of
an argument, being instead part of the psychology of discovery.ls
Against this solution - a solution no less discontinuist than the classical
hypothetico-deductivist one - there exist strictly logical considerations. If
one introduces the plausibility arguments or the 'good reasons' only in the
context of justification, which is equivalent to considering them as a first
appraisal of the hypothesis, then one gets caught in an infmite regress. For, as
is evident from the Bayesian schema (2), any such evaluation requires some
initial probabilities, and these in turn some other probabilities and so on.
Whence then should we draw the original initial probabilities which are an
indispensable prerequisite for any evaluation?
This is the first logical paralysis of hypothetico-deductivism, even in
Salmon's revised version. The only way to avoid it is to recognize that the
plausibility considerations are not something which is added to a hypothesis
after it has been invented, but are the very reasons in the light of which a
hypothesis is conceived and advanced. This is equivalent to recognizing that a
hypothesis is not an a-logical or pre-logical guess but the plausible conclusion
of an inference and hence the result of an induction. This is furthermore
equivalent to recognizing that the process of scientific research is continuous
and consists of a continuum of arguments, from the initial one from which
the hypothesis springs to the successive ones by which it gradually grows
stronger.
Besides, the argumentative nature of hypotheses can also be ascertained
by reflecting on the function that 'hypotheses' perform in the common
meaning of this term discussed in Section III. If a hypothesis is a provisional
explanation, a provisional answer to a problem, then in order to be recognized
as such, in order to exist as a hypothesis, the hypothesis must contain those
reasons which establish its plausibility as an explanation. When a researcher
thinks up a hypotheSis he may know, and actually often does not know, how
158 MARCELLO PERA

and why he has conceived that idea, i.e. how and why he made that inference;
yet he does know what inference he has made: it is precisely the inference
through which he expresses, in the first place to himself and then to others,
the reasons why he thinks that the research problem is solvable, and the facts
explainable, in that particular way. There do not exist nor can there exist
two distinct thought episodes - flrst blind invention and then the interven-
tion of the plausibility considerations; there is rather a single argumentative
act: a hypothesis springs from the very same argument which provides the
initial reasons of its plausibility.
One example, among the many, will illustrate the adequacy of this induc-
tivist point of view.
On March 18, 1755 Benjamin Franklin sent the physician John Lining,
who had asked him how he had reached the hypothesis of the identity of
nature between electrical fluid and lightning, the following extract from the
records of his experiments:
Nov. 7, 1749. Electrical fluid agrees with lightning in these particulars: (1) Giving light.
(2) Colour of the light. (3) Crooked direction. (4) Swift motion. (5) Being conducted
by metals. (6) Crack or noise in exploding. (7) Subsisting in water or ice. (8) Rending
bodies it passes through. (9) Destroying animals. (10) Melting metals. (11) Firing inflam-
mable substances. (12) Sulphureous smell. The electric fluid is attracted by points. We
do not know whether this property is in lightning. But since they agree in all the par-
ticulars wherein we can already compare them, is it not probable they agree likewise in
this? Let the experiment be made (1755, p. 524).

I.et us indicate by F the electric fluid, by L the lightning and by PI, ... ,P13
the given properties; Franklin's reasoning can then be represented in the
following form:
Fhas the properties PI. Pl , •.• ,P1l • P 13
L has the properties PI. Pl , ••• ,P12

H: It is plausible that L may also have the property P 13

and therefore that F and L may be of the same nature.
This is not a 'wild' or 'bold' conjecture; rather, this is an analogical induction,
an argument by which a plausible hypothesis is derived starting from certain
facts.
Of course the thesis that a hypothesis arises by induction does not imply
the existence of an inductive logic working as a discovery machine which can
replace the intervention of individual talent. There are no recipes for evolving
new ideas, for grasping this or that analogy, for drawing this or that inference:
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 159

the rules of logic, be it inductive or deductive, are still canons of evaluation

or arguments, not organs for the production of inferences. But the fact that
there cannot be an organ for the invention of new ideas does not imply that
the invention of a hypothesis is bound to be an irrational, ineffable act not
susceptible of a rational reconstruction. The tacit avowal of the validity of
the two above implications is another error of modern hypothetico-deduc-
tivists. As in the instance of the fallacies considered in Sections II and III,
here too we have an instance of anti-Baconian reaction going too far: in
response to the optimistic Lord Chancellor, who was all for using his tables
to put all men's wits on an equal footing, the objection must unfortunately
be raised that induction is not a method in this sense; nevertheless it has to
be conceded that scientific discovery does have an inductive procedure: the
first act of this procedure - the invention of a hypothesis - is an inference
that can assume all the forms of an inductive argument; and in a similar
fashion the successive acts - whereby the hypothesis is tested - are also
inferences with the very same forms.
We are now left with what is going to provide a further confirmation of
the validity of our conclusion, namely the second logical paralysis which
cripples the hypothetico-deductivist conception. This paralysis arises as
follows: if scientific procedure is exhaustively contained in the pair, 'con-
jectures and refutations' (or confirmations), then this procedure turns out
to be no more viable than that of the crude inductivism: for, as there is no
limit on the quantity of observations the crude inductivist has to collect,
so there is no limit on the number of hypotheses that the hypothetico-
deductivist has to put to test. If the former is compelled to observe every-
thing, the latter is forced to test everything. But this would block all scien-
tific activity: the hypothetico-deductivist or Popperian scientist would be
paralyzed exactly like his despised Baconian colleague.
The logical impasse is well illustrated in the following passage in which
R. Feynrnan describes the difficulties surrounding scientific discoveries and
the way in which they happen:

It is something like this. You are sitting working very hard, you have worked for a long
time trying to open a safe. Then some Joe comes along who knows nothing about what
you are doing, except that you are trying to open the safe. He says "Why don't you try
the combination 10.20. 30?" Because you are busy, you have tried a lot of things, maybe
you have already tried 10.20.30 (1965, p. 161).

The drawback readily meets the eye: if - as Feynrnan maintains - the

scientific method is a method whereby "first we guess, then we compute
160 MARCELLO PERA

the consequences of the guess ... and then we compare the result of the
computation to nature" (1965, p. 156), then we cannot excuse ourselves
from testing any suggestion whatsoever. Feynman should therefore not neglect
combination 10.20.30 either; if he does, it is evidently because he regards its
plausibility as too low: and after all we can hardly blame him for this, since
the one who has suggested it to him is any old Joe and not the famous Joe
the burglar wanted by the police on account of his unbridled love for safes.
But although it may be perfectly reasonable from our point of view, for
a consistent hypothetico-deductivist burglar the refusal to try a combination
is not legitimate; the same goes for a hypothetico-deductivist scientist: a
priori one guess is as good as another and no guess is more plausible than
another until it has been evaluated and hence submitted to a fIrst form of
test. Therefore the categorical imperative of the hypothetico-deductivist
is to test all conjectures. But 'all' conjectures are an infmite number; thus
the order cannot be obeyed: ad impossibilia nemo tenetur.
As an explanation of why in actual fact scientists do not get lost in this
theoretical infinity, Medawar, perhaps the only one to have clearly seen this
logical diffIculty, has advanced the following suggestion:
In real life, of course, just as the crudest inductive observations will always be limited
by some unspoken criterion of relevance, so also the hypotheses that enter our minds
will as a rule be plausible and not, as in theory they could be, idiotic. But this implies
the existence of some internal censorship which restricts hypotheses to those that are
not absurd, and the internal circuitry of this process is quite unknown. The critical
process in scientific reasoning is not therefore wholly logical in character, though it can
be made to appear so when we look back upon a completed episode of thought (Medawar,
1969, p. 53).

This is a reasonable account of what actually happens; however, it is in-

consistent with the hypothetico-deductivist and Popperian conception which
Medawar upholds. To this account one can in fact object: 1) that it sub-
ordinates the logic of knowledge to the psychology of knowledge (indeed, to
a psychological mechanism still 'quite unknown'), something which amounts
to a mortal sin for all Popperians; 2) that, once we have come to rely on the
psychology of knowledge, there is no guarantee whatsoever that the hypo-
theses which come to our mind will all be plausible, nor that the plausible
ones are fmite; fInally, 3) that even if there were ~uch a guarantee, this would
amount to acknowledging that hypotheses, far from being 'free creations',
spring up in our mind in conformity with and subject to certain plausibility
criteria and hence under the form of inductive arguments and in accordance
with an inductive logic. It is therefore not consistent to invoke an 'internal
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 161

censorship' and at the same time to maintain, as Medawar does, that "the
process by which we come to formulate a hypothesis is not illogical but non-
logical, i.e. outside logic" (1969, p. 46).
In actual fact the 'internal censorship' is the Trojan horse unwisely allowed
through the walls of the hypothetico-deductivist conception. I believe that if
reasoning is not restrictively interpreted as mechanical reasoning and if it is
recognized that it does not exclude creativity or imagination or individual
talent, no relevant objection is left to the thesis that the inductive method is
the method of science. Perhaps the last remaining obstacle is only psycho-
logical resistance to the rejection of an old (and glorious) paradigm.
Translated by
RACHEL COSTA
NOTES

* I am extremely grateful to Professor Francesco Barone for having read a former ver-
sion of this paper and for having discussed the main thesis with me. My sincere thanks
also go to Professors John North and Carl Kordig for their stimulating comments and
suggestions.
1 See Bacon (1620), p. 109. "My way of discovering science goes far to level man's wits,
and leaves but little to individual excellence; because it performs everything by the
surest rules and demonstrations".
2 See Descartes (1628), pp. 371-72: "Per methodum autem intelligo regulas certas et
laciles quas quicumque exacte servaverit, nihil unquam falsum pro vero supponet, et
nullo mentis conatu inutiliter consumpto, sed gradatim semper augendo scientiam, per-
veniet ad veram cognitionem eorum omnium quorum erit capax" .... ["by a method I
understand certain and easy rules such that whoever has employed them exactly never
supposes anything false as true, and without uselessly consuming his mental effort but
rather always gradually increasing his knowledge, will arrive at a true cognition of all
those things of which he will be capable". 'Rules for the Direction of the Native Talents',
in Descartes, The Essential Writings, trans. J. J. Biom, New York, Harper and Row,
1977, pp. 22-98. Cf. p. 31.
3 Cf. Popper (1959), p. 49: "A methodology is a theory of rules of scientific method".
According to Lakatos (1971, p. 92), "modern methodologies or 'logics of discovery'
consist merely of a set of (possibly not even tightly knit, let alone mechanical) rules for
the appraisal of ready, articulated theories". Cf. also Radnitzky (1977), p. 1: "a meth-
odology is a system of rules or recommendations about how to act in certain types of
research situations in order to facilitate achieving the aim of this activity: scientific
progress". Very important on this and related topics is Radnitzky(1979).
4 The following defmition (Enciclopedia Mondadori delle Scienze, Milan, 1967, under
the heading 'Method') combines the second and the third meanings: "the scientific
method is the set of rules for the formation of concepts and the drawing of inferences,
together with the complex of techniques of tested observations that are used in the
acquisition of knowledge".
162 MARCELLO PERA

5 The current Methodenstreit between the school of Popper and the 'anarchist' school
or the trends of thought dominated by the latter concerns the very possibility or useful-
ness of invariant rules defining the 'game of science'. And indeed this is quite a trouble-
some point; in fact I think that the explication of each one of the three meanings of
'method' leads to worrying results that could be summarized in the following 'paradox
of the scientific method': "science is characterized by its method, but the characteriza-
tion of the scientific method destroys science". Despite this paradox I do not, however,
think that the best policy in matters of methodology is that of declaring oneself to be
'Against Method' even though this is undoubtedly the best way to shake up ingrained
manners of thinking and to call to the attention of lazy minds something that does con-
stitute a real problem. On the paradox of the scientific method see Pera (1978), Part I.
6 On the thesis of the 'mental jump' see Popper (1963), p. 46: "without waiting for
premises we jump to conclusions"; Wisdom (1952), p. 49: "a hypothesis is attained by
some mental jump".
7 The most complete and sophisticated version of the Popperian model of scientific
procedure is to be found in Radnitzky (1976) and (1977).
8 In his letter of 1861 to Henry Fawcett; cf. Darwin (1903), Vol. I, p. 195.
9 As always, a number of hypotheses can 'save' the same phenomena; in our case, even
the one - which happens to be ad hoc - advanced by J. Agassi (1975, p. 152) according
to which Darwin "collected while on the Beagle as much information as he could for an
obvious technical reason: he did not hope to arrange a second visit".
10 On this matter Z. Bechler's study (1974) is rather insightful. But it is likely that
Bechler would not be in agreement with the view that Newton's optical controversies
(especially with Huygens) derive from a conflict of epistemological assumptions in
the sense being used here. Cf. the following note.
11 In my (1978), Chapter II. 5, I have presented a more analytic classification and have
provided a variety of exemplifications regarding the various functions assumptions come
to have in scientific research. Referring to the meanings (1)-(4) as indicated, I speak,
respectively, of 'fundamental assumption', 'derivative assumptions', 'local assumptions'
and 'epistemological assumptions'. As far as their logical status is concerned, it is my
claim that they have the status of a priori principles that remain invariant throughout
whole historical eras (except the fust principle which is rigid) during which they leave
their imprint on research or provide the 'image of science'. I believe that the concepts
of 'paradigm' or of 'research programmes' would become more effective if they were
more fully articulated (especially the former, which is still very much in the same posi-
tion as Thomson's atomic model if one compares the latter with Rutherford's or Bohr's
models). If one had this more articulated reconstruction of the growth of scientific
knowledge, a revolution in the sense of Kuhn would, according to the ideas being enter-
tained here, be a drastic innovation undergone by the theoretical framework together
with the forsaking of some principle or accompanied by a different distribution of their
weight. For it appears that during scientific revolutions some principles - namely the
most central and protected ones, hence the ones which in a Kantian manner we may
consider to be constitutive - remain: thus, for instance, in astronomy the (derivative)
assumption of simplicity survived even after Kepler had abandoned the (local) assump-
tion of circularity of the orbits.
12 Cf. von Wright (1957), pp. 206-209; J. P. Day (1961), especially Chapter 5.2.4;
M. Bunge (1960); R. Camap (1966), Chapter 23.
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 163

13 For a useful survey, discussion and proposal, cf. Koningsveld (1973).

14 On this point Carnap's position is different; he claims that "the hypothesis is then
tested in a manner analogous in certain ways to the testing of an empirical law" (1966,
p.230).
15 Consider for instance the following passages, which contain both points. Popper:
"the initial stage, the act of conceiving or inventing a theory, seems to me neither to
call for logical analysis nor to be susceptible of it .... Every discovery contains 'an
irrational element' or 'a creative intuition', in Bergson's sense" (1959, pp. 31-32).
Reichenbach: "the act of discovery escapes logical analysis; there are no logical rules in
terms of which a 'discovery machine' could be constructed that would take over the
creative function of the genius" (1954, p. 231). Carnap: "there cannot be an inductive
machine - a computer into which we can put all the relevant observational sentences
and get, as an output, a neat system of laws that will explain the observed phenomena"
(1966, p. 33). Hempel: "there are, then, no generally applicable 'rules of induction' by
which hypotheses or theories can be mechanically derived or inferred from empirical
data" (1966, p. 15).
16 Let H be a hypothesis and E the evidence; from Bayes' theorem
p (H, E) = P (H) . p(E, H)
p(E)
it follows that if the plausibility p(H) is zero, the posterior probability p(H, E} is like-
wise zero, even if the likelihood peE, H} is high or equal to one. On the othm hand, it
also follows from the theorem that even a low plausibility of H can increase considerably
in posterior probability: it only needs for expectation p (E) to be low.
1? Among Hanson's works on this subject, see (1958a), (l958b), (1963), (1965a),
(1965b). Achinstein has corrected Hanson in (1971), Chapters VI-VII, which contain
the most biting and documented criticism of the hypothetico-deductivist conception.
18 Cf. Salmon (1966), Chapter VII. The separation into three episodes - initial thinking,
plausibility, acceptability - is shared also by C. R. Kordig (1978): "logic is not essential
to initial psychological thought .... Plausibility and justification require good reasons.
Thus, initial psychological thought is logically distinct from both plausibility and justi-
fication" (pp. 114-115). Although I find this way of thinking unacceptable, I agree with
Kordig that the plausibility and acceptability reasons "need not, and usually do not,
change when scientific theories change" (pp. 115-116; but cf. also Kordig, (1974}).
However, I doubt that Kordig would agree with my explanation of this fact, namely
that plausibility and acceptability standards are invariant (at least within vast historical
eras) because their nature is that of a priori principles (see Note II).

BIBLIOGRAPHY
Achinstein, P.: 1971, Law and Explanation (Clarendon Press, Oxford).
Agassi, J.: 1975, Science in Flux (D. Reidel, Dordrecht - Boston).
Bacon, F.: 1620, Novum Organon, in The Works of Francis Bacon, ed. by J. Spedding,
R. L. Ellis and D. D. Heath (Longmans and Co., London, Vol. 1, 1857).
Bechler, Z.: 1974, 'Newton's 1672 Optical Controversies. A Study in the Grammar of
Scientific Dissent', in Y. Elkana (ed.), The Interaction between Science and Philosophy
(Humanities Press, Atlantic Highlands, N. J.)
164 MARCELLO PERA

Bernard, C.: 1865, Introduction d l'hude de la medicine experimentale (J. B. Bailliere

et Fils, Paris).
Bunge, M.: 1960, 'The Place of Induction in Science', Philosophy of Science, 22, 262-
70.
Carnap, R.: 1966, Philosophical Foundations. of Physics, ed. by Martin Gardner (Basic
Books, New York-London).
Cohen, M. R. and Nagel, E.: 1934, An Introduction to Logic and Scientific Method
(Routledge and Kegan Paul, London).
Darwin, C.: 1903, More Letters, ed. by F. Darwin and A. C. Seward, 2 vols. (Murray,
New York).
Darwin, C.: 1958, The Autobiography of Charles Darwin, ed. by Nora Barlow (Collins,
London).
Day, J.P.: 1961, Inductive Probability (Routledge and Kegan Paul, London).
Descartes, R.: 1628, 'Regulae ad directionem ingenii', in Oeuvres de Descartes, published
by C. Adam and P. Tannery (Vrin, Paris, Vol. X, 1966).
Feynman, R. P.: 1965, The Character of Physical Law (British Broadcasting Company,
London).
Franklin, B.: 1755, 'Letter of March 18, 1755 to John Lining', in The Papers of Benja-
min Franklin, ed. by L. W. Labaree (Yale University Press, New Haven, Vol. 5, 1962).
Hanson, N. R.: 1958a, 'The Logic of Discovery', The Journal of Philosophy 55, 1073-
89.
Hanson, N. R.: 1958b, Patterns of Discovery (Cambridge University Press, Cambridge).
Hanson, N. R.: 1960, 'More on the "Logic of Discovery" " The Journal of Philosophy
57,182-88.
Hanson, N. R.: 1963, 'Retroductive Inference', ·in Philosophy of Science. The Delaware
Seminar, Vol. I (Interscience Publishers, New York).
Hanson, N. R.: 1965a, 'The Idea of a Logic of Discovery', Dialogue 4, 42-65.
Hanson, N. R.: 1965b, 'Notes towards a Logic of Discovery', in R. J. Bernstein (ed.),
Perspectives on Peirce (Yale University Press, New Haven).
Hempel, C. G.: 1966, Philosophy of Natural Science (Prentice-Hall, Englewood Cliffs,
N. J.).
Koningsveld, H.: 1973, Empirical Law, Regularity and Necessity (H. Veenman and
Zonen, Wageningen).
Kordig, C. R.: 1974, The Justification of Scientific Change (D. Reidel, Dordrecht).
Kordig, C. R.: 1978, 'Discovery and Justification', Philosophy of Science 45, 110-117.
Lakatos, I.: 1971, 'History of Science and Its Rational Reconstructions', in R. D. Buck
and R. S. Cohen (eds.), Boston Studies in the Philosophy of Science, Vol. 8 (Reidel,
Dordrecht), pp. 91-136.
Medawar, P. B.: 1969, Induction and Intuition in Scientific Thought (Methuen, Lon-
don).
Mill, J.S.; 1843, A System of Logic (Longman, London, 1970).
Pera, M.: 1978,Induzione e metodo scientifico (Editrice Tecnico Scientifica, Pisa).
Planck, M.: 1923, 'Kausalgesetz und Willensfreiheit', in Wege zur physikalischen
Erkenntnis (S. Hirzel, Leipzig, 1933).
Popper, K. R.: 1957, The Poverty of Historicism (Routledge and Kegan Paul, London).
Popper, K. R.: 1959, The Logic of Scientific Discovery (Hutchinson, London).
Popper, K. R.: 1963, Conjectures and Refutations (Routledge and Kegan Paul, London).
 INDUCTIVE METHOD AND SCIENTIFIC DISCOVERY 165

Popper, K. R.: 1972, Objective Knowledge (The Clarendon Press, Oxford).

Popper, K. R.: 1974a, 'Autobiography', in P. A. Schilpp (ed.), The Philosophy of Karl
Popper (Open Court, La Salle, Illinois).
Popper, K. R.: 1974b, 'Replies', in P. A. Schilpp (ed.), The Philosophy of Karl Popper
(Open Court, La Salle, lllinois). .
Radnitzky, G.: 1976, 'Popperian Philosophy of Science as an Antidote against Relativ-
ism', in R. Cohen, P. Feyerabend, and M. Wartofsky (eds.), Essays in Memory of Imre
Lakatos. Boston Studies in the Philosophy of Science, Vol. 39, (Reidel, Dordrecht),
pp.505-46.
Radnitzky, G.: 1977, 'Rationality and Progress in Research' (fIrst draft). Lecture in
Erice, February 1977 (this volume).
Radnitzky, G.: 1979, 'Metodo' in Enciclopedia del Novecento, Vol. IV (lstituto della
Enciclopedia Italiana, Rome).
Reichenbach, H.: 1954, The Rise of Scientific Philosophy (University of California Press,
Berkeley).
Salmon, W.: 1966, The Foundations of Scientific Inference (University of Pittsburgh
Press, Pittsburgh).
Wisdom, J.: 1952, Foundations of Inference in Natural Science (Methuen, London).
von Wright, G. H.: 1957, The Logical Problem of Induction (Basil Blackwell, Oxford).
 VITTORIO SOMENZI

SCIENTIFIC DISCOVERY FROM THE VIEWPOINT OF

EVOLUTIONARY EPISTEMOLOGY

In the mechanism proposed by Charles Darwin to explain the origin and sub-
sequent evolution of living species, there are two essential factors: the chance
variation of hereditary characteristics and the selection by the environment of
those best suited for survival. These factors were already interpreted in Dar-
win's day as two main phases in the process by which an organism solves the
problem of its adaptation to environmental variations.
In Darwin we fmd the metaphor of an environment that 'selects' among
the diverse descendants of an individual or couple those endowed with char-
acteristics which are advantageous for their conservation and reproduction;
hence in about 1870 Samuel Butler introduced the metaphor of the organism
which 'solves the problem' of its own individual and species survival by bring-
ing about hereditable modifications in its own somatic structure or in its
behavior.
Today we include among these behavioral modifications all those that
involve the construction or improvement of exosomatic instruments (from
nests and incubators to traps and dams), these being more easily comparable
to the inventions of human technology and to their 'cultural' propagation.
The Darwinian metaphor no longer appears as such when it is made clear
that the environment is not to be understood as the anthropomorphized
counterpart of the farmers or breeders who effect an 'artificial' selection
among the natural products they deal with, but rather it is their work which
is to be seen as one of the 'natural' components of the environment in which
what are subsequently called 'domestic' animals and plants happen to live.
Similarly, the Butlerian metaphor is rendered more acceptable when it is
made clear that the subject of the natural actions of 'problem solving' is not
the single individual (as in Lamarck's two hypotheses of the voluntary
development of organs or 'habits' for the satisfaction of new necessities im-
posed by the environment, and of the hereditability of the characteristics so
acquired), but the whole evolving species to which that individual belongs.
The metaphor can then be applied also to the human species, whose
cultural progress is seen as the natural extension of the biological one, but
usually with the reservation that individual creative processes are really
individual, even if species-specific for homo sapiens, and that their 'mechanics',
167
MD. Grmek, R. S. Cohen, and G. amino (eds.), On Scientific Discovery, 167-177.
Copyright © 1980 by D. Reidel Publishing Company.
168 VITTORIO SOMENZI

being purely mental, is completely different from that of Darwinian evolu-

tion, credited with the origin of man's body and brain.
Evolutionary epistemology, as recently developed with different philosophi-
cal approaches by Donald Campbell, Konrad Lorenz, and Karl Popper, starts
off instead with a non-metaphorical identification between the individual
processes of learning, invention, or discovery, and the collective processes of
adaptation of a species to an environment in a state of change.
Popper, despite his Platonic-Cartesian view of the interactions among
World 1 of nature, World 2 of individual consciousness, and World 3 of
language and culture, even goes so far as to affirm that "novelty in evolution
can be interpreted as being the result of a kind of invention, by the organism,
of a new environment: of a new ecological niche" and that "new ideas have a
striking similarity to genetic mutations". A Darwinian type of selection
among the 'possible' ideas of World 3 can, through the action of inhibitory
neurons, give rise to the formation of a new mental construction "working
like a sculptor who cuts away and discards part of the stone in order to form
his statue".
The neurophysiologist John Eccles, co-author with Popper of the recent
book The Self and its Brain, where the above remarks appear, believes that
the sudden appearance of new ideas can be compared to the birth of a 'brain
child', even if unreliable results sometimes come from that unconscious
process from which the creative illumination seems to emerge suddenly
through the effect of resonances among patterns of neuron excitation, which
would amplify one or another among the chance fluctuations of the cerebral
electrical activity.
The Darwinian character of chance variation and of selection in the
'incubation' of new organizations of ideas is underlined (also in its negative
aspects) by Jason Brown who notes that in the process of 'unconscious
cerebration' the sudden change of an element is usually incompatible with
survival: the new element (a 'thought' seen as a genetic mutation) survives
only through a slow readaptation or reorganization of the whole structure of
the organism.
The very idea nofjust of a formal similarity but of a substantial identity
between the mechanisms operating in nature in the creation of new living
forms and those operating in the mind or brain of a creative man, seems to be
subject to these alternations of rejection or acceptance by the scientific
community according to its apparent lethality or vitality in the context of the
biology and psychology of the time. Campbell points this out well in his
chronicle of the periodical appearance, disappearance and reappearance of the
 THE VIEWPOINT OF EVOLUTIONARY EPISTEMOLOGY 169

principles of an evolutionary epistemology in the philosophical and scientific

literature of the last two centuries.
Campbell neglects Butler's contribution to the transplanting of the
Darwinian approach into the field of intellectual creativity, perhaps because
this contribution moves in the opposite direction - that of a 'humanization'
of natural creative processes - and at times has a critical or satirical intent
towards the most simplistic Darwinism. But the amount of evidence collected
by Campbell is truly astounding, as Popper notes, and such as to induce even
Lorenz into making another attempt at the foundation of evolutionary
epistemology as a 'science', for which the philosophical tradition would
supply only the problems, while the collaboration of physicists, biologists,
psychologists, and sociologists would supply the instruments for their effec-
tive solution.
Campbell's point of departure is the transition from the direct exploration
of their own environment by animals having only sense organs of a mechani-
cal and chemical type, from touch to smell, to the indirect exploration
brought about by a central nervous system's processing of the information
coming from sense organs of an acoustical or optical type, capable of using
both external sources of sound or light waves and possible internal sources,
not to mention ultrasonic waves, or electric, magnetic and electro-magnetic
fields which are qualitatively or quantitatively different from those to which
the human organism is sensible. According to Campbell, the coincidence
between the impenetrability of most objects to animal locomotion and their
opacity to the visible band of electromagnetic radiations is the basis for the
substitution of 'blind' motor exploration of the environment by its pre-
liminary exploration by the eye. (Cases of exception are fog, scarcely relevant
to the evolution of the eye, and glass, certainly irrelevant in natural ecology.)
Between 1966 and 1978 Richard Gregory developed an interpretation of
the mechanism of sight as a mechanism for the formulation of hypotheses
about the external world and the postponement of their 'tactile' verification
to a possible subsequent moment. J. Z. Young, also neglected by Campbell,
began in 1951 his own attempt to biologize epistemology; this attempt shared
with Campbell's project the reference to Kenneth Craik's pioneering work
(1943). From both of these approaches there arises a strong need to relativize
the world image constructed by our brain and expressed in our language,
along with the consequent search for general views which go beyond the
anthropomorphic limits set by these two instruments, which perhaps were
not biologically designed to describe the world and ourselves.
In Logik der Forschung (1935) Popper had already emphasized the
170 VITTORIO SOMENZI

Darwinian character of the procedure of exposing scientific theories to the

'struggle for survival' aimed at falsifying the greater part of them in their
comparison with observational and experimental data. His 'conjectures and
refutations' would correspond to the 'trial and error' of organic evolution,
except for the obvious detail that the scientist benefits from the experience
accumulated along with the errors, outliving them, while a living species,
even the human species, if it is the bearer of harmful mutations, is destined to
extinction in the environment in which these mutations do not offer any
compensatory advantages.
Popper considers the growth of scientific knowledge as the most interest-
ing result of a growth of common knowledge, sharing with traditional gnosi-
ology and epistemology - from Descartes to Poincare, from Hume to Russell
- a faith in the possibility of applying their analyses, even if at first destruc-
tive, to the construction of methods which would augment the creativity of
the scientist. In Conjectures and Refutations (1963) Popper writes that "the
method of learning by trial and error - of learning from our mistakes -
seems to be fundamentally the same whether it is practised by lower or by
higher animals, by chimpanzees or by men of science". In both it is more a
case of 'directing' rather than 'knowing', of proceeding with 'premature'
forecasts founded on the prejudice of the regularity or lawfulness of the
events under consideration.
As in nature feed-back procedures, particularly the visual ones, permit the
survival of organisms which would otherwise pay for their own behavioral
errors with their lives, so in human culture "hypotheses die in our stead". We
can add that, as animal species sometimes lose an organ by natural selection,
which had been useful in a preceding phase of their evolution, so human
civilizations can abandon the use or fabrication of certain exosomatic instru-
ments which are no longer suitable for the new environment which they have
themselves found or built.
In 'Of Clouds and Clocks' (1966) the paradigm of natural selection is
seen by Popper as the universal non-teleological explanation of teleological
phenomena, of processes apparently guided by an end. The formation of a
crystal is the first example of order generated by a disorderly movement of
molecules prone to some couplings which are more stable than others. The
next example is given by the duplication of DNA, which takes place by
selection among the nucleotides which by chance approach the 'template'
thread. Campbell proceeds with the application of the Darwinian model to
widely different phenomena, from the formation of relatively stable aggre-
gates of subatomic and submolecular particles to embryonic development,
 THE VIEWPOINT OF EVOLUTIONARY EPISTEMOLOGY 171

from the healing of wounds to the regeneration of amputated limbs. In an

intentionally extremicist and 'dogmatic' way, he also extends the schema of
blind variation, together with processes of selection and of conservation or
propagation of the selected variants, to cases in which knowledge seems to
have been obtained in another way. The extension is warranted in so far as
such processes can succeed in explaining also the distant chance origin of non-
chance procedures excogitated by scientists in order to 'shortcut' some non-
compulsory stretches along their cognitive path.
Popper puts Lamarckism and inductivism on the same plane with regard
to the passivity of the animal's or man's reception of information from the
environment. To this conception he opposes the activism of the explorations
of this environment by Darwinian chance variations of the genome and by the
blind trials of the human brain. Like animal knowledge, "all human knowl-
edge is fallible and conjectural. It is a product of the method of trial and
error". Despite their common acceptance of Lorenz's innatism as far as the
human tendency for exploration and imitation is concerned, Popper begins to
diverge from Campbell in attributing to the scientist a preliminary knowledge
of the problem, also acquired by trial and error and nevertheless sufficient to
render the search for a solution to this problem not entirely 'random',
especially when many unsatisfactory solutions have been already tried.
Popper talks about 'decoding instincts' which, if they remain unsatisfied
as in the case of color-blind children, can give rise to neuroses due to the lack
of adjustment to the visual codes of normal children; but the instinct which
remains unsatisfied in this case could also simply be that of the imitation of
the behavior of other individuals of the social group in which the child
develops.
Further differences appear in the interpretation of Kantian a priori cate-
gories, which are objectively valid for Popper, while Campbell sees them as
merely preceding experience and necessary for it, and consequently traces
back the origin of the concept of causality to contingent features of the
human brain.
Campbell succeeds in putting a great number of examples of scientific
creativity within the framework of the formation of new ideas by chance
variation of preceding ones and the selection of their optimal combinations.
It is interesting in this regard to stress the type of verbal expressions used by
the various scientists and epistemologists to describe the experience of sudden
illumination or of the new idea which, just emerging onto the level of con-
sciousness, already appears capable of completing the mosaic of the previous
knowledge with the missing 'tessera' which resolves the long-studied problem.
172 VITTORIO SOMENZI

Mach speaks of the "result of a gradual selection" which "appears as if it

were the outcome of a deliberate act of creation", and he interprets in this
way the passive attitude of scientists and artists to the arrival in their brains
of a quantity of ideas, among which it would be sufficient to simply keep the
'right' ones and discard the rest.
Poincare believes that the selective criterion of the mathematician has an
aesthetic character in the sense that only 'harmonic' combinations of ideas
have the possibility of proving under subsequent detailed analysis to be valid
as mathematical truths.
Jevons appeals to statistics to immerge the few valid ideas of the 'fortunate'
researcher in an enormous quantity of combinations which are valueless but
which must in any case be made; only in this way may chance include, among
the many, those combinations which are capable of getting through the filter
of self-criticism.
Most explicit of all seems to be Paul Souriau (1881) with his motto,
"Ie principe de {'invention est l'hasard", and with his appeal to statistical
considerations not only for the individual attempts to solve a problem, but
also in regard to all the attempts of the community of scientists dealing with
it. A reference to the procedures of artificial selection is made by him, but it
appears more clearly Darwinian - as Campbell remarks - in William James
(1880), who attributes to the excessive instability of the human brain the
capacity for spontaneous variations in its functional activity, which the
external environment only confirms or refutes.
Aside from his reference to environmental selectors rather than to the
internal mental selectors of Mach, Poincare and Campbell, James too seems
to belong to the line of evolutionary epistemology which is interpretable to-
day in the neuropsychological terms of task division or the differentiation of
functions between the left and the right cerebral hemispheres of man and
anthropoid apes: if the chimpanzees of Wolfgang Kohler and his followers
also show a capacity for 'problem solving' by mental 'insight' instead of by
'trial and error' of a manual type, this really does not exclude the possibility
that the process of a blind search for a solution to a problem takes place
within their brains during the phase of 'meditation' which Koehler inter-
preted as the reorganization of the perceptual field. At the present stage of
neuroanatomical research these primates also appear to be endowed with
cerebral lateralization, with a development Of 'linguistic' areas in the domi-
nant hemisphere which make them suitable for learning the artificial languages
purposively designed for them by man. Thus it seems to me proper to attribute
a capacity to their 'minor' hemisphere, and so much the more so to man's,
 THE VIEWPOINT OF EVOLUTIONARY EPISTEMOLOGY 173

for 'holistic' or global and immediate individuation of the solution of a

problem involving the disposition of objects or symbols, a solution which the
motor organs commanded by the dominant hemisphere will then carry out in
time-ordered steps.
The transition from a phase of overall and silent information processing,
characteristic of the 'synthetic' and 'analogical' right hemisphere, to a phase
of logical and verbal processing characteristic of the 'analytic' and 'numerical'
left hemisphere, is especially evident, in my opinion, in the personal
experiences recounted by Poincare. That is to say, pattern recognition of an
aesthetic type, brought about suddenly by the 'intuitive' hemisphere, would
permit the recognition of pOSSible 'harmonies' in mathematical reasoning
before this is translated into symbols and further specified by the cortical
activity directed by the 'discursive' hemisphere.
In a comparison with the procedures of 'artificial intelligence', Herbert
Simon (1966) notes the possibility that in the human brain unconscious
processes operating 'in parallel' alternate with conscious processes operating
'in series'. He does not, however, attribute the difference to distinct interven-
tions of the two hemispheres and prefers to hypothesize that a serial process-
ing, provided with the artifice of time sharing, can also simulate a parallel
processing in the brain. This artifice would allow the rapid comparison of
calculations, made in different instants and in different explorative directions
by the brain, with the end of choosing the most effective heuristics. The
model of 'incubation' and 'illumination' processes, typical of scientific dis-
covery, which Simon proposes in tenns of 'familiarization', 'selective forget-
ting' and 'tree-blackboard' schemata, is very suggestive and interesting in view
of an imitation of the supposed human procedures by a program for electronic
computers. In a certain sense, the task of identifying the heuristically most
suitable 'graph' is removed from the analytic or serial procedure typical of
digital computers and of the left brain, and is transferred to the synthetic or
parallel procedure typical of the right brain and of electronic 'pattern recogni-
tion' devices. Not only would the form of the physical or mental phenomenon
under study be first examined by the right hemisphere, but also the fonn of
the reasoning or of the program which the left hemisphere would then be
made to execute.
Something analogous seems to occur in the evolution of organisms when,
at the beginning of life on Earth or in present 'neutral' and 'silent' mutations,
natural selection operates not on the phenotype taken into consideration by
Darwinian theories but on the genotype, that is, on the hereditary program
represented by the sequences of nucleic acids, whose variations in composition
174 VITTORIO SOMENZI

can or cannot affect the subsequent development of the organisms which

correspond to them.
We know that the evolution theorists of the last century were faced with
the obstacle of the great phylogenetic changes on genus, family and order
levels, and so for these changes they sometimes hypothesized the intervention
of sudden macro mutations which would have been more 'creative' than those
which give rise to different varieties or races within the same species or to
different species within the same genus. In the Kuhnian terminology adopted
by Simon, these macromutations would correspond to paradigm changes
characteristic of scientific revolutions, and the Origin of Species by Darwin
would constitute a first attempt to explain the biological revolutions not
explained by Linnaeus' fixism which confined itself to admitting infraspecific
variations and, within the framework of this comparison, would have con-
demned Nature to doing only 'routine' scientific research. Continuing in the
direction of the principal taxonomic ramifications, the mathematical theory
of 'catastrophes' would today be able to explain, according to Rene Thom
and his followers, the diversifications among orders, classes and types of
organisms, which are not sufficiently motivated by twentieth century neo-
Darwinian theories.
Unlike Campbell and contrary to the role he assigns to blind trials in the
evasion from the paradigms of a crystallized science, Simon reaffirms the
continuity of scientific progress as routine research's gradual preparation for
a change in paradigm. (In Darwinian evolutionism infraspecific variations
accumulate gradually until they give rise to a true speciation, and the same
thing happens for the variations within higher systematic categories.) That is,
the creative scientist cannot avoid taking previous results into consideration
and, according to Simon, his trials are not so blind as to make him again
travel over the same paths already tried in vain by his predecessors, not even
in that proper formulation of new questions which constitute, in the opinion
of several authors, the most important phase of a scientific revolution.
Simon deals with the powerfully selective heuristics employed by the great
scientific revolutionaries who were few in number in relation to the extent of
the history of human culture, but he does not consider it necessary to imagine
for them procedures which are radically different from those adopted by
other scientists. His epistemological interests consequently do not embrace
that definition of 'artificial intelligence' which excludes its use as a functional
model of natural intelligence, as it is known to us, because it refers to the
distant possibility that the development of the right programs, structural
elements, and circuits may permit electronic computers to resolve problems
 THE VIEWPOINT OF EVOLUTIONARY EPISTEMOLOGY 175

no man could ever resolve, however intelligent and long-lived he might be.
Less removed from Campbell's radical position, which excludes true crea-
tivity in choices which are oriented by the information the individual starts
with, is the 'operant conditioning' interpretation of innovations - which are
cultural in an ethological sense - brought about through blind trials by
exemplars of different animal species and rapidly imitated by other members
of the collectivity to which these exemplars belong. From this point of view,
the greatest creativity seems to be that of the child, as it appears in the fIrst
phases of the acquisition of any language. Already in Piaget the Haeckelian
principle that ontogenesis recapitulates phylogenesis is used for a comparison
between the individual development of the child's thought with regard to
natural phenomena and the historical development of scientillc thought.
In more recent years Howard Gruber has pushed the analysis of creativity
farther in this direction of 'starting from scratch' and of exploratory curiosity,
both in the child and in the scientist, reaching the conclusion (implicitly in
agreement with those of evolutionary epistemology) that Darwin himself is
a typical example of the necessity of adding the adult factor of courage and
tenacity to the infantile factor of curiosity and creativity.
W. Koehler, too, was struck by the repetition in numerous historically
checkable cases of a link between the creativity of the scientist and a diver-
sion from his predominant concern, brought about by accidental circum-
stances like a walk or a trip, Archimedes' 'Eureka!' bath, and the less famous
half-sleep that permitted Otto Loewi twice to intuit the mechanism of
nervous action on the cardiac rhythm. But some British physicists confIrmed
Kohler's idea in such generalizing terms as to perhaps discourage further
analysis: "We often talk about the three B's: the Bus, the Bath and the Bed.
That is where the great discoveries are made in our science".
Systematic analyses were carried out later by William Dement and others
with the usual statistical criteria of experimental psychology, that is, they were
carried out on numerous subjects, but neither exceptionally gifted ones, nor
persons who were deeply motivated to solve in their sleep the uninteresting
problems put to them when they were awake. In the few positive cases found,
the result would seem to confIrm the hypothesis that the nocturnal inter-
ruption of the inhibition exercised by the dominant hemisphere on the minor
hemisphere favors the elaboration by the latter of an answer expressed in fIg-
urative terms to the problem put to the dominant hemisphere in verbal terms.

Translated by
R.MABERRY and A. TALIERCIQ
176 VITTORIO SOMENZI

BIBLIOGRAPHY

Ashby, W. R.: 1956, 'Design for an Intelligence Amplifier', in Automata Studies, ed. by
C. E. Shannon and J. McCarthy (Princeton University Press, Princeton).
Brown, J.: 1977, Mind, Brain and Consciousness (Academic Press, New York).
Campbell, D. T.: 1960, 'Blind Variation and Selective Survival as a General Strategy in
Knowledge-Processes', in Selforganizing Systems, ed. by M. C. Yovits and S. Cameron
(Pergamon Press, New York).
Campbell, D. T.: 1974a, 'Evolutionary Epistemology', in The Philosophy of Karl Popper,
ed. by P. A. Schilpp (Open Court, La Salle).
Campbell, D. T.: 1974b, 'Unjustified Variation and Selective Retention in Scientific
Discovery', in Studies in the Philosophy of Biology, ed. by F. J. Ayala and T. Dobz-
hansky (Macmillan, London).
Craik, K. J. W.: 1943, The Nature of Explanation (Cambridge University Press, Cam-
bridge).
Craik, K. J. W.: 1966, The Nature of Psychology, A Selection of Papers . .• , ed. by S. L.
Sherwood (Cambridge University Press, Cambridge).
Dement, W. C.: 1972, Some Must Watch and Some Must Sleep (Stanford Alumni Asso-
ciation, Stanford).
Eccles, J. C.: 1970, Facing Reality (Springer Verlag, Berlin, Heidelberg, New York).
Evans, R. 1.: 1975, Konrad Lorenz, the Man and His Ideas (Harcourt, Brace, Jovanovich,
New York and London).
Fogel, L. J., Owens, A. J., and Walsh, M. J.: 1966, Artificial Intelligence through Simu-
lated Evolution (Wiley, New York).
Gregory, R. L.: 1966, Eye and Brain (McGraw-Hill, New York).
Gregory, R. L.: 1980-81, 'Are Perceptions like Hypotheses in Science?', in Levels of
Reality (An International Symposium, Florence, September 1978), to be published
by Feltrinelli (Milan).
Grmek, M. D.: 1976, 'Le role du hasard dans 1a genese des decouvertes scientifiques',
Medicina nei secoli, No.2, 277-305.
Gruber, H. E.: 1973, 'Courage and Cognitive Growth in Children and Scientists', Piaget
in the Classroom, ed. by M. Schwebel and J. Raph (Basic Books, New York).
Gruber, H. E.: 1974, Darwin on Man (a Psychological Study of Scientific Creativity)
(E. P. Dutton, New York).
Hadamard, J.: 1945, The Psychology of Invention in the Mathematical Field (Princeton
University Press, Princeton).
Jaynes, J.: 1976, The Origins of Consciousness in the Breakdown of the Bicameral Mind
(Houghton Mifflin, Boston).
Kohler, W.: 1969, The Task of Gestalt Psychology (Princeton University Press, Princeton).
Lorenz, K.: 1973, Die Riickseite des Spiegels (Piper, Munich).
Mach, E.: 1943,Popular Scientific Lectures (Open Court, La Salle, Illinois).
Mahdihassan, S.: 1961, 'Keku1t!'s Dream of the Ouroboros and the Significance of This
Symbol', Scientia No.6, 187-195.
Medawar, P. B.: 1967, The Art of the Soluble (Methuen, London).
Newell, A., Shaw, J. C., and Simon, H. A.: 1963, 'The Processes of Creative Thinking', in
Contemporary Approaches to Creative Thinking, ed. by H. E. Gruber, G. Terrell, and
M. Wertheimer (Atherton Press, New York).
 THE VIEWPOINT OF EVOLUTIONARY EPISTEMOLOGY 177

Piaget, J.: 1978, Behavior and Evolution (Pantheon Books, New York).
Poincare, H.: 1905,La valeur de la science (Flammarion, Paris).
Poincare, H.: 1908, Science et methode (Flammarion, Paris).
Popper, K.: 1963, Conjectures and Refutations (Routledge & Kegan Paul, London).
Popper, K.: 1968, The Logic of Scientific Discovery (Hutchinson, London).
Popper, K.: 1973, Objective Knowledge (Oxford University Press, Oxford).
Popper, K., and Eccles, J. c.: 1977, The Self and Its Brain (Springer, Berlin).
Pribram, K. H.: 1976, 'Language in a Sociobiological Frame', Annals of the New York
Academy of Sciences 280, 798-809.
Rumbaugh, D. M. (ed.): 1977, Language Learning by a Chimpanzee (Academic Press,
London, New York, San Francisco).
Simon, H. A.: 1966, 'Scientific Discovery and the Psychology of Problem Solving', in
Mind and Cosmos, ed. by R. G. Colodny (University of Pittsburg Press, Pittsburgh);
reprinted in Models of Discovery. Boston Studies in the Philosophy of Science,
Vol. 54 (Reidel Dordrecht, 1977), pp. 286-303.
Simon, H. A.: 1973, 'Does Scientific Discovery Have a Logic?' Philosophy of Science,
No.3, 471-480. (Reprinted in Models of Discovery. Boston Studies in the Philos·
ophy of Science, Vol. 54, Reidel, Dordrecht, 1977, pp. 326-337).
Toulmin, S.: 1971: 'Brain and Language: a Commentary', Synthese 22,369-395.
Toulmin, S.: 1972, Human Understanding, Vol. 1 (Clarendon Press, Oxford).
Wertheimer, M.: 1959,Productive Thinking (Harper, New York).
Young,1. Z.: 1951,Doubt and Certainty in Science (Clarendon Press, Oxford).
Young, J. Z.: 1978,Programmes of the Brain (Oxford University Press, Oxford).
Zaidel, E.: 1978a, 'Concepts of Cerebral Dominance in the Split Brain', in Cerebral
Correlates of Conscious Experience, ed. by Buser and Rougeul-Buser (Elsevier-North
Holland, Amsterdam).
Zaidel, E.: 1978b, 'Lexical Organization in the Right Hemisphere', Ibid.
 DEREK DE SOLLA PRICE

THE ANALYTICAL (QUANTITATIVE) THEORY OF

SCIENCE AND ITS IMPLICATIONS FOR THE
NATURE OF SCIENTIFIC DISCOVERY

Within the last few years a new level has been reached in a germinating field
of study that has developed from quantitative sociology of science, research
on science indicators statistics, general bibliometrics, and citation analysis.
From these advances I feel it is now possible to put forward, albeit tentatively
and with many reservations and much uncertainty at several places in the
technical detail, a comprehensive analytical theory of science. By this I mean
a conceptual framework that is consistent and extensive and can be related
both quantitatively and qualitatively to several places in the historical and
philosophical e~amination of science. It is the function of this paper to
explore the apparently far-reaching implications of this theory for the nature
of the concepts implied. This will be done in developing one topic and
discussing briefly two other separate researches in the field that display this
new approach to the explicand of 'Science of Science' or 'Science Studies',
viz:
(1) A study of the Ups and Downs in the History of Science and Tech-
nology 1 shows how quantification of almost anything in a time series can
throw light on what otherwise might be apparent only to a very competent
Toynbee-like historian. It also gives an objective method for commenting on
the periodization of science and suggests that the Scientific Revolution has
a precursor-like role and that the Industrial Revolution may be merely an
artifact of historiographical convenience.
(2) A study of Cumulative Advantage Processes by statistical mathe-
matics 2 gives results that go from a very simple probabilistic model to yield
quantitative laws in agreement with empirical evidence, and thereby explain
the peculiarities of all well-known regularities in scientometrics and biblio-
metrics. This gives a conceptual basis for the sociological hierarchies prevailing
in the scientific community.
(3) Recent discoveries by Griffith and Small show that science can be
mapped by using the technique of co-citation analysis. The resultant map is
surprisingly two-dimensional, and this indicates that irrespective of the map-
ping procedure, the structure of science can be ascertained in a very graphic
and provocative fashion. One important result is that the paper atoms cluster
into sub-discipline molecules each corresponding to an invisible college; there
179
M. D. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 179-189.
Copyright © 1980 by D. Reidel Publishing Company.
180 DEREK DE SOLLA PRICE

is no other level of aggregation. This result enables one to conjecture that

science can be modelled after a cooperatively solved jigsaw puzzle. It can be
shown that this model has important features that seem to provide rather
more exactly the sort of features often spoken of as Kuhnian paradigms,
which it thereby explains and extends. Another important feature of this
evidence is that it strongly suggests that scientific instruments and methods
procedures have a much more important role than usually allowed for by
historians and philosophers of science. It may well be that the true impor-
tance of instruments lies not in their testing of theories but in the provision
of new sense data not expected nor indeed desired by the paradigmatists.
Now I wish to develop the first topic.

I. MOTIVATION FOR THE QUANTITATIVE STUDY OF THE

UPS AND DOWNS IN THE PULSE OF SCIENCE AND TECHNOLOGY

My working hypothesis 3 is that the objectivity and transnational character

of basic science lend to its historical development a much larger element of
determinism and of imperviousness to local socio-economic factors than one
is accustomed to elsewhere in human affairs. It follows that a vital task of
the historian of science and technology is to analyze such quasi-automatic
secular change as proceeds regardless of particular causes, for only then can we
dissect out those non-automatic and significant events that require special ad
hoc explanation. We need to perceive and understand regularity of behavior
before we can get to second-order explanation of the deviations therefrom.

II. PREVIOUS WORK IN STATISTICAL PERIODIZA TlON

Taking the grand sweep in statistical historiography to measure the pulse,

so to speak, of particular sciences and for particular countries is an old and
doubtless influential tradition in our field. The importance of the early work
by Rainoff4 and by Sorokin 5 and others has been well summarized by Mer-
ton. 6 Yuasa, Tomita and Hattori 7 have added a wealth of data from Oriental
sources, and Simonton 8 has given the data the more sophisticated statistical
methodology and control that has long been needed for reliability estimates.
All these previous studies have sought to squeeze from quantitative data the
maximum information on the changing deployment of science and technology
from field to field and from nation to nation. The difficulty is large and the
intrinsic errors considerable, for one is trying to evaluate many different
 THE ANALYTICAL THEORY OF SCIENCE 181

causes acting in concert, and these are seen against a background of random
fluctuation that is huge in the crude data. Statistically, as a rule of thumb,
if one generates data by counting a population of events sorted into pigeon-
holes, a count of N events implies a fluctuation of magnitude ±VJV. If one
has as many as 100 events in a group the random error is therefore 10% and
one cannot be sure of the reality of distinctions from one group to another
beyond such a limitation.
To avoid such problems, the present research demands only the minimum
from the data, the main trend of an overview of all scientific and technolog-
ical activity without any disaggregation by nation or by field. This ensures the
maximum number of events in each annual group, and to make the fluctua-
tions even less, we have combined many sources of data and applied well-
known smoothing procedures to the time series.

III. METHODOLOGY

The starting point was an examination and hand-count, page by page, through
a large number of chronologies and histories of science and technology,
recording each event that was given a precise or an approximate date. For
control purposes a similar but smaller study was made of general history. Ihe
year counts were then smoothed by taking a running weighted mean, the
weighting factors being given by cos 2 (90° xm/4), m = 0, ±l ,±2,±3) so-that
it ran for three years on either side of the target year. Zeros being absent
from the smoothed series, the logarithms of this mean were taken, a linear
regression against time computed, and the deviations from the regression
found. These deviations then gave for each source the amount by which it
exceeded or fell short of a linear increase in the logarithm, i.e. of regular
exponential growth which is the gross deterministic behavior. The deviations
from all sources were then averaged to give the results shown in Figure 2, and
this was further given a grand smoothing by a running weighted average, this
time taking in 10 years on either side of target, to give the fmal product
exhibited in Figure 1. The data was found to be sufficiently numerous for
some confidence only from 1500 to the present time, though any individual
source yielded information only over part of the entire range. A similar in-
vestigation for the General History control confirmed that the regularities of
Figure 1 were peculiar to the history of science and technology; systematic
fluctuations of this sort were far less well marked and quite different for the
general history events.
 0.5
.....
00
N
COPERNICUS
SCIENTIFIC
REVOLUTION
)( 2.0 INDUSTRIAL
GALILEO
REVOLUTION
",

x 1.5-

"
steady
exponential 0 I . \ . \ '/ ., 1 /0: ~ {" ~ .y 'i'\ :/ \'. :I
growth
" "

BISMARK
-:- 1.5
BOER
CROMWELL
WAR
-:- 2.0 ININIININII
3D YEARS
WAR FRENCH (,
AMERICAN
REVOLUTIONS

-0.5+1;---;--~~--'----:-:~--"'----r----'---;------r--~
1500 1600 1700 1800 1900 2000
Fig. 1. Ups and downs in the pulse of science and technology.
 0.5

x 2.0

x 1.5-'

o·
\
+1.5

+2.0 ~-

- 0.5 -l-,---;---r-----r----,----r---:-::-C-----r----:~--,---~;_-
1500 1600 1700 1800 1900 2000 ....00
w
Fig. 2. Ups and downs; fine detail.
184 DEREK DE SOLLA PRICE

IV. INTERPRETATION

Each particular history or chronology had, as expected, its own exponential

growth rate, this depending partly on the acceleration of science itself and
partly on the idiosyncratic perspective of the compiler or historian, concen-
trating more or less on the distant past compared with the more recent. The
deviations of the logarithmic indicator were however very much the same
from one source to another, correlating well, and fitting together into an
extensive time series that looks somewhat irregular when seen in the fine
detail of Figure 2, but much more systematic in the highly smoothed version
of Figure 1.
It might be supposed that systematic deviation above and below regular,
even exponential growth gives one the pulse of historians' interest rather than
that of the events themselves. There exist, however, several statistical studies
of particular scientific field bibliographies, and others of the total literature
and the highly cited papers selected from it. These - for example the careful
study of Comparative Anatomy by Cole and Eales (Figure 3) - show just the
same pulse fluctuations and one is therefore driven to suppose that the pre-
sent study indicates very much more the intrinsic character of science than
it charts the prejudices and predilections of historians. For this reason we
have been able to use the recent citation data (Figure 4) to extend the histor-
ical study to the present day (dotted line on Figure 1).

V. HISTORIOGRAPHICAL IMPLICATIONS

It must be remembered that the indicator derived charts the secular and sys-
tematic swings of the exponential growth rate of science; it is not a measure
of the quantity of discovery and invention. It may be rather surprising that
over the entire range, the rate does not vary by more than about a factor of
two above or below the average long-term trend which we expect on theore-
tical grounds. Within these limits, I maintain the indicator gives one a measure
of scientific activity that agrees well with the historians' intuition of relatively
active and inactive periods.
Amongst the obvious expectations for eras of high activity one fmds very
clearly indicated the peaks of the scientific and the industrial revolutions as
well as localized outbursts in the times of Copernicus and Galileo (I use the
names only as surrogates for their periods). The major wars and social up-
heavals of history all correspond to localized troughs that are very clear.
Highly unexpected (at least to one historian) is the enormous trough
 180

>-l
~
<lJ
::t:
CII t!1
>-
........ >
Z
UI
c: >
t"'
o
'.j:; 0<
<lJ >-l
,~
:::0
n
::l >
t"'
." a.
dQ' - >-l
0 ::t:
.;., ~
t!1
CII 0
..c ~
E 0<
~ 50
0
40 ."
U>
30 n
;;
20 Z
n
10 t!1

1550 1600 1650 1700 1750 1800 1850

COMPARATIVE ANATOMY PUBLICATIONS BETWEEN 1540 AND 1860

• ..-
00
(adapted from Cole & Eales, 1917) VI
186 DEREK DE SOLLA PRICE

1906-9 1925-31 1950

b-------I-20%
1945-49

7"1.

1880 90 1900 10 20 30 40 50 60
WWI WWII
1913-19 1939-44

Five yearrunning average of number of heavily cited

papers per year by year of publication. All papers
cited at least ten times in any single year during
the period 1961-1972
1000+---------------~--------------------------~~~

x2 in
3.6 years

100+---------------~----------------_+~~----~--~

10+---------------+----~~~~+---------------~--~

I I
1880 1890 1900 1910 1920 1930 1940 1950 1960

Fig. 4. Growth and acceleration of scientific production in modern times.

 THE ANALYTICAL THEORY OF SCIENCE 187

indicating a post-Scientific Revolution slump of more monumental propor-

tions than any other event, and the rather weak and diffuse low profile char-
acter of the Industrial Revolution compared with other peaks. From the first
one must suppose that more is at stake than Newton being a hard act to
follow; from the second it seems likely that the Industrial Revolution may be
rather more of a compartmentalizing convenience of the historian than an
actuality.
I take the conventional view of historians of science to be that, ancient
and medieval groundings notwithstanding, modern science began with bursts
of activity in the fifteenth and sixteenth centuries, which led to the mid-
seventeenth century Scientific Revolution, setting the tone for all later devel-
opments, including a postponed revolution in chemistry and still more post-
poned events in other fields.
The effect on me of this new quantitative data for the general secular
changes is to make me revise this convention. It would appear much more in
keeping with the general pulse if one supposed that the early isolated peaks
for Copernicus, Galileo and the Scientific Revolution represent precursor-like
events, and that the main trend leading to modern times began only towards
the end of the eighteenth century. I suggested therefore that the (Lavoisier)
revolution in chemistry is not any sort of 'postponed' event, but a marker
of a period when all of the modern movement would have had its 'natural'
and expected beginning. There was therefore a premature inception through
the accident that astronomy comes out surprisingly neatly and had already
been successfully mastered by Ptolemy and the followers of mathematical
astronomy. Chemistry and biology were not late, it was astronomy and
mechanics that happened early and then suffered a lag during the compara-
tively dull first half of the eighteenth century.
The most recent period is best understood in terms of this general indicator
as a somewhat typical post-war recovery from a trough, leading to an over-
shooting and a contemporary recovery to oscillation in the neighborhood of
the long-term average growth rate. Thus the modulation of steady exponential
growth seems to correspond reasonably well with both historic and modem
intuitive estimates of general scientific and technological activity.

NOTES

1 This study was presented at the International Symposium on Quantitative Methods in

the History of Science (Berkeley, August 1976).
188 DEREK DE SOLLA PRICE

2 Published in Joumal of the American Society for Information Science 27 (1976),

292-306.
3 D. de S. Price, 'Toward a Model for Science Indicators', in Toward a Metric of Science
(Wiley-Interscience, New York, 1978).
4 T. J. Rainoff, 'Wave-like Fluctuations of Creative Productivity in the Development
of West-European Physics in the Eighteenth and Nineteenth Centuries', Isis 12 (1929),
287-319.
S P. A. Sorokin, Social and Cultural Dynamics. II: F7uctuation of Systems of Truth,
Ethics, and Law (American Book Company, New York, 1937).
6 R. K. Merton, 'The Sociology of Science: an Episodic Memoir', in The Sociology of
Science in Europe, ed. by R. K. Merton and J. Gaston (Southern Illinois University Press,
Carbondale, Ill., 1977).
7 Yuasa, M., 'The Shifting Center of Scientific Activity in the West', in Science and
Society in Modem Japan, ed. by S. Nakayama, D. L. Swain, and E. Yagi (University of
Tokyo Press, Tokyo, Japan, 1974). T. Tomita and K. Hattori, 'Some Considerations of
Quasi-Quantitative Analysis of the History of Science in Japan by Key-Words - Trial
of Quantitative History - " presented at International Congress of the History of
Science, Tokyo, 1974. T. Tomita and K. Hattori, Review of 'History of Science Society
of Japan (ed.): Nihon Kagaku-GijutfU-shi Taikei (History of Science and Technology in
Japan), 25 Vols., 1964-1970" Japanese Studies in the History of Science, No.9 (1970),
164-167. T. Tomita and K. Hattori, 'Compilation of a Thesaurus and Total Index for
Nihon Kagaku-GijutfU-Shi Taikei by Means of a Computer', Japanese Studies in the
History of Science, No. 11 (1972), 41-65. T. Tomita and K. Hattori, 'Outline of a
Thesaurus of Nihon Kagaku-Gijutsu-Shi Taikei with Heading List of Classified Key-
words', Japanese Studies in the History of Science, No. 12 (1973), 15-38.
8 D. K. Simonton, 'Invention and Discovery among the Sciences: A P-technique Factor
Analysis', Journal of Vocational Behavior 7 (1975), 275-281. D. K. Simonton, 'Socio-
cultural Context of Individual Creativity: A Transhistorical Time-Series Analysis',Jour-
nal of Personality and Social Psychology 32 (1975), 1119-1133. D. K. Simonton,
Private communications, September 12, 1975; September 19,1975; October 5, 1975;
and January 3, 1976. D. K. Simonton, 'The Causal Relation Between War and Scientific
Discovery. An Exploratory Cross-National Analysis', Journal of CrofS-Cultural Psy-
chology 7 (1976), 133-144.

BIBLIOGRAPHY

Darmstaedter, Ludwig: Handbuch zur Geschichte der Naturwissenschaften und der

Technik. 2d ed. (Springer, Berlin, 1908).
Darmstaedter, Ludwig, and Du Bois-Reymond, R.: 4000 Jahre Pionier-Arbeit in den
exacten Wissenschaften (Stargardt, Berlin, 1904).
Feldhaus, Franz M.: Lexicon der Erjindungen und Entdeckungen auf den Gebieten der
Naturwissenschaften und Technik in chronologischer Uebersicht mit Personen und
Sachregister (Winter, Heidelberg, 1904).
Garrison, Fielding H.: An Introduction to the History of Medicine with Medical Chron-
ology, Suggestions for Study and Bibliographic Data. 4th ed. (Saunders, Philadelphia
and London, 1929).
 THE ANALYTICAL THEORY OF SCIENCE 189

Pledge, H. T.: Science Since 1500, A Short History ofMathematics, Physics, Chemistry,
Biology (HMSO, London, 1939).
Walden, Paul: Chronologische Uebersichtstabellen zur Geschichte der Chemie von den
iiitesten Zeiten bis zur Gegenwart (Springer, Berlin, Goettingen, Heidelberg, 1952).
Williams, Neville: Chronology of the Modem World: 1763 to the Present Time (Mckay,
New York, 1966).
 GABRIEL GOHAU

DIFFICULTIES INHERENT IN A PEDAGOGY OF

DISCOVERY IN THE TEACHING OF THE SCIENCES

One can classify the whole range of pedagogical practices in two groupS:
on the one hand, those where the teacher alone speaks: I shall call these
expository methods: on the other, those where the teacher questions his
students: I shall call these interrogative methods.
Expository techniques are evidently those of text-books, as well as of
popularizing articles ... and lectures. In the classroom, they now meet with
a degree of disfavor. One has tended to label them 'dogmatic' since child
psychology has shown us that the young person is not an adult in miniature,
but someone needing to build his own knowledge progressively, by way of
successive re-adjustments.
Interrogative techniques came into being to respond to this need. There
the student is constantly solicited, so that the teacher is forced to follow, to a
certain extent, the course of his interlocutors' thought, with its hesitations,
deviations, even its temporary regressions. The archetype of these methods
is of course the Socratic dialogue, where the teacher manages, by way of
questions, to destroy the semblance of knowledge in the disciple, and replace
it by true knowledge.
This reference to Socratic maieutics shows that interrogative methods have
not waited for the development of child psychology to command the atten-
tion of the best pedagogues. Without aspiring to retrace the history of peda-
gogy, it may be said, very roughly, that the expository method seems to pre-
vail in periods when culture is being preserved, while the interrogative method
would be that of times of cultural creation, when previously acquired knowl-
edge is being questioned.
Thus, the Renaissance criticizes the dogmatic pedagogical methods of
the Middle Ages. To make a literary allusion, we can say that the expository
method is that of the first teacher of Rabelais's Gargantua, and the interroga-
tive method that of the second teacher. Rabelais jeers at the "great Sophist
doctor called Master Thubal Holophernes" who taught the alphabet so well
to his pupil that the latter "recited it by heart backwards". The second
teacher, Ponocrates, on the contrary, puts his pupil through all kinds of
exercises: for example, ''while passing through meadows or other grassy
places, inspecting the trees and the plants", he gathers some to study at
191
M. D. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 191-210.
Copyright © 1980 by D. Reidel Publishing Company.
192 GABRIEL GOHAU

home, and compare what he observes with what the ancients said about
them.l
I have chosen on purpose this short extract from Gargantua because
Rabelais alludes in it to the observation of nature (here plant matter). Now I
must speak of the pedagogy of science teaching, in which observation (as well
as experimentation) is a fundamental component. In this way the opposition
between the interrogative and the expository methods is governed in the fmal
analysis by an opposition between a process which leads to progressive dis-
covery by way of repeated observation and experiment, and a technique
whereby one recites the rational exposition constructed by the teacher, where
the observation of objects and phenomena is replaced by the reproduction of
schema. That is why this method is sometimes called 'bookish': for it pro-
ceeds, as books do, by way of exposition and schema.
But opposition is found as well at another level. The interrogative method
builds up knowledge, that is to say laws, theories, and general concepts, using
observations and experiments as a starting-point. It proceeds therefore by
induction, moving from the particular to the general. By the contrary method,
it is usual for the teacher, who unaided sets out knowledge, to start from
principles whose particular cases he deduces. If therefore he introduces
experiments into his teaching, he puts them at the end of the lesson as an
illustration of the theory. His method is deductive.
Finally, in science teaching, this distinction between induction and deduc-
tion is more fundamental than the exposition-dialogue opposition or the
opposition between a class without experiments and a class with experiments.
For today few teachers reject dialogue altogether, and all are obliged by min-
isterial instructions to experiment and to have their pupils handle apparatus
(during periods of practical work).
Because it aims at a construction based on the pupils' observations, the
inductive method may pride itself on following the thought of the young in
its development. Moreover it tries to reproduce the progress of science, mov-
ing from particular facts to laws and general theories: - a fact which wins it
a certain prestige in the field of science pedagogy. However, the relationship
between pedagogical method and the progress of science is not so Simple.
The place of induction in the progress of science has been discussed by
several of the previous speakers. I will limit myself to referring you to their
papers, adding that here I understand by induction, in a very wide sense, any
process of scientific discovery.
Induction, thus interpreted, evidently diverges from that of philosophers
(amplifying or generalizing induction). The logician R. Blanche 2 humorously
 THE TEACHING OF THE SCIENCES 193

compares the first to a high jump (in the sense that it invents, and rises above
already known facts), the second to a long jump, for it simply extends what
is already established fact.
Thus defined, inductive methods tend to rediscover the progress of science,
of scientific discovery. They are therefore of greater concern to the historian
of science than deductive methods which, starting from principles, aim only
at the logical exposition of a theoretical edifice the construction of which is
supposedly completed.
However, their utilization poses an epistemological problem, for given a
state of ignorance on the science teacher's part vis-a.-vis the history and epis-
temology of his subject, inductive methods have led to absurdities in the
belief that it was enough to reverse the traditional deductive method, and
establish an active pedagogy of discovery. To demonstrate this, I shall ex-
amine the method used in the French teaching system since the end of the
war, and known by the name of the technique of rediscovery.
Proposed by Inspector-General Brunhold, this technique ties up with an
inductive tradition begun in the 1880's, at the time of the passage of the lay-
school laws, a fact which, in passing, emphasizes that a revival of inductive
methods accompanies periods of progress.
In a text of 1948,3 Charles Brunhold thus defines his objective: it is in
no way concerned with "repeating the historical progress of this or that
research", but only with "retracing, with the means at the teacher's disposal
today ( ... ) the results of such research". The author specified that if his
reflection had led him to opt for rediscovery rather than a stricter historical
method, it was because of "the necessity for many of our pupils, who intend
to take up scientific or technical careers, to absorb, prior to any specialized
training, the total picture of fundamental knowledge": something which, he
considered, would not have been possible with purely historical methods in a
study time of reasonable length.
In other words: 'rediscovery' aims at the acquisition, by means of the
'spirit' and 'method' of the history of science, of this 'indispensable knowl-
edge' which was previously instilled by expository processes. A praiseworthy
undertaking, if one can give the young the same learning in a manner more
formative for their minds, by following more closely their mental develop-
ment.
But the fundamental ambiguity of the method in question comes from the
fact that it believes it legitimate to erase the obscurities and gropings of his-
torical progress. "Everything holds together when the whole is built" said
Gaston Bachelard 4 to emphasize that after discovery all the uncertainties on
194 GABRIEL GOHAU

the path of research are wiped out by the one who knows. Then verification
(or confirmation) experiments, capable of testing the theory, can be set up.
This means that the path which leads from the hypothesis to its confirmation
is not the exact opposite of the one which leads from preliminary observa-
tions to the formulation of the hypothesis. We find once more an opposition,
typical in the history of science, between science-in-the-making and estab-
lished science. I would not insist on this point known to all of you if 'redis-
covery' were not ignorant of this primordial distinction.
For indeed, at the prompting of Brunhold, appointed Director of French
secondary education, teachers and text-books reversed the order of exposi-
tion of theories and experiments. This took place in such a way that an
experiment, which was used until then to illustrate a lesson, at the end of a
chapter, was suddenly placed in the introduction, the lesson deriving from an
'interpretation' of the handling of apparatus. To give an idea of the applica-
tion of this method, here is a short extract from the draft syllabus presented
in 1966 by the General Inspectorate of natural science. It will be seen that
when I state that the same experiment which formerly illustrated the lesson
today serves to introduce it, in no way do I exaggerate. I shall take as an
example lesson 20, entitled 'chlorophyllian syntheses'.

"Practical work:
Analyses of experiments showing
- that a green plant can grow in a purely mineral medium,
- that certain elements are necessary in this medium ( ... )
Experiments on an aquatic plant revealing a release of oxygen ( ... )
Experiments leading to relative measurement of the intensity of release. A
study of the influence of one factor on this intensity ( ... )
An expression in graph form of the results.
Lesson: chlorophyllian gaseous exchanges."

The final version made the wording much less rigid by eliminating practical
work-lesson division, and giving more freedom in the choice of experiments.
But the spirit remained the same. That is, in the General Inspectorate's intent,
the class is in particular invited to perform and analyze a series of complicated
experiments for the purpose of deducing from them the notion of chIoro-
phyllian exchanges. All this, last but not least, in two hours of practical work
and one hour of teaching.
A comparison between this extravagant list and the time allotted for
studying it would make one smile if it did not reveal a completely erroneous
idea of the scientific process. Thus, to imagine that a scientific notion can
 THE TEACHING OF THE SCIENCES 195

be established by a series of class experiments, one must have a particularly

empiricist conception of science, and take as aberrations the hesitations and
gropings of its history.
This outline actually implies:
(1) That the verification experiments can be presented as first data. In
other words, that if science has performed, on the same subject, many far less
conclusive experiments before arriving at these, what delayed unduly the
sound understanding of the phenomenon was just a regrettable material hitch.
The obscurities and slow ways of science and with them all the history of
scientific thought belong to a kind of pathology of research.
(2) That class experiments are demonstrative and are consequently real
verification experiments (in the strong sense of the term).
These two propositions are the mark of an empiricist thought, but of an
exaggerated empiricism, quasi-caricatural, totally unconnected with the
philosophical empiricism of the Viennese school, for example. S No epis-
temologist could uphold them and I am going to show that they ignore the
least disputed data of the history of science: empiricism postulated by redis-
covery is teratology pure and simple!
Let us look for a start at the first statement. Could a verification (or con-
firmation) experiment,6 historically posterior to the hypothesis it tests,
logically precede it? One must reply emphatically, No.
In order to prove this I shall start with the very fine example given by
M. D. Grmek in his work on Claude Bernard's toxicological research. 7 This is
an interesting case because the experiment, which later was going to become
crucial, was performed before the researcher formulated the corresponding
hypothesis. Now, that experiment was first of all completely misunderstood
by the renowned physiologist, who formulated an explanation (a false one)
in flagrant contradiction with his result.
We know that blood poisoned by carbon monoxide (CO) takes on a bright
red color, very similar to that of oxyhemoglobin (a combination of oxygen
with hemoglobin). This led the first observers to believe that they were dealing
with the same compound and to take blood poisoned with carbon monoxide
for blood rich in oxygen. "Carbon monoxide poisons by preventing the arte-
rial blood from becoming venous", declared Claude Bernard in April 1856,
maintaining this confusion. The error was legitimate and the statement would
have been in no way surprising had not the famous biologist previously shown
that blood loses its oxygen in the presence of CO. If CO displaces oxygen, it
is therefore not oxyhemoglobin which colors the blood of a poisoned animal
bright red. Such is the elementary reasoning that we make, we who know of
196 GABRIEL GOHAU

the existence of carboxyhemoglobin (a combination of carbon rnonoxide and

hemoglobin).
Yet, Claude Bernard for all his greatness did not reason thus. Later, when
he had solved the question, this experiment of the displacement of oxygen by
carbon monoxide became crucial. It allowed the verification of the hypoth-
esis. But as long as the hypothesis is not stated, the experiment is insignificant
- i.e. without signification. If, in performing it, we want our pupils to dis-
cover in a few moments the wlution that took years (of reflection and new
experimentation) to mature in Claude Bernard's mind, we are evidently
turning the experiment into a simple pretext for the teacher's exposition of
the hypothesis. To give the pupils the illusion that they have themselves
discovered the solution is pure deception.
My demonstration, however, is incomplete. It is obvious, indeed, that
Claude Bernard, in the example we have studied, commits a logical error in
formulating a conclusion directly contradicted (that is to say falsified) by the
experiment. Therefore one objection to my argument is that if the crucial
experiment cannot be perceived as such before the hypothesis tested by it is
formulated, it is because of a mere mistake in the reasoning. One might add,
then, that the teacher's role, enlightened by the knowledge of the result, is to
avoid such errors and to enable the young ... to reason more accurately than
Claude Bernard.
But the direct transition from crucial experiment to theory cannot be
obtained by reasoning alone: it is not a simple induction. In the case in ques-
tion a strictly inductive procedure would consist in bringing together all
experiments where the blood takes on a bright red color: what was spontan-
eously done earlier, by Claude Bernard's predecessors, and by himself at the
beginning of his research. To discover carboxyhemoglobin, one must make
a detour, take the opposite course to this cumulative work. In short: the two
cases of bright red coloration must be separated according to whether the
color is caused by oxygen or carbon monoxide. Contrary to the inductive
method which associates similar cases, we must dissociate them. The inven-
tion of a hypothesis always marks a certain break with the ideas of the time.
It is an act of creation, an 'upward leap'.
Now to the second point: is a classroom experiment, which is a simplified
variant of laboratory experiments, as demonstrative as these?
In expository techniques, classroom experiment served to illustrate the
lesson and not to verify it. The teacher actually knew that what he said had
been proved a long time before by experiments far more carefully carried out
than those he could perform in class. If, exceptionally, he had been obliged
 THE TEACHING OF THE SCIENCES 197

to fulfil this verifying (or corroborating) function, he would have known that
he could not be satisfied with those coarse 'tinkering' manipulations con-
cerned only with illustrating his exposition. An example will help us to clarify
this point.
Not long ago the world of natural science teachers in French high schools
was swept by a protracted debate about the use of methylene blue to show
up incidences of tissue dehydrogenase. I leave on one side details on the
theses opposing each other. It is enough for our purpose to know that one of
my friends, a biochemist by profession, performed a careful series of experi-
ments to resolve the question at hand. His answer 8 was that methylene blue
is in fact discolored by cases of dehydrogenasis and therefore enables them to
be shown up. All teachers who used methylene blue to show that the first
stage of cellular oxidation is a case of dehydrogenasis were immediately
reassured: the experiment they used was conclusive.
So be it. Yet, if Claude Bourgeois's sensitive experiments had been found
necessary to establish a correct 'interpretation' of our summary classroom
experiment, was that fact not also proof that that coarse manipulation was,
by itself, not demonstrative? Now, when Bourgeois carried out his study, all
natural science teachers in France used that non-conclusive experiment to
allow their pupils to discover (!) the presence of enzymes extracting hydrogen
from cellular metabolites.
Whether the interpretation given was exact or not is of no interest. Cer-
tainly, it was and it is evidently all the better. But, had it been false, our
pupils would none the less have 'rediscovered' the existence of the dehydro-
genases. Classroom experimenting, therefore, proves absolutely nothing,
because it is non-'falsifying', since the teacher teaches only exact theories -
or at least those which pass as such at the level of secondary school education.
It is therefore out of the question that the experiment should 'fail'. Let us
imagine that the experiment leads to a negative result: it is so crudely per-
formed that this is frequently the case. What does the teacher do? He evi-
dently will not refute the theory he believes to be exact, but blame the
experiment which has 'failed'.
Consequently, the experiment cannot, by rights, falsify the theory. Its
testing value is strictly worthless. It holds no information on the theory. The
pupil is deceived if he is allowed to believe that he has proved - or worse:
that he has discovered - the theory. When the experiment yields the hoped-
for result he is skilfully guided towards the correct interpretation ... which
he must underline in red. When it gives another result, one cannot help con-
fessing to the deception, while stating all the same, the 'correct' conclusion,
198 GABRIEL GOHAU

guiding the student, in the best circumstances, to an understanding of what-

ever secondary influences, improperly handled, brought about the failure.
I take an example which will show, in an almost caricatural way to what
degree class experiment is non-demonstrative. The progressive generalization
of reflex reactions in a spinal frog, parallel to an increasing concentration in
acid of a solution of acidic water, bears the name ofPflliger's laws, after the
German physiologist of the last century who is reported to have established
them. Roughly speaking, it is admitted that the animal ftrst withdraws the
limb that is immersed in acidic water; that next, when the solution is more
concentrated, it withdraws at the same time the symmetrical limb (the two
hind legs when a rear leg is immersed); next again, that it shakes all four
limbs.
As a matter of fact, it frequently happens that generalization is produced
in another way, ftrst the rear leg that has been stimulated (of course!), then
this leg and the foreleg situated on the same side, then the four limbs. Now
the 'laws' have maintained only the frrst type of generalization. So, if the
fortunes of experimentation allow us to obtain the second type of reaction,
we are obliged, in spite of all, to enunciate the would-be 'law of symmetry'.
Conftdentially, I shall admit that faced with the frequency of the hetero-
dox result, I have long bewailed my wretched talent as an experimenter.Until
the day I learned that skilful manipulators obtained the same result. So, this
'law', falsifted thousands of times or more in high school classes in France
and surely elsewhere, has never been doubted. Secondary school (even
university) textbooks piously transcribe it, teachers repeat it impressively
while lamenting their own clumsiness (mine, however, is real!). Can one show
more obviously that classroom experiment plays no falSifying role, therefore
no corroborating one, since, as K. R. Popper shows, the degree of corrobora-
tion of a test is directly related to the degree of falsiftability of the theory
tested. Classroom experiment is exclusively illustrative.
All textbook writers know it very well moreover and choose their experi-
ments according to the subject to be illustrated, not the reverse. An authen-
tically empiricist procedure would require a start from complex reality with a
view to dissecting it. A concern for a truly heuristic method would start from
raw documents, with the aim of bringing out the problems and drawing up
concepts, frrst rough, then more and more refmed.
Instead of that, textbooks and classroom lessons never present anything
but already worked out facts, objects freed from dross. They claim to re-
create the historical progress of science, but in fact they set out from its
present state. They pretend to make us discover laws and theories, but right
 THE TEACHING OF THE SCIENCES 199

away they put before our eyes experiments done with a view to proving these
laws after they had been stated. This is doubly incorrect.
On the one hand, the experiment is 'parachuted'. Like Athena, it springs
fully armed from the teacher's brain. The traditional expository method is
called dogmatic, but is one less dogmatic when one explains an experiment
than when one explains a theory? In both cases, the teacher presents a fully
worked-out knowledge, sparing his pupils the effort of analysis which allowed
the progressive discovery of that knowledge.
On the other hand, the pupil is allowed to believe that the implication
uniting theory to experiment is reciprocal: which is nonsensical from the
point of view of formal logic. The experiment, a particular fact, is the con-
sequence of the theory, a general statement, but the reverse is obviously
absurd.
Finally, the method of rediscovery is altogether dogmatic, but insidiously
so: it imposes theory as much as the expository method. Even more so, since
it claims to 'deduce' it from an experiment the pupils have seen. Cannot its
hidden dogmatism, its cryptodogmatism, be called bi-dogmatism since, not
content with imposing the teacher's knowledge, it claims to base it on experi-
mentation?
The danger of this technique lies in the fact that the pupil, when he learns
later on that the knowledge instilled by his teachers was imperfect, will be
tempted to lose all belief in science, and to reject the whole scientific edifice.
He will perhaps take refuge in parapsychology or will throw himself into the
arms of one of those sects which flourish today. We should fear that, dis-
appointed with the false picture of science we give them, our pupils turn
towards hoaxers or quacks.
This severe criticism of the present state of inductive techniques may
appear excessive, especially for those who still live under the yoke of exposi-
tory methods. But precisely, the lesson must be learned without indulgence,
from a failure out of which French pedagogy has not yet found its way. Now
let us try to show what would be a veritable pedagogy of discovery which
would not turn its back on the history of science.
If one wants a pupil to discover, one must put him in a context of genuine
research. In the pedagogy of rediscovery, the situation of 'researchers' is
altogether artificial: thanks to a few summary experiments brutally hurled
before their eyes, they are required to refind a major scientific theory, which
took centuries of groping to work out. What is absurd, is not to believe the
pupil capable of performing research, but to think that in a few hours he will
rediscover a great concept of biology or physics.
200 GABRIEL GOHAU

The pedagogy of rediscovery was on the right path, but its ambition was
excessive. The Nuffield Foundation project, tried out in Great Britain,has far
better gauged possibilities of the young in the field of experimental research.
This project provides for 16-18 year old ('A level') students to devote a
tenth of their study time to 'a project or independent work' which is a
veritable program of experimental research. 9
The student may tackle a new problem, but most often he will be asked
to modify a technique he has already used and whose results he has studied.
For example, he will apply to a different animal a method of analysis studied
previously in class. He will be able to compare his results with those of the
bibliography. This type of research is clearly relevant to what Kuhn calls
'normal science': applying an experimental technique that one has mastered
to new materia1. 10
This research evidently requires only a modest amount of initiative. How-
ever, it offers original results, as opposed to rediscovery which simply repeats
things long known. From this point of view, the independent work of the
Nuffield project is far more satisfying for the student's mind than the pseudo-
discoveries of so-called inductive methods, for if the research involved is
modest, the discovery is real.
Mter all, this is basically quite normal, for the newest discoveries do not as
a rule start from fundamental questions. When he turned his attention to the
rotatory power of tartrates, Pasteur did not realize that he was going to lay
the foundations of a new science. Major discoveries are often solutions to
small problems ... which one has been able to carry out to the end. 'Normal
science' is the only one to have a real existence. What Kuhn calls the change of
'paradigm' has an existence only retrospectively. Pasteur introduced such a
change through his perseverance in following up the implications of his discov-
ery of molecular dissymmetry. At what moment his research comes out of the
framework of normal science is not easy to say, and in any case can be known
only after the event, when the consequences of his discoveries are known.
Finally, the pedagogy of rediscovery approaches problems inside out: it
believes that, under the pretext that a few experiments become a posteriori
crucial, one has the right to isolate them, to extract them out of their tissue
of obscure, stubborn researches and to reduce all scientific activity to those
turning points of its history. But if one wishes the pupil really to discover,
to produce an original result, it is impossible to choose a crucial experiment
for him: the already quoted example of Claude Bernard has shown us ade-
quately that an experiment is crucial only if one has in mind the new idea
one wishes to test.
 THE TEACHING OF THE SCIENCES 201

The authors of the Nuffield project achieved what Brunhold had not
dared: a dissociation of the acquisition of scientific knowledge from appren-
ticeship in its method. A piece of research conducted by the pupil cannot aim
at discovering a major concept or theory. The two activities are quite different
if not opposed. They must be separated.
In the case of young children (5 to 13 years) this same project makes
generous allowance for student activity and initiative. 11 A few French teach-
ing establishments are experimenting in France in the same direction, under
the guidance of V. Host and J. P. Astolfi (in biology). The aim of these
scientific activities is to create an experimental attitude in the young. The
main point is that the students be brought to ask each other questions, then
to try to resolve them by themselves through experimentation. There is no
thought of teaching them science, but only of making them understand what
constitutes research.
But naturally this dissociation leaves untouched the problem of the
acquisition of knowledge. One may suggest placing side by side an active
apprenticeship in the spirit of research and a scientific teaching of the tradi-
tional type, presenting science in its most rational form. It seems to me that
under such a system one would lose part of the benefit of inductive methods.
It is possible to present acquired knowledge in a new way which avoids the
artifices of rediscovery techniques.
The great inconveniences of deductive methods is their dogmatic nature:
scientific knowledge is from the start presented under its present guise of
perfection. There is lacking in them, if not real induction, at least progres-
sive construction, either following the history of the discoveries, or finding
again the most general concepts starting from partial knowledge.
Genetic psychology has shown us that the pupil needed to assimilate
knowledge, to digest it in some way. Now deductive methods present it in a
form ... which is indigestible. We have a mania for wanting to teach from the
start our most abstract conceptions. Would it not be possible to distinguish
levels in the acquisition of knowledge? Respiration is first an alternative
movement, then it is a gaseous exchange; then it is transportation of gas,
fmally, it becomes an exchange of electrons.
The advantage of drawing out progressively the abstract concepts is
evidently to make their acquisition easier, by taking them to pieces. But it
also consists in associating one's pupils with the way they were worked out.
Certain general concepts appear as means of unifying fields of science until
then independent: the theory of evolution is of this kind. In a well-conducted
lesson the pupils will be brought, not to discover these unifying theories, but
202 GABRIEL GOHAU

at least to understand their genesis, and to reconstruct it actively, under the

teacher's direction.
At this stage of the work, experimenting seems of little help. But it would
be stupid and dangerous to wish to experiment all the time. It was precisely
the fault of rediscovery to want to find everything again by manipulation. It
was a reaction against old methods, a wish to rehabilitate direct observa-
tion at the expense of bookish knowledge. We have behaved like Rabelais's
contemporaries. Textbooks had too long removed teachers from mani-
pulation and observation, and some thought that to know one had only to
see.
It is time that the book was rehabilitated in its turn. The necessary develop-
ment of an experimental attitude which I was speaking about earlier, would
be dangerous if it was accompanied with a complete disdain for the book.
Replying to a pupil each time he poses a problem, 'You only need to experi-
ment by yourself', would be as stupid as to say systematically to him, 'Look
it up in a book'.
An unconditional respect for bookish learning deprives one of personal
apprenticeship, but excessive recourse to one's own observation may very
well give the mind a feeling of security in its prejudices. In fact, what is to be
condemned, is not the use of the book but the exclusive usage of the text-
book.
The textbook brings ready-made answers to questions which do not occur
to the pupil. In this sense it sterilizes all curiosity. But the book is not neces-
sarily a textbook. The child, especially the adolescent, can be put in touch
with works from which they must make a resume, extract a general idea....
This kind of activity is as educational as experimentation. One must be wary
of works which give the pupil nothing but schemata: as for reality, they offer
nothing but abstract models constructed by scientists. But if the book con-
sulted forces him to construct the schema himself starting from rough out-
lines, scattered elements, partial schemata ....
Moreover, the rejection of the textbook coincides with the abandonment
of the formal lecture, for both have the same defect: they have an answer for
everything. They are in fact too well made. The teacher who prepares a
lesson, just like the author of the textbook, prunes what he has learned
retaining only the essential part. But can one understand this essential portion
if one has not followed the process and manner of its working out? Often a
good resume is unintelligible to someone who has not read the text it is
supposed to present in condensed form. Evidently there is no question of
asking the pupil to pass first through the teacher's learning before acquiring
 THE TEACHING OF THE SCIENCES 203

the basics of knowledge. But the paths of apprenticeship must not be too
direct: one understands well only what one teaches.
Let me specify that when I speak of a book as the main instrument of
information (alongside the teacher and personal observation), I do not mean
to exclude modern aids to the transmission of knowledge. We do not live in
Rabelais's time, and our pupils can learn by way of records, the cinema or
television, as well as books. I have absolutely no intention of neglecting
audiovisual techniques, and I do not share the scorn with which too many
teachers still treat them.
However, the cinema and television must be dealt with in the same way as
books, and all I have said previously applies to them. In other words, teaching
will not be modernized because the teacher or the textbook have been re-
placed by a television screen which will broadcast knowledge in just as
dogmatic a form. (Even more dogmatic, for it suppresses all dialogue and
increases the distance between the one teaching and the one taught: a
mediocre teacher speaking in front of thirty listeners is often more effective
than an excellent teacher on the air reaching (?) millions of viewers.)
The question of knowing how a television program can replace a formal
lecture i.s outside my topic, since ... we are considering how to do away with
the formal lecture. Vis-a-vis our problem (how the pupil can build up his
knowledge with the help of documents), audiovisual techniques have the
same status as books. What applies to the one extends to the others. A tele-
vision program that has been too well processed is no more valuable than a
textbook or a formal lecture.
I said earlier that to understand a problem, one must teach it. Bf'cause
to assimilate processed knowledge, one must oneself have practiced its pro-
cessing. The scientist who hears a colleague's paper (or reads his article)
understands it better than the public because he discovers what lies behind
it: he can picture the experimentation and the vicissitudes which have led to
the stated result.
This remark can be extended to the book: to understand a textbook one
must be capable of writing it. And with regard to television: to watch a pro-
gram with awareness, one must have produced some. Learning by pictures
is first of all learning what a picture is. The pupil can therefore build up his
scientific learning ACTNELY by other means than mere experimentation.
However, if he learns his lesson from books or audiovisual techniques,
if he initiates himself to experimentation through personal research, does
he not lose practically all prolonged contact with his teacher and peers
alike? Techniques of rediscovery have accustomed us to classes by dialogue,
204 GABRIEL GOHAU

where the teacher directs and controls the pupils' instruction. Is all dialogue
impossible?
Undoubtedly the dialogue of our so-called 'active' classes is illusory. It
resembles certain Platonic dialogues where the pupil is content to punctuate
the master's discourse with approving remarks: 'as you say', 'certainly', 'you
are right'.12 G. Leroy, a Belgian educationalist, in an excellent little book 13
gives examples of the kind of dialogue in use in our classes. He shows that the
teacher's skill consists in asking sufficiently precise questions - they are called
'closed' - so that the pupil necessarily gives the expected answer. When a
mathematics teacher says: 'I write a2 and a X a. What does a2 stand for?',
one can only reply: a X a.
According to Leroy, there are four or five times as many narrow ques-
tions as broad ones. Moreover, even when the question is apparently open,
the teacher often expects only one answer. Thus, the teacher who asked,
'How does the compression of a gas occur?' wanted the answer, 'through
liquefaction'. Pupils gave various answers the least logical of which was
certainly not that of the boy who simply said, 'With a pump,!14
However, in spite of these criticisms, the fact remains that.dialogue exists.
Should one not proceed for dialogue in the same way as for experimentation
and induction, the two other characteristics of the method of rediscovery:
preserve, but appreciably improve them?
Now dialogue, in a veritable pedagogy of rediscovery, would have a funda-
mental role to play: to allow the expression of the pupils' 'representations'.
Teachers forget that pupils are not virgin wax on which are marked all the
impressions inscribed thereon. J. Piaget's school has shown for a long time
now, that the minds of the young traverse, mutatis mutandis, the important
stages of the history of thought, from the first civilizations to contemporary
science. Such misapprehension introduces a fundamental ambiguity in the
teacher-pupil dialogue, for the pupil burdens with affective relationships and
egocentric representations what the teacher expresses at a purely rational
level.
In other words, the aim of dialogue ought to be to rid children's minds of
what Bachelard called epistemological obstacles. 15 Teachers believe they
form their pupils' minds by loading them with the results of science. But they
are unaware that these minds are already full of pre-scientific representations,
magical conceptions, phantasms which encumber them. Let us begin by rid-
ding them of these. Let us perform what Bachelard calls a 'psychoanalysis of
the objective mind'. For the dialogue I propose is not essentially different
from that which develops between the psychoanalyst and his patient.
 THE TEACHING OF THE SCIENCES 205

A simple example of an epistemological obstacle is that which made

Claude Bernard confuse oxy- and carboxyhemoglobin. The same obstacle -
confusion of like things - caused Cetaceans to be classified among Fish -
until the time of Linnaeus. Our pupils, who rely on rough analogies, commit
similar blunders. At 11-12 years of age, it is very difficult to say if snakes are
Vertebrates, they seem so close to worms. It is not certain that the appren-
ticeship of zoological classification divests them in an enduring manner of
that mistake.
In physiology, namely the role of oxygen in the respiration of tissues,
I was surprised to find, among pupils of 17-18 years of age, expressions of
the kind: the organ is 'regenerated' or 'made healthier'; one witnesses the
'renewal of cells', or again: oxygen provides 'a natural energy'. The function
of oxygen is manifestly valorized and carried from the rational level to a
quasi-moral level: that of the healthy and the unhealthy. Vitalistic ideas
underlie the representation of the banal phenomenon of cellular oxidation.
In the minds of twentieth century adolescents, oxygen is not very far
removed from the vital pneuma of the ancients.
Geology also provides demonstrative examples of such reluctance of the
mind to conceive modern notions. Eighteenth-century naturalists believed
that the topography of the globe had hardly changed since its origin. It is
common (Celsius, Maillet, Buffon ...) to reduce the history of the earth
to the progressive withdrawal of a primeval ocean which covered our highest
mountains. Now many young people have this same idea. A 14-year-old boy
evokes the time when 'the sea covered Europe' as if it was a single event
which had preceded the formation of the continents. To support his state-
ment, he adds this 'proof which shows he has no idea of topographical
changes: "Look at Alsace", he says, "salt potash is found there; as Alsace is
higher than the Parisian region one deduces that if Alsace happens to be sub-
merged, the same thing applies to the Parisian region".
A rigorous argument, and a convincing one ... in so far as one endorses
the unchanging nature of mountains. But when Celsius traced marks on the
shores of the gulf of Finland, he was in bondage to the same assumption.
Something which is more curious and more disquieting: similar resistances
are found among students. They have studied the principles of geology ...
and are very capable of reciting them, but if questions are put in a slightly
indirect way, lo! pre-scientific representations resurge, the demons reappear.
Year after year, when questioning 20-year-old students, I come up against
these.
They all know, of course, that if layers of ground are undulating, it is
206 GABRIEL GOHAU

because they have undergone lateral deformations (folds). However, when

they are shown such a deformation in a slightly unfamiliar light, one obtains
answers like: 'mound', 'accumulation of sediments', 'outlier', where they
should have found an anticline. In other words: deformation of sediments
has been interpreted as a mere local accumulation. Our students, who do have
a knowledge of tectonics, reason as Buffon did, who likened a syncline to a
sedimentary basin.
Another example, which concerns the same students. Their teachers have
taught them about continental drift and the formation of the oceans through
the opening of 'rifts'. When questioned on this matter, they will explain with
plenty of detail how the Atlantic opened out more than a hundred million
years ago. But if they are presented with a document showing that such a
region of the ocean-bed had then emerged, they immediately think of a
marine regression ... of several thousand metres: without realizing that in
fact the Atlantic did not then exist. Now the history of geology shows that if
vertical movements of the earth have proved difficult to imagine (changes in
coastline being attributed exclusively to variations in the volume of water
- cf. supra), lateral shifts have been recognized with even greater difficulty.
To allow that our hills, valleys and especially mountains have not existed
from time immemorial is already difficult. But to think that our continents
may have grown closer together or farther apart, is something beyond under-
standing. The most learned mind resists that idea, as Wegener's contemporaries
still did.
All these examples are drawn from my personal experience. Every teacher
may have been able to make the same observations as long as he has taken
care, however little, to listen to his pupils and possessed some knowledge of
the history of his discipline. But how many science teachers have had the
curiosity to know the rudiments of this history?
Most feel they would have wasted their time in such a pursuit, and that it
is better to follow the developments of contemporary science. They are not
entirely wrong, since the history of science hardly offers them, at present,
something which would be of use to them: a history of epistemological
obstacles. The history of science, such as it is generally studied, is doubtless
useful to the student. It is liable to be dangerous for the high school pupil if
it crowds his mind with false notions. To want to teach Stahl's chemistry to
adolescents, before initiating them in Lavoisier's pneumatic chemistry would
be unwise. Yet, does not the young person retain for a long time the idea that
a body which burns loses something which vanishes in the air? Does not he
preserve the idea that air and water are elements, and more generally that a
 THE TEACHING OF THE SCIENCES 207

natural body is simpler than an industrial product? Now these are the obstacles
which have impeded the birth of modern chemistry.
To dispel that sloth, that clumsiness of thought, in the varied forms of
resistance to scientific progress, which one comes up against in every genera-
tion in the course of its apprenticeship, that slowness of mind which causes
them to confuse things which look alike, and find simple what is familiar:
such could be the task of a history of those obstacles. But the historian would
have to be certain of the collaboration of psychogeneticists and teachers who
would help him detect the same obstacles in the manner of thinking of young
people, so as the better to recognize them.
Once such obstacles are identified, solutions would have to be invented to
overcome them. For the teacher is not a collector of obstacles. Certainly, to
bring to the young person's level of consciousness the complex of analogies,
of false images which delay his acquisition of scientific knowledge, would
already be something decisive. Clumsiness of thought is all the more active
when it remains unconscious. However, the history of science offers us also a
choice method of overcoming them by highlighting the ways in which one
clears obstacles. Bachelard called these 'epistemological acts' .16
I am going to try to give an example, relating to the cellular theory.
Cellular division allows us to understand the transition from the single cell
which is the egg, to that complex assembly of cells which is a pluricellular
creature. Now the mechanism of this division has been understood only
belatedly in relation to the elaboration of the cellular theory. To explain the
development of the living creature, biologists for a long time preserved old
ideas inherited from the eighteenth century. Around 1830, the question
asked was whether cells are formed by the enlargement of a pre-existing germ
or whether new cells are incorporated through feeding and conveyed by the
blood. The slothful mind prefers to believe in their pre-existence rather than
explain the genesis of cells (or else he evokes a mysterious spontaneous
organization, starting from a homogeneous matter).
But in the same period (1800-1940) 'lower' animals were studied actively
and, among other things, animal colonies were discovered in the hydrozoa
and the tunicates .... 17 Now the organism is after all only a colony of cells,
and cellular multiplication is not far removed from the burgeoning of the
colony from the original polyp. Again in the same period it is noticed that
ringed worms (Annelida) are formed from an undivided larva by the sprout-
ing of successive rings. Cannot these observations, so much easier to make
than that of cellular division, have provided a model for it? Do they not
constitute the way towards the epistemological act which overcame the
208 GABRIEL GOHAU

obstacle of the pre-existing germ? Could they not be used to help young
people to understand cellular division? The study of such questions would
help a new pedagogy of the sciences to be thought out.
In being satisfied with turning upside down the traditional method, and
transforming the experiment of illustration into a pseudo-experiment of dis-
covery, the pedagogy of 'rediscovery' has thought itself capable of reforming
science teaching without the expense of methodological and epistemological
reflection. The result is disappointing. Experiments are performed but in a
totally artificial way, since the theories are hidden in the experiments and
smuggled in. The pupil's mental development is followed, but answers are
whispered to him . . . to such a point that some teachers write them in
advance in their lesson notes. Experimentation and dialogue, each one emptied
of substance, have brought under suspicion any recourse to books. At a time
when television gets ready to supplant books, teachers are still in the 'pre-
Gutenberg' era.
An authentic pedagogy of discovery is indispensable, but it would require
teachers to be trained as researchers. How can one initiate young people in
experimental research if one has had only a purely theoretical education?
It would also call for the history of science to help them in the development
of a history of epistemological obstacles. Since this meeting provides me with
an international audience, I especially want my appeal to be heard.
To conclude on a modest note, I shall say that the pedagogy of discovery
should aim more at strengthening than at instilling the spirit of research. For
it is not certain that a taste for enquiry, an intuition for hypothesis, the
genius of demonstrative experiment are entirely shaped by school exercises.
My friend Evry Schatzman, a French astrophysicist, wrote recently that to
form a scientific mind, "early contact with experimentation is very important.
A very great number of experiments may be conducted by the children them-
selves". After quoting a few examples, he added that "this can take place in
the school setting but also within the home environment". Later, he also
expressed the belief that "the spirit of research is not acquired through
contact with a teacher", for the latter "can beget neither curiosity nor a
taste for discovery". 18
To add a polemical touch, I would readily declare that one would be
satisfied if the school preserved the spirit of discovery, for very often, alas!,
present-day teaching sterilizes the child's curiosity. May masters and parents
promote the flowering of this curiosity, may they goad it on at every turn
and not only during science lessons.
Does not the unravelling of detective stories encourage the maturing of the
 THE TEACHING OF THE SCIENCES 209

scientific mind? Some of Edgar Allan Poe's short stories - I am thinking pre-
cisely of The Murders in the Rue Morgue 19 apply the procedure of science
exactly to the successful pursuit of a search for a murderer: observation aimed
at finding clues, the working-out of a hypothesis in breach of common sense
(which expects a murderer to be a human being, when the murderer is in fact
here an orang-utang), then verification of that hypothesis (through a small
advertisement ).
After all, what if the scientific spirit were to be formed by the reading of
good detective stories?

Translated by
MARGARET ROUSSEL

NOTES

Rabelais, F., Gargantua. 1534. [The Histories of Gargantua and Pantagruel, trans. by
J. M. Cohen, Penguin, Baltimore, 1969.]
2 Blanche, R., L'induction scientifique et les lois naturelles (Presses Universitaires de
France, Paris, 1975).
3 Brunhold, C., Esquisse d'une pedagogie de fa redecouverte dans f'enseignement des
sciences (Paris, 1948).
4 Bachelard, G., Le rationalisme applique, (Presses Universitaires de France, Paris,
1949).
5 On logical empiricism one may consult the following:
Camap, R., 'The Aim of Inductive Logic', in Nagel, Suppes, Tarski (eds.), Logic, Meth·
odology and Philosophy of Science (Stanford University Press, Stanford, 1962) and
Hempel, C. G., Philosophy of Natural Science (Prentice Hall, Englewood Cliffs, 1966).
6 Popper, K. R., The Logic of Scientific Discovery (Hutchinson, London, 1959).
7 Grmek, M. D., Raisonnement experimental et recherches toxicologiques chez Qaude
Bernard (LibI. Droz, Geneva, 1973).
8 Bourgeois, C., 'Utilisation du bleu de methyJene pour mettre en evidence les deshy-
drogenases respiratoires', Bull. A.P.B. G. (Paris, 1966).
9 Nuffield A·Level Biological science, 14 volumes in all; on independent work: Triker,
B. J. K. and Dowdeswell, W. H. Projects in Biological Science (Penguin Books, Har-
mondsworth,1970).
10 Kuhn, T., The Structure of Scientific Revolutions (University of Chicago Press,
Chicago, 1962).
11 cr. Nuffield Junior Science (Collins, London, 1967). Emphasis is put on 'problem
solving'. "The role of the school is not to teach results but to use the scientific process
of research as an educative tool".
12 Plato, The Early Dialogues; for example, The first Alcibiades.
13 Leroy, G., Le dialogue en education (Presses Universitaires de France, Paris, 1970).
14 Lazerges, G., 'Techniques de la classe', Special number of the Bulletin de f'enseigne·
ment public au Maroc, May 1956; also Paris, 1959.
210 GABRIEL GOHAU

15 Bachelard, G., La formation de l'esprit scientifique (Vrin, Paris, 1938).

16 Bachelard, G., L 'activite rationaliste de la physique contemporaine (Presses Univer-
sitaires de France, Paris, 1951).
17 Gohau, G., 'Pn!curseurs fran~ais de la theorie cellulaire en botanique', Congres
Soc. Savantes (Paris, 1976). See also J. Roger's excellent paper in this volume, 'Two
Scientific Discoveries: Their Genesis and Their Destiny, pp. 229-237.
18 Schatzman, E., 'Peut-on former l'esprit scientifique?', Cahiers pedagogiques, No.
141 (Paris, 1976).
19 Poe, E., The Murders in the rue Morgue (1841).
 VINCENZO CAPPELLETTI

DISCOVERY AND VOCATION

Does vocation exist within naturalistic knowledge? It seems so, if one goes
through the autobiographies of Einstein, Freud and Jung. Einstein was born
in 1879, and when he was close to seventy, he wrote a long, autobiographical
essay for the volume that Schilpp's 'Library of Living Philosophers' 1 devoted
to him. He asserts that he has no doubts ("es ist mir nicht zweifelhaft") that
our thinking goes on for the most part unexpressed in words and uncon-
sciously. Otherwise, our wonder about some experiences could not be ex-
plained. "This 'wondering' ["dies 'sich wundern'''] seems to occur when an
experience comes into conflict with a world of concepts which is already
sufficiently fIxed in US".2
A wonder of such nature [Einstein continues] I experienced as a child of four or five
years, when my father showed me a compass. That this needle behaved in such a deter-
mined way did not at all fit into the nature of events, which could not find a place in the
unconscious world of concepts (effect connected with direct 'touch') .... I can still
remember - or at least believe I can remember - that this experience ['Erlebnis'] made
a deep and lasting impression upon me. Something deeply hidden had to be behind
things. 3

At twelve, a second, different wonder: in a school text -dealing with Euclidean

geometry, the encounter with "assertions, as for example the intersection of
the three altitudes of a triangle in one point, which - though by no means
evident - could nevertheless be proved with such certainty that any doubt
appeared to be out of the question".4 These were decisive experiences:
Einstein derived from them an absolute aversion towards coercive, mnemonic
study as well as a need for structural, constructive understanding, going from
the more to the less essential and evident, and a claim for freedom in univer-
sity studies.
It is, in fact, nothing short of a miracle that the modern methods of instruction have not
yet entirely strangled the holy curiosity of enquiry; for this delicate little plant, aside
from stimulation, stands mainly in need of freedom; without this it goes to wreck and
ruin without fail. 5

Even beasts of prey would lose their voraciousness if they were forced to
devour continuously under the threat of the whip - it is Einstein speaking. In
211
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 211-226_
Copyright © 1980 by D. Reidel Publishing Company.
212 VINCENZO CAPPELLETTI

the course of the autobiography the experiences of childhood and adolescence

stand out more and more. At the age of sixteen - he reports - he found
himself before a paradox, which was later included in many introductory
explanations of the relativity theory:
If I pursue a beam of light with the velocity c (velocity of light in a vacuum), I should
observe such a beam of light as a spatially oscillatory electromagnetic field at rest. How-
ever, there seems to be no such thing, whether on the basis of experience or according to
Maxwell's equations. From the very beginning it appeared to me intuitively clear that,
judged from the standpoint of such an observer, everything would have to happen
according to the same laws as for an observer who, relative to the earth, was at rest.
For how, otherwise, shOUld the first observer know, i.e., be able to determine, that he is
in a state of fast uniform motion?6

Young Einstein compared actual situations with each other, as he had done
when his father had shown him a compass with the needle moving with no one
'touching' it: and the comparison became a problem, a problem of identity,
analogy or difference of situations, which are similar in some aspects or op-
posite in others. The 'fundamental' physical concept that sixteen-year-old
Einstein was considering was the 'field': it surely revived the memory of the
compass needle that he had seen when he was four and brought into conscious-
ness; but in the unconscious the memory of a childhood observation stated in
rational terms represented the dynamic factor. And also the other experience
of the boy Einstein - his contact with geometry and its proofs - influenced
the genesis of relativity. Einstein asserts that he had been looking for a 'uni-
versal formal principle' , like the principle of thermodynamics, which states the
impossibility of constructing a perpetuum mobile, supposing the laws of na-
ture to be what they are. An essential step towards axiomatics, towards defini-
tions that, like the geometric ones, represent an absolute beginning: a cogni-
tive stage later examined in Geometry and Experience, published (in German)
in 1921. 7 The principle of special relativity, of value equal to the above-men-
tioned principle of thermodynamics, is the following: "The laws of physics
are invariant with respect to the Lorentz-transformations (for the transition
from one inertial system to any other arbitrarily chosen system of inertia)". 8
In essence, a basic principle of natural laws, like that of thermodynamics,
excludes the existence of perpetual motion. But let us pick up the thread of
our argument and try to draw a conclusion. Einstein's autobiography shows a
singular continuity between scientific observations of childhood and adoles-
cence, and the discoveries of adulthood. The premise on which this continuity
is based is wonder: the habit of asking why and how each phenomenon
diverges from what is traditionally or customarily considered lawful, regular.
 DISCOVERY AND VOCATION 213

Freud's autobiographical writings number more than one. 'On the History
of the Psychoanalytic Movement'9 starts with the years following his M.D.
degree, that is, his journey to Paris in October 1885, his friendship with
Breuer and subsequent estrangement, and his first works. So far, nothing of
interest for us. The Autobiographical Study, dated 1925,10 however, includes
some interesting suggestions. He entered gymnasium, the 'Sperl Gymnasium'
in Vienna, when he was nine, and in the following seven years, he says
(actually only six, according to the investigations of his biographers who
succeeded in checking such minute details) he was at the top of his class. No
precocious disposition towards medicine.
I was moved, rather, by a sort of curiosity, which was, however, directed more towards
human concerns than towards natural objects, nor had I grasped the importance of
observation as one of the best means of gratifying it. 11

Among his favorite authors - during adolescence, of course - were Darwin

and Goethe. To find other significant information we must take a step back-
wards and leaf through a short article, 'Some Reflections on Schoolboy Psy-
chology', that Freud wrote in 1914 for the fiftieth anniversary of his old
school, renamed 'K. K. Erzherzog-Rainer Realgymnasium' .12 The memory of
the school, he wrote, revives when, walking through the streets, one comes
unexpectedly on one of the schoolmasters, by now well on in years; but the
boy too has become a man.
And I seemed to remember that through the whole of this time there ran a premonition
of a task ["die ganze Zeit von der Ahnung einer Au/gabe durchzogen war") till it found
open expression in my school-leaving essay as a wish that I might during the course of
my life contribute something to our human knowledge. 13

This premonition of a personal commitment to the progress of science has

the same sharp relevance in Freud's account, as Einstein's wonder. The two
autobiographies refer both to a kind of scientific knowledge that to the
child means contrast between uniformity, mental or practical, and novelty.
This is the wonder before the magnetic needle which moves without being
touched. This is the feeling for progress and the wish to contribute to it.
In Memories, Dreams, Reflections of C. G. Jung (entitled more briefly
and effectively Ma vie in the French edition 14) the founder of analytical
psychology illustrates the symbolic dimension of the child psyche through
fragments of dreams, moments of family life and games. Among the oldest
memories, was the moment when he "became conscious of smelling", aware
of perceiving. Then a long dream, again of preschool age and expressed through
instinctual entities but fllied with symbolism. Then again the experience of
214 VINCENZO CAPPELLETTI

alienation while playing with his playmates that "alienated me from myself',
surely due to difficulties connected with symbolism, when coming from a
private sphere of meanings and going back to a public one of intersubjective
references. And yet another, the experience of fire, of its ineluctable 'burn-
ing', relived thirty years later, coming back to the same places, in the garden
of his father's house. At school, the contact with a teacher for whom

algebra was a perfectly natural affair, to be taken for granted, whereas I didn't even
know what numbers really were. They were not flowers, not animals, not fossils; they
were nothing that could be imagined, mere quantities that resulted from counting. 1S

And the scandal of the transitive law of equality: if a = b and b = c, a = c. It

was difficult enough for him to understand how it was possible to write a = b,
and not a = a only. The school teacher, who, unlike Carl, considered algebra
self-evident, had never felt any wonder before the diversity of existing things.
The failure in mathematics and drawing was not the only reason that alienated
young Jung from primary school: there was also his "sense of smallness in the
face of the vast world". 16 Therefore Carl Gustav considered a fortunate event
the accident (when he was twelve) that prevented him from attending school
and allowed him "to plunge into the world of the mysterious. To that realm
belonged trees, a pool, the swamp, stones and animals, and my father's
library".17 But another fortunate event returned him, recovered, to 'reality':
it was his father's anxious preoccupation because the boy was running the
risk of not being able to earn his own living. "I was thunderstruck. This was the
collision with reality. 'Why, then, I must get to work!' I thought suddenly".IB
Being can be attained.

Wonder, vocation for knowledge, feeling for the difference between under-
standing and learning: these are the characteristic features of the first contact
with science - or with experience lived through as science - of three great
discoveries of reality. Let us immediately make one observation: during the
same years in which Einstein and Freud, as children, were opening their
minds, the one to 'Wunder' and the other to 'Ahnung einer Aufgabe', 'grown
up' science vindicated the necessity of these fundamental situations for the
knowing subject. Ernst Mach, one of the protagonists of scientific innovation
at the end of the nineteenth century, offers an example of this. The Science
of Mechanics published in 1883 19 is contemporary with little Einstein's
wonder at the compass needle moving without being touched. It is an extrin-
sic circumstance, which will acquire significance in what we are about to say.
With Mach science became historiography in order to regain its own origins
 DISCOVER Y AND VOCA nON 215

and the evolutionary rhythm of theories, from 'facts in themselves' to axio-

matics, from perceiving to stating. Mach is representative of something entirely
new in the historiography of scientific thought: he appears as the historiogra-
pher acting as philosopher, the man who reads texts to satisfy reason's de-
mands. A complex environment rich in influences surrounds him. Psychology
knocks at the door of scientific thought, a psychology still wavering between
physics, neurophysiology, and autonomous conceptual development: it is
Wundt's elementistic psychology, Fechner's psychophysics, the first Gestalt
psychology, the beginning of analysis. Measurements are performed, signifi-
cant laws are established, as Weber's and Fechner's law on the logarithmic ratio
between increment of stimulus intensity and increment of sensation intensity.
It seems, however, that sensation cannot be reduced to movement. In 1872 and
1880, two lectures of the physiologist Emil du Bois-Reymond, entitled Uber
die Grenzen des Naturerkennens, and Die sieben Weltrdthsel, respectively,20
denounce the aporias of the universe as constituted of a single substance: the
space of Descartes' Principia philosophiae (1644), the atom of Laplace's Essai
philosophique sur les probabilites (1814). Mach intuitively perceived the
absurdity of denying the qualitative - sensation is qualitative - from which
our knowledge of the world originates. He, too, remembers his youth:
I have always felt it as a stroke of special good fortune, that early in life, at about the
age of fifteen, I lighted, in the library of my father, on a copy of Kant's Prolegomena to
Any Future Metaphysics. The book made at the time a powerful and ineffaceable impres-
sion upon me, the like of which I never afterwards experienced in any of my philosophi-
cal reading. Some two or three years later [this autobiographical passage is taken from
The Analysis of Sensations 21 ] the superfluity of the role played by 'the thing in itself'
abruptly dawned upon me. On a bright summer day in the open air, the world with my
ego suddenly appeared to me as one coherent mass of sensations .... 22

In Mach sensation defmes completely the whole world, from which

mechanicism had ejected it. And, of course, it becomes multiform to avoid
the difficulties of the amorphous atom: from sensation it becomes complexes
of 'elements':
For us, therefore, the world does not consist of mysterious entities, which by their
interaction with another, equally mysterious entity, the ego, produce sensations, which
alone are accessible. For us colours, sounds, spaces, times, ( ... ) are provisionally the
ultimate elements, whose given connection it is our business to investigate. 23
'Connection' is in fact the other primitive term of Mach's theory, through
which the theory departed from the alleged self-sufficiency of sensation
without being, however, really adequate to build the structure of reality
from dissimilar elements.
216 VINCENZO CAPPELLETTI

Let us turn to the analogies between Mach's Science of Mechanics and the
autobiographical accounts mentioned earlier. In Mach, too, wonder has a
precise cognitive relevance:
But the aim of my whole book is to convince the reader that we cannot make up proper-
ties of nature with the help of self-evident suppositions, but that these suppositions must
be taken from experience. 24
This is the clearest programmatic statement to be found in the whole of
Mach's work. Sensation reveals the world, which, being what in fact is, is
undeducible. Sensation precedes analysis, since the world is given before it
can be asserted. Static moment, acceleration, inertia, the impossibility of
perpetual motion: all of these are mechanical qualities, which cannot be
deduced, which become engraved in our representations. They are "a treasure-
store which is ever close at hand and of which only the smallest portion is
embodied in clear, articulate thought".25 And how about wonder? Wonder is
the appearance, in the cognitive process, of novelty and unpredictability of
nature. Mach quotes Aristotle, not the Aristotle of 'thaumazesthai',26 but the
author of the mechanical writings, explaining why the lever appears wonderful:
If a thing takes place whereof the cause be not apparent, even though it be in accordance
with nature, it appears wonderful (. . .). Such are the instances in which small things
overcome great things, small weights heavy weights, and incidentally all the problems
that go by the name of 'mechanical'. (... ) To the apories (contradictions) of this charac-
ter belong those that appertain to the lever. 27

The lever contains, embodies in itself one of the world's peculiarities, the
static moment:
As a matter of fact, the assumption that the equilibrium-disturbing effect of a weight
P at a distance L from the axis of rotation is measured by the product P' L (the so-called
statical moment) is more or less covertly or tacitly introduced by Archimedes and all his
successors.2 8

Let's read again the pseudo-Aristotelian Mechanica that we already men-

tioned. Wonder is excited by events of which we do not know the cause:
these events conflict with others, already known. We wonder in front of
novelty, paradox. Novelty may conflict with another, or join another, and
still remain the same: and perhaps it is in the latter case that its greatest
charm lies, rather than in the former, which risks assuming as measure non-
being rather than being. But the importance of the contrast is not to be
underestimated, especially as far as education is concerned. We must remem-
ber that just this contrast aroused little Einstein's wonder before the compass
needle movement.
 DISCOVERY AND VOCATION 217

We found wonder in Einstein's autobiography, wonder in Mach: the same

fundamental situation in a child taking his first steps along his cognitive
path, as well as in an adult calling for science to know itself. The teaching
of sciences and the philosophy of science agree on a common postulate. If
it were not so, if pedagogy and epistemology were not symmetrical with
respect to thought which is trying to understand the reality of nature, we
could draw out of the adolescent not the free and autonomous subject which
will contribute to the development of knowledge but an automaton repeating
behaviors imposed by a society concerned with the preservation of patterns
and interests.
Wonder is mentioned a second time in the Science of Mechanics, when
Mach speaks of Stevin, the great sixteenth century Dutch builder and physicist
(working in statics). Mach writes:
The service which Stevinus renders himself and his readers, consists, therefore, in the
contrast and comparison of knowledge that is instinctive with knowledge that is clear, in
the bringing the two into connection and accord with one another, and in the supporting
the one upon the other. The strengthening of mental view which Stevinus acquired by
this procedure, we learn from the fact that a picture of the endless chain and the prism
graces as vignette, with the inscription "Wonder en is gheen wonder", the title page of
his work Hypomnemata Mathematica. As a fact, every enlightening progress made in
science is accompanied with a certain feeling of disillusionment. We discover that that
which appeared wonderful to us is no more wonderful than other things which we know
instinctively and regard as self-evident; nay, that the contrary would be much more
wonderful; that everywhere the same fact expresses itself. Our puzzle turns out then to
be a puzzle no more; it vanishes into nothingness, and takes its place among the shadows
of history. 29
This is the case of wonder vanishing: and it is natural that it vanishes after
having given rise to the problem in a thought pretending to have exhausted
reality. Thought must explain, restore in itself the organic unity of being:
this is the reason why, at times, it goes through the Jungian 'sphere of mystery'
and sees common experience in a new light; other times it reexamines this or
that notion of traditional knowledge. The lowest price for stopping wonder is
a proof starting from premises already known. This is the case with Stevin,
reported in Mach. Mach, instead, like the grown-up Einstein, had to face the
revision of the most rooted assumptions of science.
To support the comparison between Einstein and Mach, let us recall what
is well known elsewhere, namely the profound influence that Mach exercised
on the discoverer of relativity. Einstein writes in his autobiography:
Even Maxwell and H. Hertz who in retrospect appear as those who demolished the
faith in mechanics as the final basis of all physical thinking, in their conscious thinking
218 VINCENZO CAPPELLETTI

adhered throughout to mechanics as the secured basis of physics. It was Ernst Mach who,
in his History of Mechanics, shook this dogmatic faith; this book exercised a profound
influence upon me in this regard while I was a student. I see Mach's greatness in his
incorruptible skepticism and independence; in my younger years, however, Mach's
epistemological position also influenced me very greatly, a position which today appears
to me to be essentially untenable. For he did not place in the correct light the essentially
constructive and speculative nature of thought and more especially of scientific thought;
in consequence of which he condemned theory on precisely those points where its
constructive-speculative character unconceivably comes to light, as for example in the
kinetic atomic theory. 30

There is no place in Mach for thought: and, in fact, the third component of
his construction, 'connection', appears as weak, groundless, with respect to
'sensation' and 'element'. Nonetheless, for Mach, also scientific reasoning
teaches how to read, how to interpret the sensation: and in the course of this
reading, as we saw, wonder fades into evidence. The latter, since it is capable
of absorbing any novelty, is before and beyond the marvelous and causes
its disappearance. Evidence is the organ of the 'great whole'. "Galileo and
Huygens used to constantly alternate the consideration of the particular
phenomenon and the great whole", he writes in a passage of the Science of
Mechanics, which is reminiscent of a passage in Plato's Phaedrus concerning
Hippocrates' method. Sensism becOlp.es a program, a password, the equiva-
lent to idealism.
Mach's epistemology and the embryonic pedagogy of science found in
the Science of Mechanics or derivable from it, assumed a psychology. Other-
wise, the postulate of sensation would have remained hazy and, above all,
devoid of connections with experimental research. Well, Mach was a psy-
chophysicist, the author of Lessons of Psychophysics, 31 convinced supporter
of Fechner's theory of parallelism, but not of his work on the unique, un-
knowable substance of which physical and psychical facts would be attributes.
"My natural bent for the study of these questions" - he noted from Prague
in November 1885, presenting The Analysis of Sensations - "received the
strongest stimulus twenty-five years ago from Fechner's Elemente der Psy-
chophysik (Leipzig 1860)".32 And in the preface to the fourth edition, in
November 1902:
If there is" no essential difference between the physical and the psychical, we shall hope
to trace the same exact connection, which we seek in everything that is physical, in the
relations between the physical and the psychical also. 33

In short, the characteristic common to all things is quantity, that is the

possibility of measurement. But if sensations exist, and not the sensation,
 DISCOVERY AND VOCATION 219

if elements exist, and not the element, if the structure of the world is rich
and different, then quality must exist besides quantity, the 'what' as well as
the 'how much'. And here we have, in addition to the 'exact connection' -
that is the 'connection' quantitatively connotated - another relationship
among phenomena: 'symmetry', which is discussed in one of the Popular
Scientific Lectures. 34 Mach is always ingenious and undetermined, and his
theoretical wavering is the expression of his need to connect different cate-
gories: but the frequent lack of connection favors the partial interpretations
- either phenomenalistic or reductionistic, subjectivistic or prerelativistic -
of his work. At the end of his life he, too, considered himself as a man of nov-
elties not arranged in a system, as subject of a vocation towards innovation:

Through the constant autoanalysis or criticism of myself since my younger years, I was
oriented towards the direction that today is called relativistic, and I could have brought
forward all these things; instead, subordinating my own ideas, I first of all tried to attain
an overall view on the future and to surpass the limits of the past, eluding the sphere of
influence that naturally the great thinkers create around their names. This is the reason
why I was concerned also with general problems of the physiology of sense and with
psychology. After all, we live only once, and I wanted, within my own limits, to have
as much of the world as possible; I could not fill up my life with only one thought, to
save my strengths. 35

But we should not forget that the background of Mach's psychology still
remained psychophysics. And the historian, always curious about details, can
confirm it, on the basis of still another circumstance: Mach's friendship with
Joseph Breuer, the first to dissent from Freudian dynamic psychology.
The Analysis of Sensations sets Breuer's figure in a light different from
that of Ernest Jones' Freudian 'biography'. Breuer was a neurophysiologist
of great merit and marked originality: his research on the function of the
otoliths in the vestibular apparatus were prompted by sagacious hypotheses
and by a physicalistic presupposition which, through Ernst Wilhelm Briicke,
physiologist in Vienna, was connected with the school of Milller, and in par-
ticular with du Bois-Reymond and Helmholtz. In 1875 Mach had published
Grundlinien der Lehre von den Bewegungsempfindungen,36 and Breuer's
investigations were the development of a subject well known to him:

Breuer, in a later piece of research, has made it probable that the sensations of progressive
acceleration vanish very much more quickly than those of angular acceleration and that
perhaps the organ of the former, at any rate in human beings, is atrophied. Further,
Breuer finds that, except for the semicircular canals, B, the otolitic apparatus, 0, with
its plane of sliding corresponding to the planes of the semicircular canals, is the only organ
adapted to the signalizing of progressive accelerations and position simultaneously.37
220 VINCENZO CAPPELLETTI

Breuer's name is not among those included in the bibliographical sec-

tion of treatises of nervous system physiology, unjustly, as one can judge
from the ideas reflected in Mach's quotation. What is interesting to note
here, however, is that Mach's closeness with Breuer and his friendship for
him (in 1894 Mach was among those proposing him for membership in
the Vienna Academy of Science) clarify the nature of the quarrel between
Breuer and Freud, its background and its fmal reasons, beyond minute
anecdotes. Breuer, as well as Mach, was a psychophysicist opposed to the
new depth psychology and to what it presupposed: a shift of psychological
interest from sensation - so puzzling to du Bois-Reymond - to representa-
tion, and the need for a sign and symbol hermeneutics. Jones should have
grasped and developed this motive emerging here and there in the first volume
of the 'biography', for instance when he scrupulously gives Breuer credit for
having formulated some energetic concepts that were later included in Freud's
'metapsychology' :

In the paper on 'Theory' which he (Breuer) contributed to the Studien, one notes the
fundamental importance he attached to the idea that the basis of hysteria was an ab-
normal excitability of the nervous system, so that an excess of free energy, that could
not be disposed of, was available for conversion into somatic symptoms. 38

Paradoxically, Fechner, the Fechner of Revision der Hauptpunkte der

Psychophysik 39 could represent a dialectical means between opposed require-
ments and trends. No one, however, took this fundamental work seriously,
and full attention was always given only to the Elemente, with the well-known
formula:
S =K log R (S, sensation; R , stimulus)
and, in general, with the suggestion for methods of measuring psychical pro-
cesses. 40 ·

Psychophysics, however, was and still is inadequate to sustain both the epis-
temological and the pedagogic constructions: even more so if sensation, per-
ception are regarded not as moments in the process of knowing, but as the
matrix and single normative source of knowledge. We recalled that the per-
ception of the motion of a body untouched by anyone impressed the mind of
little Einstein and awakened a 'wonder' that would vanish only through the
field theory implicit in the general theory of relativity. There is continuity
between perceiving and knowing in Einstein, but also in Mach as in the fol-
lowing passage from the Science of Mechanics:
 DISCOVERY AND VOCATION 221

Still, great as the importance of instinctive knowledge may be, for discovery, we must
not, from our point of view, rest content with the recognition of its authority. We must
inquire, on the contrary: under what conditions could the instinctive knowledge in
question have originated? We then ordinarily find that the very principle, to establish
which we had recourse to instinctive knowledge, constitutes in its turn the fundamental
condition of the origin of that knowledge. And this is quite obvious and natural. Our
instinctive knowledge leads us to the principle which explains that knowledge itself, and
which is in its turn also corroborated by the existence of that knowledge, which is a
separate fact by itself.41

Mach the epistemologist was more open, less unilateral than Mach the psy-
chologist, and did not hesitate to acknowledge that sensation is already
reason. Mach and Freud were ignorant of each other - on Freud's side the
only hint of some relevance is in his letter to Fliess of 12 June 1900 - but
the memory, perhaps remote, of his great contemporary can be found in a
very beautiful autobiographical passage that Mach wrote in his last years, in
which he speaks of 'constant autoanalysis'. We already quoted it, and we refer
again to the citation. 42 Owing to Mach's waverings, Ehrenfels - who asserted
psychical 'Gestaltqualitdten' (form qualities) that were after all quite different
from Wertheimer's perceptual 'Gestalt' - could indicate as their source
Mach's Analysis. 43

Other science creators also show the same precocious inclination towards
the cognitive activity which in their lives soon assumed the centrality and
intensity of a vocation. Scientific vocation expresses itself in attitudes that
both epistemology and historiography regard as necessary for the growth of
the doctrinal organism: wonder at experience, the understanding of the
postulate objectivity of knowledge, the devotion to research. To become
objective knowledge, the ideal moment of vocation must promote the inter-
pretation of the perceptive datum. To favor this process is the task of peda-
gogy of science, which is also a pedagogy of perception.
First of all, perception must be attributed to the human subject, and must
be seen as physis and logos, both necessary. If pe{ception lacked the connec-
tion with nature, man would never refer it to reality in itself, with an attitude
which, if it is a naive identification of what is sensed with the other than the
self, it still originates from a legitimate inference from the sensed to the
existing, from the self to reality. On the other hand, if perceiving were not
thought, perception would never become wonder, that is, a problem. Wonder,
problem, resulting from the presence, and not the absence of conformity.
Let us be more explicit. Uttle Einstein, before the magnetic needle moving
without anyone touching it, had to perceive not only a phenomenon different
222 VINCENZO CAPPELLETTI

from others which are omnipresent, but also the doubt as to whether that
omnipresence of touching could be the formal cause of motion: doubt, his
own and any other analogous one, presumed the dialectic between the
all explaining thought and one principle that can explain only something.
Einstein must have been struck by the intuition that 'to be cause' and 'to
touch' are different principles, the former being more general and necessary
than the latter. He who marvels at something doubts traditional explanations,
not the existence of an explanation. Wonder is expectation: the expectation
to understand experience, as in Jung, or, as in Stevin, the expectation to
restore lawfulness.
However, perception, to be at the same time nature and thought, must
be more than quantity, or, using different terms, more than energy, motion.
There has never been a single attempt to give a quantitative interpretation
of thinking, which, on the contrary, is entirely qualitative, or, to specify
the polysemanticity of the term 'quality', structured. The rule organizing
the parts, which we call structure, produces judgment as unity of subject,
predicate and verb, or of terms with the relations connecting them. In
Identite et realite, Emile Meyerson effectively stated the priority of 'com-
prehension' over 'extension'. Each rational act unifies the diverse: and in
doing so, it assimilates diversity to identity. But the unity-identity must be
ascertained in the thing: otherwise, the legitimacy of the act, with which
sensation is seen and interpreted in terms of objectivity and world, would
cease. This is the inadequacy of psychophysics, and the reason for the revi-
sion done by Fechner in his last years. A historian of psychology, E. G.
Boring, distinguished a 'psychology of content' and a 'psychology of act',44
and included Fechner and Wundt in the former, Brentano, Meinong and
Ehrenfels in the latter. But the psychology of the last thirty years of the
nineteenth century can be better distinguished in psychophysics and psy-
chology of representative activity. Act and content cannot be heterogeneous
and Boring's scheme is liable to render them such. If the act and the content
of the psychical process do not fmd a way of corresponding to each other,
then the intentional relation to the object, in which Brentano perspicaciously
detected the peculiarity of the psychical, fails.45

Psychophysics, therefore, cannot meet the requirements of a pedagogy of

perceptive activity, which is essential, we repeat, to the pedagogy of physical
and natural sciences. Hence the latter is oriented towards 'Gestalt' and cogni-
tive psychology. Owing to some reports of Wertheimer and Kohier,46 in the
pre-World War I years, the 'Gestalt' disrupted the traditional scheme of the
 DISCOVERY AND VOCATION 223

subject's perceptive reflection of the object: that "elementary relationship

that to most people seems an unquestionable fact".47 In his report, Uber
unbemerkte Empfindungen und Urteilstiiuschungen (1913), Kohler notes that
elementary sensations presume "an exacerbated process of isolation"; and
also in this case J. Muller's law of specific energies leads one to state the
discontinuity between stimulus and sensory response. However, if perception
is structured so that perception can in fact be perception of something, also
the processes preceding it must have a structure 'isomorphic' to the percep-
tive one. In Max Wertheimer, isomorphism is not a descriptive notion, but a
cosmological statement, ontological, we might say. We refer to the twofold
moment, classificatory and explicative, that Duhem, Mach's contemporary,
thought should be distinguished in any scientific theory 48 in order to put in
the first place what he calls 'natural classification'. Wertheimer's isomorphic
relation and Einstein's invariance of the physical law , however, show how the
development of every classification is linked with an attempt at explanation.
In Productive Thinking,49 a classic of contemporary psychology, Wertheimer
would have defined the terms of still another problem, though similar to that
of the relationship between stimulus and perceptual shape: "the transition
from a blind attitude to understanding in a productive process", in other
words, intellectual creativity. He writes:

Those were wonderful days, beginning in 1916, when for hours and hours I was fortunate
enough to sit with Einstein, alone in his study, and hear from him the story of the dra-
matic developments which culminated in the theory of relativity. During those long
discussions I questioned Einstein in great detail about the concrete events in his thought.
He described them to me, not in generalities, but in a discussion of the genesis of each
question. 50

Here comes again the thought experiment of running after a light beam,
here we find the consideration of the Michelson-Morley experiment, the need
to define simultaneity, the intuition of invariance, the identification of the
invariant with the velocity of light, the relativistic postulate. Perception - the
child's perception of the magnet and others, that, perhaps, have not been
recalled - as thought: dense, unexpressed but effective. And thought as
perception. In 1918 Eddington reported to the Royal Society that the devia-
tion (from a straight line) of the light coming from a star, due to the Sun in
the relativistic hypothesis, had been confirmed within the limits of experi-
mental accuracy.
Vocation for knowledge, and rational construction: these make a discovery
surpass the range of chance and transmute it into necessity, necessity of being
224 VINCENZO CAPPELLETTI

so for a hypothetical world which could identify with the existing world.
Pedagogy is applied to vocations so that they become constructions, to the
invisible deus in nobis in order to achieve the deus inter nos, knowable by all
people and recognized as such. There must be a school willing to be a school
of science, indeed a society which sees in the birth and development of scien-
tific vocations an element capable of granting its meaning and continuance.
Not all the great creators of science wrote their memoirs: but the lack of
an autobiography can be compensated by biography, if it is compiled keeping
in mind that it should not be a collection of events and circumstances but a
cognitive integration of the scientific document. Historiography is and has
long been active: Gediichtnissreden, etoges, obituaries, even precede the
historiography of ideas in the Machian sense, and their importance and abun-
dance of intrinsic value goes far beyond the colorless sequences of names,
dates, discoveries that were so frequent not so long ago. Biography is more of
a guesswork than autobiography, but it is better protected against the risk of
'protective memories', and more generally, of self-deception. Together with
the essay, biographical historiography competes for the space of the novel,
which is not exhausted but threatened by the possible languishing of the
imagination devising it.
Biography, or else, biographical historiography, has already proved, even
without the verification of the autobiographical account, that in many cases
vocation lies behind the scientific discovery. The discovery which is reflected
in axiomatics - it is now necessary to differentiate it from discovery with
no further connotation - is always coming from a distance: its trace is also
the trace of memory and the word through which it expresses itself is like a
Janus-faced image looking at the same time at things and their origin, in the
microcosm of the human subject.

Translated by
FRANCESCA PARDI LEVI

NOTES

1 A. Einstein, 'Autobiographical Notes', in Albert Einstein Philosopher-Scientist, ed. by

P. A. Schilpp (Tudor Publishing Co., New York, 1949), pp. 1-95.
2 Ibid.,p.9.
3 Ibid.
4 Ibid.
S Ibid., p. 17.
 DISCOVERY AND VOCATION 225

6 Ibid., p. 53.
7 A. Einstein, Geometrie und Erfahrung (Springer, Berlin, 1921); Engl. transl., 'Geome-
try and Experience', in Sidelights on Relativity, II (Methuen, London, 1922).
8 A. Einstein, 'Autobiographical Notes', p. 57.
9 S. Freud, 'Zur Geschichte der psychoanalytischen Bewegung', in Gesammelte Werke
[referred to hereafter as G. W.), 18 vols. (Imago, London, 1940-1968). Vol. 10, pp. 43-
113. Translated as 'On the History of the Psychoanalytic Movement', in The Standard
Edition of the Complete Psychological Works [referred to hereafter as S.E.), ed. by
James Strachey, 24 vols. (Hogarth Press, London, 1953-1974). Vol. 14, pp. 7-66.
10 S. Freud, 'Selbstdarstellung', in G. w., Vol. 14, pp. 31-96 (S.E., Vol. 20, pp. 1-74).
11 S. Freud, 'An Autobiographical Study', S.E., Vol. 20, p. 8.
12 S. Freud, 'Zur Psychologie der Gymnasiasten', in G. W., Vol. 10, pp. 203-207 (S.E.,
Vol. 13, pp. 239-244).
13 Ibid., p. 205 (S.E., p. 242).
14 Erinnerungen, Triiume, Gedanken by C. G. Jung, ed. by A. Jaffe (Zurich-Stuttgart,
1962); Engl. trans!., Memories, Dreams, Reflections (W. Collins Sons & Co., Glasgow,
1977); French transl., Ma vie (Paris, 1966); Ital. transl., Ricordi, sogni, riflessioni di
C. G. lung (Milan, 1978). (The following notes refer to the English edition.)
15 Ibid., p. 43.
16 Ibid., p. 45.
17 Ibid., p. 47.
18 Ibid.
19 ~. Mach, Die Mechanik in ihrer Entwicklung historisch-kritisch dargestellt (Brock-
haus, Leipzig, 1883); seven editions were published while the author was still alive, the
last one of which in 1912. Afterwards, there were two posthumous editions, one in
1921, edited by J. Petzoldt, and one in 1933, edited by L. Mach; English transl. by
Thomas J. McCormack, with the author's approval: The Science of Mechanics: A Critical
and Historical Account of Its Development (Open Court Publ. Co., La Salle, Ill., 1960 6 )
(the following notes refer to this English edition, unless otherwise stated).
20 E. du Bois-Reymond, Uber die Grenzen des Naturerkennens - Die sieben Weltriithsel
(Veit, Leipzig, 1891).
21 E. Mach, Die Analyse der Empfindungen und das Verhiiltnis des Physischen zum
Psychischen (Fischer, Jena, 1896). The last edition, while the author was still alive, is
dated 1911. English transl., The Analysis of Sensations, and the Relation of the Physical
to the Psychical, translated from the first German edition by C. M. Williams; revised and
supplemented from the fifth German edition by S. Waterlow (Dover, New York, 1959):
the following notes refer to this English edition.
22 Ibid., p. 30, Note 1.
23 Ibid., pp. 29-30.
24 E. Mach, The Science of Mechanics, p. 27.
25 Ibid., p. 37.
26 Met. I, 2.
27 E. Mach, The Science of Mechanics, p. 13; the Aristotelian passage is in Mechanica
847 abo
28 E. Mach, The Science of Mechanics, p. 19.
29 Ibid., pp. 40-41.
30 A. Einstein, Autobiographical Notes, p. 21.
226 VINCENZO CAPPELLETTI

31 E. Mach, 'Vortriige liber Psychophysik', Oesterreichische Zeitschrift fUr praktische

Heilkunde 9 (1863).
32 E. Mach, The Analysis of Sensations, p. xxxvi.
33 Ibid., p. xii.
34 'Die Symmetrie', in E. Mach, Populiir-wissenschaftliche Vorlesungen (Leipzig, 1903 3 ),
pp. 100-116; English transl., Popular Scientific Lectures, transl. by Th. J. McCormack
(Open Court Publ. Co., Chicago, 1910 3 ), pp. 89-106.
35 Quoted in Ludwig Mach's Preface to the ninth edition of Die Mechanik ... , (Leipzig,
1933; reprinted Darmstadt, 1963, p. xvii).
36 E. Mach, Grundlinien del Lehre von den Bewegungsempfindungen (Engelmann,
Leipzig, 1875).
37 E. Mach, The Analysis of Sensations, p. 140.
38 E. Jones, The Life and Work of Sigmund Freud. 2 vols. (Basic Books, New York,
1953-1957; Vol. 1,1953, p. 302).
39 G. Th. Fechner, Revision der Hauptpunkte der Psychophysik (Breitkopf and Hartel,
Leipzig, 1882).
40 E. G. Boring, A History of Experimental Psychology (Prentice Hall, Englewood
Cliffs, N.J., 19502 , pp. 273-296).
41 E. Mach, The Science of Mechanics, pp. 36-37.
42 See Note 35.
43 Ch. Ehrenfels, 'Ober Gestaitqualitiiten', Vierteljahreschrift for wissenschaftliche
Philosophie 14 (1890), 242-292; M. Wertheimer, 'Experimentelle Studien liber das
Sehen von Bewegung', Zeitschrift fiir Psychologie 61 (1912), 161-265.
44 E. G. Boring, op. cit., pp. 447ff.
45 F. Brentano, Psychologie vom empirischen Standpunkt (Duncker and Humblot,
Leipzig, 1924, I, pp. 125f).
46 M. Wertheimer, Experimentelle Studien; W. Kohler, 'Uber unbemerkte Empfindun-
gen und Urteilstauschungen', Zeitschrift fur Psychologie 66 (1913), 51-80.
47 W. Kohler, Dynamics in Psychology (New York, 1940).
48 P. Duhem, La tMorie physique: son objet, sa structure (Riviere, Paris, 1914 2 , pp.
31f).
49 M. Wertheimer, Productive Thinking (Harper and Bros., New York, 1959 2 ).
50 Ibid., p. 213.
 PART II

CASE STUDIES
 JACQUES ROGER

TWO SCIENTIFIC DISCOVERIES:

THEIR GENESIS AND DESTINY

If discovery is the strong beat of scientific life and the instrument of scientific
progress, it is not easy to defme or explain. The very word, at least in several
European languages (de-couvrir, s-coprire, dis-cover, ent-decken) is deceptive,
for it indicates the simple act of moving aside an obstacle, of removing the
veil which 'covered' a truth that lay at hand, and which was simply waiting
to be at last looked at: a defmition more tempting in anatomy than elsewhere
and yet disputable.
There is no question here of reviewing all the possible kinds of discovery in
the very wide variety of their epistemological nature and their historical
circumstances. I wish to study only two precise discoveries, their genesis,
their realization and their destiny. I am here concerned with an account I
have already written! and which I now take up from a different perspective.
These two discoveries are (1) that of the ovarian vesicles and the phenomenon
of ovulation in vivipara, and (2) that of the spermatozoa. The first is attri-
buted to the Dutch anatomist, Regnier de Graaf, and dated 1672; the second,
to the Dutch microscopist Antoni van Leeuwenhoek, and dated 1677. They
were made in the same scientific field - at least according to modern de-
finitions - in the same country, and within an interval of five years. These
external similarities allow us to isolate more easily the differences and try to
account for them.
Let us briefly recall the facts. Until about the middle of the seventeenth
century two theories on the generation of animals and man shared the
allegiance of the scientific world. According to the first, which comes from
Aristotle, the male provides with his semen the 'form' of the being to come,
while the female, with the menstrual blood, provides the 'matter' of the
embryo. According to the second theory, which comes from Hippocrates,
male and female produce a liquid semen, the mixture of which, or 'conceptus'
forms the embryo. In the middle of the seventeenth century, Descartes
presented a 'corpuscularist' and mechanicist version of the theory of the two
semina (De [ormato [oetu, published posthumously in 1664), while William
Harvey gave a fresh lustre to the Aristotelian theory (Exercitationes de gener-
atione animalium, 1651).
However, Harvey corrects Aristotle in that he does not believe in the role
229
M. D. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 229-237.
Copyright © 1980 by D. Reidel Publishing Company.
230 JACQUES ROGER

of the menstrual blood. According to him the male semen fertilizes the uterus
of the viviparous female, and this uterus produces an egg, exactly as the ovary
of the ovipara does. On the one hand, therefore, the so-called 'female testi-
cles' are only lymphatic glands and play no part in the act of generation. On
the other hand, however, vivipara are born from an egg as ovipara are, which
justifies the well-known formula inscribed in the frontispiece of Exercita-
tiones: "Ex ovo omnia." For Harvey viviparous females have eggs but no
ovaries.
Harvey had hardly any following in his refusal to assign a role to the
'female testicles', but the idea of the egg of vivipara was very successful from
1660, each exponent, however, putting forward his personal opinion as to
the origin and formation of that egg. Things became clearer in 1667, with
the contribution made by the great Danish anatomist, Niels Stensen, alias
Steno. Steno, who, like Harvey, first thought that the egg was formed in the
uterus, discovered eggs in the female testicles of a dog-fish - a species known
since Aristotle's time as ovo-viviparous. He immediately concluded "that
women's testicles are analogous to the ovary" and produce eggs which then
pass into the uterus. While Steno continued his observations (which were
published only in 1675), a Dutch anatomist, Jan van Horne, published in
1668 a short treatise on the male genital organs, to which he added a note on
the female organs, where he stated: "Female testicles correspond to the ovary
in the ovipara, given that they contain perfect eggs, full of body-fluid
[humeur] and wrapped in a thin membrane." Let us note that three years
before, in a short anatomical treatise, van Horne had already spoken of those
ovarian vesicles and of "the dual duct whose purpose is to evacuate the body-
fluid which is prepared thereiIl.. " In 1668, therefore, he had not just dis-
covered them - they had, in fact, already been described by Vesalius,
Falloppio and others - but, after taking them for reservoirs of female semen,
he had read Steno and decided to call these vesicles eggs. Van Horne died in
1670, before being able to write the great treatise he had been contemplating
on these questions.
From 1670 onwards, work was done nearly everywhere on the ovarian
vesicles. In 1671, the Dutchman Kerckring published a treatise of a few pages,
Anthropogeniae Ichnographia, where he produced, among other things, the
picture of a human embryo already well-formed - too well-formed - which
he claimed to have discovered in an egg three days after conception. In Paris
and in London, eggs were discovered in cows' ovaries. At the end of the year,
Regnier de Graaf very briefly set out anew van Horne's ideas which he
claimed as his own. In short, everyone by then believed in the eggs ofvivipara,
 TWO SCIENTIFIC DISCOVERIES 231

but no one had yet proved anything about them. As Gallois wrote at that
time in the Journal des Savants (21 March 1672), two things are certain:
there are vesicles in female testicles and what encloses the embryo in the
womb has the shape of an egg: but "it remains to be seen if those vesicles,
which are attached to the female body, detach themselves, and if that kind of
egg where the embryo takes shape is one of those detached vesicles."
The reply to these questions had in fact just been given by Regnier de
Graaf in his treatise: De mulientm organis generationi inservientibus. Thanks
to methodical observations on female rabbits, de Graaf established that after
copulation, ovarian vesicles change into yellow bodies. Furthermore, he had
managed to discover developing eggs in the Fallopian tubes, which developed
in equal number to the yellow bodies in the corresponding ovary. It was thus
shown that each vesicle was an egg, the shedding of which left a scar in the
form of a yellow body, that the female testicles were ovaries and that the
Fallopian tubes conducted the detached eggs from the ovary to the uterus.
The 'egg doctrine', as it was then called, had thus firmly been established. As
early as 1679, according to the Journal des Savants, scarcely favorable as it
was to the idea, it had become "something so common at present that there is
practically no new Philosopher who does not accept it today." The existence
of eggs in viviparous females was no longer seriously doubted in time to
come.
Let me end this first historical account with two remarks. First, that
Regnier de Graaf's demonstration fell far short of the success that his meth-
odological rigor ought to have garnered, even if that rigor (I shall return
to the point) is imperfect. In the immediate spread of the new doctrine,
Kerckring's loud assertions or William Harvey's remote authority, had far
more weight. No doubt, later, but only later, things became clearer. The
second remark is that, as we know, ovarian vesicles or de Graafs follicles, are
not themselves the eggs of vivipara: they hold eggs which are released at the
moment of ovulation. As early as 1681, Malpighi had an inkling of this mis-
take and even believed, but wrongly, that he had discovered the veritable
little egg, which he christened ovule. All through the eighteenth century, the
ovule escaped the attention of the world of research, and the 'egg doctrine'
was none the worse for it. We know that it was only in 1827 that von Baer
discovered the ovule inside de Graafs follicle.

Just as the discovery of the egg of vivipara was a long and complicated pro-
cess, that of the spermatozoa was a simple and quick one. It is described in
a letter which van Leeuwenhoek sent in November 1677 to Lord Brouncker,
232 JACQUES ROGER

the secretary of the Royal Society of London. He recounted therein how a

medical student, called Ham, had discovered these 'animalcules' while examin-
ing under a microscope the sperm of a man suffering from gonorrhoea. Ham
thought they were "born of a sort of putrefaction." Van Leeuwenhoek him-
self, however, had found them in the sperm of a healthy man, and proposed
to consider them as a normal part of semen, without, however, attributing to
them any function in the process of generation. Afterwards he found them in
the semen of various animals" dog, rabbit, cod, still without comprehending
what their use could be. The public was informed only belatedly, first by two
papers given by Huygens at the Academy of Sciences in Paris, in July 1678,
followed in August by two articles in the Journal des Savants, which, inciden-
tally, did not mention van Leeuwenhoek, but only his young compatriot,
Hartsoeker, who indeed was trying to get the merit for the discovery. Finally
van Leeuwenhoek's obsen·ations appeared in the Philosophical Transactions
of March 1679. It was in the same year that the Dutch scholar began to think
that the animalcules played a part in generation: that is, in the perspective of
his time, he believed they contained the preformed germ of the animal to be
born. He had few followers, although as early as 1691, a synthesis between
ovism and animalculism was attempted by a Scot, named Garden, which met
with some success in spite, one must point out, of fierce opposition from van
Leeuwenhoek himself, who refused to believe in the existence of eggs.
Leibniz, for reasons in line with his philosophy, adopted van Leeuwenhoek's
system, which, however, hardly survived its author's death in 1723. In the
middle of the eighteenth century, when almost everyone believed in the
existence of the eggs of vivipara, which no one had seen, no one knew what
to do with the spermatic animalcules which everyone could observe.

In the light of these two historical accounts, we can make a few reflections.
The flISt will be to note the difficulty of saying who has made a discovery.
It has been shown that it was almost impossible to say who discovered
oxygen. Let us take here the simpler of our two examples. Who discovered
the spermatozoon? First possible reply: the student Ham, who was the
first to see it. But Ham did not see a spermatozoon: he saw a microscopic
animalcule born, according to him, of putrefaction. Second answer: van
Leeuwenhoek, who observed systematically the sperm of various animals
and found everywhere spermatic animalcules, which he took for the active
agent in generation. But he believed that the animalcule contained the pre-
formed embryo. In French, the very word spermatozo'ide appeared only
in the nineteenth century, progressively taking over from kindred names:
 TWO SCIENTIFIC DISCOVERIES 233

spennatozoaire, spennatobion, spennatuie and spermatozoon this last

word was retained in the English language. In a manner of speaking, to arrive
at today's spermatozoon one had to await the theory of cells, chromosomes
and DNA.
It is even less easy to say who discovered the egg ofvivipara. Vesalius and
Falloppio described the vesicles of the 'female testicles'; Riolan described a
tubal pregnancy; Harvey wrote ''Ex ovo omnial"; Steno and van Horne
equated the egg and the ovarian vesicle; Regnier de Graaf showed that there
was a deftnite relationship between the two, but took the vesicle for the
ovule; Malpighi had an inkling of the error, but did not fmd the ovule which
he was moreover looking for in the yellow body; after him it took more than
a century before von Baer saw that ovule. It is hardly necessary to add, from
von Baer's ovule to the ovule as we know it today, the same progress had to
be made as for the spermatozoon.
The second remark will therefore be that it is impossible to study a dis-
covery without bringing in the question of its interpretation. The 'discovery'
of the ovule or of the spermatozoon is in fact an uninterrupted process which
may have begun with Regnier de Graaf and van Leeuwenhoek, but has not
stopped since and has doubtless not ended.
Now this link between a discovery and its interpretation exists right from
the start and singularly complicates the study, for even when the discovery is
simple - particularly so in the case of spermatozoon - its interpretation
demands an understanding of practically the whole intellectual situation of an
era: and the discovery cannot be separated from this general situation, unless
arbitrarily.
Let us note, however, that the two discoveries in question are of two
types, and maybe two borderline cases of the relationship between a dis-
covery and its interpretation. In the case of the egg in vivipara, the interpreta-
tion precedes the discovery, and so to speak demands it. Rather than the 'dis-
covery', one should perhaps speak of the 'demonstration' of a 'truth' already
afftrmed in exalted quarters and generally desired. Actually, Regnier de
Graars 'demonstration' in no way demonstrated that ovarian vesicles were
eggs. One could even perfectly well interpret it within the terms of the theory
of two spermata, since it was enough to say that each vesicle j::ontained, then
poured into the tube, the 'dose' of maternal seed necessary for the formation
of an embryo. Now no one put forward this interpretation among the 'new
Philosophers', for the presence of an egg had to be demonstrated, and the
ftrst acceptable experiment answered the purpose. The case of the spermato-
zoon is the strict inverse of the other. No one, not even Ham or van
234 JACQUES ROGER

Leeuwenhoek, expected such a discovery. An interpretation had therefore to

be improvised, which would comply with the requirements of contemporary
scientific thought. This took some time, and finally led to a lasting failure.
Let us note again that in spite of their proximity in space and time, the
two discoveries were not made in the same sector of scientific activity, in the
same epistemological or ideological field. For us, the ovule and the spermato-
zoon are the male and female gametes the union of which forms the initial
cell of a new living being. The existence, the production, the anatomical and
biochemical structure of each are relevant to the same scientific question, the
physiology of reproduction. But that unity of the scientific field exists only
in so far as the same methods and the same instruments are used in research.
It did exist, therefore, at the time when the normal study of the male and
female reproductive organs was undertaken by means of anatomy. It is indeed
by an anatomical study that Regnier de Graaf established the function of
ovarian vesicles, but van Leeuwenhoek's discovery breaks the unity of the
scientific field, for it was made by means of a microscope. Now, as we shall
see in a moment, micrography at the end of the seventeenth century was not
simply a particular method of inquiry, using a new instrument: it was a world
apart and an autonomous field of research. The laws of discovery, if one may
use the term, cannot be the same in the two fields, and the ideological implic-
ations are just as different as the methods.

After these preliminary remarks it is easier for us to analyze the history of

those two discoveries and to explain their destiny. The important moment in
the 'discovery' of the egg of vivipara arrived when anatomists, like Steno and
van Horne, 'decided' that 'female seed' was to be replaced by 'eggs.' That
decision was made prior to Regnier de Graafs demonstration and determined
the whole sequence of events.
To pass from the seed to the egg, anatomists were impelled first of all by
analogy. The mechanicist rationalism of the seventeenth century emphasizes
the unity of the ways of nature, and the world of ovipara, already very vast,
had just been enriched, particularly as a result of the work of Redi, with all
the insects which had previously been thought to be produced by spontane-
ous generation. Mammals became a scarcely comprehensible exception, and
the case of ovo-viviparous fish made the transition easier. In this context,
Harvey's "Ex ovo omnia" became a seductive slogan.
But the passage from liquid seed to egg also corresponds to an evolution of
mechanicism in the second half of the seventeenth century. For sixteenth-
century doctors, liquid seed was animated by spirits or vegetative souls. As
 TWO SCIENTIFIC DISCOVERIES 235

for Descartes, he had imagined it to be made up of corpuscles, the disposition

of which, complying with the laws of movement, formed the structure of the
new living creature. In the second half of the seventeenth century these two
interpretations became equally impossible. Everyone, or almost everyone,
rejected the Aristotelian 'form' or the 'souls' theorized by the doctors of the
old school. But it was no longer believed, either, that the laws of movement
were capable of creating a well-ordered structure. For most scholars, for
example for Robert Boyle, who gave a lengthy explanation of the idea, the
structure must pre-exist, and the laws of movement can only regulate the way
it functions. Without a pre-existing structure, the laws of movement were
unable to do away with chance, that is to say disorder. These initial structures
must therefore have been created by God and it was obvious that they would
be much better protected in an egg than in a liquid seed. That such a train of
thought renewed the ancient myth of the primordial egg was certainly of
service to the doctrine in the sight of the learned scholars, but must have
played only a minor role in persuading the 'new Philosophers.'
One might even wonder if the cosmological revolution did not playa part
in that case. In the traditional geocentric cosmos, the heavens continually
brought vivifying influences to bear on the earth. Without even evoking the
fanatics of astrology, let it suffice to recall here the Aristotelian formula:
'Homo generat hominem, atque sol.' The earth, having become a planet
moving round in infmite space, could no longer count on these beneficent
influences, and the fragile living structures needed protecting: a protection all
the more necessary as the Nature of the Mechanicists had lost the generous
fecundity with which sixteenth century scholars had endowed it, and which
had counted spontaneous generation as just one of its visible attributes. A
passive and mechanical nature can only destroy, at best preserve, order: it can
no longer create it. All order dates from the very moment of creation and it
must be preserved.
At all levels of thought and imagination, the transition from liquid seed to
egg was becoming necessary. It is therefore not surprising that the egg of
vivipara was 'found' and so eagerly adopted.

But then the world of the microscopists, which developed during the last
quarter of the seventeenth century, did not wholly partake in the spirit of the
time. Let us note right now that the microscope, practically a contemporary
of Galileo's telescope, remained unused for a long time. When it began to be
used, it was used by specialists of the technique itself, who examined with
their instrument the most varied objects. Van Leeuwenhoek's observations
236 JACQUES ROGER

were typical of such a disorder, which was already found in Hooke. At that
time, however, the microscope was above all a 'great purveyor of marvels.' It
showed the unexpected, the improbable, something which could not have
been discovered by analogical reasoning. It accompanied or encouraged in the
observer the idea that nothing is impossible, and that the human mind cannot
conceive, much less pretend to order, the infmite variety, the infmite richness
of creation. In this sense the discovery of spermatozoa was not a chance
phenomenon. It was prepared, or at least made pOSSible, by the existence of a
certain mentality in the particular scientific sector where it happened.
Such a particular state of mind was scarcely in keeping with the one pre-
vailing at the time. More precisely, it heralded that of the naturalists primarily
dedicated to observation, so numerous in the first half of the eighteenth
century. In this sense the micrographers of the 1670's were witnesses to the
European 'crise de conscience. ' They came a little early, and in spite of the
curiosity they aroused, van Leeuwenhoek's work suffered because of this
fact. The discovery of the spennatozoon, like many others, suffered in parti-
cular from the difficulties of verification. Many people, who had never
handled a microscope, debated about it. But once the fact was established,
interpretation presented far more important difficulties. It was in conformity
to the spirit of the time that van Leeuwenhoek imagined an embryo pre-
fonned in the spennatozoon. But that animalcule, which one saw being born
and dying, presented no real guarantee of safety to a pre-existing structure.
Besides, its movement led it to be considered an animal. By what extra-
ordinary 'metamorphosis' could that animal become a man, at a time when
'equivocal generations' had gone out of fashion? Finally, the enonnous
number of spennatozoa, the immense majority of which were condemned to
a purposeless death, was in too violent contradiction of the idea of a thrifty
nature and a wise God, creator of a well-ordered universe. As a contemporary
said: 'that makes a good deal of seed lost!'
It is therefore not surprising that the two discoveries, coming to light in
different areas of research, placed in a totally different relationship vis-a-vis
their interpretation, and therefore the spirit of the time, coming within the
purview of two different epistemological types, had opposite destinies. One is
allowed to think that many discoveries have a position between these two
extremes, but that the history of their genesis and success lends itself to the
same kind of analysis.

Translated by
MARGARET ROUSSEL
 TWO SCIENTIFIC DISCOVERIES 237

NOTE

1 Cf. Roger, J., Les sciences de la vie dans la pensee fram,aise du XV/lIe siecle: la
generation des animaux de Descartes a l'Encyclopedie (Colin, Paris, 1971), esp. the 2nd
part, Chapter II, pp. 255-323.
 RICHARD TOELLNER

LOGICAL AND PSYCHOLOGICAL ASPECTS OF THE

DISCOVERY OF THE CIRCULATION OF THE BLOOD

A. THE GENERAL PROBLEM

Traditional history of science consists mainly of the history of discoveries.

Especially in some particular fields of science and their respective disciplines,
it proceeds from a generally acknowledged and presently valid doctrine of its
subject matter and poses questions leading back into the past: How did such a
body of knowledge come about chronologically? History thus prefers to ask
at what time, in which place, in which country, by what person a certain
observation was made for the first time; when, where and by whom an experi-
ment was undertaken, a theory outlined or a technique first applied, in short,
a discovery or an invention first made. The criterion for the importance of a
discovery is in any case the relative significance that is attributed to the 'fact'
or technique discovered or the theory outlined within the framework of
science that the historian works from. 1
Such a history of science regarded as a history of discoveries has its many
merits in the processing of historical material. Inasmuch as it is looking for
sources, fmds them, and makes them available to historical researchers in
historical-critical editions, its work is irreplaceable. But as soon as traditional
history of science depicts the path of science as a one-way street, where
scientific reason progresses unwaveringly from one discovery to the next until
it reaches the point where we stand today, approaching the truth by never-
ending progress, and as long as it follows immanent scientific logic, then it
merits the criticism voiced by modern scientific history, that this kind of
presentation does not throw light on the true history of science, but that it
conceals, obscures, and even distorts it. This is why it is of no intellectual
value for today's science. The main objection of modern historians to the
presentation of history as an avenue of poplars running towards us is directed
against the practice of taking one's own point of view as an absolute value
and assuming one's own criteria unreflected as the standard for the judgement
of history. A judgement is, however, the fmal aim of any historical interpreta-
tion and it is accordingly the first prerequisite of a just view of history that
our standards be corrected by the standards that govern history itself. Each
historical event has a value in its own right and may not be reduced to a mere
239
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 239-259.
Copyright © 1980 by D. Reidel Publishing Company.
240 RICHARD TOELLNER

case history. This is, however, what happens in a history of discoveries, where
the discovery as such is reduced to a date, to one station on the road to one's
own preconceived aim. 2
When modern history of science asks for new discoveries, inventions, and
insights, in short, for the latest news in the history of science, then it is not
interested in new facts as mere facts, then it is not primarily concerned with
the question, by whom, when, and where something new was discovered, but
it is interested in the question in which historical context this new fact could
emerge as a new fact, what the historical prerequisites of this discovery were,
in what way and under which historical conditions it could come into effect.
Modern history of science asks, as does the topic of the second course of the
International School of History of Science, for the 'Logical, Psychological,
Cultural and Social Aspects of the Scientific Discovery. '
But it is only the enquiry into all these single aspects and the adequate
linking of each result in research into one over·all picture that transform the
isolated date of chronological science, the pure fact of case history into an
event of history of science. Only the presentation of a discovery as a histor·
ical scientific event, into which the patterns of ideas, the content and aims of
preliminary science have blended as well as the cultural and social conditions
of today's life, only such a presentation that has explored all this and takes it
into consideration, will truly be able to offer any information on the specifi·
cally 'new' aspect of a scientific discovery, can really describe how a new
insight developed, in which way and by which means it came into effect or
had no effect at all.
It is only against the background of Scientifically immanent as well as
sociocultural factors that the figure of the discoverer and his psychology can
claim the importance that is due to it as the most important factor of the
historical event of the discovery. In this context I fully agree with Bernard
Cohen, when he says:

It is my thesis that in studying science, just as in every other area of creative activity, the
historian must take account of the special qualities of individual genius, that an aware·
ness of the temperament and personality of each scientist is for the historian of equal if
not greater importance than the general character of the age and the particular environ-
ment of ideas in which the scientist worked. 3

There are, of course, discoveries in the history of science that have become
without any doubt the basis for today's scientific view of the world, and their
effect is such an overwhelming one that traditional history of science seems
to be correct in its opinion that only the discoveries really matter and that
 DISCOVERY OF THE CIRCULATION OF THE BLOOD 241

the historical conditions for their development and their influence are of
minor importance. To name a few examples, one can refer to discoveries
which now constitute our physical conception of the world linked to the
names of Copernicus, Galilei, Kepler and Newton. No doubt that, as far as
medicine is concerned, such an exceptional discovery, on which a completely
new physiology and medicine was based, must be the discovery of the circula-
tion of the blood by William Harvey.
It is not by chance that all the examples mentioned are taken from the
particular era when ancient-medieval science was replaced by modem science.
The eighteenth century in which the new science had its breakthrough under-
stood this process as "an emancipation of science from the chains of author-
ity", as a farewell of science to blind faith in authority, as "a departure of
man from self-inflicted infancy" as Kant's classical remark goes,4 as "a
liberation of scientific research from any preconceived views, from ideological
dogmatic Authority" as the physicist and historian of science Walther Gerlach
called it. 5 The new basic postulate of science is now: Scientific statements are
no longer to be verified, validated and authenticated by calling on the
authority of the Ancients, but by referring to autonomous reason and its
deductions on the one hand and to autonomous experience, that is, one's
own observation, research and experiment on the other.
It may be easily understood that with such a deep change in the structures
of thinking of science it is not the mere results of this historical process as
pura facta that constitute the major value of appreciation for modem science
but rather the historical conditions of their possibilities, their complicated
genesis and the way they come into effect.
I thus propose to demonstrate this general process by dealing with
the special case of Harvey with its logical and psychological implications
that illustrate the dispute Harvey had not directly but indirectly with
Descartes about the nature of cardiac motion. This debate represents only
a minor episode in the history of seventeenth century medicine and has
not hitherto received much attention in the Harveyan or Cartesian litera-
ture. 6
However, problems and insights may emerge from the study of a single
historical event which could have a more general and lasting character and
may still affect us today. In this sense, we shall attempt to illustrate how
Descartes and Harvey, two great scientists, deal with experience and reason
for the correct description and explanation of a physiological phenomenon.
Furthermore, we will also try to show how the dispute between the founders
of modem philosophy and physiology heralds the great controversy over the
242 RICHARD TOELLNER

proper reason-experience relationship which Kant brought to a temporary

conclusion.
Finally, we shall examine the seeming paradox that Harvey, empirically
investigating, while deeply rooted in late medieval Aristotelianism, was in the
end correct in contrast to Descartes, the creator of the modem scientific
world-view. Harvey's theoretically less-than-revolutionary circulation scheme
started - as Rothschuh has convincingly demonstrated - the "process of
autocatalytic increase in knowledge and unfolding of problems." Such a
development radically changed a two-thousand-year-old medicine "so that we
can properly speak of a post-Harveyan era in medicine." 7
By contrast, Descartes' physiological conceptions were quite revolutionary
but soon abandoned by his closest followers because they were impossible to
prove experimentally. Therefore, Leibniz would speak of Cartesian physics
and Haller of Cartesian physiology with contempt, both regretting the formu-
lation of such 'physical fables.'8 However, Leibniz, the founder of German
rationalism, and Haller, the originator of modem experimental physiology,
would be inconceivable without Cartesianism. Moreover, Descartes' influence
was decisive in the establishment of the new theory of blood circulation.
In order to understand the prevailing conditions under which the dispute
on cardiac motion took place between Descartes and Harvey, it is essential to
describe the points of departure of both authors. The Venetian humanist,
Cesare Cremonini, has become, at le.ast since Bertolt Brecht, a symbol for
those people who refuse to accept empirical facts because they do not fit into
their world-view. 9 In a letter written to Kepler and dated August 19, 1610,
Galileo mentions that Cremonini has steadfastly refused to take even one
look through the telescope to see the planets and their satellites. Whether the
Aristotelian Cremonini, conscious of the unpredictable consequences
associated with the introduction of the Copernican system, acted wisely or
not is unimportant. He certainly closed his eyes to a sensually perceptible
proof of the new astronomy, "seeking the truth not in the world or in nature
... but in his acquaintance with texts", 10 as Galileo reproachfully stated.
In refusing to recognize a certain relationship, one has, in a sense, already
tacitly accepted it. This point is made clear by the Bolognese mathematician
Magini, who in the interests of the old cosmology simply demanded of
Kepler: "The four new satellites of Jupiter must be eliminated and for-
gotten.',11 The aim was not to deny the observed circumstances and attempt
to label the perception of the phenomena erroneous. Rather, the suggestion
was to disregard such facts because "nothing can possibly exist which must
not be."
 DISCOVERY OF THE CIRCULATION OF THE BLOOD 243

The example emphatically demonstrates the strength of the new insight

which emerges in the early seventeenth century. Such a novel concept pro-
posed that observations and experiments derived from nature were more valid
than the knowledge derived from books and schools. The sentence, "No other
proof can be more convincing than perception and personal observation",12
reflects the conviction of a man who replies to "those who believe it to be a
crime to doubt the authority of the ancients." The answer is "that no dogmas
can suppress the obvious facts and no old traditions stifle the work of nature
because nothing is more ancient and of greater authority than nature
herself."13 The man who spoke these words and appears to be, like Galileo,
an adversary of Aristotle, is no other than the Aristotelian William Harvey.14

B. THE NATURE OF THE CARDIAC MOTION

I Harvey's Position
Harvey was a contemporary of Francis Bacon and Galileo Galilei, men whose
names we link with the introduction of the empirical and inductive method,
and thus the beginnings of modern scientific research. Although Harvey must
have encountered Galileo at Padua and, as a court physician to James I, could
not have ignored the activities of Lord Bacon, there is no proof that they ever
influenced him directly .1 5
The appreciation of experiments, observations and experience was, there-
fore, a general characteristic of a period in this history of science deeply
polarized by the followers and detractors of Aristotle. Harvey's preferences
for the 'sensus' (perception) and 'autopsia' (personal observation) as con-
trasted with 'ratiocinium' (the fruits of reason) are not so much characteristic
of the man himself but rather of his time. Harvey's confidence in experience
was therefore more a contemporary presupposition than a self-acquired
insight which allowed him to become the immortal discoverer of the circula-
tion.
The discovery was aided by Harvey's preference for anatomical studies.
These investigations were carried out during his student days in Cambridge
and Padua, and continued until his death. From the notes which Harvey made
and the manuscripts of his anatomical lectures - in 1615 he held a chair of
anatomy and surgery in London - the stages of his discovery become
apparent.
At first, Harvey carried out careful anatomical observations of the venous
valves which his teacher, Fabricius ab Aquapendente, had described in Padua.
244 RICHARD TOELLNER

Then he went on to study the anatomical differences between arteries and

veins, the structure of the heart and especially the cardiac valves, which made
him question the traditional physiology ascribed to the central organ.
Performing many animal experiments, Harvey attempted to observe care·
fully the cardiac motions, noting that the auricular contractions alternated
with those of the ventricles. Moreover, he realized that the ejection of blood
did occur during systole when the chambers contracted rather than in diastole
when the walls were distended. During systole, the apex of the heart arose,
touching the chest wall and generating a beat which was synchronous with
the pulsations of the arteries. Therefore, the throbbing of the heart or systole
actually coincided with the diastole of the arteries.
The nature of the cardiac motions which could now be divided into an
active systolic contraction and a passive diastolic dilatation became the start·
ing point for new inquiries. Taking into account the valvular mechanics
already elucidated by Galen, the thought emerged that the blood entered the
ventricles from the auricles during the dilatation of the former. This blood
was subsequently ejected during the ventricular contraction into both the
large artery (aorta) and the arterial vein (pulmonary artery). If one estimated
the volume of blood which was discharged to the periphery during each
ventricular systole, and multiplied such an amount by the number of heart
beats, the resulting volume was so great that it could not possibly be all used
in the periphery of the body for nutrition and structural replacement as
postulated by the older physiology.
At this stage of his observations and thinking, Harvey declares:

In consequence, I began privately to consider if it [the blood] had a movement, as it

were, in a circle. This hypothesis I subsequently verified, rmding that the pulsation of
the left ventricle of the heart forces the blood out of it and propels it through the
arteries into all parts of the body's system in exactly the same way as the pulsation of
the right ventricle forces the blood out of that chamber and propels it through the
artery-like vein into the lungs. 16

The conceptual framework from which Harvey drew such a conclusion is

reflected in the passage which follows his description of the circulation. He
wrote: "We have as much right to call this movement of the blood circular as
Aristotle had to say that air and rain 'emulate the circular movement of the
heavenly bodies.,,17 With this sentence, Harvey reveals himself not only as an
individual versed in Aristotelian meteorology and cosmology, but, what is
more important, he seems to partake of Aristotle's metaphysical foundations.
Harvey's world is still the Aristotelian cosmos in which circular motion
 DISCOVERY OF THE CIRCULA TION OF THE BLOOD 245

occupies a special place among all other movements. Moreover, life is a dis-
tinctive process within the sub1unar, terrestrial sphere, characterized by birth
and death, growth and decay, change and motion, all of which must be under-
stood and derived from first philosophical principles. Thus, as the sun's move-
ments and warmth are the source of all motions, mixtures and transforma-
tions of the macrocosrnic elements, the heart with its contractions and beat is
the fountain, origin and highest principle of the body or microcosm.
Therefore, Harvey can state:

Therein, by the natural, powerful, fiery heat, a sort of store of life, it [the blood 1 is re-
liquefied and becomes impregnated with spirits and (if I may so style it) sweetness. From
the heart it is redistributed. And all these happenings are dependent upon the pulsatile
movement of the heart. This organ deserves to be styled the starting point of life and the
sun of our microcosm just as much as the sun deserves to be styled the heart of the
world. For it is by the heart's vigorous beat that the blood is moved, perfected,
activated, and protected from injury and decay.18

We know from Harvey's notes that he was convinced as early as 1616 "that a
steady flow of blood takes place in a circular manner with the help of a
cardiac pu1sation."19 However, only twelve years later did he finally decide
to acquaint the public with his theory. Although repeated observations and
experiments had convinced him of the correctness of his conclusions - they
coincided with the principles of Aristotelian-Galenic medicine - Harvey fore-
saw the effects which his new theory was to have. Indeed one of his early
adherents in Germany, Paul Marquard Schlegel, wrote that

the unheard of scheme which was directed against the general concepts prevailing for
centuries brought about a great commotion. There were hardly any physicians who, after
hearing of Harvey's discovery, did not decry the work as complete nonsense and urge its
banning from the schools. 20

Harvey had expected such a reaction, as reflected in the cautious words he

used on the occasion of dedicating his work to the "doctissimis medicis
collegis suis amantissimis. " The book, entitled Exerr:itatio anatomica de motu
cordis et sanguinis in animalibus, [hereafter, De motu cordis] appeared in
1628, published by Wilhelm Fitzer in Frankfurt. The publisher excused him-
self for the numerous printer's errors, adducing 'unfavorable times' - the
Thirty Years' War which was ravaging Germany.

2. The Cartesian Position

Harvey's delayed publication of the new circulation theory caused not only
246 RICHARD TOELLNER

confusion and fear; one representative of the French landed gentry immedi-
ately greeted the book as "by far the most important and useful discovery in
medicine.,,21 The author of this laudatory statement was the Sieur du Perron,
Rene Descartes, born in Touraine in 1596, who had just emigrated to Holland
when the De motu cordis appeared.
Descartes had gone to a country in which both economy and science were
thriving, and tolerance as well as peace prevailed. There, at his leisure and in
the seclusion and safety of that land, he proposed to do those things which he
had planned as a young man. In his search for truth, Descartes declared:
I intended to spend a great deal of time in necessary preparations in order to eradicate
from my mind all previously adopted and detrimental convictions. Moreover, I wanted
to collect a lot of experiences as material for subsequent conclusions, always practicing
my self-prescribed method so as to acquire greater skills in using it. 22

Thus, Descartes spent twenty years in Holland establishing a new philosophy

and physical science. In 1650, he died in Sweden because Queen Christina did
not understand that one should not summon a philosopher at 5 A.M. during
wintry Swedish nights in order to receive philosophy lessons. Descartes did
not survive such calls. He caught pneumonia and, true to himself and his
prinCiples, refused treatment, dying of the disease, to the satisfaction of the
attending physicians. Nevertheless, they never forgave him for having perished
without their assistance. 23
Rene Descartes sought the truth with an unequalled seriousness of mind.
He looked for unquestioned knowledge, certainty and confidence. Dis-
appointed because of the contradictions and uncertainties inherent in all
book-learning and science, Descartes probed in 'the book of the world',
nature, for more precise and unambiguous answers.24 However, he discovered
that the perceived phenomena, appearances and relationships in nature were
just as equivocal and muddled as the theories and interpretations about them.
Therefore, Descartes decided to search for truth within himself rather than in
nature or in previous authorities.
Descartes defined as true only that which could be recognized as clear and
distinct, together with certain statements which could successfully resist any
possible doubt. Thus, the philosopher began systematically to doubt every-
thing he had previously considered to be certain and true, including God, the
soul and the world.
Everything I have hitherto believed to be true, I received from or through the senses, but
I sometimes discovered that they had deceived me. Now wisdom means that one should
never trust those who have cheated us, even when the deception happened only once. 25
 DISCOVERY OF THE CIRCULA nON OF THE BLOOD 247

Hence, everything which is possibly certain falls prey to the methodological

doubt, for example, the internal and external sensory impressions mediated
through body and environment. The empirical sciences and even mathe-
matics, whose principles appear to be clear and true, were included in the
sceptical approach. When everything seemingly secure and certain was en-
gulfed in doubt, Descartes reached the 'rock of certitude', the 'fundamentum
inconcussum' which he had set out to find. Descartes declared: "For while I
doubt, I cannot doubt that I doubt what I think. Because I think, therefore, I
exist."
Taking the certainty of his own existence as a thinking entity - the 'res
cogitans' - as his basis, Descartes recovers all those elements, such as God and
the world, he had previously dismissed in his scepticism. "Archimedes only
required a firm and immovable point to mobilize the whole earth", wrote
Descartes, "and therefore great things may be expected if I find something
firm and unsha\<.able no matter how small.,,26
This so-called 'fundamen tum inconcussum' which Archimedes searched
for had apparently been found by Descartes, who was determined to change
the world. With the help of deduction, he wanted to separate the object - res
extensa - from the perceiving subject. The history of philosophy, much to its
detriment, has not hitherto adequately recognized that Descartes' thinking
was strongly geared toward practical verification. The great Cartesian polemic
against traditional metaphysics is not carried out in order to erect a new meta-
physical framework, but to arrive at a new physical science. In truth, the
metaphysical foundations of the Cartesian method - "to guide reason prop-
erly and search for truth in science"27 - only serve to fasten his method
more securely in order that man may become the 'maitre et possesseur', the
master and owner of nature through recognition of his own autonomous
reason. 28
Descartes succeeds in his efforts to reduce nature, its components and
events, into matter and motion. This is accomplished by a rigorous separa-
tion and differentiation between the thinking subject and the objective world,
subject and object, spirit and matter. Thus, everything in nature can be
explained 'more geometrico' and controlled as well as constructed by reason.
Descartes' new, purely mechanical physics allows in theory a technical
approach to nature, and he was convinced that it was not farfetched to
imagine the organism as a mechanical instrument.
I presume, that the body is merely a figure or machine made out of clay which God casts
into a form closely resembling us. Therefore, he not only gives externally such a machine
the color and form of our limbs, but furnishes inner parts necessary for walking, eating
248 RICHARD TOELLNER

and breathing. Finally, it is endowed with all our functions which we presume are
derived from matter and the disposition of the various organs. 29

Hence, Descartes constructs his physiology on the basis of mechanical models

provided by his contemporaries: the clock, organ, and hydraulic devices. In
view of such physiological ideas, one can understand the reasons why
Descartes - at first on hearsay only - enthusiastically accepted the discovery
of the circulation by the English physician William Harvey. However, after
studying in detail the book, De motu cordis, Descartes discovered that,
although the movement of the blood seemed to be confirmed by his own
criteria, the cardiac motions were explained quite contrary to his ideas.
Descartes believed in the connection of heart and circulation as one ther-
modynamic unit which functioned as follows: through the venous openings,
two large drops of blood fall into both cardiac chambers where theyexperi-
eGce an explosive dilatation thanks to the heat contained in a residual
fraction of fermented blood. The ebullition causes the distillation of various
blood fractions, the uppermost being that of the fine 'animal spirits.' The
pressure accompanying the explosion closes the venous valves and forces an
opening of the arterial orifices, whereby the hot and divided blood is ejected
towards the periphery at the height of the cardiac diastole or dilatation. The
more subtle fraction reaches the brain through the carotid arteries, traveling
from there to the hollow nerves where it becomes the nervous spirit matter
responsible for sensation and motion. 3o

C. THE CONTROVERSY

1. The Opposing Viewpoints

From the above description of the Harveyan and Cartesian positions, it is
clear that Descartes, in contrast to Harvey, denies the active contraction of
the heart. Seemingly traditional, the philosopher holds fast to the idea that
the ejection of blood occurs during diastole, and that such diastole is the
result of an expansion of heated blood. Descartes also borrows the concept of
a 'calor innatus' - an innate heat - from the ancients. Thus, he undeniably
adopts many ideas from traditional physiology although he attempts to
understand them according to his own principles. The 'calor' of the heart,
which Descartes never calls 'innatus', is not the traditionally implanted heat
but rather a 'lightless fire.' This he conceived to be something "that makes
the hay warm when it is stored before being dry, and what causes young
 DISCOVERY OF THE CIRCULA TION OF THE BLOOD 249

wines to fennent.,,31 Heat is, therefore, a physically conceived ferment, the

rapid motion of subtle particles of matter occuring in the undivided world of
dead and living things. Heat is the result of motion, not a principle of vitality.
For Harvey, the "pulsating motion of the heart" was the "origin (princi-
pium] of life.,,32 Although Harvey later claimed, on the basis of his embryo-
logical studies, that the blood itself was the cause of its own and the cardiac
motions, this shift has no bearing on the matter. What is important is that the
cardiac capacity for active contraction was, whatever its cause, a substantial
fonn. According to Aristotelian concepts, it was an effect of the soul, a
'spiritus vivificus' which Descartes, referring to Harvey considered an occult
qUality.33
Descartes could not take such an entity, derived from ancient qualitative
physics, as the cause of motion for his mechanical system. Rather, he was in
need of a mechanical power which he found in heat. Such a cause was
sufficient to explain the pressure, boiling, and production of the various
blood fractions. At the same time, the thermic theory necessarily molded the
Cartesian concept of cardiac motion, a fact freely acknowledged by
Descartes:
For, if we presume that the heart is moved in the manner described by Harvey, one has
not only to consider a special power causing motion - whose nature is difficult to
explain - but also accept the presence of other forces capable of changing the qualities
of the cardiac blood. 34

Thus, the question concerning the correct description of cardiac motions and
their cause becomes the cornerstone for the entire Cartesian physiology.
Descartes himself declares
that it is of the greatest importance to ascertain the true origin of the cardiac motions.
Without them, it would be impossible to learn anything concerning physiology since all
other bodily functions depend on the movement of the heart. 3S

Indeed, the entire question becomes the touchstone for Descartes' philos-
ophy. "I explain ... everything that belongs to the motions of the heart
quite differently than Harvey", Descartes announces on February 9, 1639, to
his close friend Mersenne, "but if what I have written is false, I will concede
that the rest of my philosophy is likewise an error ."36 What these sentences
reflect is the hitherto overlooked seriousness with which Descartes viewed his
so-called 'fables' or hypotheses, and the reason for his stubborn opposition to
Harvey's ideas on cardiac motion which were based on observation and
experiment.
If one compares the Cartesian and Harveyan schemas of cardiac movements
250 RICHARD TOELLNER

as well as their origin and effects with those valid today, Harvey's theory
emerges as the correct one in all essential aspects, while Descartes' ideas
were all erroneous. Such a fact constitutes a devastating indictment of the
Cartesian hypothesis. Now, if one asks why Harvey arrived at the correct
solution while Descartes blundered, the answer will have to be sought in the
relationship between 'experimentum et ratio.'

2. The Origin of the Can troversy

Both Harvey and Descartes, as we have seen, experimented with widely
diverging results. One can explain such a discrepancy by stating that Harvey
made careful and precise observations at the proper time and on the proper
subject while Descartes did not. However, if one scrutinizes the whole
question in more detail, both authors appear to have observed similar facts in
the same fashion, describing the rhythmical change between cardiac expan-
sion and contraction. Although certain details remained controversial
between both authors, one can quickly recognize that the real issue was not
about what could be seen or perceived, but how those events were to be inter-
preted. Therefore, the controversy between Harvey and Descartes concerning
the nature of cardiac motions does not arise because of observational dis-
crepancies but lies in the differing interpretations of the perceived pheno-
mena.
This is made plain by Descartes' understanding of two experiments which
actually refute the idea of a passive cardiac dilatation, confirming instead
Harvey's theory of an active contraction of the heart. In the first experiment,
a heart is taken out of the organism and deprived of all blood supply, and
continues to contract for some time. Therefore, the pulsation cannot be
ascribed to the expansion of blood in the cardiac chambers.
Descartes does not deny the experiment but objects to the conclusion
drawn from it. Being fond of log fires, he explains the phenomenon with the
help of an analogy. The situation, he says, is similar to the burning of a green
piece of wood still quite impregnated with moisture as the heart is saturated
with blood. When the log burns, he declares, one can see
that the vapors escaping from the interior of the wood emerge through the narrow
crevices of the bark because of the heat. Their escape can be likened to a wind, since the
bark begins to swell and show a series of fracture lines. Once the bark breaks at a certain
point, the swelling disappears because the trapped moisture has been released. Shortly
thereafter, as new vapor is formed, the bark arches again, intermittently releasing the
vapor through the existing opening. This phenomenon occurs rhythmically and can be
compared to the pulsations of an isolated eel's heart. 37
 DISCOVERY OF THE CIRCULA TlON OF THE BLOOD 251

In the second experiment, a fmger introduced into the cardiac ventricle

through a hole at the apex is able to feel the systolic contraction. Again
Descartes does not deny the phenomenon but gives it a different interpreta-
tion. He declares that the finger is not squeezed by the contracting walls of
the ventricle but rather by the tendinous fibers which exist between the
papillary muscles and the valves in the chamber. These fibers are being dis-
tended during the ventricular dilatation. It is only logical that Descartes con-
cludes his interpretation by declaring that these experiments prove only "that
our experience provides us often with opportunities for error if we do not
examine carefully all possible reasons.,,38
Seldom has Descartes expressed himself as clearly about the relationship of
experience and reasoning as in the above statements. They also depict the
Cartesian viewpoint about observation, experience and experiment. Hence,
the conflict between Descartes and Harvey about the cardiac motions proves
to be, in the end, a dispute about the proper sequence of experience and
reason, 'experimentum et ratio.'
Descartes' experiments and observations were obviously only designed to
confirm his rational conclusions about the structure and function of the
organism, and not to challenge these views. When a certain observation or an
experiment seemed to contradict such ideas, Descartes was always prepared -
on the basis of his fundamental principles - to doubt those empirically
acquired results. For him, only the conclusions which reason clearly and
distinctly recognizes were true, while the phenomena were by themselves,
confusing and ambiguous.
Objects placed before our eyes reveal only their complex secondary
qualities. Phenomena can be adequately grasped only if the primary qualities
of things - their geometrical characteristics such as motion and form - are
clearly distinguished. Such recognition is necessary before one can appreciate
the secondary qualities and recognize the existing geometrical and mechanical
relationships. Descartes wanted absolute certainty of knowledge and this is
only possible through 'more geometrico', the method of rational revelation.
Descartes does not trust the senses but rather relies on reason.
Because Descartes builds into his system only those observations and
experiences which can be mechanically interpreted, he incorporates uncriti-
cally and without hesitation many of the old ideas pertaining to anatomy,
physiology and pathology. Such an assimilation was carried out without any
consideration of the context, diametrically opposed to his own, in which
these ideas had originated and been operational. Descartes thus overlooked
the fact that his contemporaries, still thinking within these older cosmological
252 RICHARD TOELLNER

frameworks, were bound to misunderstand him. Moreover, he did not realize

in his scorn for history that language, the most effective historical tool, would
become a Trojan horse for his system. Precisely because he adopted older
concepts and ideas, Descartes allowed his opponents to infiltrate his system
and attempt to destroy it from within. Therefore, the history of philosophy
has been able to demonstrate, step by step, how Descartes remained a
prisoner of the history which he believed he had shed so determinedly by his
methodological approach. In the end, he became the victim of his own
fiction.
In contrast, Harvey based the certainty of his knowledge primarily on
'sensus, autopsia et experimentum' - perception, personal observation and
experience, which he placed before 'ratiocinium' or reason. 39 In so doing,
Harvey remained completely within the Aristotelian framework of thought,
even though he contradicted individual tenets of the system because he re-
cognized the phenomena more correctly through more precise observations
and experimental skill.
However, Harvey used his new insights into cardiac motion and the circula-
tion of the blood the better to illuminate the Aristotelian truth of life as a
special process within the entire cosmic activity and the inextricable rela-
tionship between form and matter, body and soul, force and motion. The
unquestioned acceptance of cardiac contraction as a vital phenomenon
particularly allowed the establishment of the new concepts. Harvey, of
course, came into conflict with his own physiology because of novel empirical
findings, thus creating problems which he himself could not solve.
The theory of the. circulation of the blood contradicted almost all those
functions previously ascribed to the blood. Harvey's most serious adversary,
the Parisian anatomist Riolan the Younger, clearly pointed out the difficulties
for traditional physiology, pathology and therapy which resulted from Har-
vey's scheme. Riolan's presentation was sharp and made in the spirit of self-
righteous conviction which certain people adopt when they are called upon to
defend holy, centuries-old beliefs and theories.
Riolan forced Harvey to answer the question whether traditional knowl-
edge or personal observation and experience had priority. Placed in that
position, Harvey took a decisive step, declaring that his observations were
correct even though they merely corresponded with some facts of anatomical
research. Said Harvey, "Before analyzing the reasons and the purpose of the
circulation, one must first make sure that it exists.,,40
Let me close by summarizing in a few sentences what I wanted to illustrate
by referring to the example given (the discovery of the circulation of the
 DISCOVERY OF THE CIRCULATION OF THE BLOOD 253

blood and the controversy regarding the nature of cardiac motion) in connec-
tion with our overall subject (logical and psychological aspects of discovery).
(1) Harvey was only able to make his discovery because he was completely
rooted in the old way of thinking. Coming from ancient-medieval philosophy,
Aristotelian physics and Galenic physiology, he and his contemporaries could
easily regard two important preconditions for his doctrine of circulation as
self-evident and thus posing no problems:
(a) the theory of circulation as such,
(b) the active contraction of the heart as a logical sequence of its facultas
movendi which as vis pulsifica, is a substantial form of the heart.
(2) Harvey's method of careful anatomical research and observing phy-
siological experiments was nothing new in his time; the empirical research of
anatomical structures and - in initial stages - of physiological functions is
the general characteristic of Renaissance medicine and science.
(3) New is the result of his hypothesis, that ensues conclUSively from the
observed phenomena: the quick circulation of blood in the body. He is well
aware of the novelty of this result. But apparently this does not make him
proud, but hesitant and uncertain, for this result does not fit at all into his
traditional conception of physiology. The function of a quick circulation of
blood in the body is not only utterly obscure to him, but it also contradicts
the functions ascribed to it by the old physiology in all major aspects.
(4) It was only when Riolan pointed out the undisputable dilemma and
showed that his doctrine of the circulation of the blood would completely
jeopardize the old physiology and medicine that Harvey tried to develop new
ideas about the function of circulation. It is, however, much more important
that he insisted on his observation being correct and that he defended it
against the reproach of having contradicted the authority of the Ancients by
declaring it to be a statement of nature itself. Nature is thus made the
supreme court of appeal "for nothing is older and of greater authority than
nature."
(5) Harvey can only justify his method and its result by referring to the
authority inherent in the old way of thinking. He plays off this authority of
the Ancients against the even older authority of nature. Nature gives him the
certainty necessary to believe in his fmdings. These can only be legitimized
by the old way of thinking.
Descartes is quite different. He achieves certainty about his cognition by
his new "method to guide the thoughts in a safe way." The new self-assurance
of the powers of reason had now discovered the general questionability of
every old authority as well as of the sensory cognition of nature and now
254 RICHARD TOELLNER

developed 'more geometrico' a new system of the world that is for the
organism a machine theory of the living, a mechanical physiology. Starting
from a new method of thinking Descartes arrives at a new pattern of thinking.
He realizes that this innovation is a radical one and is proud of it. The
authority of the Ancients and of nature have to yield to the authority of
reason.
What fascinates Descartes when contemplating Harvey's theory of circula-
tion is the possibility of a mechanistic interpretation. Anatomical and physio-
logical observations confirm his machine theory. The doctrine of circulation
fits into his system and it is not by pure chance that the recognition of the
doctrine of circulation and the establishment of a mechanistic physiology are
closely connected.
The one thing about Harvey's doctrine of circulation that had to displease
Descartes was Harvey's assumption about the motor of this circulation. The
active contraction of the heart could only be explained and understood by
using the old pattern of thinking of Galenic physiology and Aristotelian
physics. This is why he had to explain cardiac motion by a new mechanistic
theory. History proved, however, that his theory and observations were
wrong. His self-assured reason had led him astray; Harvey was right in the
end.
The clock of historical reason runs in a different way than Descartes could
have dreamed of. It is our task to investigate it if we wish to contribute to the
self-knowledge of science.

Translated by
I. MAGNUS and G. B. RISSE

NOTES

1 Cf. Toellner, R., 'Wissenschaftsgeschichte', Wolfenbiitteler Renaissance-Mitteilungen 1

(1977),60-63.
2 Cf. Toeilner, R., 'Mechanismus-Vitalismus: ein Paradigmawechsel? Testfall Haller', in
Die Struktur wissenschaftlicher Revolutionen und die Geschichte der Wissenschaften,
ed. by A. Diemer (Hain, Meisenheim am Glan, 1977), pp. 62-64.
3 Cohen, I. 8., 'Newton's Personality and Scientific Thought', Actes du Be Congres
international de l'Histoire des Sciences (1956) (Gruppo italiano di storia delle scienze,
Florence, 1958), Vol. I, p. 195.
4 Kant, I., 'Beantwortung der Frage: Was ist AufkHirung?' (Konigsberg, 1784), A 481:
"AufkHirung ist der Ausgang des Menschen aus seiner selbstverschuldeten Unmiindigkeit".
5 Gerlach. W., 'Die Kapazitiit hat das Wort', in Autoritiit - Was ist das heute?, ed. by
 DISCOVERY OF THE CIRCULATION OF THE BLOOD 255

G. Lehner (Ehrenwirth, Munich, 1965), pp. 121-135: "Mit der Befreiung der Natur-
forschung von jeder Voreingenommenheit, von weltanschaulich-dogmatischer AutoriHit
... wird die neue Naturwissenschaft groB". (pp. 125-126)
6 Cf. Toellner, R., 'The Controversy between Descartes and Harvey regarding the Nature
of Cardiac Motions', in Science, Medicine and Society in the Renaissance. Essays to
Honor Walter Pagel, ed. by A. G. Debus (Science History Publications, New York, 1972),
Vol. 2, pp. 73-89; Bernoulli, R., 'Descartes' Grundgedanken in medizin-historischer
Sicht', Gesnerus 35 (1978),44-53.
7 Rothschuh, K. E., 'Die Entwicklung der Kreislauflehre im Anschluss an William
Harvey. Ein Beispiel der "autokatalytischen Problementfaltung", in den Erfahrungs-
wissenschaften', Klinische Wochenschrift 35 (1957), 605-612. Cf. p. 611.
8 Leibniz and Haller said in criticism of Descartes that he tended constantly toward the
speculative. "Mons. des Cartes n'est tombe icy dans l'erreur que par ce qu'i! se fioit trop
11 ses pensees." (Leibniz, G. W., 'Discours de metaphysique', in Die philosophischen
Schriften, ed. by C. I. Gerhardt (Weidmannsche Buchhandlung, Berlin, 1875-1890).
Vol. 4 (1880), p. 443. [" ... Descartes has fallen into error here only because he had too
much confidence in his thoughts ... ", 'Discourse on Metaphysics', in Leibniz, Philosoph-
ical Papers and Letters, trans. and ed. by L. E. Loemker. 2nd ed. (Reidel, Dordrecht,
1969), pp. 303-330. Cf. p. 315.
"Deux autres romans physiologiques de Descartes demontrent qu'on peut connoitre
la bonne methode de rechercher la verite, et suivre celle qui lui est la plus contraire."
(Haller, A., Art. 'Physiologie' in Supplement Ii l'Encyclopedie by Diderot and D'Alem-
bert, Vol. 4, Amsterdam, 1777, 349.) [Two other physiological works of Descartes show
that one can know the right method of looking for the truth and follow the one most
contrary to iLl
9 Brecht, B., Leben des Galilei (Suhrkamp, Berlin, 1957), Act 4, scene 8, pp. 57 -73.
(The Life of Galileo, trans. by D. I. Vesey (Methuen, London, 1963).
10 Kepler, J., Johannes Kepler in seinen Briefen, ed. by M. Caspar and W. von Dyk.
2 vols. (Oldenbourg, Munich, 1930), Vol. 1, p. 353.
11 Cf. Olschki, L., Geschichte der neusprachlichen wissenschaftlichen Literatur. 3 vols.
(Winter, Heidelberg, 1919-1927), Vol. 3: Galileo und seine Zeit, 1927, p. 220, Note 3.
12 Harvey, W., Exercitationes duae anatomicae de circulatione sanguinis, ed. by K. J.
Franklin (Blackwell, Oxford, 1958), p. 137ff.: "Nulla alia certior demonstratio ad fidem
faciendam adduci poterit, quam sensus et Cti'TO\jJi.c,".
13 "Aut approbatas opiniones relinquere indignum putent; et per tot saecula traditam
disciplinam, veterumque auctoritatem, in dubium vocari nefas putent; his omnibus
respondeam, facta manifesta sensui nullas opiniones, naturae opera nullam antiquitatem
morari: natura enim nilii! antiquius, majorisve auctoritatis." Ibid., p. 136.
14 Cf. Lesky, E., 'Harvey und Aristoteles', Sudhoffs Archiv 41 (1957), 289-316 and
349-378. Pagel, W., William Harvey's Biological Ideas. Selected Aspects and Historical
Background (Karger, Basel and New York, 1967), pp. 28-47.
15 Pagel,op. cit., pp. 20-23.
16 Harvey, W., Exercitatio anatomica de motu cordis et sanguinis in animalibus (Frank-
furt, 1628), ed. by K. J. Franklin, (Blackwell, Oxford, 1957), p. 165: "Coepi ego met
mecum cogitare, an motionem quandam quasi in circulo haberet: quam postea veram
esse reperi, et sanguinem e corde per arterias in habitum corporis et omnes partes
protrudi et impelli a sinistri cordis ventriculi pulsu, quemadmodum in pulmones per
256 RICHARD TOELLNER

venam arteriosam a dextri; et rursus per venas in venam cavam et usque ad auriculam
dextram remeare, quemadmodum ex pulmonibus per arteriam dictam venosam ad sinis-
trum ventriculum ... ".
17 "Quem motum circularem eo pacto nominare liceat, quo Aristoteles aerem et pluviam
circularem superiorum motum aemulari dixit." Ibid.
18 "Ibi calore naturali, potenti, fervido, tanquam vitae thesauro, denuo colliquatur,
spiritibus et (ut ita dicam) balsamo praegnans; inde rursus dispensatur: et haec omnia a
motu et pulsu cordis dependere. Ita cor principium vitae et sol microcosmi, ut propor-
tionabiliter sol cor mundi, appeUari meretur; cujus virtute et pulsu sanguis movetur,
perficitur, vegetatur, et a corruptione et grumefactione vindicatur." Ibid.
19 "Unde perpetuum sanguinis motum in circulo fieri pulsu cordis." Harvey, W., Prae-
lectiones anatomie universalis. De musculis, ed. by G. Whitteridge (Livingston, Edinburgh
and London, 1964), Fol. 80 v.
20 Schlegel, P. M., De sanguinis motu commentatio, in qua praecipue in Joh. Riolani,
V. C. sententiam inquiritur (Hamburg, 1650), Praef., I.
21 Descartes to Beverwick, 5 July 1643: "Quippe, quamuis circa sanguinis circulationem
cum Heruaeo plane consentiam, ipsumque vt praestantissimi illius inuenti, quo nuUum
maius & vtilius in medicinii esse puto, primum auctorem suspiciam, tamen circa motum
cordis omnino ab eo dissentio". Descartes, Oeuvres, published by Charles Adam and
Paul Tannery, 13 Vols. (Cerf, Paris), 1897-1913, Vol. 4 (1901), p. 4. [Indeed, although
I clearly agree with Harvey about the circulation of the blood, and believe him to be
the first author of that most excellent discovery, which I think is greater and more useful
than anything else in medicine, I nevertheless disagree with him completely about the
movement of the heart.)
22 "et que ie n'eusse, auparauant, employe beau coup de terns a m'y preparer, tant en
deracinant de mon esprit toutes les mauuaises opinions que i'y auois receues auant ce
terns la, qu'en faisant amas de plusieurs experiences, pour estre apres in matiere de
mes raisonnemens, & en m'exerc:;ant tousiours en la Methode que ie m'estois prescrite,
affin de m'y affermir de plus en plus." Descartes, Oeuvres 6 (1902), 22 (Discours de la
Methode, 1637, referred to hereafter as Discours).
23 For many similar judgments, HaUer cited here: "Des D(escartes) Vorurtheile kosteten
im das Leben, er woUte im Seitenstiche sich nicht die Ader Mnen lassen", G6ttingische
Gelehrte Anzeigen, 1773, Zugabe, 371.
24 "Mais apres que i'eu employe quelques annees a estudier ainsi dans Ie liure du monde,
& a tascher d'acquerir quelque experience, ie pris vn iour resolution d'estudier aussy en
moymesme, & d'employer toutes les forces de mon esprit a choysir Ie chemins que ie
deuois suiure." Descartes, Oeuvres 6 (1902), 10 (Discours, 1637). ["But after I had
spent a few years studying the book of the world, and in trying to acquire some experi-
ence, one day I made a resolution to also study of myself and to employ aU the forces
of my wits in choosing the paths I ought to foUow". 'Discourse Concerning the Method',
in Descartes, The Essential Writings, trans. by J. J. Blom (Harper, New York, 1977),
pp. 114-164. Cf. p. 120. Referred to hereafter as Discourse.)
25 "Nempe quidquid hactenus ut maxime verum admisi, vel a sensibus, vel per sensus
accepi; hos autem interdum fallere deprehendi, ac prudentiae est nunquam illis plane
confidere qui nos vel semel deceperunt." Descartes, Oeuvres 7 (1904), 18 (Meditationes,
1641).
26 "Nihil nisi punctum petebat Archimedes, quod esset firmum & immobile, ut integram
 DISCOVERY OF THE CIRCULATION OF THE BLOOD 257

terram loco dimoveret; magna quoque speranda sunt, si vel minimum quid invenero quod
certum sit & inconcussum." Ibid. 24 (Meditationes, 1641).
27 Subtitle of Discours de la Methode: Pour bien conduire sa raison, et chercher la verite
dans les sciences.
28 Descartes said of the foundation of his physics: "elles m'ont fait voir qu'il est pos-
sible de paruenir a des connoissances qui soient fort vtiles a la vie, & qu'au lieu de cete
Philosophie speculatiue, qu'on enseigne dans les escholes, on en peut trouuer vne prati-
que, par laquelle connoissant la force & les actions du feu, de l'eau, de I'air, des astres,
des cieux, & de to us les autres cors qui nous enuironnent, aussy distinctement que nous
connoissons les diuers mestiers de nos artisans, nous les pourrions employer en mesme
fa«:;on a tous les vsages ausquels ils sont propres, & ainsi nous rendre comme maistres &
possesseurs de la Nature." Descartes, Oeuvres 6 (1902), 61-62 (Discours, 1637). ["For
these notions have made me see that it is possible to reach various knowledge very useful
to life, and that, instead of the speculative philosophy taught in the schools, one can find
another practical philosophy by which we can know - as distinctly as we now know the
different professions of our artisans - the force and action of fue, water, air, the stars,
the heavens, and all the other bodies that surround us; and thus, in the same way that we
employ artisans, we could also employ all these things for all the uses for which they are
fit, thereby rendering ourselves like masters and possessors of nature." Discourse, op. cit.,
p.154.)
29 "Ie suppose que Ie Corps n'est autre chose qu'vne statuti ou machine de terre, que
Dieu forme tout expres, pour la rendre la plus semblable it no us qu'il est possible: en
sorte que, non seulement il luy donne au dehors la couleur & la figure de tous nos mem-
bres, mais aussi qu'il met au dedans toutes les pieces qui sont requises pour faire qu'elle
marche, qu'elle mange, qu'elle respire, & en fin qu'elle imite toutes celles de nos fonc-
tions qui peuuent estre imaginees proceder de la matiere, & ne dependre que de la dis-
position des organes." Descartes, Oeuvres 11 (1909), 120 (Traite de [,Homme, 1633).
30 Cf. Descartes, Oeuvres, VI, 46-53; XI, 123, 127; XI, 228-245. Rothschuh, K. E.,
'Rene Descartes und die Theorie der Lebenserscheinungen', Sudhoffs Archiv 50 (1966),
25-42.
31 "ie me contentay de supposer que Dieu formast Ie cors d'vn homme, entierement
semblable a l'vn des nostres, tant en la figure exterieure de ses membres qu'en la con-
formation interieure de ses organes, sans Ie composer d'autre matiere que de celle que
i'auois descrite, & sans mettre en luy, au commencement, aucune arne raisonnable, ny
aucune autre chose pour y seruir d'ame vegetante ou sensitiue, sinon qu'il excitast en
son coeur vn de ces feux sans lumiere, que i'auois desia expJiquez, & que ie ne conceuois
point d'autre nature que celuy qui echaufe Ie foin, 10rsqu'on I'a renferme auant qu'il
fust sec, ou qui fait bouillir les vins nouueaux, lorsqu'on les laisse cuuer sur la rape."
Descartes, Oeuvres 6 (1902),45-46 (Discours, 1637). ["I contented myself with assum-
ing that God formed the body of man entirely similar to our body, as much in the
exterior figure of its members as in the conformity of its interior organs; and that he
composed this body only from the matter I had described, and without placing in it, in
the beginning, any reasoning soul, or any other thing to serve therein as a vegetative or
sensitive soul, except that he excited in its heart one of those fires without light that I
had already explained and conceived as having the same nature as the fue that heats hay
when one has stored it away before it was dry, or that activates new wines when one
allows them to ferment over sediment". Discourse, op. cit., p. 143.)
258 RICHARD TOELLNER

32 "Mot us et pulsus cordis", "cor principium vitae et sol microcosmi", Harvey, De

motu cordis, 165.
33 Descartes to Plempius, 23 March 1638: "Nam cum, vt explices quo pacto cor in
hominis cadauere ab anima absente moueri possit, confugis ad calorem & spiritum viui-
ficum, tanquam animae instrumenta, quae in virtute eius hoc agant, quid, quaeso, aliud
est quam extrema velie experiri?" Descartes, Oeuvres 2 (1898), 65. [For when, in order
for you to explain how the heart can be moved in the corpse of a man by the absent
soul, you have recourse to heat and the quickening spirit as instruments of the soul,
which act by virtue of it here, what else is there, I ask, than to try to want extreme
things.]
34 "Or en supposant que Ie coeur se meut en la fay on qu' Heruaeus Ie decrit, non seule-
ment i! faut imaginer quelque faculte qui cause ce mouuement, la nature de laquelle est
beau coup plus difficile a conceuoir, que tout ce qu'i! pretend expliquer par elle; mais i1
faudroit supposer, outre cela, d'autres facultez qui changeassent les qualitez du sang,
pendant qu'il est dans Ie coeur." Descartes, Oeuvres 11 (1909), 243-244 (De la forma-
tion de l'animal, 1647 -1648).
35 "Et neantmoins il impo:te si fort de connoistre la vraye cause du mouuement du
coeur, que sans cela il est impossible de rien s<;:auoir touchant la Theorie de la Medecine,
pource que toutes les autres fonctions de l'animal en dependent, ainsi qu'on verra
clairement de ce qui suit." Ibid., p. 245.
36 Descartes to Mersenne, 9 February 1639: "i'ay escrit Ie mesme qu'Herueus, a cause
de la circulation du sang, qui leur donne seule dans la veue, i'explique toutefois tout ce
qui appartient au mouuement du coeur d'vne fa<;:on entierement contraire a la siene ...
ie veux bien qu'on pense que, si ce qui i'ay escrit de cela, ... se trouue faux, tout Ie reste
de rna Philosophie ne vaut rien." Descartes, Oeuvres 2 (1898),501.
37 "Praeterea me mini me alias vidisse, cum ligna viridia vrerentur, vel poma coqueren-
tur, vapores vi caloris ex eorum partibus interioribus emergentes non modo per angustas
corticis rimas exeundo ventum imitari, quod nemo non aduertit, sed etiam interdum ita
dispositam esse partem corticis, in qua tales rimae fiunt, vt aliquantum intumescat prius-
quam rima aperiatur; quae deinde rima aperta confestim detumescit, quia nempe omnis
vapor ilIo tumore inclusus affatim tunc egreditur, nec nouus tam cito succedit. Sed paulo
post, vapore alio succendente, pars eadem corticis rursum intumescit, & rima aperitur, &
vapor exit, vt prius. Atque hic modus saepius repetitus pulsationem cordis, non quidem
viui, sed eius quod hie habeo ex anguilla excisum, perbelle imitatur." Ibid., pp. 67-68.
38 "Et toutesfois cela ne prouue autre chose, sinon que les experiences mesme no us
donnent souuent occasion de nous tromper, lors que nous n'examinons pas assez toutes
les causes qu'elles peuuent auoir." Descartes, Oeuvres 11 (1909), 242 (De la formaton
de I 'animal, 1647-1648).
39 "Astronomiae exemplar non hic imitandum est; ... Sed (sicut quis, eclipseos causam
perquirens, supra lunam sisteretur, ubi sensu causam discerneret, non ratiocinio) sensi-
bilium, quae sub sensum cadunt, nulla alia certior demonstratio ad fidem faciendam
adduci poterit, quam sensus et OtVTOl/liot." Harvey, De circulatione, 137-138. ["The ex-
ample of astronomy is not to be followed here .... But (as one seeking the cause of an
eclipse would be placed above the moon to discover that cause by sensation and not by
reckoning) with regard to sensible things, that is, the things that come under the senses,
it will be impossible to bring forward any surer demonstration to induce belief than the
actual sensation and seeing for oneself." Harvey, De circulatione (Oxford, 1968), p. 47.]
 DISCOVERY OF THE CIRCULATION OF BLOOD 259

40 "Ad eos qui circulationem repudiant, quia neque efficientem, neque finalem causam
vident, cui bono? (de quo adhuc nihil adjunxi ... ): Prius in confesso esse debet, quod
sit, antequam propter quid, inquirendum." Ibid., pp. 135-136. ["With regard to those
who repudiate the circulation because they see neither its efficient nor its fmal cause, I
have to date added no reply to their query 'Who benefits?' ... First, one ought to admit
what should be investigated rather than the reason for such further study." HarveY,De
circulatione (Oxford, 1968), pp. 44-45.
 LUIGI BELLONI

THE DISCOVERY OF DUODENAL ANCYLOSTOMA AND

OF ITS PATHOGENIC POWER *

Duodenal Ancylostoma (hookwonn), was discovered in Milan in 1838,1 one

year after the death of Giovanni Rasori (1766-1837) and in the same city
where, from the beginning of the century, he had advocated his high-sounding
doctrine of the 'counterstimulus.' This discovery represents, therefore, the
most tangible sign of the disengagement from what Antonio Cazzaniga
(1885-1973) so aptly defmed as "the great crisis of Italian medicine in the
early 18oo's."2 In other words, this represented the way out of that blind
alley that vitalistic systems had gotten themselves into.
It is well known that overcoming the systematic and typically aprioristic,
speculative trend was due substantially to the new 'observation'-medicine so
well summarized in Laennec's De ['auscultation mediate, which appeared in
Paris in 1819. In accordance with the lines set forth in De sedibus et causis
morborum per anatomen indagatis (1761), Laennec and his followers strove
to identify in the living being the lesion of that organ which constituted the
seat and cause of disease, as meant by Morgagni. Auscultation, in addition to
Auenbrugger's percussion (1761) and the old practice of palpation, integrated
physical semiotics, considered as an essential instrument for the study of
patients. Clinical medicine conceived in these terms called for the systematic
necroscopic control of dead patients, in order to establish the correlation
between postmortem fmdings and clinical symptoms, said correlation being
borne out by adequate statistics. Medicine that is based essentially on the
study of patients and corpses is, first and foremost, a 'hospital'-medicine,3
and such was the case also in Milan where it was introduced, above all, by
Domenico Gola (1797-1867), a physician working at the S. Ambrogio ad
Nemus Hospital which opened in 1823, the same year the first volume of the
Italian translation of Morgagni's De sedibus appeared in Milan.
Just a bit later, observation-medicine was introduced into the old and
glorious 'Ospedale Maggiore' of Milan, most probably thanks to Francesco-
Enrico Acerbi (1785-1827), himself widely-known through Manzoni's The
Betrothed as a supporter of the living contagion theory. Proof exists that
Acerbi had been practicing observation-medicine at the Ospedale Maggiore
ever since 1825, and "that at that time he was maintaining a scientific
correspondence with Laennec himself"; but both researchers were soon to die
261
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 261-279.
Copyright © 1980 by D. Reidel Publishing Company.
262 LUIGI BELLONI

from pulmonary tuberculosis, undoubtedly caused by the close and continual

contact with patients required by the new semiotics.
This new method, introduced by Acerbi, was welcomed with enthusiasm
by some young Ospedale Maggiore physicians, such as Napoleone-Massi-
miliano Sormani (c. 1800-1848) and Alessandro Gambarini, who gave us a
splendid description of 'tiger heart' [a fatty, degenerated heart in which the
fat is disposed in the form of broken stripes] in 1838.4
However, the most important representative of the anatomo-clinical
approach at the Ospedale Maggiore was Angelo Dubini (1813-1902), whom
we can consider at the same time one of the greatest pathological-anatomists
in Italy in the 1800's, though such an activity was only one aspect of his
complex figure as a physician. Born in Milan in 1813, Dubini graduated in
medicine from Pavia University in 1837 with a dissertation on anatomical
technique (Trattato di antropotomia), sponsored by Bartolomeo Panizza
(1785-1867), who represented the best of Pavia's Medical School in the first
half of the nineteenth century. In fact, during those decades, Panizza was the
scientific heir of A. Scarpa, a professor of medicine, who reflected the uni-
versity's scientific splendor at the end of the eighteenth century. Panizza
passed on this heritage to all those connected with the renaissance of biology
at Pavia University in the 1860's. Giulio Bizzozero, Camillo Go1gi and Battista
Grassi were among the foremost.
Once back in Milan, Dubini started his career at the Ospedale Maggiore,
but returned to Pavia as assistant to G. Corneliani, professor of the Medical
Clinic for the period 1839-1841, and "during this two-year period he gave a
free course in auscultation."
In November 1841, Dubini began a long tour of the main European
medical centers. In Paris, he attended the courses of G. Andral, who was
carrying on with Laennec's program superbly. During the period of his
travels, Dubini scrupulously visited the most important museums of normal,
comparative and pathological anatomy of the main French, English, Belgian,
Austrian and German cities, taking diligent notes too. His keen interest in
pathological anatomy is also testified to by his virtually daily practice of per-
forming autopsies. In his paper De/1'arte di fare /e sezioni cadaveriche
(1847)5, Dubini sets out the method he had been observing "for almost ten
years, i.e., by never opening fewer, and often more, than three hundred
corpses a year." Such a systematic devotion to the autopsy table was
rewarded shortly after his return to Milan from abroad.
In fact, in November 1842, he was able to confirm his discovery of a "new
intestinal worm in humans", which he had already come across in the same
 DISCOVERY OF DUODENAL ANCYLOSTOMA 263

hospital back in May 1838 while dissecting the body of a female peasant who
had died from croupous pneumonia. Dubini made his discovery public in
April 1843 6 and designated this new worm as the Agchylostoma [sic!]
duodenale, a designation that underlined its hooked mouth and 'habitat' in
the human organism. In this work Dubini emphasized the very high incidence
of this worm, which "even if it has not been detected or described by other
workers, is to be found in at least twenty corpses out of a hundred that are
dissected in order to fmd it." So, the problem then became to discover the
"reasons that have delayed its discovery up to now"; and it is worthwhile
giving the entire explanation of this as left by Dubini:
For a few years now, namely since most recent anatomo-pathological studies undertaken
in cases of typhoid fever or tuberculosis with intestinal ulcerations,? we have been dis-
secting the intestine to investigate the state of the mucous membrane. Doctors of old,
who either observed the intestine intact or did not always open it completely, could have
no knowledge of a worm that was in any case so small as to be undetectable to fingers
that had squeezed the initial intestine loops by chance. However, let us admit it, modern
practice was not the most suited to detect it either, as doctors usually opened the in-
testine and rinsed it with abundant water before examining the inner membrane. In this
way its mucous product was washed off and settled to the bottom of the wateI together
with the helminths [or intestinal worms) housed therein. Finally, the considerable
yellow-pink or ash-grey, and less-than-transparent mucus in which they are usually
wrapped, coupled with then small size, undoubtedly constituted a considerable obstacle
even for the most watchful observers. Yet, it is sufficient to separate the intestine from
the mesentery, segment by segment, an arm's length at a time, thereby unfolding its
loops; cut it lengthwise by means of an enterotomy [incision into the intestine); lay it
over one of the corpse's thighs and carefully examine the mucus contained therein. This
mucus should then be removed to expose the underlying membrane by using the back of
the same instrument, no washing being necessary. When, in the aforementioned circum-
stances, considerable thick, ash-grey or reddish mucus exists in the duodenum or
jejunum, it will not be difficult, provided it is carefully observed, to see small worms
wrapped in it, curled very much around themselves, transparent for the front quarter of
their length, yellowish, reddish and brown for then hind three quarters and marked in
the middle by a black spot which is the stomach that is normally filled with blackish
matter.

This passage is indicative both of the reasons that prevented earlier anatomists
from observing the worm and of the reasons that led Dubini away from attri-
buting any pathogenic action to it. The worm is, in fact, usually found free in
the duodenum-jejunum lumen and coated with a thick and "mostly yellow-
pink colored'" mucus, which is present in the lumen itself in greater
quantities. Sometimes the worm 'adheres' to the mucous membrane "by
means of its hook-shaped oral extremity." The mucous membrane "may
appear as normally constituted, i.e., slate-color-like with many block or red
264 LUIGI BELLONI

dots or also simply veined." The meaning of this dotting, for us quite
obvious, was the source of deep and prolonged concern for Dubini, who was
partially misled by cases in which the dots were either not accompanied by
worm-infestation or were localized in other areas that were not the Ancy-
lostoma's 'habitat' (as, for example, the case in which "the dotting was con-
fined to the elliptical plates of the ileum"). In one case, "this dotting was
found in the jejunum and there amidst the minute slate-colored dots, striae of
red points could be seen, which evidently had given rise to a color deeper
than slate in other parts of the mucous membrane."s In his Entozoografia 9
Dubini also recalls these 'slate-colored dots' which "seem to have originally
been red dots."
Dubini then pointed out - both in the worm and in the human organ
holding it - various indications of a hematophagous· aggressiveness of the
worm towards the organ, but preferred to adopt a very cautious attitude to-
ward the hypothesis of the hookworm's pathogenic activity. This perplexity
was increased both by the very high frequency (20% incidence at least) with
which the worm appeared in deliberately undertaken autopsies, as well as
by the presence of other lesions that might serve as causa mortis. What is
more, "the variety of diseases that caused those individuals to die prevents us
from establishing a relationship between the character of the disease and the
presence of the worm. I could only observe that deteriorated constitutions,
the cachectic [related to a general lack of nutrition and wasting occurring in
the course of a chronic disease or emotional disturbance] , the diarrheic, the
emaciated, or the anasarcous [a generalized inf:tltration of edema fluid into
subcutaneous connective tissue] were those in which it was found more
often." Dubini concludes that the presence of the worms "certainly cannot
be indifferent to the organism, even though - as is for trichocephalus [a
genus of worms related to the trichina worm] - they do not give rise to parti-
cular symptoms."
Extremely suggestive is the anatomo-clinical parallel relationship that
Dubini established 10 at the end of an autopsy he carried out on 5 July 1843,
in which the most conventional of anaemic patterns appeared - starting from
a 'tiger heart' - in a woman "not drained of blood by natural or artificial
losses" and carrier of "100 or more agchilostomas" [sic!].
The 'agchilostoma' -later changed to 'ancylostoma' - was to become the
object of special treatment in Dubini's parasitology treatise, Entozoografia
umana per servire di complemento agli studj d'anatomia patologica 11
published in 1850. This is now a classic treatise thanks also to its fifteen
copper plates, that were reprinted in Edoardo Perroncito's treatise, I parassiti
 DISCOVERY OF DUODENAL ANCYLOSTOMA 265

dell'uomo e degli animaU utiU12 published in 1882 and again in 1902 as a

second edition. We read in the Entozoograjia that
in some cases the number of hookworms is so prodigious and the quantity of mucus
acting as a bed, and perhaps pasture, for them is so large that where there is no other
apparent lesion, it would be quite natural to believe the disease and death due merely to
their presence and to admit without any hesitation a vermination of ancylostoma just as
Alibert did for taenia [genus of tapeworms] ,ascaris [A scaris lumbricoides: an intestinal
parasitic worm, similar to earth-worm lumbricus], trichocephalus . ...

For the purpose of a critical history of science, it is very interesting to study

the reasons which prevented Dubini from declaring decisively that the worm
he had discovered exhibited hematophagic activity, especially in view of the
fact that he was a convinced supporter of the 'live contagion' theory at a time
when anticontagionism was predominantP As early as 1842, at the time of
his stay in Paris, when he was "called in as an eye-witness to Gruby's micro-
scopical observations and his excellent therapeutical experiments in con-
nection with the tinea favosa [ringworm] ", Dubini 14 confirmed that it
derived from the discovery by Agostino Bassi (1773-1856)15 in Lombardy
of the cryptogamic etiology of the muscardine disease among silkworms.
Dubini's experiments on the transmission of scabies on himself also date back
to his Paris stay.16 For his tests Dubini would extract the mite from scabied
skin and, having deposited it upon his own skin, would observe its aggressive
action, in particular its penetration and digging of a burrow or groove. 17
Why then was Dubini so reluctant to claim a similar aggressiveness for
hookworm as well? In addition to the above mentioned reasons, it ought to
be added that worm pathol~gy had fallen into disfavor at that time. Reaction
to the vitalistic systems was occurring not only at an 'official' level - namely
from observation-medicine quarters - but also on the level of 'heretical'
medicine, for which Milan was just as much a feeding ground for its develop-
ment, as counterstimulant therapy proved to be truly harmful with its
excessive use of bleeding and large doses of strong drugs. Consequently, many
people preferred those phYSicians or healers who gave substantially harmless
treatment, with a liturgical or doctrinaire charge aimed at winning over the
patient. I am alluding to homeopathy, neomesmerism, Priessnitz's hydro-
pathy and so forth, all methods which were a far cry from poisoning or
bleeding the patient.
A representative example of the disorientation of the times is the 'popular
lecture', n volgo e la medicina, which the surgeon, Giovanni Raiberti (1805-
1861) published at his own expense in 1840-1841 to defend the practice of
bleeding against the "vulgar delirium about the healing art" and above all,
266 LUIGI BELLONI

against homeopathic medicine. Side by side with this main delirium, Raiberti
did not forget either neomesmerism nor hydropathy, while he considered two
other therapeutic methods as eclipsed. He wrote:
Our age - I am referring only to those systems within the reach of popular intelligence
and favor - has witnessed the birth, life and death of two crazy methods of healing all
diseases: Bucellati's vermifuge method and Le Roi's purgation method. The former
(Bucellati), having grown up in our lands, was not brilliant enough to be known abroad
and so his reputation was confmed to our city. The other (Leroi), because he came
from Paris, was known over the whole of Europe as if he were a fashion plate. Both of
them had their followers, both boasted of miracles, and had some martyrs. Hence they
passed on, as all errors do, leaving no traces other than a number of victims and a few
epigrams.

The 'Bucellati' mentioned by Raiberti is, in fact, the physician Luigi

Bucellati 1B from Piacenza who settled in Milan in 1816, where, for a number
of years and with the help of a notable number of publications, he success-
fully advocated a therapy that rejected the practice of bleeding which he
"considered to be the cause of most diseases." Non-iatrogenic diseases - as
they might be called today - were on the contrary, related to an "unhealthy
gastric stimulus" and to 'saburrae' [decomposition of food in the stomach]
and 'vermination' in particular, to the extent that Bucellati restricted his
therapy to purgative and anthelmintic drugs. According to his own state-
ments, the "abundant discharges of truly fetid matter" especially if mixed
with numerous worms or fragments of worms, did not fail to influence the
patient favorably and sometimes imparted a 'very favorable course' to the
disease.
As agreeable as this appeared to many patients who were terrified by the
counterstimulant theory, Bucellati's approach easily laid itself open to serious
objections, first of all, that
expelled worms are also observable in an illness that we know is not attributable to these
parasites but due to other causes, and in this case the evacuation of worms throws no
light on the origin of the illness, but only notifies us of a complication which is quite
secondary and inconclusive.

These cautious words are Dubini's19 and, to a great extent could also be
adapted for the case of hookworm, especially since this worm was found free
in the intestinal lumen, more or less like an ascarid. As we know today, this is
a typical post-mortem fmding - migration of the worms - but at that time
it was the decisive element that put Dubini off the track from attributing a
definite pathogenic action to the worm he had discovered.
 DISCOVERY OF DUODENAL ANCYLOSTOMA 267

Dubini published his discovery in April 1843. Two years later, in 1845, it
was highlighted by the zoologist, Karl-Theodor-Ernst von Siebold (1804-
85),20 in the course of a summary review of the progress made in the field of
helminthology in the previous two years. In 1847, Franz Pruner (Pruner Bey,
1808-82)21 mentioned the high incidence of hookworm in corpse dissections
carried out in Egypt where, unlike Lombardy, corpses could be dissected
immediately after death. The parasite adheres to the mucous membrane of
the duodenum by means of the hooks on its sucking apparatus thereby giving
rise to ecchymosis [black-and-blue spot] .
After moving to Egypt in 1850, Wilhelm Griesinger (1817-68)22 and
Theodor Bilharz (1825-62)23 went even further. The enterorrhagia [intes-
tinal hemorrhage] that accompanied ecchymosis in one corpse dissected on
17 April 1852 emphasized the analogy between hookworms and leeches. The
so-called Egyptian chlorosis [synonym for ancylostomiasis, i.e., hookworm
disease] , a severe and fatal cachectic disease, widely endemic in Egypt, may be
considered as a 'hookworm-originated disease' (,Anchylostomenkrankheit').
The Egyptian hookworm firmly anchors itself to the mucous membrane
of the duodenum and sucks abundant quantities of blood, gradually putting
the infested patient into an acute state of anaemia until he dies ('progressive
pernicious anaemia'). The Lombard hookworm, on the other hand, is much
less dangerous and usually prefers to remain free in the intestinal lumen, as
Dubini had discovered it. Basically, this was the opinion of Giacomo Sangalli
(1821-97),24 Professor of Pathological Anatomy at the University of Pavia,
who stated in 1866: "for some years I have found the hookworm in nearly
half the corpses dissected in my school". 2S This same opinion was set forth
again by Sangalli 26 some ten years later in 1876.
The Brazilian hookworm showed itself to be just as aggressive as the
Egyptian as its blood sucking causes the so-called intertropical anemia or
tropical chlorosis. In these cases, autopsy findings were quite like those for
Egyptian chlorosis, as Otto Wucherer (1820-73)27 found in Bahia, from
1865 on.
On 20 May 1877, Prospero Sonsino,28 a Florentine physician who prac-
ticed in Egypt, informed the Societa Medico-Fisica Fiorentina about the
observations he had made in connection with hookworm chlorosis following
Griesinger's thesis during autopsies performed in Cairo. On 30 September
of the same year, Carlo Morelli,29 influenced by Sonsino, found in Florence
a hookworm in the corpse of a woman who had died from 'progressive
pernicious anemia', thereby finding the hookworm in Italy for the first time
outside of Lombardy_
268 LUIGI BELLONI

However, it may be worth remembering that we have now entered the

crucial period in which the 'live contagion' theory is taking hold of its own.
The famous discoveries of Louis Pasteur (1822-95) and Robert Koch (1843-
1910) showed that minute living beings were the cause of serious diseases in
higher animals and humans. 1876, the year Koch produced his fundamental
work on the carbuncle bacillus, marked a turning point also in the biological
study of hookworm. Karl Georg Friedrich Rudolf Leuckart (1823-1898)30
set out his biological study of hookworm in the second volume of his treatise
Die menschlichen Parasiten und die von ihnen henUhrenden Krankheiten
which had just come out that same year. As the human-hosted worm was not
found in Germany, Leuckart shifted his research to the Dochmius trigona-
cephalus in dogs, carefully investigating its development, embryo formation,
egg hatching, larva freeing and the successive moults up to the fully grown
organism.
Similar to dog Dochmius is Dochmius Balsami 31 of cats which was dis-
covered in September 1876 by Battista Grassi (1854-1925), who at the end
of his fourth year of medicine at the University of Pavia, was spending his
summer holidays in his home town of Rovellasca (in the province of Como).
Upon his return to Pavia that autumn, Grassi continued studying the worm in
Leopoldo Maggi's (1840-1905) comparative anatomy laboratory, where he
worked alongside the assistant, Corrado Parona (1848-1922)32 and extended
the research itself to duodenal hookworm. In a short article published 18 May
1878 in the Gazzetta medica,33 the two authors summarized what they had
discovered about the worm's embryology.
The egg of Ancylostoma, both in the body of the female worm and in its
free state within the intestine had been illustrated in those very plates of
Dubini, who had failed, however, to exploit his own observation. First of all,
Grassi and C. Parona, using systematic anatomical research, noticed that the
presence of the worm in the human intestine is always accompanied by the
presence of eggs at various stages of segmentation up to the morula. "Eggs in
the process of segmentation (yet never beyond the morula stage) ... " could
be found even in fresh feces and vomit just passed by the carrier, and these
eggs could be easily distinguished from the eggs of other parasitic helminthes.
Once outside the human body, the eggs continue to develop, and the two
researchers observed embryo formation, larva hatching and "skin changing at
least twice." It then goes into another larva that goes through another moult.
To complete their research, Grassi swallowed - with impunity - "a big pill
full of hookworm eggs in the segmentation stage."
Grassi's and C. Parona's researches are fundamental both from a biological
 DISCOVERY OF DUODENAL ANCYLOSTOMA 269

and clinical point of view, as they indicated, in the microscope examination

of feces, the way to recognize the wonn's presence in living man. 34 Grassi
immediately set to work exploiting this fact by working with Ernesto Paron a
(1849-1902),35 brother of Corrado, and who was assistent head of the
Department of Internal Medicine of Pavia University, under the direction of
Francesco Orsi (1828-1900).
Thanks to stool examination, the two young research workers succeeded
in detecting the wonn's presence in some hospitalized "cases of serious
anemia" with "obscure and inadequate" etiology. They also observed that the
adult worm, while nonnally not found in feces of affected patients, is
expelled upon administration of an anthelmintic drug, i.e., a powder consist-
ing of santonin, calomel and resin of jalap thanks to which they considered
"the hope of obtaining an effective treatment not so far-fetched."
One particularly instructive case concerning a twenty-two year old woman
is worth reporting in full, following the version given by Filippo Silvestri36 in
1925, and probably inspired by Grassi himself.
As a student in his final year of Medicine, he (Grassi) was attending the internal medicine
course given by Professor Orsi from Pavia. There was a very seriously anemic woman
among the patients hospitalized in the clinic. Professor Orsi, a man of remarkable clinical
intuition, would often stop in front of this patient's bed with his students. According to
the Professor, the patient was suffering from an incurable anemia, the cause of which he
was unable to identify. As a consequence of this diagnosis, Grassi thought that it might
be a case of ancylostomiasis. He examined the patient's feces and discovered a consider-
able number of hookworm eggs, but he kept the discovery to himself. Some days later,
while Professor Orsi was doing the rounds, he again stopped in front of the patient's bed
and again called his students' attention to her case. Having learnt that Grassi had exam-
ined her feces, he asked him ironically if "he himself had found the cause for the anemia
by peering through the hole" (jokingly referring to the microscope eyepiece). To the
astonishment of the listeners, Grassi smiled and answered modestly that he had, in fact,
found the origin and that it was due to hookworm. Orsi then remembered having seen an
Egyptian woman who had suffered from serious anemia and that he had supposed that
the anemia be due to hookworm. However, he did not really believe Grassi. Nonetheless,
he allowed him to administer a vermifuge to the woman, and a certain number of hook-
worms were expelled, but the treatment was started too late and after some days the
patient died. The autopsy confumed beyond doubt that death was due to Ancy-
lostomiasis. This led Orsi to exclaiming that silly women are right in treating any disease
as if it were caused by worms.

So it would appear that, according to this renowned physician, any hel-

minthic etiology of disease was "a matter of silly women", just as Bucellati's
system had been, which, as it had fallen into disrepute, probably had an
inhibiting effect on Dubini.
270 LUIGI BELLONI

So, "Egyptian chlorosis occurs among humans in Italy too, seeing as the
hookworm sucks blood in our country as well", Grassi stated in a paper dated
the following November in 1878.37
His work with the Parona brothers had been essential to this end and had
to be improved upon by therapeutic evidence which would be the only proof
to confirm the cause and etiological relationship between hookworm and
anemia. Anemia had to be treated by administering an anthelmintic drug
capable of ridding the patient of hookworm. In the hands of Giovan Cosimo
Bonomo (1687)38 topical treatment, aimed at killing the acarus played an
important role in asserting the acarus-based etiology of scabies. 39

The search for hookworm carriers in Turin was started by Bonaventura

Graziadei (1852-1935),40 who practiced copromicroscopical tests on the
patients hospitalized in a number of hospital and medical school wards and,
in particular, on patients admitted to the two medical clinics - general and
propedeutical - directed by Luigi Concato (1825-1882) and Camillo
Bozzolo (1845-1920), respectively.
In June 1879, Graziadei detected two cases, both of whom were oven
attendants at a brick-makers, just as many of the cases examined by Grassi
and E. Parona had been. These two gave Bozzolo the spur to conclude his
introductory clinical medicine course with a lecture on L 'anchilostomiasi e
l'anemia che ne conseguita (anchilostomoanemia).41 In this lecture, besides
emphasizing the importance of the work carried out by Grassi and the Paron a
brothers, Bozzolo also put special emphasis on a laboratory examination that
was later to prove very valuable in evaluating the degree of anemia in its oscil-
lations. This test, in fact, consisted in determining the index of hemoglobin in
a blood sample by means of a chromocytometer which had just been set up
by Giulio Bizzozero (1846-1901).42 The attempts at therapy also failed in
Bozzolo's two cases: " ... we know no means capable of disabling, killing and
expelling hookworm and in our clinic, we have tried almost every anthelmintic
drug in vain." In addition to Graziadei as copromicroscopist, and Ignazio
Fenoglio, the chromocytometrist, we fmd that Bozzolo also mentioned in his
lecture Edoardo Perroncito (1847-1936), a pathologist at the School of
Veterinary Medicine of Turin, who contributed in measuring hookworm eggs.
By now the ground was cleared for solving the St. Gothard Tunnel patho-
logy riddle that was gradually, though tragically, taking shape.4 3 Tunnel
excavations, which had started back in 1872, were about to be brought to an
end. However, there was talk that an endemic, cachectic disease had broken
out among the workers, which, from distant times, had been considered a sad
 DISCOVERY OF DUODENAL ANCYLOSTOMA 271

attribute of mine workers: the so-called 'miner's cachexy'. Disabled for work,
many sick miners returned home and went to their local hospitals. Cachectic
veteran miners from the tunnel were also hospitalized in the general internal
medicine clinic of Turin directed by Concato. Subjected to copromicroscope
examination by Perroncito,44 they were shown to be hookworm carriers.
This clinical diagnosis was confirmed by a post-mortem fmding in
February 1882. The case was a severely cachectic miner who had died quickly
as a result of serofibrinous peritonitis follOwing an intraperitoneal transfusion
of defibrinated blood,45 attempted 'in extremis.' During the course of the
autopsy, solemnly performed by Professor of Pathological Anatomy,
Francesco-Vittorio Colomiatti (1848-1883), in the presence of a large and
authoritative audience,
more than 1500 hookworms were found in the duodenum and jejunum, many of whom
exhibited their hematophagic properties very clearly. Their intestines contained various
quantities of very well-preserved erythrocytes [mature red blood cell].

While political newspapers were announcing the fatal epidemic to the

public, provoking questions in Parliament, and relief measures from the
government, a very interesting controversy developed on a scientific level.
Encouraged by the first success, Perroncito extended his research, becoming
more and more convinced that the St. Gothard anemia was chiefly due to
worm infestation and to hookworm first and foremost. (He also admitted the
varying presence of Anguillula intestinalis and stercoralis discovered by
Normand in Cochin China [south-east Asia] in 1877 and again detected by
Grassi and the Parona brothers in Pavia in 1878.)
Perroncito carried out further basic in vitro research on hookworm larval
metamorphosis and the ideal conditions for its larval development. The
worm's larvae need a period of a free life in the expelled feces and the ideal
surroundings from this point of view (moist ground!) explain the endemic
spread of the worm within certain human groups (farmers, brick-makers,
miners) and therefore point out the path to prevention which was to be
taken.
Perroncito also tested the action of various anthelmintics, on hookworm
larvae raised in vitro and opted for the male fern plant which had already
been known for some time as a teniafuge. This drug was in fact tested on
miners hospitalized in Concato's clinic, and other research workers also
followed suit. Foremost among these was Ernesto Parona who, after some
initial doubts and failures, eventually worked out the ideal treatment and
tested it - as we shall see - on a very large number of cases.
272 LUIGI BELLONI

Thanks to Bozzolo, thymol46 was also identified as another specific

an thelrnintic.
The solution of the therapeutical problem at the same time marked the
victory of the Perroncito-supported contag;onist doctrine, over the localist
theory to which on the other hand Bozzolo and hygienist Luigi Pagliani
(1847-1932) adhered. After having brought their research into the heart of
the mines - on the spot, as it were - these two put forward their conclusions
during a famous meeting of the Reale Societa Italiana di Igiene, in Milan on
10 April 1880.47
Obviously, the two authors had found hookworm in cachectic miners, but,
instead of considering the disease as attributable entirely to that fact, they
considered the worm as a complementary disease-causing factor "that con-
siderably aggravated the already sad conditions inside the tunnel."
To facilitate understanding of the Bozzolo-Pagliani position, I should like
to recall, albeit briefly, that Robert Koch's (1843-1910) work on the cholera
parasite was about to be announced (1884) and that it was just this work that
enabled the contagionist doctrine to prevail for this disease rather than the
localists', as sustained by Max Pettenkofer (1818-1901), father of modern
hygiene. According to Pettenkofer, cholera - a disease which had decimated
Europe in successive waves since the 1830's - depended on a local chrono-
logically determined disposition, in other words, on a whole array of environ-
mental factors that converged in certain times and places only to produce a
cholerigenous effect on the population (epi-demia). According to Petten-
kofer, it was up to new hygienics to exploit the methods of chemistry and
physics in order to quantify these various environmental factors, soil and
underground water features and so on, in short to evaluate The Air, the Water
and the Places, just to recall the title of the famous Hippocrates treatise.
A whole series of undoubtedly abnormal environmental conditions con-
verged deep inside the tunnel: worker concentration, poor ventilation, the
stagnation of mine gas, and exploded dynamite powder in particular, the
heaping up of excrementa, lack of light, rise in temperature, and so on. Each
of the factors got worse as the work of the tunnel proceeded, thereby
explaining the concern Bozzolo and Pagliani felt when faced with such
conditions that have only occurred in recent years, this being one of the first times man
has gone so deeply into the innermost bowels of the earth, bearing his demolishing bit
with titanic courage into the very core of the rock, more than 7000 meters inwards.

To a certain extent their concern reflects our own for the partially unknown,
environmental conditions, to which astronauts are subjected today.
 DISCOVERY OF DUODENAL ANCYLOSTOMA 273

The controversy over contagionism or localism was soon settled by the

brilliant results obtained with specific anthelmitic therapy, that was soon to
be made available with two highly effective drugs: male fern (perroncito) and
thymol (Bozzolo). The disease does neither improve nor disappear by re-
moving a cachectic worker from the tunnel environment: this only occurs if
he is freed from hookworm. The process of recovery could be documented in
numerical terms as well, thanks to hemoglobin quantification with Bizzo-
zero's chromocytometer.
Ernesto Parona - whom we left sometime back intent on working with
Grassi at the Internal Medicine Clinic of Pavia - was called upon to direct the
Municipal Hospital of Varese in November 1879. He kept this post until
September 1885, during which time he applied his pioneer experiences in
order to consolidate the helminthic etiology of St. Gothard anemia and work
out a therapy for it, inasmuch as that the hospital he directed was among
those chosen to treat cachectic ex-tunnelers at the government's expense. In
his Relazione intomo alla cura dei minatori del Gottardo occolti a carico del
R. Govemo nel Civico Ospedale di Varese, recorded in August 1884 and
printed in Varese in 1885, Parona was in a position to supply case histories 48
that were also impeccable from a laboratory point of view. They included
249 formerly cachectic tunnelers who had been discharged fully cured thanks
to the anthelmintic treatment, based primarily on male fern.
Once the St. Gothard Tunnel anemia problem - together with its relative
prophylactic and therapeutic implications - had been solved, the more wide-
spread problem of anemia cases among miners working in permanent mines
came to the fore. In the localism versus contagionism controversy, the latter
won. Yet, it was not a quick victory and after the St. Gothard Tunnel case, it
had to prevail mine by mine. Among the other early victories, was that of the
Lercara Friddi (province of Palermo) sulfur mines, thanks to the efforts of
Alfonso Giordano (1842-1915),49 who, having detected the presence of
hookworm in anemic miners,so was able to produce some cases recovered
under therapy of male fernS! alone as early as February 1882.
In this way the apostle of Sicilian miners takes the lead among those who
gave a humanitarian and social tone to one of science's conquests. The
enormous practical value of all this was fully recognized about twenty years
later when, having learned from the St. Gothard tragedy, Man could celebrate
a great new victory over nature in digging the Simp Ion Tunnel, without
having to pay any toll whatsoever to hookworm infestation. This time it was
the Piedmontese physician, Giuseppe Volante (1870-1936) who led the
battle by virtue of his role as 'Enterprise Physician - Southside.' The Simplon
274 LUIGI BELLONI

Tunnel inauguration was celebrated in Milan in 1906, with an imposing series

of ceremonies, culminating with the First International Congress on Occupa-
tional Diseases. Thanks to Luigi Devoto (1864-1936) the first clinic for
occupational diseases was soon to be set up. Inaugurated on 20 March 1910,
it secured a leading position for the same town where Angelo Dubini had first
discovered the duodenal Ancylostoma in 1838.

Translated by
R. MABERRY and A. TALIERCIO

NOTES

* All medical definitions have been taken from Stedman's Medical Dictionary, 22nd ed.
(Williams and Wilkins, Baltimore, 1972).
1 L. Belloni, 'La scoperta dell'Ankylostoma duodenale', Gesnerus 19 (1962), 101-118;
'La medicina a Milano dal Settecento aI1915', in Stona di Milano (Fondazione Treccani
degli Alfieri, Milan, 1962), Vol. 16, pp. 991-997. The exhaustive Bibliography of Hook-
worm Disease (Rockefeller Foundation, International Health Board, New York, 1922)
is particularly worthy of mention.
2 A. Cazzaniga, La grande crisi della medicina italiana nel primo Ottocento (Hoepli,
Milano, 1951), cf. p. 122. This important historical study was also issued in instalments
in Vols. 4, 5 and 6 (1948-50) of Castalia (Milan).
3 E. H. Ackerknecht, Medicine at the Paris Hospital 1794-1848 (Johns Hopkins Uni-
versity Press, Baltimore, 1967).
4 L. Belloni, 'Per la storia del cuore tigrato. I contributi portati nell'Ospedale Maggiore
da A. Gambarini (1838) e A. Dubini (1843-44)', L 'Ospedale Maggiore 44 (1956),
252-8.
5 In Gazz. med. di Milano 6 (1847), 319-22, 328-31 and 388. Shortly after its publi-
cation, this paper was published both as an abstract of 30 pages (Milan 1847), and in a
"second edition with many additions and remarks" in Enciclopedia anatomica (G.
Antonelli, Venice) together with a reprint of Trattato di antropotomla.
6 'Nuovo verme intestinale umano (Agchylostoma duodenale), costituente un sesto
genere di Nematoidei proprii dell'uomo',Ann. universali di med. 106 (1843), 5-l3.
'Notizia di un nuovo verme degli intestini umani costituente un settimo genere di
Nematoidei proprii dell'uomo', Gazz. med. di Milano 2 (1843), 153-54.
7 Dubini is referring to studies carried out by Paris writers on tuberculosis, abdominal
typhus (typhoid fever) and intestinal ulcers of the two diseases. See Ackerknecht's work
mentioned in Note 3.
8 'Alcune avvertenze di anatomia patologica medica', Ann. universali di med. 113
(1845), 268-71.
9 p. 109. See Note 11.
10 A. Dubini, 'Abito cereo: palpitazioni di cuore con rumore di soffio perisistolico:
tensione e dolore all'epigastrio: lingua pallida, liscia e asciutta: poi subdelirio, stupore,
e nell'ultimo giorno 6 respirazioni per minuto: battendo ancora il polso 78 volte nello
 DISCOVERY OF DUODENAL ANCYLOSTOMA 275

stesso spazio di tempo. Un centinaio di agchilostomi nell'intestino duodeno e nel

digiuno, screziatura gialla della sostanza del cuore, edema dei polmoni', Gazz. med. di
Milano 2 (1843), 345-7.
[Pale in color; heart palpitations with perisystolic murmur; tension and pain in the
epigastrium; tongue pale, smooth and dry; then subdelirium, stupor, and on the last day
six breaths per minute, pulse rate 78 beats per minute. About 100 hookworms in the
intestinal duodenum and jejunum, yellow speckling in the heart structures, pulmonary
edema].
11 Entozoografia umana per servire di complemento agli studj d'anatomia patologica
con tavole XV in rame rappresentanti i vermi proprii dell'uomo Molti de' quali disegnati
dal vero, ingranditi e notomizzati, ed altri tolti dalle opere di Treutler, Werner, Brera,
Rudolphi, Bremser, H. e G. Cloquet, Vogel, Schmalz, Eschricht, Curling, E. Wilson,
Owen, Leblond, Lebert, Dujardin, Delle Chiaje, Blanchard, ecc. Nella Quale trovasi
classificato un nuovo genere di vermi intestinali proprii dell 'uomo, non ancora descritto
nei trattati elmintologici. Seguita da un 'appendice sui parassiti esterni del corpo umano
tanto animali che vegetabili parimenti rappresentati con tavole del dottor Angelo Dubini.
Opera alia Quale venne aggiudicato il premio Dell'Acqua per l'anno 1848 della Com-
missione a cio eletta nell 'Ospedale Maggiore di Milano. Edizione riveduta dall'Autore.
Milan, Societa degJi editori degJi Annali universali delle scienze e dell' industria, 1850.
[Human endozoography to complement studies in pathological anatomy with fifteen
copperplates representing worms peculiar to man. Many of these are drawn from life,
enlarged and dissected, and others are taken from the works of Treutler, Werner, Brera,
Rudolphi, Bremser, H. and G. Cloquet, Vogel, Schmalz, Eschricht, Curling, E. Wilson,
Owen, Leblond, Lebert, Dujardin, Delle Chiaje, Blanchard, etc. A new genus of intestinal
worms peculiar to man is classified here, one not yet described in other helmintic
treatises. This is followed by an appendix on plant and animal parasites outside the
human body represented by Dr. Angelo Dubini's plates. This work was awarded the
Premio dell'Acqua for 1848 by the specially elected Commission of the Ospedale
Maggiore of Milan. Edition revised by the author.] Entozoografia was issued in instal-
ments in Ann. universali di med. 131 (1849), pp. 502-78; 132 (1849), pp. 5-119, 449-
524; 133 (1850), pp. 5-92, 225-395. [This is the full title of Dubini's work. - C.R.F.]
12 Dedicated by the Author "Ad Angelo Dubini, illustre e fortunato cultore della paras-
sitologia." [To Angelo Dubini, famous and successful expert on parasitology.]
13 E.H. Ackerknecht, 'Anticontagionism between 1821 and 1867', Bull. of the History
of Med. 22 (1948),562-593.
14 'Sulla natura vegetabile della Tigna vera 0 favosa', Gazzetta medica 1 (1842),65-8.
15 L. Belloni, Documenti Bassiani, Milan 1956. - 'La scoperta di Agostino Bassi nella
storia del contagio vivo', in Actes du VI/~ Congres International d'Histoire des Sciences
(Florence-Milan, 3-9 September 1956), Gruppo italiano di storia delle scienze, pp. 897-
909. - 'Agostino Bassi nella storia del contagio vivo', in Actes du Symposium Inter-
national d'Histoire des Sciences (Turin, 28-30 July, 1961), Florence, Gruppo italiano di
storia delle scienze, 1964, pp. 54-63. - 'La scoperta di Agostino Bassi e i suoi primi
sviluppi in patologia umana (dermatomicosi)" in Atti del XLI Congr. Naz. d. Soc. Ital. di
dermatologia e sifilografia (Milan, 29 September 1956) and Minerva Dermatologica 31
(1956),544-8. - Furthermore, I refer the reader to the special issue, 'Studi su A. Bassi'
(1956) of the Archivio storico Lodigiano, despite its numerous misprints which are not
the fault of the authors.
276 LUIGI BELLONI

16 1. Castaldi, 'II contributo di Angelo Dubini aile conoscenze causali della scabbia e
l'importanza della sua Entozoografia per la Storia della microbiologia', Riv. di storia d.
scienze med. e natur. 28 (1937), 204-8.
17 A. Dubini, Entozoografia umana (Milano, 1850), p. 390.
18 E. Benassi, 'Un medico sedicente novatore del primo Ottocento: il Dott. Bucellati e i
suoi successi terapeutici', in Minerva Med. 42 (1951) 220-225 of the 'Parte varia.'
19 A. Dubini, Entozoografia umana (Milano, 1850), p. 277.
20 'Bericht liber die Leistungen im Gebiete der Helminthologie wahrend des Jahres 1843
und 1844', Archiv for Naturgeschichte 2 (1845), 202-255; see especially pp. 220-221.
21 Die Krankheiten des Orient's vom Standpunkte der vergleichenden Nosologie be-
trachtet von Dr. F. Pruner (Palm and Enke, Erlangen, 1847), pp. 244-245: "Selten
offnet man eine Leiche in Aegypten ohne Individuen von einer oder auch mehreren
Arten zu finden. Der Ascaris lumbr., Ascaris und Oxyuris vermicularis und Trichoceph-
alus dispar finden sich zu Haufen schon in den Gediirmen der Kinder. Unter den
Erwachsenen sind es besonders die kachektischen, wassersiichtigen und skrofulosen
Subjekte, we1che ausserdem an Anchylostoma duodenale im Zwolffingerdarme leiden,
wo dieser Parasit seinen vierfacherigen Saugeriissel mit 40 Haken an die Schleimhaut
heftet, wodurch er Ecchymosen veranlasst." [One seldom opens a corpse in Egypt
without finding worms of one or more kinds: ascaris lumbr., ascaris, oxyuris vermicularis
and trichocephalus dispar are found in swarms even in the intestines of children. In
adults it is especially the cachectic, dropsical and scrofulous who suffer from ancy-
c1ostoma duodenale in the duodenum where this parasite sticks its four-sided proboscis
with forty hooks into the mucous membrane where it causes ecchymosis.]
22 'Klinische und anatomische Beobachtungen uber die Krankheiten in Egypten', in
Archiv for physiologische Heilkunde, issued in many instalments, see particularly Vol.
13 (1854), pp. 555-561: 'Anchylostomenkrankheit und Chlorose.' - Please refer also
to Griesinger's own short communication mentioned in Note 27.
23 'Ein Beitrag zur Helminthographia humana, aus brieflichen Mittheilungen des Dr.
Bilharz in Cairo, nebst Bemerkungen von Prof. C. Th. v. Siebold in Breslau', Zeit. t
wissensch. Zoologie 4 (1852), 53-76 and plate V. See also p. 56: "Die Stelle, wo ein
solcher Strongylus sass, ist durch eine linsengrosse Ecchymose bezeichnet, in deren Mitte
ein weisser Fleck von Stecknadelgrosse bemerkbar. Dieser weisse Fleck ist in der Mitte
durchbohrt von einem nadeldicken bis in das submucose Bindegewebe dringenden
Loche. Manchmal zeigt die Schleimhaut flache Erhabenheiten von Linsengrosse und livid
braunrother Farbe, welche eine zwischen Tunica mucosa und muscularis im Bindegewebe
befindliche, mit Blut geflillte Hohle und darin zusammengeringelt den lebenden von Blut
voll gesogenen Wurm (bald ein Mannchen, bald ein Weibchen) enthalten." [The place
where such a strongyle was fixed is indicated by a lentil-sized ecchymosis in whose center
a white spot the size of a pinhead can be seen. This white spot is pierced in the middle by
a hole the thickness of a needle, which penetrates into the submucous connective tissue.
Sometimes the mucous membrane shows smooth protuberances the size of a lentil and a
livid reddish brown color where there is a blood-filled cavity (between the mucous tunic
and the muscle of the connective tissue) containing a curled-up living worm (sometimes
male, sometimes female) sucked full of blood.]
24 G. Sangalli, 'Geografia elmintologica: Anchilostoma e trichina', Giomale di anatomia
e fisiologia pat%gica 3 (1866), 100-106.
25 "More often than in the duodenum, I would find it in the first few loops of the
 DISCOVER Y OF DUODENAL ANCYLOSTOMA 277

jejunum, where it would lie stuck to the mucous membrane by means of a thick and
copious greyish mucus, that is always to be found in helminthiasis of this kind." The
worm would be simply stuck to the mucous membrane and not fixed to it by its perioral
hooks, as Bilharz found in Egyptian chlorotic patients. "Bilharz attributes the chlorosis
to the sad effects of this hemorrhage, chlorosis being very frequent in Egypt. In our
countries this phenomenon due to the effects of hookworm has not yet been observed,
but I was able to convince myself of such a possibility. In fact the worm sometimes
adheres loosely to the mucous membrane with its mouth; and sometimes, on removing
it, a reddish spot is observable. This is due to the blood spreading in the mucous
membrane provoked by the worm's biting. Once I found a large ecchymosis in the
mucous membrane and the hookworm firmly flxed in the middle of it by means of its
mouth."
26 G. Sangalli, 'Sopra alcuni punti controversi di elmintologia' Mem. d. R. lstit.
Lombardo (Classe di scienze matem. e natur.) 13 (1876), 349-61: " ... I am not able to
give a reasonable explanation for the important fact that the worm, equipped with a
four-hooked mouth both in Egypt and in Italy, reveals in Egypt its fatal property (where
it is widespread) by thrusting its head into the mucous membrane of the intestine where
it lives, thereby causing hemorrhages, and consequently serious illnesses, while, in our
country, it only exceptionally exhibits a similar behavior and is, therefore, virtually
harmless." At this point, it may be appropriate to point out that, when using the term
'virtually harmless', Sangalli meant the case of a seriously anemic, 57 year old male
subject who, once autopsied, revealed 700 hookworms without any other detectable
recent or past perforation of the intestinal mucous membrane: "Yet, I admit that the
anemic condition exhibited by that corpse, and therefore the cause of the patient's
death, would not have been clear to me, had I not admitted some fatal influence of the
worms on the intestine, hence faulty chylification and hematopoiesis." In some rare
cases, then, among the many in which it is hosted in the human organism, the hookworm
has an anemia-producing action, that could, however, be ascribed to altered intestinal
absorption, rather than to a blood~ucking mechanism. We know that the longer the time
gap between patient death and corpse dissection, the more numerous are the worms
which detach themselves from the mucous membrane and are free in the intestine lumen.
On the basis of this observation, we can now explain a phenomenon that at that time
was construed as a difference in hematophagic aggressivity between the Italian hook-
worm and the Egyptian one.
See also Sangalli's 'Annotazioni critiche sull'anchilostoma duodenale', Rendiconti d.
R. lstit. Lombardo 11 (1878),460-467.
27 'Ueber die Anchylostomenkrankheit, tropische Chlorose oder tropische Hypoamie',
Deutsches Archiv fill' klinische Medicin 10 (1872), 379-400. - The importance of
Wucherer's research had already been highlighted by W. Griesinger, 'Das Wesen der tropi-
schen Chlorose', Al'chiv der Heilkunde 7 (1866), 381.
28 P. Sonsino, 'L'Anchilostoma duodenale in relazione coll'anemia progressiva per-
niciosa',Imparziale [Florence] 18 (1878),227-234.
29 C. Morelli, 'Intorno ad un caso d'anemia progressiva con Anchilostoma duodenale',
Lo Sperimentale 41 (1878),27-39.
30 R. Leuckart, Die mensch lichen Parasiten und die von ihnen hel'iUhrenden Kl'ank-
heiten. 2 vols. (Winter, Leipzig-Heidelberg, 1876), Vol. 2, pp. 410-460 and 480.
31 In honor of Giuseppe Balsamo-Crivelli (1800-1874), Professor of Natural History
278 LUIGI BELLONI

and subsequently of Zoology and Comparative Anatomy at Pavia University. Balsamo-

Crivelli continued the studies of A. Bassi on the cryptogam that caused muscardine
disease in silkworms and identified it in the Botrytis genus as a new species, which he
had formerly named paradoxa and subsequently Bassiana.
32 'Di una nuova specie di Dochmius (Dochmius Balsami). Nota del dott. Corrado
Parona e di Battista Grassi, studente del 5~ anno di medicina; presentata dal M.E.
professor Giacomo Sangalli', Rendiconti d. R. Istit. Lombardo 10 (1877), 190-5 and
one plate (meeting of March 15, 1877). - L. Maggi, 'Sugli studj di C. Parona e G.B.
Grassi intorno all'Anchilostoma duodenale Dub, Rendiconti d. R. Istit. Lombardo 11
(1878),428-36. Observations of Prof. P. Pavesi are set forth on pp. 436-438 and lastly
on pp. 438-9: G. Sangalli, 'Contro osservazioni alle osservazioni del Prof. P. Pavesi alia
lettura del professor Maggi.'
33 'Intorno all'anchilostoma duodenale (Dubini). Annotazioni di Battista Grassi,
laureando in medicina, Corrado dott. Parona e Ernesto dott. Paron a', Gazz. med.Ital. -
Lombardia 5 d.s. VII (1878), pp. 193-6.
'Intorno all'anchilostomiasi. Osservazioni dei dottori Grassi Battista e Parona Ernesto,
con un Appendice embriologica, dello stesso dott. Grassi e dott. Corrado Parona', Ann.
universali di med. e chir, 247 (1879),407 -24 and one plate.
34 "Diagnosing ancylostoma is very easy; to do so rapidly, it is sufficient to observe a
stool or vomit sample diluted with any menstrua under at least a 90-x microscopic
magnification. When the sample matter is fresh, hookworm eggs appear just at the stage
of segmentation. When the sample is stale, also embryos and larvae appear."
35 Dr. Ernesto Paron a was the author of Discorso Gracco·Maroni·Brocca. Inaugurando
Ricordo Perenne nell'Ospedale Fate·Bene·Fratelli. Milano 30 Maggio 1904. Novara, 1905.
36 In the miscellaneous booklet, Discorso (Onoranze a Battista Grassi), (Bardi, Rome,
1925), pp. 33-34.
37 'Intorno ad una nuova malattia del gatto, analoga alla clorosi d'Egitto (anemia da
anchilostomi) dell'uomo', Gazz. med. Ital. - Lombardia, 5 d.s. VII (1878), pp. 451-454.
38 1. Belloni, 'La medicazione topica nella scoperta della etiologia acarica della scabbia',
Simposi Clinici, Vol. I, 1964, fasc. 3, pp. XXI-XXVI.
39 This represented a perspicuous model which the contagionists frequently referred to.
So, for instance, Bilharz asks himself "ob unser Distomum Haematobium zur Dysenterie
in derselben Beziehung steht, wie Acarus scabiei zur Kriitze" [whether our distomum
hematobium (obsolete term for trematode worms that are parasitic in the blood) stands
in the same relation to dysentery as acarus scabiei (synonym for sarcoptes scabiei, or
mange mite) does to scabies]. See p. 76 of the work mentioned in Note 23.
40 N. Valobra, 'Bonaventura Graziadei (1852-1935)', Giornale della R. Accademia di
Medicina di Torino (Parte II: Memorie originali) 49 (1936), 24-35.
41 In Giornale internazionale delle scienze mediche 1 (1879), 1054-1069 and 1245-
1253.
42 G. Bizzozero, '11 cromo-citometro. Nuovo strumento per dosare l'emoglobina del
sangue', Afti d. R. Accademia delle Scienze di Torino 14 (1878/9), 899-942 and plate
XII. Bizzozero's paper was presented by M. Lessona at the May 11, 1879, meeting.
43 1. Belloni, 'L'anemia del Gottardo', Gesnerus 29 (1972),33-44.
44 E. Perroncito, 'Osservazioni elmintologiche relative alla malattia sviluppatasi
endemica negli operai del Gottardo', in R. Accad. dei Lincei. Mem. d. Classe di scienze
fisiche, matem. e natur. 7 (1880), 1-55 and plate. - La malattia dei minatori dal S.
 DISCOVERY OF DUODENAL ANCYLOSTOMA 279

Gottardo al Sempione. Una questione risolta (C. Pasta, Turin 1909). - La maladie des
mineurs du St. Gothard au Simp/on. Une question resolue. Appendice (Etablissement
typographique nationale, Turin, 1912).
45 Lava, 'La trasfusione del sangue per la cavita del peritoneo', in L 'osservatore. Gazzetta
delle Cliniche. Organo utI d. Soc. di med. e chir. di Torino 16 (1880),81-82.
46 Many case histories in favor of thymol therapy were produced by B. Graziadei, 'II
Timolo nella cura dell'anchilostomanemia', Gior. d. R. Accad. di medicina di Torino
30 (1882), 821-855.
47 'L'anemia al traforo del Gottardo dal punto di vista igienico e clinico', Giorn. d. Soc.
Ital. d'Igiene 2 (1880), 276-346.
48 For the first cases studied by E. Parona in the Varese Hospital, see his 'L'anchilo-
stomiasi e la malattia dei minatori del Gottardo', Annali universali di Med. e chir. 253
(1880), 177 -202 and 464.
49 On the fiftieth anniversary of his death, Giordano was commemorated on June 28,
1965 at Palermo University by Giovanni Frada and Giuseppe Pontieri. See Folia medica
48 (1965), fasc. 10.
so E. Parona, 'L'anchilostomiasi nelle zolfare di Sicilia', Annali universali di med. e chir.,
277 (1886), 464-468.
51 From Giorn. d. R. Accad. di medicina di Torino 30 (1882), 90: "On behalf of Dr.
Giordano of Lercara, Perroncito wishes to communicate that Giordano has found hook-
worms in the sulfur mine workers suffering from anemia and that upon subjecting the
hookworm to male fern treatment, the patient recovers from the anemia." (From the
minutes of the meeting of 24 February 1882.)
 SALVO D'AGOSTINO

WEBER AND MAXWELL ON THE DISCOVERY OF THE

VELOCITY OF LIGHT IN NINETEENTH CENTURY
ELECTRODYNAMICS

I. INTRODUCTION

In 1846 Wilhelm Weber (1804-1891) discovered that a characteristic velocity,

almost equal to the velocity of light, was a significant constant in the theory
of electrodynamics. James Clerk Maxwell (1831-1874) accepted this dis-
covery within his theory of the electromagnetic field (1861-62) and used it
as important evidence for his electromagnetic theory of light. This is interest-
ing for a number of reasons.
The fact that this same discovery has been accepted and highly esteemed
within two theories as different as Weber's action-at-a-distance electrody-
namics on the one hand and Maxwell's theory of the electromagnetic field
on the other, presents a problem regarding the relationship between the
empirical contents of this discovery and the value attributed to it within each
respective theory. That these two theories - so different as to be considered
antagonistic to each other - are based upon the very same discovery and
both, therefore, accept that set of instrumental operations, rules and laws
leading up to that discovery, implies that the two theories defmitely have
some ground in common.
K. F. Schaffner in his 'Logic of Comparative Theory Evaluation' has
recently proposed 1 an explanation of the fact that some experimental results
"seem to have a common element from any of the competing theories' point
of view". The latter point arouses a certain interest, among others, in connec-
tion with a recent, particularly important, thesis in the field of philosophy
of science: Kuhn's well-known tenet on paradigmatical leaps and on the
incommensurability of paradigms. The fact that the aforementioned two
antagonistic theories have something fundamental and important in common
(which may be called 'an overlap') may somehow question the validity of
Kuhn's conception at least in this particular case.
Furthermore another interesting aspect regarding this historical event of
Weber's discovery is how it may contribute to change the usual concept of
'discovery' within the more abstract levels of physics theory. In physics, as
in other scientific diSciplines, 'discovery' usually tends to be related to the
idea of 'something new'. In addition, what is new is often connected with
281
MD. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 281-293.
Copyright © 1980 by D. Reidel Publishing Company.
282 SALVO D'AGOSTINO

hard facts. In other words, a new phenomenon or a new effect is often

'discovered'. However, this characteristic of newness should be changed when
applied to the event dealt with in this paper.
This aspect of newness at such a sophisticated level as that of Weber's
discovery should refer to a 'research program' (as defined by Lakatos) and
should be interpreted within its context rather than in isolation. Should this
not occur, the element of newness would be lost. In fact, with regard to the
historical event discussed in this paper, it would seem more appropriate to
underline the feature of 'relevance levels' rather than that of newness or
unexpectedness.
Therefore, by not emphasizing the unexpectedness of the discovery,
another feature becomes more significant, i.e. the importance of the scale of
scientific values that belong to a 'research program' rather than merely being
produced by a single theory.
In conclusion, the importance of this discovery should not be related to a
specific theory (since it is still valid even when the initial theory within which
it originated has been modified) but to the metrological research program
which flourished in physics in the mid-1800's.

II. WEBER'S DISCOVERY OF A CHARACTERISTIC VELOCITY

IN HIS THEORY OF A FUNDAMENTAL LAW OF
ELECTRIC FORCE

In 1846 Weber formulated a theory which was particularly influential in

German electrodynamics throughout the second half of the nineteenth cen-
tury.2 This theory soon conflicted with other electrodynamic theories and
particularly with those of Helmholtz related to the acceptance of Weber's
Law of electric force. However, this theory was well accepted within scientific
circles until almost the end of the century.
Weber's researches were inspired by the ideas of Andre Marie Ampere, who
made the reduction of magnetism into galvanism possible by postulating that
any magnetic effect could be simulated by a suitable arrangement of currents.
This challenging idea, when applied to natural magnets or lodestones, clashed
with the difficulty of accepting that currents in the microscopic world could
flow continuously without dissipation of heat.
The evolution of Ampere's ideas led him to consider magnetic forces as a
modification of electric forces due to the velocity of electric charges. How-
ever, he did not develop the theory, and left it for future researchers. Weber's
'Electrodynamische Massbestimmungen', a long and detailed study on the
 DISCOVERY OF THE VELOCITY OF LIGHT 283

theory of interactions between moving charges, represents the completion of

Ampere's ideas.
The focal point of Weber's theory was his law of force between electric
particles, which he called the 'Fundamental Law of Electric Forces'. By this
law he succeeded in explaining all the electric and magnetic phenomena
known at that time, which ranged from electrolysis to forces between cur-
rents, up to electromagnetic induction discovered by Faraday in 1831.
Weber's Law of Force 3 was a function of the relative velocity and of the
acceleration of the particles and also contained a non-reducible constant
(henceforth indicated as Cw). This constant, in Weber's opinion, was a velo-
city to which he attributed a particular importance. According to Weber's
law, in fact, should electric particles move at this velocity, any electrodynamic
effect between electric currents would cease (in this sense Cw was a 'limit'
velocity). Should higher velocities be reached, the forces would reverse their
signs (in this sense Cw was also a 'critical' velocity).
Before Weber's discovery of Cw no physicist (with the exception of Karl
Friedrich Gauss) had ever suspected that such a characteristic velocity might
be linked with electricity. This is understandable if we consider that previously
only accelerations and not velocities had been accepted as theoretically mean-
ingful quantities in classical mechanics, hence also in physics. It was in this
context that Weber's discovery represented something new.
Weber's law was not directly verifiable. From this point of view this law
was even less verifiable than Ampere's law, which, being a law of force be-
tween 'current elements', represented a differential law that had no empirical
meaning unless integrated twice. 4 In reality Weber had derived his Law of
Force from Ampere's Law under certain conditions. The conditions under
which Weber's law was consistent with Ampere's law were Theodore Fechner's
hypotheses (1845) that electric currents consisted of a double motion of
particles of both signs moving in opposite directions along the same wire. 5
In this respect, Weber's Fundamental Law of Force was even more elemen-
tary and remote from the empirical level than Ampere's and thus it could be
considered a 'theory within a theory'.
However, the velocity Cw with its somewhat empirical meaning of a limit
and critical velocity was part of that law. The direct measurement of Cw by
an experiment was understandably unfeasible due to both instrumental and
conceptual difficulties, which would have given the potential experiment the
aspect of a Gedankenexperiment.
By overcoming this difficulty and doing so within the metrological pro-
gram which he and Gauss so strongly defended, Weber demonstrated 6 a high
284 SALVO D'AGOSTINO

degree of talent. This program supported the establishment of a rational order

within the system of electric and magnetic units, in contraposition to the
proliferation of practical units which spread out primarily to suit the needs
of technicians.
Weber and Gauss maintained that magnetic and electric units should be
arranged within a system and be defined in connection with units of mechan-
ical force, mass and time.
They called this system of units an 'absolute system'. The requirement of
Weber and Gauss was quite in line with the mechanistic conceptions of their
era, which gave a privileged status to mechanical quantities.
On the other side, this requirement represented a turning point in the
development of physics and rendered electricity more scientific. In fact,
rather than being simple transcriptions of proportionality relationships be-
tween quantities measured in arbitrary units, electric laws themselves in the
form of algebraic equations were used to defme units, thereby acquiring a
meaningful structure of their own.
This need for order and rationality had already been recognized by the
'Decimal System Metric Commission' in the definition of space, mass and
time units at the time of the French Revolution. Although this step was then
important for its technical and social aspects, it had no significant conse-
quences for the evolution of physics theory. Instead, the aforementioned
need showed its full potentiality in the mid-1800's when it was extended to
the problem of electric units.
In fact, thanks to the new conception of a system of units and the sub-
sequent form of electric and magnetic laws, Weber succeeded in measuring
the value of Cw in the form of a 'conversion factor' among different units.
Weber's achievement was based upon the following considerations: the
limitation (implied in the conception of a system) of the validity of each
written law within a specific system allows one to write the same law within
different systems and to measure the numerical constants therein such as
Cw. 6
Weber's law was written in the system of electrostatic units and any law
derived from it (such as Ampere's Law of Force between currents) would
therefore also be written in the same electrostatic units. This written form of
Ampere's law contained the constant Cwo But it was also possible to choose
another system by convention - the so-called electrodynamic system - in
which Ampere's Law was used as the 'Primitive Law'. Within this system,
Ampere's Law was written through unitary constants and thus the compari-
son between the two different writings of Ampere's Law in the two different
 DISCOVERY OF THE VELOCITY OF LIGHT 285

systems allowed Weber to conclude that Cw had also the meaning of conver-
sion factor between units of current in the two systems. 7
In more general terms Weber's procedure can be summarized thus: once
the idea is accepted that the written algebraic expressions of physical laws
are valid within a specific system of units, any difference in the value of
constants is a mere metrological 'accident'. However, it is this very 'accident'
which allows one to measure the value of Cw in the form of a 'conversion
factor', a ratio between the units of the same physical quantity in the two
systems.
Through this method, Weber, together with Rudolf Kohlrausch, an expert
experimentalist, succeeded in surmounting the difficulty of measuring Cwo
Towards this goal he also invented a new instrument, the ballistic galvano-
meter.
Depending on the choice of units, the result of the measurement 8 was:
Cw = 439450.10 6 mm sec-I (static and electrodynamic units, respectively).
Also, Cw = 310740.10 6 mm sec-I (static and electromagnetic units, respec-
tively).
The second value was equal to the speed of light as Weber and Kohlrausch
soon realized, but they attributed no particular significance to this new result,
which they considered 9 as a mere coincidence.
About five years later, Maxwell considered this equality as one of the most
important elements supporting his electromagnetic theory of light.
This point underlines the weight that research programs have in the evalua-
tion of empirical data. The idea that electricity could consist of particles and
involve remote forces as one of its fundamental features, oriented Weber's
theory and his evaluation of the meaning of Cw and implied a sharp differen-
tiation between light (already interpreted within the context of a quasi-elastic
wave theory) and electric phenomena. This differentiation tended to reject or
dismiss as irrelevant any evidence that the two might be connected through
Cwo
Yet Weber's conception of Cw influenced later developments in physics
for its other aspects. His theory introduced the concept of a characteristic
velocity in electrodynamics and this concept set a precedent lO that Einstein
certainly could not have ignored in his well-known statement that a limit
velocity exists in electrodynamics and in physics in general.
It needs to be emphasized that Weber's result stems from a metrological
approach to electrodynamics and physics.
History of physics seems to have overlooked that it is at this point in the
mid-1800's that metrology becomes an essential component of physics. The
286 SALVO D'AGOSTINO

manner in which physics laws are written influences physics theory as 'syntax'
conditions 'semantics' in general.

III. MAXWELL'S TRANSFORMATION OF WEBER'S DISCOVERY

In the early mid-1800's, English physicists highly esteemed German electro-

dynamics for its theoretical and experimental standing. In particular, the high
class of German scientific instruments, designed mostly by Gauss and Weber,
and the accuracy of German experiments were valued by English scientists. 11
German metrology spread in England mostly through the work of the
'Committee on Electrical Measurements' appointed in 1862 by the 'British
Association for the Advancement of Science', with the task of defining 12 the
best system of electric units. Maxwell and William Thomson (1824-1907)
were from the beginning outstanding members of the Committee.
While working on this Committee, Maxwell also prepared his articles and
memoirs on the theory of the electromagnetic field and the electromagnetic
theory of light. It is not surprising therefore that he was interested in the
metrological significance of Cwo However, Weber's interpretation of Cw as
'a limit' and 'critical' velocity and his conception of a fundamental electro-
dynamic force acting-at-a-distance, finds no place in Maxwell's theory of the
electromagnetic field and electromagnetic theory of light.
Maxwell, however, used Weber's metrological concepts in order to define
two 'absolute', complete, and independent systems for all electromagnetic
units. In addition he developed a detailed theory of so-called 'dimensions' of
units and of the homogeneity of laws of physics. 13
According to Maxwell's electromagnetic theory, his particular choice of
systems of units and his theory of dimensions, Weber's Cw represents 14 the
Conversion Factor between electrostatic and electromagnetic units of charges
(or currents) and, at the same time, the velocity of the electromagnetic
waves. Its numerical value equals 15 the velocity of light in aether, within the
errors. He maintained that this equality represented important evidence 16 in
support of his statement that electromagnetic waves and optical waves are
oscillations of the same aether and that light consists of electromagnetic
waves.
Weber's concept of a velocity of particles in motion was thus transformed
substantially by Maxwell, since Weber's Cw became a propagation velocity,
the velocity of both light and electromagnetic waves.
The first demonstration that Weber's Factor is the velocity of light is
presented by Maxwell in his 1861-62 memoir, 'On Physical Lines of Force'.
 DISCOVERY OF THE VELOCITY OF LIGHT 287

For his demonstration he exploits a double analogy between the theory of

elastic waves on the one side, and optics and electromagnetism on the other.
This method of demonstration was soon abandoned and already by the time
of Maxwell's 1865 memoir, 'A Dynamical Theory of the Electromagnetic
Field', the equality between Weber's Factor and the velocity of light had been
established mainly on the basis of consideration relating to metrology.
In his principal work, 'A Treatise on Electricity and Magnetism' (1873),
Maxwell presented absolute units metrology as a general theory and attri-
buted an essential function for his electromagnetic theory of light to it.
According to this theory physics symbols represent a number and a dimen-
sion and consequently the need for homogenizing factors such as Weber's
Factor arises in the equations.
On the contrary, for Weber symbols were just numbers and equations were
algebraic expressions. In Maxwell's metrology, Weber's Factor is implicit in
every conversion of units between the two systems, i.e. this Factor has now
become a structural quantity in the expression of electric and magnetic laws.
This new role ascribed to Weber's Factor now enabled 17 Maxwell, in his
Treatise, to establish its identity with the velocity of light almost automati-
cally without the need of difficult and at times convoluted arguments raised
by elastic and electric analogies.
To understand the reach of Maxwell's theory let us analyze the difficulties
he encountered in this achievement.
To put it briefly, Maxwell had only selected one aspect among the theore-
tical meaning of Weber's Cw, namely the fact that it was a conversion Factor.
He disregarded that Cw indicated a characteristic velocity. 18
Consequently, the whole impact of Maxwell's optical theory was almost
totally supported by the metrological meaning of Cw: its value, as measured
with laboratory equipment by Weber and Kohlrausch, represented for Max-
well the only existing measure of the velocity of electromagnetic waves
obtained by an electromagnetic procedure (a more 'direct' measurement of
this velocity was carried out by Hertz, only about twenty years later, in
1888).
Maxwell's position placed Weber's and Kohlrausch's measurement in a
somewhat peculiar light. It made them out to be the first scientists to have
measured the velocity of light by an electromagnetic procedure in the form of
a conversion factor, non-deliberately.
Among other things, Maxwell, by attaching importance to metrology, had
overcome traditional positions, which considered decisions relating to the
selection of units a matter of conventional choice and therefore theoretically
288 SALVO D'AGOSTINO

insignificant. In order to assure a firmer foundation for his metrological

position, Maxwell felt compelled to give a privileged status to his choice of
systems of units, affirming that only one choice was truly scientific, i.e. the
one regarding electrostatic and electromagnetic systems. By denying the
legitimacy of other systems, Maxwell implicitly included the electrodynamic
system so widely exploited by Weber among them.
Though successful as it was in Maxwell's hands, the metrological method
for the introduction of the velocity of light into electromagnetic theory was
not followed by Heinrich Hertz.
Years later it was harshly criticized 19 by Arnold Sommerfeld, who com-
pared the program of absolute systems to 'Procrustes' bed', within which the
choice of independent electromagnetic units was thought to be unnaturally
restricted.
These aspects of the development of Maxwell's science hardly fit in the
historiographies of linear development where theoretical assurances and
unquestionable experimental bases are considered the norm. On the contrary
these aspects lead to the conception of a dynamics of science as a flexible
project aimed at overcoming difficulties and reaching a favorable outcome.
Maxwell's struggle to link optics with electricity through Weber's Factor
exemplifies the above point. However, although Maxwell could not rely on
theoretical assurances and unquestionable experimental bases, he had refer-
ence points (like Weber's Factor) and indications of directions in which he
developed his electromagnetic theory of light.

IV. THE EMPIRICAL CONTENTS OF THEORIES AND

THE PROBLEM OF 'AN OVERLAP' BETWEEN WEBER'S
ELECTRODYNAMICS AND MAXWELL'S ELECTROMAGNETISM

The historical event we have dealt with so far now enables us to discuss
further the problem regarding the empirical contents of a theory, with partic-
ular reference to the case at hand.
Both Weber and Maxwell accepted the usual electromagnetic phenomena
and experiments of their times as a phenomenological basis for their theories.
One of these is the experiment which charges and discharges a capacitor
suitably inserted in a circuit (the so-called 'electric transient' experiment)
and the typically uneven distribution of electrostatic and electromagnetic
forces in various parts of the circuit. Probably both scientists would have
agreed that the electromagnetic forces indicated by the rotation of a magnetic
needle set up by a current flowing in the wire are correlated with the variation
 DISCOVERY OF THE VELOCITY OF LIGHT 289

of the density of charges upon the capacitor plates indicated by the displace-
ment of the leaves of an electroscope and consequently with the variation
in electric forces. In order to quantify these experiments and observations
Maxwell and Weber probably would also have agreed on methods for measur-
ing such forces, on the criteria for defining charge and force units. These
methods and criteria involved Ohm's and Kirchhoff's laws. Therefore they
would have agreed on a whole series of strictly correlated experimental
methods and laws, identifiable with the 'overlap area' of the two theories.
It would be just as wrong to consider this 'overlap area' as consisting of
mere observational contents as it would be inaccurate to consider it as being
purely theoretical.
While theories exist in this area, they are not comparable, however, with
Weber's and Maxwell's two rival theories. The theoretical contents of the
'overlap' may be located closer to the level of the 'observables' analyzed by
Margenau. 2o In the same sense, these contents may be considered as consisting
of weaker theories, or instrumental theories according to Bunge. 21 In other
words one could affirm that this overlap also contains a theoretical com-
ponent because it is made up of a whole complex of procedures and rules and
therefore its contents can be defined 'quasi-empirical'. 'Overlap' assertions
are not theory-neutral yet their theoretical contents are not of the same level
as those of Weber's and Maxwell's theories.
This point should be further analyzed. One should explain how the 'over-
lap' contents, substantially those on which Weber's discovery is based, can be
included in two such radically different theories.
In this analysis I have been substantially guided by K. F. Schaffner's con-
siderations in his 'Outlines of a Comparative Theory Evaluation'. 22
According to Schaffner, a scientific theory consists of statements and rules
of different kinds. Some statements concern non-logical primitive terms
whose Significance is related to previous theories (Le. to preliminary, or back-
ground knowledge). This is defmitely true if we look at a word like 'molecule'
as Maxwell used it in his kinetic theory; or 'electric current' as it was used in
its early meaning (according to Ampere). Besides, there are some statements
which establish how entities and/or processes described by axioms of the
theory are in the end related to the realm of the observable. For purposes of
convenience and in accordance with Schaffner's usage, let us refer to both
statements as Type-C statements (sometimes also referred to as correspond-
ence rules).
According to Schaffner, Type-O statements, on the other hand, describe
inter-subjective experiences such as, 'the appearance of light and dark fringes
290 SALVO D'AGOSTINO

in an eye-piece', a 'sound sensation' or 'curved trajectories on a photographic

plate'.
In our case, a Type·O statement might consist of the affirmation that,
while a capacitor, suitably incorporated in an electric circuit is being charged
or discharged, a magnetic needle is deviated from a steady state position by
a force that increases as the needle is moved along the wire away from the
capacitor.
Type-O expressions or t,~rms appear in the preceding statement with terms
like 'magnetic needle', 'capacitor', and also logical connection and disjunc-
tion relationships as well as theory-neutral mathematical terms. According
to Schaffner, Type-O terms have two possible meanings: (1) a principal
meaning connected to a referent, be it an object or an operation (e.g. needle
deviation, etc); (2) a secondary meaning attributable to their being closely
linked with theoretical entities by means of correspondence rules or Type-C
statements.
A Type-C statement in our case might read as follows: the deviation of a
magnetic needle is caused by a magnetic force, i.e. by the perturbation of
electrostatic forces which is due to charge motion and is described by Weber l
fundamental law .
However, the term 'needle deviation' could also have been co"elated to
Maxwelll theory.
To summarize, it may be stated that the observational basis for a Type-O
expression, namely its principal meaning, can be shared by two contrasting
theories. However, this does not prevent these expressions from also including
theoretical elements which permit that indeterminacy of meaning by which
the same term appears as an empirical basis in two different and even con-
trasting theories. (In short, the observable is not the object of the theory, but
merely a referent which allows a partial significance to be given to some of
the theoretical terms.)
Naturally these observations also deal with the old problem of crucial
experiments and the pretension that they discriminate between contradicting
theories. It is exactly in this respect that Schaffner uses his thesis with refer-
ence to the Michelson-Morley experiment. As far as our case is concerned, we
may state that Weber and Maxwell adopted theoretical strategies which,
however divergent they might have been, were both compatible with the
'overlap area' known in their times.
This was possible because the 'overlap area' was flexible and compatible
with Maxwell's and Weber's operations since it was quite far removed from
both data concreteness and concept rigidity. Fundamentally, follOwing
 DISCOVERY OF THE VELOCITY OF LIGHT 291

Schaffner, this characteristic of the 'overlap area' has been connected with
the role of Type-O statements.

V. CONCLUDING REMARKS: DISCOVERY IN PHYSICS BELONGS

TO THE REALM OF THE QUASI-EMPIRICAL AND ITS NOVELTY
AND RELEVANCE FEATURES ARE MEANINGFUL ONLY IN THE
CONTEXT OF A RESEARCH PROGRAM

Not even after a detailed analysis is it possible to place Weber's discovery

within the realm of the observable or, in other words, to affirm that its
'contents' belong to the realm of the empirical. If one wished to insist on its
empirical nature, one would say that the discovery is concerned with the
fact that electromagnetic forces are weak compared to electric forces. How-
ever, this quasi-quantitative statement can contain a precise meaning only
when metrological criteria are determined (Le. when it is clearly stated how
forces deriving from equal charges shall be compared, be they static or in
motion). The metrological procedure and all the pertinent laws and opera-
tions are then included as major aspects of the statement, thereby jeopardiz-
ing any pretension of an empirical nature. In brief, in Schaffner's words this
is a Type-O statement, which contains terms having secondary meaning in
the aforementioned sense.
We should like to qualify this affirmation more clearly: one could say that
the velocity of light can be calculated in electrodynamics by carrying out
some experiments and measurements of charges, currents, forces, etc. Now
the above statement should be substantially restricted adding that this occurs
only if one chooses by convention absolute electrostatic and electromagnetic
units as Maxwell did.
The same results cannot be obtained if we opt for a different system of
units, such as the previously mentioned Weber electrodynamic system.
In addition, as it is well known, the so-called Giorgi System introduces
the speed of light from the 'outside', without deriving it from metrological
reasoning.
In conclusion, with regard to relevance features of the discovery of Cw,
Weber emphasizes the critical characteristic velocity aspect. The novelty lies
therein, since no such velocities existed in Newton's mechanics. The fact that
Weber refused to ascribe any significance to its near-equivalence with the
velocity of light, lowers the relevance level of his discovery. On the contrary
Maxwell emphasizes the most phenomenological aspect of Weber's discovery
i.e. the metrological aspect, and ignores (or virtually ignores) its significance
292 SALVO D'AGOSTINO

as characteristic velocity. However, he grants a high degree of novelty and

relevance to its relationship with the velocity of light. In fact, it can be said
that Maxwell attaches an enormous relevance to this second aspect, due to
the fact that the equivalence is one of the keystones of his electromagnetic
theory of optics. It is precisely this very same quasi-empirical nature of Cw
which allows this freedom of choice within the metrological program, and
enables Weber and Maxwell to shift the novelty and relevance features into
a different aspect of the discovery.
It might be interesting to compare the meaning of 'discovery' here iden-
tified in the more advanced theories of physics with its meaning in other
scientific fields.

Translated by
JUDITH A. BOFFA

NOTES AND REFERENCES

I Kenneth F. Schaffner 'Outlines of a Logic of Comparative Theory Evaluation with

Special Attention to Pre- and Post Relativistic Electrodynamics', in R. Stuewer (ed.),
Minnesota Studies in the Philosophy of Science, Vol. 5 (University of Minnesota Press,
1970), pp. 311-351.
2 W. Weber 'Elektrodynamische Massbestimmungen iiber ein allgemeines Grundgesetz
der elektrischen Wirkung', in W. Weber's Werke (Berlin, 1893), Vol. 3, pp. 25-211.
K. H. Wiederkehr, Wilhelm Edward Weber, Erforscher der Wellenbewegung und der
Elektricitat (Wiss. Verlagsges., Stuttgart, 1967).
3 W. Weber's Werke, op. Cit., pp. 597 -608. Weber's Law of Force is the following:

- ee' [ 1 1
r' - -Cw'
where e and e' are the charges and r is their distance. When the relative motion is uni-
form this breaks down to the simpler form:
ee' [ 1 dr']
-;z 1- Cw' --;uz .
It is evident that Cw represents the velocity dr/dt for which the force becomes null.
The force reverse its sign for dr/dt > Cwo
4 E. Whittaker, A History of the Theories of Aether and Electricity (Nelson, London,
1951),2 vols., Vol. 1, p. 201. Salvo D'Agostino, 'La scoperta di una velocita quasi uguale
alla velocita della luce nell'elettrodinamica di W. Weber (1804-1891)" Physis 3-4
(1976),297-318.
5 Whittaker,op. cit., p. 201.
6 W. Weber and R. Kohlrausch, 'Ueber die Elektricitiitsmenge, welche bei galvanischen
 DISCOVERY OF THE VELOCITY OF LIGHT 293

Strom en durch den Querschnitt der Kette fliesst' (1856), Weber, op. cit., pp. 597 -608;
Wiederkehr,op. cit., pp. 140, 141; D'Agostino, op. cit., p. 309.
7 D'Agostino,op. cit., pp. 305-306.
8 Wiederkehr,op. cit., pp. 140-141.
9 Weber, 'Elektrodynamische Massbestimmungen insbesondere elektrische Schwingun-
gen', Werke, Vol. 4, pp. 107-241; p. 157ff.; D'Agostino, op. cit., p. 312.
10 D'Agostino,op. cit., p. 310.
11 J. C. Maxwell praises Weber's Electrodynamic Theory in his 1855 'On Faraday's
Lines of Force' as "a professedly physical theory ... which is so elegant, so mathema-
tical, etc.... " In 'A Dynamical Theory of the Electromagnetic Field' (1864), the theory
as developed by W. Weber and C. Neumann is considered "exceedingly ingenious and
wonderfully comprehensive ... it has served to guide the speculations of one who has
made so great an advance in the practical part of electric science, both by introducing
a consistent system of units in electrical measurement, and by actually determining
electrical quantities with an accuracy hitherto unknown".
12 R. Sviedrys, 'Physical Laboratories in Britain', in Historical Studies in the Physical
Sciences, Vol. 7, pp. 405-436, p. 425. S. D'Agostino 'Esperimento e teo ria nell'opera
di Maxwell', Scientia 113 (1978),453-467, p. 454. Also: D'Agostino 'Experiment and
Theory in Maxwell's Work' (English Translation), ibid.
13 J. C. Maxwell,A Treatise on Electricity and Magnetism, 3rd ed. Preliminary, Sections
1-6; also Sections 620-628.
14 Ibid., Section 786.
15 Ibid., Section 787.
16 Another evidence is, for Maxwell, the transversality of electromagnetic waves.
Treatise, Section 791. The equality between the square root of the dielectric constant
for melted paraffin, on one side, and its index of refraction, on the other (the so-called
Maxwell's Law) is not considered by Maxwell satisfactorily confrrmed by the existing
experiments. Treatise, Section 789.
17 See for details: "D'Agostino 'Esperimento e teoria ... ' ".
18 A related consideration to this aspect was given by Maxwell in a section of his
Treatise (Sections 769-770).
19 A. Sommerfeld, Physikalische Zeitschrift 36 (1935), 814, 820; D'Agostino, 'Esperi-
mento e teoria', p. 463.
20 Henry Margenau, The Nature of Physical Reality (McGraw-Hill, New York, 1950),
especially: Chapters 6, 7.
21 Mario Bunge, Philosophy of Physics (Reidel, Dordrecht, 1973), pp. 39-179.
22 Schaffner,op. cit., pp. 311-330. Schaffner's arguments are exemplified in his essay
by the Lorentz-Einstein theory of Relativity and the Michelson-Morley experiment. In
this paper I use 'overlap' in the same sense as Schaffner.
 HOWARD E. GRUBER

COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY,

AND THE CASE STUDY METHOD*

"Historical facts are, in essence, psy-

chological facts." - MARC BLOCH,
The Historilln 's Craft. 1

But what kind of psychology?

When we approach the subject of creative scientific thought as a problem
in cognitive psychology, a host of questions arise: What is the nature of a
cognitive structure that can grow through purposeful work? What is the
relation between the organization of knowledge in the individual and the
organization of knowledge in various scientific collectivities? Does the organi-
zation of the individual's multiple purposes correspond to the organization of
his knowledge? Does the study of how a person thinks contribute to our
understanding of what he thinks? And vice versa?
The study of the thinking of scientists is peculiarly liable to the error of
psychological oversimplification. The aim of science is the unification of
diverse strands of knowledge in a smaller number of principles. Through pro-
longed effort, the product appears in public as a miracle of Simplicity, co-
herence, and compression. But the process by which this result is achieved,
far from displaying these same attributes, is complex, many-sided, sometimes
incoherent, and often greatly extended in time. Since every scientific effort
opens up a number of potential pathways for future investigation, the
avoidance of chaotic pseudo-growth and the successful organization of work
require the operation of certain regulatory mechanisms that can steer activity
toward steadfast elaboration and development of comprehensible conclusions.
But the more one fragments behavior and intellectual work, the less pro-
nounced these directional properties appear. By their fragmentary approaches,
therefore, both disciplines - psychology and history of science - have evaded
the necessity of studying these regulatory mechanisms.
Psychologists and historians of science have followed two rather different
pathways toward this overSimplification. Psychologists have focused their
attention on processes rather than contents, hoping to arrive at a general -
that is, content-free - picture of intellectual processes. Indeed, they have
295
MD. Grmek, R. S. Cohen, and G. amino (eds.), On Scientific Discovery, 295-322.
Copyright © 1980 by D. Reidel Publishing Company.
296 HOWARD E. GRUBER

often deliberately done their utmost to expunge meaning from the material
they present to the persons whose mental processes they are studying: non-
sense syllables or strings of unconnected words to be remembered; problems
to be solved having nothing to do with the prior knowledge or real concerns
of the person - these are still the commonplaces of experimental psychology.
Even when more meaningful contents are considered, as in various attempts
at 'content analysis', the approach taken is often statistical in nature - sub-
stituting the counting of the occurrences of ideas for a thorough and reflec-
tive examination of the structure of the particular ideas under investigation
as seen within the framework of the larger structure of the individual's
knowledge.
When we turn to the history of science, the picture seems, at least to this
occasional visitor in that domain, exactly the opposite. Maximal attention is
given to the contents of thought, and almost none of all to the process. More-
over, the most powerful vector is toward treating thought as highly specialized.
Monographic treatments tend not to deal with the person's thinking as a
whole, but to treat of particular achievements in isolation from each other.
Consider one example, which I choose not because it is poor, but because
it is excellent work, Camille Limoges' La selection naturelle: etude sur fa
premiere constitution d'un concept (1837-1859).2 Limoges' examination of
Darwin's initial construction of the theory of evolution through natural
selection is restricted to Darwin's 'Transmutation Notebooks'. Limoges does
not take up the interplay of the contents of these notebooks with the note-
books on 'Man, Mind, and Materialism', which Darwin kept during the very
same period of time. 3
Thus, the prevailing tendencies in both fields are expressions of three
questionable assumptions: (1) the contents and processes of thought can be
understood in isolation from each other; (2) diverse intellectual processes
can be studied in isolation from each other; and (3) individual ideas can be
studied in isolation from other ideas.
In recent years, efforts to consider the structure of knowledge and the
processes of thought together have grown in several quarters. The work of
Jean Piaget and other exponents of genetic epistemology is a foremost
example. At the level of concern for biological adaptation, Piaget's work has
always been highly process-oriented. But in his extended researches on
children's thinking, it is probably fair to say that, unlike other psychologists,
Piaget has been primarily concerned with the contents and structure of
knowledge. Only in recent years, especially under the leadership of Barbel
Inhelder, has the Geneva school devoted much attention to specifiable thought
 COG NITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 297

processes such as the strategies children use in solving Piagetian problems. In

that group, the work of Christiane Gillieron is especially noteworthy, both
for its attention to the interplay of contents and processes and for its effort
to specify the procedures and strategies individuals use in solving a given class
of problems. 4
In the field of information processing and artificial intelligence, there has
been a notable shift of emphasis away from concern for content-free processes
and toward dealing with ways of also representing the structure of knowl-
edge. A useful paper by Allen Newell 5 insists on the need for the develop-
ment of such more holistic methods of understanding thought.
In my own work and that of my students, we have found that cognitive
case studies must strive toward a consideration of the individual's cognitive-
economy-as-a-whole, and that we must view it as a system in constant develop-
ment, hence we call ours an 'evolving systems approach'. Although we re-
cognize that revealing the cognitive-economy-as-a-whole may be an unattain-
able ideal, we believe that trying to study it is a fruitful venture. For example,
Martha Moore-Russell, in her examination of John Locke's notebooks, has
been able to show convincingly that the structure of Locke's ideas changed,
over a ten-year period, in previously unnoticed ways, and that understanding
these changes clarifies apparent contradictions in Locke's thought. 6 We would
be happy if our cognitive case studies contribute to re-uniting the study of
the structures of knowledge and the processes of thOUght.
In the present paper I want to take up some of these structures and pro-
cesses, especially the individual's network of enterprises, images of wide
scope, and regulatory schemes. Most of the evidence and thought I will
introduce is based on intensive case studies of one scientist, Charles Darwin.
Unfortunately, both psychologists and historians of science have avoided
such cognitively oriented case studies. It is therefore necessary to begin with a
discussion of the case study method. Let us examine its underlying logic
with regard to one central point: what kind of general conclusions can we
hope to draw from the study of unique cases?

I. THE CASE STUDY METHOD

It is common in discussing the case study method to preserve an aura of due

scientific prudence by stressing the point that single cases cannot lead to any
safe generalizations. The special role for cases is the production of hypotheses
to be tested by methods permitting sounder inductions.
I cannot let this position stand unchallenged. The case study method is
298 HOWARD E. GRUBER

certainly not in a privileged position \>lith regard to hypotheses: all methods

produce them in the hands of a person who likes to make them. Nor can we
allow ourselves to say that the case study method produces 'nothing but'
hypotheses.
It would be better to throw caution to the winds, to insist that all induc-
tions are unsound if they are nothing but inductions, that an N of 100 is no
better than an N of 1, since any generalization based on N cases can be
exploded by the (N + 1)th case.
All scientific enterprises must ultimately involve the testing of hypotheses
and the construction of theories by the blending of appropriate logical and
empirical methods, and the case study must make its special contribution
through the kind of hypotheses it can lead to and can test. In short, these
must be hypotheses and theories concerned with the understanding of the
whole person as he develops over long periods of time. The essence of the
empirical argument must reflect this emphasis on contextual wholes extended
in time. The hypotheses advanced and tested, and the way in which they are
tested, must do justice to the individual case, or we would have done better
to remain within the more comfortable confmes of the laboratory.
One possible way of dodging this need to understand the whole person
would be to argue that the study of one case is merely a preliminary to the
study of hundreds: naturally, we study them one at a time, but when we have
many, we can scan them for certain regularities and employ familiar inductive,
statistical methods. Unfortunately, the more time and energy is expended on
accumulating data from many subjects, the less is available for the detailed
examination of the structure of a single subject's behavior and thought. More-
over, the statistical study of cases is definitely not the case study method
unless it can be done in a way that permits the comparison of whole-properties
of individuals. And the way of thought necessary for such comparisons must
grow out of the study of individual cases, N = 1.
The greater part of scientific thought and inquiry addresses itself to the
search for universal truths. Since in reality things are rarely if ever exactly
the same in one place as in another, such truths have a peculiar price. Instead
of flat statements of empirical universality, we must restrict ourselves to
conditional statements of the form, 'if X is the case, Y is always the case', or
to statements of a functional relationship, such as 'the greater the value of
X, the greater the value of Y, or Y =I(x), etc.
If the task of science were only to uncover such law-like relationships, we
might restrict our empirical work to the experimental method, whose great
strength lies in the systematic manipulation and observation of small groups
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 299

of well-defined variables. But our task goes far beyond that. We want to
understand how these variables function in relation to each other, in some
coherent system, such as a solar system of many bodies, an individual organ-
ism, an ecospace, or a society. Such systems may be thought of as organized
structures: no functional relationships occur outside such structures, all
functional relationships occur within such structures. See Figure I for a
simple example.

Fig. 1. The significance of a variable depends on the structure of which it is a function.

Movement 'up' in the visual field corresponds to movement away from the observer, if
the moving object is on the ground or on a floor, as at a. But movement 'up' corresponds
to movement toward the observer, if the moving object is in the sky or on a ceiling, as at
b.

I am not merely saying that we must repeat our in vitro experiments in

vivo to make sure that relationships observed in the laboratory can also be
observed in nature. If we want to understand any relationship, we must
understand the system or structure in which it occurs. We may have the
option of choosing to say either (1) that law-like relationships are expressions
of the functioning of the system in which they occur or (2) that such systems
are only particular combinations of relationships. But this is not an option
we can exercise intelligently unless we grasp the relation between laws and
systems. While the experimental method is at its strongest in establishing
lawful relations among small groups of variables, the case study method is at
300 HOWARD E. GRUBER

its strongest in permitting study of the interplay of complex groups of factors.

But certain additional considerations enhance the importance of the case
study method.

Sense of Proportion
In trying to understand the case, we have a chance of being led to attend to
factors that are important in at least that case. This means that we may pro-
tect ourselves from wasting a career on something that is never important.
Moreover, attending to that factor in just the ways necessary for understand-
ing the one case may lead to understanding how that factor is important in
other cases. (But not necessarily in the same way, as I will discuss below.)

The Importance of Unique Events

There are, after all, a large number of interesting events in the world that are
unique, and whose main interest for us lies precisely in this uniqueness. One
major class of unique events (if the term class can be applied here!) is com-
posed of those human achievements called 'creative'. Another is the evolution
of each new species, which occurs only once on our planet. Still another is
the emergence of a new form of society.
It may be said that the task of science is only to search for those features
of such unique events that are general. Even if this were agreed upon, it would
not absolve us of the need for case studies, since we could not know in
advance which features were general and which unique, or if - as I believe
is a more promising line of attack - the uniqueness of the event lies in the
occurrence of a novel configuration of the whole. We cannot therefore
simply collect many cases and pin them like butterflies for display, singling
out this or that feature for comparison of contrast. Each case must first be
studied in its own right.
But there is a stronger argument. The unique case is of scientific interest
in its own right. If our goal is to understand and describe objective reality or
different parts of it, there may be important and interesting instances in
which all we can do is to understand and describe the individual case. Suppose
we should have reason to believe that there are several ways of forming new
solar systems, and that our own was formed in a unique, unrepeatable way.
Would that belief make our solar system any less interesting an object of
inquiry? (The study of our solar system, stretching over millennia of human
existence, is indeed an example of a 'case study'!) Suppose the creative
processes of Sigmund Freud or Charles Darwin are unique and unlike anyone
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 301

else's. Would that make them any the less interesting? Would we have to turn
studies of such individuals out of the scientific fold? Or, on the other hand,
do we need all the tools of science to understand these exceptional events?

Few Interesting Empirical Inductions Possible

Thanks to the efforts of experts in experimental design, we have been warned

against the dangers of careless inductions stemming from poor techniques
(unrepresentative samples, badly controlled experiments, etc.). Thanks to the
efforts of philosophers, we have been adequately warned of the logical
pitfalls in the process of straightforward induction, and various alternatives
have been proposed, such as the hypothetico-deductive method. Yet, in
psychological research enormous efforts are still expended on what is funda-
mentally a set of inductive procedures - that is, the attempt to arrive at a
generalized descdption of a group of phenomena. From a finite set of observa-
tions that do not contradict statements such as Y = [(X) we hope to conclude
that the equation is true for all values of X and Y. With all its difficulties, this
method resides securely in the heart of experimental science, and orderly
thought would be inconceivable without it.
Nevertheless, in the case study method applied to innovative processes
we must give up all hope of generalizing from our cases well studied to some
larger population, at least so far as the most interesting questions go. There
are a few matters that may yield to time-honored methods such as the
accumulation of similar instances. For example, almost every interesting
creative innovation we can observe seems to have taken a considerable time
for its construction. This is a safe statement no matter what position we take
on the importance of sudden insights. But what happens during the time
elapsed? How does the innovative system function and evolve? Were it not
a self-contradictory sentence, we could only say, 'uniqueness is the rule'.
Creative processes may have certain similarities when looked at from the
outside, but there is no a priori reason to presuppose that in the essentials
of their inner systemic functioning any two creative processes are alike.
The very fact that we have chosen our cases for their individuality and
uniqueness means that we have no possibility of defining a population or a
sample. This does not prevent us from studying something a little bit like one
aspect of creativity, for example, divergent thinking, 7 in a random sample of
midwestern American college sophomores. But there is no defensible way of
generalizing inductively from anything we might learn in such studies to the
processes involved in unique and extraordinary events. Nor need we shed any
302 HOWARD E. GRUBER

tears over this loss. There does remain something constructive that we can do,
a point I shall expand in the second half of this paper.

What Kind of Generality Can We Expect from the Case Study Method?

We bring to each case a repertoire of skills, methods, questions, expectations,

and above all, ideas about how such evolving systems work. This last is no
commitment to a belief that they all work in the same way, but rather that
there are certain types of question that we can fruitfully ask about them all.
Is this kind of heuristic generality enough, or all we can hope for in his-
torical inquiry? If the task of historiography were simply to write clear
narratives, the question of generalizability might not arise. But if we hope
to understand the history of science, we must link our narrative to some
general ideas about intellectual processes as they occur in historical contexts.
It is only reasonable to expect psychological science to provide some of the
necessary insight.
When I began my own efforts in this domain, I hoped to bring together
two fields of inquiry, history of science and cognitive psychology, in a
fruitful new synthesis. To my dismay, as indicated above, I found that in
neither field is inquiry organized or theory constructed in such a way as to
make this synthesis possible. In a word, on both sides of a high disciplinary
fence, fragmentation and atomism prevail. Specialists know a great deal about
numerous subjects. Indeed, the new methods of citation analysis permit the
description of such specialties as organized communities, or international
'villages' within which communication is intense but between which it approxi-
mates zero. One can, moreover, very well imagine a historical research project
that would trace the growth and differentiation of such villages, their geo-
graphical spread, their eventual decline, etc. - all without ever asking any
important questions about the actual work and thought actually going on in
one village or in one person's head. Such is the power of statistical technicity
that it can distract our attention from the very object of our attention. It is
the organization of knowledge as a whole - as he understands it - that
guides the work of the individual. Therefore, a powerful new tool for sketch-
ing that wide organization may be of great service. The difficulty with statis-
tical tools arises only when they are utilized (as is all too often so) to detect
general trends, averages, etc., and to minimize attention to unusual events.
Insofar as we are concerned with innovation and creativity it is just these
unusual events that interest us - the rara avis who crosses into a habitat not
frequented by his kind, for example. s
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 303

We take specialization so much for granted that we forget that even in the
mind of the expert, the specialist, there is some organization of knowledge as
a whole - first of all, his own knowledge and secondly, his conception of
human knowledge in general. These organizations guide the work of the
individual, and their psychogenesis ought to be of far greater interest to us
than has been the case until now. In childhood, both in spontaneous play and
in organized education, specialization is absent - making its appearance
mainly at a point approximately the equivalent of advanced university train-
ing. In adolescence, especially in creative adolescents, there is commonly a
period of romantic, often passionate search for cosmic truths, for some
framework to understand the meaning of life, to justify the impending
specialization that the young person sees looming before him, to find his own
identity by placing his own future in some context. 9 Then, of course, in life
as we know it, disciplined specialization, careerism, and the pursuit of realistic
goals - all take hold of the person. Later in life, often enough, this narrowing
ebbs and gives way to a renewed search for meaning and coherence. In a
happy few this search is active throughout the life span, and it has been often
said (although not yet well documented) that great innovations are made by
those who bring together previously disparate disciplines - Le., by those who
do not accept the alienated condition of science as a world of non-communi-
cating villages.
In many scientific lives then, and not necessarily only those attaining
heights of great eminence, the appearance of narrow specialization is only a
moment in a life-history that is really far more complex. To understand all
this movement we need to look carefully at the individual's organization of
knowledge and at his organization of purposes. We need to consider them as
they develop over the life-history.
We may also choose to look at these organizations of knowledge and
purpose which are shared by individuals in face-to-face scientific communities
such as research teams, in 'invisible colleges', or in paradigmatic 'villages' of
the kind Derek de Solla Price discusses in his paper for this volume. But no
matter which focus we choose, it is plausible that some of the most interest-
ing intellectual phenomena we will wish to understand are unique events.
To pursue this goal we will have to struggle toward a method for con-
structing what may be called the theory of the case. We will need to develop
the necessary structural and systemic acumen. We will of course be guided by,
and in the course of our work will hope to contribute to, some very general
ideas about such structures and systems. The kind of generalization we hope
for, then, may not be about the objective world, so much as it is about our
304 HOWARD E. GRUBER

evolving way of describing it. Even if every innovative event in the history
of science were incurably unique, our ways of describing these events would
have something in common, we would ask similar questions about them all.
Imagine, for example, a motor mechanic in a world inhabited by a multitude
of ingenious inventors. He wants to understand how each new machine
works. But they are all different. Suppose each one uses a different fuel. He
may still ask, "How does this one get its energy?" The ways in which the
machines are alike may be of some interest, but such generalizations would
not satisfy our mechanic. He would want to know, 'How does this one work?'
That is the kind of question and the kind of answer we may hope for in the
study of unique systems producing new knowledge.

Related Methods

By now the reader may very well be asking: 'Isn't he flogging a dead horse?
Isn't he merely making a plea that we go on doing what we already do? What
does he want?'
To answer these questions I will now briefly survey a number of related
approaches, each of which might seem to produce cognitive case studies,
but does not.
I begin with Stillman Drake's critique of the historiography of science,
in which he voices concerns similar to my own.l0 First, he characterizes
recent shifts in historiography of science as a change from maximizing the
role of a few great revolutionary pioneers to minimizing the role of the
individual in history. In dealing with someone seemingly great, the strategy
of the new style is to "attribute as much of his thought as possible to his
predecessors, and to grant as little as possible to his own originality ..." and
thus to "discern the great underlying causes... [of] the slow, uneven but
continuous emergence of modern science from ancient philosophy" .11 Then
Drake points out the danger of error in using the historical context to explain
an individual's thought without deep study of the documentation of that
person's life: the "concatenation of ideas within an individual mind mayor
may not be identical with one selected as characteristic of a given society" .12
He suggests that a careless importation of our own general concerns into our
description of the thought of the individual being studied may lead to serious
mistakes. For example, if we are interested in the role of extraphysical ideas
in Galileo's thinking, we need not necessarily look only or even chiefly to
philosophy; in Galileo's case musical theory may have been more important,
a point which Drake develops in some detail. His father was a musician and
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 305

both father and son were interested in musical theory. In sum, Drake charac-
terizes history of science as having been the study of the development of
disembodied ideas, detached from the individuals who think them. He makes
a plea for psychologically oriented biographical work, and it is clear that he
does not mean psychoanalytic biography so much as the kind of cognitive
case study that is the subject of the present paper. He explains that he has
deliberately chosen the title, Galileo Studies, Personality, Tradition· and
Revolution, rather than Koyre's 'Galilean Studies', to bring home the point
that the thinking person is at the center of his concerns.
In spite of this orientation, it seems to me that Drake himself has not
come very far in carrying out such a program. His own work, seminal as it is,
is only a set of sketches moving in the direction of a cognitive case study. The
situation, then, is that in the case of Galileo, a figure of the first order of
magnitude, and an unusually well-documented case, nothing like a cognitive
case study has yet appeared.
I have often asked myself, how could it happen that a novice and an
amateur like myself might have the opportunity to write the first article
on the thousands of pages of notebooks Charles Darwin kept during the
Beagle voyage? How could it happen that Darwin's transmutation notebooks
and his notebooks on man, mind, and materialism lay almost unexploited in
the Cambridge University Library until a psychologist knowing little about
the history of science came along? And now, how does it happen that my
student, Martha Moore-Russell, has a similar opportunity to open up the
almost virginal notebooks of John Locke? If it is true in the case of such
figures, there must be many others waiting. Somehow, our disciplines,
psychology and the history of science, have not evolved in such a way as to
pay close attention to the inner workings - in all their richness and in their
complex development - of creative minds, even when the documentation is
available. That we do not yet have the conceptual tools to do this work, and
must improvise them as we go along, means that, for some reason, we have
not been asking the kind of questions that generate cognitive case studies.
A similar situation obtains in other fields. When I read biographies, such as
Clark's recent excellent biographies of Einstein and of Russell, I am invariably
disappointed at how little attempt is made to examine the actual thought pro-
cesses of a person whose greatest distinction lay in those thoughts. How indeed
could Clark address himself to the thinking that went into Principia Mathe-
matica in the scant twenty-two pages he devotes to that chapter in Russell's
life? It would seem as though the biographer is all too easily distracted or
bemused by 'objective' and external events in the life of the thinking person,
306 HOWARD E. GRUBER

distracted away from that thinking which is at the heart of an intellectual's

concerns.
Writing about literary biography, Hilton Kramer has made a similar com-
ment: "The biographer proceeds on the assumption, which nowadays seems
to be shared by the reader, that it is possible to give a true account of a
writer's life without giving an account of his writings". 13 And a review of a
recent psychoanalytically oriented biography of Beethoven, while lauding the
book as a landmark of Beethoven scholarship, closes with the wistful remark
that a biography giving a full treatment of the man and his music has yet to
be written: "Gripping as the story of the life is, one sometimes is inclined to
ask with a touch of impatience, what was he (Beethoven) really doing all that
time?,,14
When we tum to that variety of biography now sometimes called 'psy-
chobiography' or 'psychohistory' - that is, psychoanalytically oriented
biographies, a similar situation prevails. Not enough detail of the actual
workings of the creative thought process is given for us to understand what
went on; personality is considered as a given which can be used to explain the
person's work; and typically, some variant of psychoanalytic theory is also
premissed as a given. This attitude is expressed clearly in Saul Friedlander's
Histoire et psychanalyse, IS a very competent exposition of work of this type.
Since the emphasis is placed on seeking causal mechanisms originating in early
childhood, and since this period of life is almost certain to be poorly docu-
mented, a great deal of the theoretical work in such studies must go into
filling the gap with long-range inferences. This focus on obscure origins,
coupled with relative inattention to the creative work itself, makes inevitable
the strategy of using the study as a demonstration of the efficacy of a pre-
existing general theory, rather than as an opportunity to test empirically new
theoretical work.
The heavy reliance on psychopathology in the establishment of psycho-
analytic theory plays an important role in this genre. Freud, in his classic
study, Leonardo da Vinci and a Memory of His Childhood,16 established the
pattern of seizing upon a fragment of the creative person's life and using it
to divine the psychosexual well-springs of the individual's creativity, con-
necting infantile origins with mature achievements without looking at the
actual process by which those achievements were arrived at. In a study of
Darwin by a leading psychoanalyst we see a similar pattern. I refer to
Phyllis Greenacre's work, The Quest for the Father: A Study of the Darwin-
Butler Controversy, as a Contribution to the Understanding of the Creative
Individual. 17 A supposed life-long neurosis is taken for granted; an obscure
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 307

incident in Darwin's later life (the relation between Darwin and Butler began
with Butler's letter to Darwin in 1863, when the latter was 54 years old) is
seized upon; some psychosexual connection is made between a highly specu-
lative reconstruction of the two men's infancies on the one hand and the
vicissitudes of their later relationship on the other; and the whole is served up
as an explanation of Darwin's creative work. All without opening up a single
notebook from the massive documentation left behind by Darwin, and worse
still, all without any serious discussion of the contents of Darwin's (or for
that matter Butler's) thinking.
I do not mean by these remarks to pass any judgment on the general
relation between neurosis and creativity. I only question how work of this
sort can ever be expected to shed any light on the actual thought processes
involved in creative scientific work. Suppose we grant that Darwin's almost
cosmological aspirations and his seeming interest in the obscure origins of
species can plausibly be explained as expressions of yearning for the lost
mother (she died when he was ten years old) and curiosity about his own
origins. This does little to explain how his thinking developed, how he went
from one idea to another.
If we try to explain the theory of evolution itself as an expression of
personality, we stumble on the difficult fact that two men as different as
Alfred Russel Wallace and Charles Darwin - different in class background,
family life, temperament, and even philosophical point of view - hit upon
approximately the same theory. What we need to do is to show how each
individual, through his own struggles, arrived at a viable intellectual adapta-
tion. Darwin and Wallace travelled different roads for most of their lives, the
same theoretical road for a while, and then their paths diverged (Wallace, a
spiritualist, rejected the idea of the origin of human mentality through na-
tural evolutionary processes). If personality is relatively fixed, which parts of
these two men's life cycles were expressed in their works? In short, a theory
that explains thought as the re-enactment of early experiences cannot explain
creative work; such work has as its predominant characteristic, not re-enact-
ment but purposeful growth.
Perhaps a more promising approach is to suggest that every individual
contains several potential or virtual personalities: rather than pre-existing
and causing the person's work, they are brought alive and even generated
as needed by the specific situations in which the individual finds himself.
These situations are in good part determined by the fortunes of the work,
so central in the lives of the kind of people we are discussing. Just as a brave
act may make a man brave (Mao Tse Tung?), a successful theoretical effort
308 HOWARD E. GRUBER

may make a person more reflective. Thus, an evolving systems approach takes
a special view of the relation between the person and his or her work: both
evolve continuously throughout the life history, and they co-exist in a fruit·
ful dialectical relationship which is the motor of growth. This is not the
position taken in psychoanalytically oriented studies: neither Freud's work
on Leonardo, nor Erikson's seminal study of Luther, 18 nor Frank Manuel's
fascinating psychobiography of Issac Newton. 19 These works all have a com·
mon logical structure: personality is the explanans and creativity is the
explanandum.
These efforts represent a search for causal laws, while the approach to the
psychology of creative scientific thought that we are proposing is systemic.
Just as we are sceptical of the value of searching for one great moment of
insight, or intellectual conception, we are sceptical of the value of searching
for the causes of adult creativity in one infantile experience. There is not
some one special moment in the functioning of an evolving system that
explains all later moments. Each moment prepares the way for a new set of
possibilities, and in the system's interaction with its milieu, a new choice is
made, and so on indefmitely.
If the reader will retrace his path a few pages he will note that I spoke
of Darwin's seeming preoccupation with origins. In reality, Darwin made
little or no effort to trace out the particular origins of any organism or
group. He was preoccupied with describing the function of the system of
nature in order to account for its perpetual originality. Marc Bloch wrote
disparagingly of the 'idol of origins'. It cannot be thought that this great
historian was uninterested in the past, but from the evolving systems point
of view the past has no special privileges.

II. THE ORGANIZATION OF KNOWLEDGE AND

THE ORGANIZATION OF PURPOSE

To understand creative scientific thought as the working of an evolving

system, we must attack the joint problems of the organization of knowl·
edge and the organization of purpose. To be sure, our data will never be
more than behavior of the thinking person - including in that behavior
not only the fmished products of his work but also notebooks, conversa·
tions, correspondence, experiments, reading, etc. In spite of its richness,
this behavior is in principle a product of a far more complex structure: we
can think many thoughts but utter only a few of them. In going over the
Darwin notebooks, understanding and clarifying what he actually wrote was
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 309

a major part of the task, but in the long run we face the more difficult one of
reconstructing the underlying structure giving rise to it.

Behavior is Choice

Not only is behavior - even defined in my present, relaxedly cognitive way -

a drastic simplification of the possibilities latent in this underlying structure,
but also, behavior is sequential. It is a good approximation to say that we can
only do or say or write one thing at a time.
This necessary simplification in the behavioral record of the functioning of
a cognitive system should not tempt us into the error of describing the struc-
ture itself in the same simplistic terms. Darwin might choose to eliminate
homo sapiens from the Origin of Species, but we know from his early note-
books that this species was very much a part of his thinking whenever he
wrote about evolution. We need to describe cognitive structures in such a way
as to permit statements like: 'Darwin wrote only XY, but he was thinking
XYZ'. Or, 'Darwin in 1859 wrote X 3 , knowing that in earlier versions he had
written Xl and X 2 , aspiring to an improved version X 4 , and all the while
thinking XYZ'.
We come, then, to a central feature of the kind of structure we are looking
for: to be effective, out of its manifold possibilities, out of its pluralism, it
must perpetually re-organize and regulate itself in order to produce singular
outcomes. This is why the examination of the organization of knowledge and
the organization of purpose are ineradicably intertwined.
As we pursue our task a bewildering assortment of possibilities confronts
us. On the side of substantive knowledge we may list: cognitive maps, seman-
tic networks, associative chains, lists of themes, collections of images and
metaphors, etc.; and each one of these approaches produces a great profusion
of exemplars. On the side of procedural knowledge, there is an equally
bewildering array of conceptions: schemes, operations, transformations,
strategies, tactics, etc. When we tum to the organization of purposes, again a
profusion of possibilities awaits us: task hierarchies, projects, enterprises,
plans, hopes, visions, etc. Even the above tripartite division - substantive
knowledge, procedural knowledge, purpose - lends only a nominal and
transitory order to the array. Another person's purposes, or some representa-
tion of them, readily become part of my substantive knowledge, and all the
other classificatory boundaries are" crossed with equal ease.
How can we put some order in this seeming chaos? The answer I propose
took me by surprise when I first began to think of it: we should not! Consider
310 HOWARD E. GRUBER

the analogy of a chart room. Within any given chart, there is a marvelous
order, and what is more, a point-for-point correspondence between the chart
and some natural object, such as an island, or a continent, or a starscape. We
would not lightly change the relationships so carefully depicted within any
chart. For example, we would not, without considerable thought about our
map of the earth, interchange Greenland and Australia. Between charts, how-
ever, there is no such stable ordering. If we are investigating a biogeographical
hypothesis, we may overlay maps of the distributions of species on a map
depicting climatic changes. If we are investigating a certain geological hypoth-
esis, we do not hesitate to bring the coastlines of distant continents together.
I do not, of course, mean to suggest that cognitive structures are all maplike
in the sense of being spatialized representations of spatial phenomena.
Sometimes our charts may be ordered merely to make them accessible for
any purpose, as alphabetically. At other times they may be ordered for a
particular purpose. We might be tempted to speak of a collection of 'local'
organizations. But some of the representations we are discussing are hardly
local - they may be quite inclusive, even cosmic. It is better to speak of
special organizations, each one a representation constructed for some purpose.
A first phase, then, in deepening our grasp of the organization of knowl-
edge and purpose is to make a set of such 'charts'. Or at least to develop ways
of making them as needed. As I have argued, our approach to the problem of
representation must be pluralistic. We are not looking for one master chart
containing the individual's entire knowledge and purposes. In a recent exhibi-
tion at the Museum of Modern Art in New York City, a quarter of Tokyo,
Shinjuku, was depicted by means of more than 12 'experience maps' (geo-
graphical trails of different individuals, odor distributions, affective quality,
etc.). No one of them could be taken as capturing the 'essence' of Shinjuku.
Rather, the ensemble of maps represented some part of the complexity of a
few individuals' experience of that quarter. This approach corresponds to
trends in the field of information processing and computer simulation of
cognitive processes, where such representations are called 'frames' or 'scenes'
deliberately in order to emphasize their multiplicity.2o
We can expect the number of possible representations to be indefinitely
large: from the interplay of already existing representations of substantive
knowledge, procedural knowledge, and purposes a new representation can be
generated at any time. For example, the words 'take a closer look at that!'
describe an intention to enlarge a representation and to map more detail
onto it. From recent work on imagery, we know that the subjective 'size' of a
representation is a meaningful variable that has an experimentally measurable
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 311

effect on problem solving processes. 21 To this internal interplay of represen-

tations must be added representations arising from the action of new informa-
tion upon the system.
It may well be argued that underlying all these representations there is still
another structure that gives rise to them. Or, as I prefer, that the set of these
representations constitutes the whole of the cognitive system, that they
contain procedures within them for doing work on themselves, on each other,
and on new data. For the moment, probably a long one, this is not a theoreti-
cal choice we need to make. Representations such as these are what we have
some access to. Our next move must be to examine some of them, to which
I now turn.

The Individual's Network of Enterprises

Creative scientific thought is protracted work organized in a group of enter-

prises. Each enterprise, or undertaking, is itself a group of tasks. Some but
not all of these tasks are problematical. There is a mistaken tendency to
identify scientific thought with problem solving. The latter is only an occa-
sional process going on within the context of a larger group of cognitive
processes which the individual must orchestrate in order to do effective work.
Psychology has neglected the study of work, and there exists no theory for
conceptualizing a working life-time. It is proposed here that a person's work
is organized in a branching network of enterprises. The existence of such a
structure facilitates diverse simultaneous or parallel activities, occurring within
the same span of time and varying in the degree of their dependence upon
each other. This structure gives the individual choice as to the sequence and
timing of different facets of his work, permits him to re-activate a dormant
enterprise when he cannot progress along some other line, and gives continu-
ity to his total pattern of work.
I choose the term enterprise for three reasons. First, I want to distinguish
between enterprises and tasks. A task generally has definite limits. To take
an extreme case, digging a ditch across a field is a task - when the field is
crossed at the required depth, the task is done. Operating a farm is an enter-
prise. It is continuous, it has no definite limits, and at any time its scope is
subject to revision. An enterprise is usually a group of tasks. Indeed, having
decided to pursue some enterprise, the individual must first shape it by dis-
covering the tasks it entails. Of course, it often happens that a new enterprise
evolves out of work on a particular task while the latter is still clearly a part
of some other enterprise. Enterprises tend to be self-perpetuating: by the
312 HOWARD E. GRUBER

time the person has completed the feasible tasks entailed in the enterprise at
one point in time, a new collection of tasks has emerged to replenish the
inventory of work to be done.
Second, I want to distinguish between enterprises, tasks, and problems.
Not all tasks are problems. Again the ditch: for most farmers, digging a ditch
is not usually a problem, although admittedly sometimes a problem may be
encountered, such as a big rock. A creative life is one in which a group of
enterprises are organized in a structure that permits the individual to achieve
a creative end, or a series of them. Each enterprise entails a group of tasks,
some of which turn out to be problematical. In some cases a task may be
undertaken specifically in order to expose any problems it may involve, and
in the solution of those problems the enterprise as a whole will be stressed
and shaped. Solving a series of problems does not necessarily make a creative
life, as any crossword puzzle addict knows. The problems solved must be
organized as members of a coherent enterprise, leading to some novel and
effective product.
Third, I want to distinguish between enterprise and paradigm. In Kuhn's22
usage, a paradigm is a highly evolved, shared way of working, which only
emerges when considerable progress has been made along certain lines, and
when a socialized pattern of education, communication, and criticism has
developed. The enterprises of which I speak may be paradigmatic, pre-
paradigmatic, or in some obstinate cases even post-paradigmatic. Most likely,
in a creative life, some enterprises are shared and others quite unique to the
individual. I would add, however, that even if every enterprise in a given
person's network was a shared paradigm, the network as a whole would cer-
tainly be patterned and orchestrated in a unique way by that person, and
creative results might ensue.

Darwin's Network of Enterprise

As with any ongoing, evolving system, the choice of a beginning is arbitrary:

purposes evolve out of purposes. Taking Darwin as he appears during the
Beagle voyage, his network of scientific enterprises has two main branches,
biology and geology. This is clearly reflected in his notebooks of that period,
which divide easily into those two categories, geology having by far the
greater weight. (There is really a third branch, that of polymathic scientific
traveller and literary natural historian, modelled after his then hero, Alexander
von Humboldt. This is reflected especially in his Diary 23 kept during the voy-
age, which provided the basis for his first major publication, the celebrated
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 313

Journa[24 of the voyage. The diary and Journal are important because there
we see much of Darwin's imagery, and also because his observations and
reflections on homo sapiens wherever he went in his circumnavigation are the
beginning of another major branch of the network, his concern for the evolu-
tion of psychological processes in homo sapiens and other animals.)
The geological enterprise took definite shape during the Beagle voyage and
remained at least sporadically active throughout his lifetime, or from 1831-
1882. The biological enterprise, although having somewhat earlier beginnings
in his boyhood, grows more slowly during the five years of the voyage. By the
end of the voyage or shortly thereafter, both these major branches have been
transformed so that they are part of one much larger enterprise, the search
for one unifying theory of evolution and evolutionary approach to all of
nature; Darwin's interest in the human species and all other enterprises are
incorporated in this all-embracing enterprise.
From 1838 on it is clear that we must speak of three major branches of
the network: geology, biology, and psychology (or better, perhaps, sciences
of man). Darwin's quasi-literary interests recede into the background, but
do not disappear, giving rise to the much re-worked second edition of the
Journal,25 to his biography of his grandfather Erasmus Darwin,26 and to his
own autobiography;27 these writings are spread out over some 40 years.
Darwin's network of enterprise is characterized by a very high order of
continuity. All branchings are foreshadowed early, before they become mani-
fest as distinct enterprises. Once a branch is begun, it never becomes per-
manently dormant again.
I will give one example of the emergence of a 'minor' branch. During the
voyage, in 1835, Darwin became interested in the way in which organisms
transform the physical characteristics of the earth, the theme that life makes
land. At that early time, he worked out his highly successful theory of the
formation of coral reefs through the action of the coral organism, or rather,
through its interaction with a group of geological and climatic processes.
Shortly after the voyage, in May 1837, he presented this theory to the London
Geological Society. In November of 1837, among his manifold activities, he
read another paper before the Society, this time on ''worms forming mould,"
as he wrote in his personal journal. 28 His interest in worms continued through-
out his life, and in 1881 he published his book, The Formation of Vegetable
Mould through the Action of Worms, with Observations on Their Habits. 29 As
is evident from the title, he used this work to report on aspects of two differ-
ent branches of his network of enterprise - on the one hand, his peculiar
blending of geological and biological considerations in the "life makes land"
314 HOWARD E. GRUBER

motif; on the other hand, his prolonged search for rudiments of intelligent
behavior in lower organisms.
In addition to its continuity, Darwin's network of enterprises is also
remarkable for its great coherence. From 1838 on, almost everything he does
is relevant to this evolution enterprise. Nevertheless, he conducted many of
his enterprises in such a way that they would be valid contributions to the
science of his day, independent of (and for a long time, without mentioning)
their relevance to his views on evolution.
As I have described in Darwin on Man, he was a pastmaster at the separa-
tion of issues when this would serve his purposes. His eight-year study of
barnacles was certainly guided by evolutionary theory, but when he published
his four classic monographs on the subject, he did not refer to this theoretical
base. On the other hand, when he wrote the Origin of Species 30 he exploited
the material he had so laboriously worked up. Because he could not solve
fundamental genetic questions, he managed to separate the issue of the cause
of variation from the issue of the role of variation in evolution. Because he
felt the time was not ripe in 1859, when he wrote and published the Origin,
he managed to separate the question of human origins from evolution in
general. He kept his silence for twelve years, publishing the Descent of Man 31
in 1871.
Without question, then, Darwin's clear grasp of his own network of enter-
prise permitted him to plan his work purposefully, to concentrate his thinking
on different subjects in a flexible and adaptive way, and to time his publica-
tions strategically. But the protracted parallel activity of different branches of
the network also serves other functions, perhaps more profoundly related to
the process of creative work. When diverse activities are coeval they have a
greater chance of influencing and enriching each other. In Piaget's language,
the mutual assimilation of schemes is one of the major mechanisms of intel-
lectual growth. Resuming work on a lapsed enterprise permits the magnifica-
tion of this effect: a technique or a style of thought developed in one context
may be seen as useful in another. In Darwin's case, as Ghiselin and I have
independently pointed out, his formulation of the theory of coral reefs bears
a striking formal resemblance to the theory of natural selection, as he con-
structed it three years later. 32 The transfer of ideas from one domain to
another is facilitated by the organization of work in separate enterprises
which can lie fallow and be re-activated in complex temporal patterns. In
general I believe it gives a more faithful picture of Darwin's thinking to em-
phasize his purposeful, conscious control of his own intellectual work, but
it should not surprise us to find major exceptions. The above example is a
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 315

striking case. The parallels are very close, and even the language is sometimes
almost identical, but nowhere does Darwin indicate any conscious awareness
of the similarity between the two theories.
In their seminal work, Plans and the Structure of Behavior, Miller, Galanter,
and Pribram 33 begin by introducing the distinction between Image and Plan.
Images have to do with the organization of knowledge and plans refer to the
organization of behavior. Almost the entire book is about plans: "A Plan is
any hierarchical process in the organism that can control the order in which
a sequence of operations is to be performed".34 When one part of a plan is
completed, the system as a whole (i.e., the larger plan of which it is a part)
is so organized as to guide the inception of the next part of the plan (or
subroutine, or schema - terms that are practically equivalent to plan). In an
important sense, the plan is conceived of as closed - it has a built-in mecha-
nism for detecting when execution up to some criterion has been completed,
and at each level of the hierarchy the same closure characteristic can be
found. Thus, plans that exist control behavior, but plans are not in themselves
generative. Nor are plans equivalent to purposes: rather, plans are what we
must have in order to carry out our purposes.
In contrast, the concept of enterprise is open-ended and generative. The
network of enterprises describes the individual's organization of purposes. Of
course, procedural knowledge (plans) is embedded in every enterprise. But
when the running off of a plan comes up against obstacles, new procedures
must be invented. How the individual decides whether to struggle with such
difficulties or to shift to some other activity is regulated by the organization
of purposes as a whole.
Images of Wide Scope
Although Miller et al. were aware of the need for a system capturing the inter-
play between knowledge and action, they abandoned the Image and pursued
Plans. I believe this may have been because they conceived of the 'image of
the world' as singular, as one great organization embracing all the individual's
knowledge. This is probably a self-defeating idea, unrealizable and well worth
abandoning. But in the years since this work appeared great effort has been
expended on ways of representing knowledge. As I have indicated above, I
believe that a pluralistic approach is most promising. Each individual must
have at his disposal a number of modalities of representation. Systems of
laws, taxonomic systems, and thematic repertoires - such as those explored
by Gerald Holton 35 - are all pertinent. In the present essay I take up only
one such idea, images of wide scope.
316 HOWARD E. GRUBER

To contribute to our pluralistic scheme, we need some conceptual tool less

comprehensive than a single image of the world. On the other hand, we need
something more comprehensive than images which simply represent concrete
objects, as has been the case with the images studied by experimental psy-
chologists. For a long time, under the repressive influences of behaviorism
and positivism, psychologists neglected the whole study of imagery. In the
recent 'cognitive revolution' there has been a resurgence of interest in it. For
the most part, however, the images dealt with are the rather narrow images of
concrete objects, objects of the sort that would be accessible to the mecha-
nisms of ordinary perceptual experience. Francis Galton 36 was one of the
great initiators of this style of work, with his famous 'breakfast table ques-
tionnaire' in which he asked many eminent individuals, including his cousin,
Charles Darwin, to describe the vividness and clarity of their visual imagery,
using the familiar breakfast table as a point of departure. Galton wrote that
he was "amazed" at the "feeble powers of visual representation" he found
among the scientists he studied. Nevertheless, he clung to the view that the
"visualising faculty" is important in creative thought, including science, and
that methods of training it should be sought. In any case, Darwin's response
to the questionnaire,37 together with his images of wide scope which I will
discuss below, suffice to place him in the category to which Galton referred
as follows: "The highest minds are probably those in which it (the visualising
faculty) is not lost, but subordinated, and is ready for use on suitable occa-
sions" .38
It is almost 100 years since Galton's foray into this complex subject. Al-
though we know a great deal more about visual imagery taken in the narrow
sense, we know very little about the interrelationships of different types of
representation and their roles in the cognitive economy as a whole. This may
well be due to an overwhelmingly empiricist view of the nature of thought. If
thought begins with sense data, and relics of these data in the form of sensory
images are the main stuff of thought, it is of no great consequence to ignore
more abstract forms of representation. Even Piaget and Inhelder,39 although
they have introduced a valuable distinction between operative and figurative
imagery, remain within the domain of such concrete, quasi-perceptual objects.
I have chosen the term 'image of wide scope', rather than 'metaphor' or
'organizing metaphor' to emphasize their flexibility and generativity, and
their regulative function. I use 'images of wide scope' to emphasize their
constructive function in the making of ideas. The metaphor points out the
likeness between the things being compared. The image is 'there' - available
to the thinker for his inspection. It is complex, so that when he inspects
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 317

it he may surprise himself by discovering unsuspected properties and relation-

ships. Such images are at least partially independent of the particular meta-
phoric use to which they are put, while the metaphor only comes into being
in the comparison. For example, Darwin's image of war existed independent
of his particular use of it in the theory of evolution through natural selection.
He had considerable opportunity to observe aspects of war during his travels,
especially the wars of extermination waged by the Spanish against the Indians
and by the British against the Aborigines. He used the image of war in describ-
ing the shaping of a coral reef by the action of the sea.40 Thus, the image
existed independent of the specific metaphoric comparison it entered into.
In discussions of metaphor it is common to distinguish between the subject
and the modifier. For example, in one phase of Darwin's argument, natural
selection was a subject and war was a modifier. In this sense, images of wide
scope are potential modifiers. It is probably true that almost anything could
serve such a function; but for a given person, some materials are 'potentiated',
charged with significance, available to this person for that function.
In his paper for this volume, Science and Analogy, J. D. North has ad-
dressed himself to similar problems, but in a quite different way. He is inter-
ested in sets of resemblances between rather carefully worked out scientific
theories about different things (e.g., Newton comparing light and sound).
The success of the operation of making such an analogy depends on fmding
a high degree of isomorphism between the entities or theories considered. In
contrast, images of wide scope are loosely coupled to their subjects in two
senses. First, as already indicated, they can be joined to different subjects in
different metaphoric uses. Second, there is no demand for a tight fit: when
the thinker notices a mismatch between image and thing imaged, he can seize
on this imperfection to illuminate a point. For example, Darwin could note
that artificial selection operates over far shorter periods of time than natural
selection, and produces correspondingly less spectacular results; or he could
notice that artificial selection, because it is focussed on narrowly defmed
human objectives, is less powerful a force than natural selection. Blind to any
such objectives, the latter operates on the fitness of the whole organism. In
spite of these limitations, the image of artificial selection stood Darwin in
good stead. It brought out the quasi-experimental character of his theory, and
it accentuated the cumulative effects of the continued operation of selection
for a particular trait.
It is because of these ubiquitous mismatches that we need models as well
as metaphors and images. The image of wide scope is quasi-perceptual, in
some way linked to something that really exists. But no thing is exactly like
318 HOWARD E. GRUBER

anything else, nor is it often conveniently like anything else in all the ways we
might need for some particular scientific purpose. For this reason, a versatile
repertoire of images is valuable for exploring the properties of the phenom-
enon that interests us. When we have done this exploration, we may know
enough about the phenomenon to construct a model. Because the model is
completely uncoupled from any reality, we can build into it any properties
we like, thereby coupling it as tightly as possible to our knowledge of the
phenomenon or process in question. In the cognitive economy as a whole,
each modality of representation has its contribution to make.

Darwin's Images of Wide Scope

At least six important images were entailed in the construction of Darwin's
theory of evolution through natural selection. For reasons of space I will only
discuss four of them here. Each had its specific contribution to make to
Darwin's argument. The reader is referred to my other writings for more
detail. 41 Artificial selection and war, referred to above, are the best known of
these images. They are not necessarily the most important.
The image that first caught my attention was the tree of nature. "Organized
beings represent a tree, irregularly branched", Darwin wrote in his first trans-
mutation notebook, in July of 1837, some fourteen months before he finally
constructed a solid sketch of the theory of evolution through natural selec-
tion. As I have written elsewhere:

Among all his metaphors, Darwin's image of the tree of nature as an irregularly branch-
ing tree certainly deserves pride of place. It appears early in the B-notebook ... and is
then quickly redrawn to bring out Darwin's thought more precisely. Over the years,
Darwin drew a number of tree diagrams, both trying to perfect it and to penetrate it -
to learn what his own imagery could tell him. In a highly formalized version, the tree
diagram is the only figure in the Origin, and Darwin refers to it over and over, through-
out the book.42

The tree diagram helped Darwin see and formulate a number of points. It
depicts speciation and evolutionary divergence. It explains why gaps (so-called
'missing links') among contemporaneous species are not an argument against
evolution, since, in a branching structure, continuity in evolutionary time can
be associated with great gaps among contemporaries. Darwin used the tree
image to explore and expound his profound conviction that living nature is
irregular. Viewed in another way, any tree diagram is a model of exponential
growth; combined with a constraint, such as a limit on the number of organ-
isms, a formal principle of selection necessarily follows. By formal I mean
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 319

that although no mechanism is specified, the occurrence of selection follows

as a conclusion from the premises. Darwin had reached a point in the growth
of his thinking where he grasped this formal principle of selection at least two
months before he read (or re-read) Malthus and experienced his insight into
the mechanism of natural selection.
Even from this brief account one can see the generative power of the tree
image, how densely it is packed with potential meanings. But, it must be
added, such images and their families of possibilities do not exist in vacuo.
The tree image was constructed by the one person who could explore it and
discover its possibilities. He could do so because he worked from a unique
point of view, the growth of which in the years preceding 1837-38 I have
described in the book, Darwin on Man.
The fourth great image is the tangled bank of the final paragraphs of the
Origin of Species. If the tree image captures the wild and explosive power
of life as it evolves, the tangled bank captures the feeling of the endless
complexity of interrelationships, the surging confusion and struggle among
contemporaries, changing from moment to moment, always bringing new
juxtapositions and possibilities into play. In a clearly recognizable form, the
image of the tangled bank occurs very early in Darwin's thought, during the
first year of the voyage of the Beagle, or 27 years before he wrote the Origin.
It is remarkable that the images of artificial selection and war have been so
widely taken as carrying the essentials of Darwin's thought. These are only
the simplifying images, dividing the world into losers and winners, or dividing
behavioral responses into failures and successes. For Darwin himself, the
complexifying images of the irregularly branching tree and the tangled bank
were more primordial and at least equally important to him. The most charac-
teristic thing about Darwin was the energy and patience with which he exerted
himself to comprehend a very dense and complex view of the system of evolv-
ing nature, without making the least sacrifice before the gods of simplicity.
These aspects of nature are reflected in his own thought, for it was in the
intricate interplay of all these images that he formed his theory.
In this essay I have focussed attention on two great regulating systems
within which thinking takes place, the organization of knowledge and the
organization of purpose. There is, of course, a third, the organization of affec-
tivity, which would be essential to include in any thorough case study. In my
essay in Aesthetics in Science 43 I have discussed the way in which the neglect
of an essential part of Darwin's imagery corresponds to certain constrictive
attitudes towards aesthetic experience and affectivity as they relate to scien-
tific thought. The relation between the organization of knowledge and the
320 HOWARD E. GRUBER

organization of affectivity is a central theme of Camille Burns' case study of

the development of Mary Wollstonecraft's thinking.44
The reader who is acquainted with the book, Darwin on Man: A Psycho-
logical Study of Scientific Creativity, may be puzzled by the difference in
approach between that work and the present essay. The two approaches are
complementary. In the book, I trace out the details of a thought process as
Darwin moved from one structuring of ideas to another, especially during the
period 1837-1838, when he worked out the theory of evolution through
natural selection. Here, I have tried to outline some of the regulatory mecha-
nisms that must guide such work.
We see that the cognitive economy as a whole is full of complexity. There
are many types of structures, and each one has a dense and complex internal
structure. In the past, psychologists have sought to explain thOUght and other
cognitive processes by searching for a few simple and very general laws. As I
have tried to show, we must re-calibrate our intellectual aspirations, learn
how to look for and how to describe organized complexity. This will be a
long and hard effort.
It is perhaps not too much to say that the scientific culture that oversim-
plifies Darwin is part of a larger civilization that has elevated fragmentation
and Simplification to high principles for the conduct of life. Is a job interest-
ing and complex, placing a demand on the intellect and character of a person?
Break it up into many jobs that will make no such demands! Is some nuance
of nature unnecessary to the life of this society of simplified human beings?
Uproot the tree, fill in the marsh, cover the earth with cement!
Nor is it too much to say that in the struggle toward something better for
our descendants we need a theory of intellectual functioning that enjoys and
does justice to human complexity.

NOTES
* I am grateful to Martha E. Moore-Russell and to Doris Wallace for their helpful com-
ments on a draft of this paper.
1 Bloch, M., The Historian's Craft (Random House, New York, 1953), p. 194.
2 Limoges, C., La selection naturelle: etude sur la premiere constitution d'un concept
(1837-1859) (Presses Universitaires de France, Paris, 1970).
3 Gruber, H. E., Darwin on Man: A Psychological Study of Scientific Creativity; to-
gether with Darwin's Early and Unpublished Notebooks, transcribed and annotated by
Paul H. Barrett (E. P. Dutton, New York, 1974; 2nd ed. University of Chicago Press,
Chicago, 1980).
4 See for example, Gillithon, Ch., Decalages et seriation (Archives de Psychologie,
Monograph No.3, 44, 1976).
 COGNITIVE PSYCHOLOGY, SCIENTIFIC CREATIVITY 321

5 Newell, A., 'You Can't Play 20 Questions with Nature and Win', in Visual Information
Processing, Chase, W. G. (ed.) (Academic Press, New York, 1973).
6 Moore-Russell, M. E., 'John Locke: The Development of a Philosopher as a Person-
in-Society', Proceedings of the Sixth Annual Symposium of the Jean Piaget Society
(Philadelphia, 1976).
7 Psychologists attempting to fractionate creativity into component abilities have la-
belled one factor 'divergent thinking', the ability to generate a large number of potential
solutions to a problem, usually a simple problem such as, 'Think of as many uses as
possible for a brick'. See Guilford, J. P., The Nature of Human Intelligence (McGraw-
Hill, New York, 1967).
8 Employed with sensitivity, statistical tools might be helpful in detecting unusual
events, so that those of us who wish could pursue the rare bird wherever it flies, and
study it close at hand.
9 See Erikson, E. H., Childhood and Society, second edition (W. W. Norton, New York,
1963); Inhelder, B. and Piaget, J., The Growth of Logical Thinking from Childhood to
Adolescence (Basic Books, New York, 1958); Gruber, H. E. and Voneche, J. J., 'Re-
flexions sur les operations formelles de la pensee', Archives de Psychologie 44 (1976),
45-55.
10 Drake, S., Galileo Studies: Personality, Tradition, and Revolution (University of
Michigan Press, Ann Arbor. Mich., 1970).
11 Ibid., p. 3.
12 Ibid.,p.5.
13 Kramer, H., 'Writing Writers' Lives', New York Times Book Review, May 8, 1977,
p.3.
14 Steinberg, M., in New York Times Book Review, December 25, 1977, p. 21, review
of Solomon, Maynard, Beethoven (Schirmer/Macmillan, New York, 1977).
15 Friedliinder, S., Histoire et psychanalyse: essai sur les possibilites et les Iimites de la
psychohistoire (Seuil, Paris, 1975).
16 Freud, S., Leonardo da Vinci and a Memory of His Childhood (W. W. Norton, New
York, 1964) (originally pub!. 1910).
17 Greenacre, P., The Quest for the Father: A Study of the Darwin-Butler Controversy,
as a Contribution to the Understanding of the Creative Individual (International Univer-
sities Press, New York, 1963).
18 Erikson, E. H., Young Man Luther: A Study in Psychoanalysis and History (W. W.
Norton, New York, 1958).
19 Manuel, F. E.,A Portrait of Isaac Newton (Belknap Press of Harvard University Press,
Cambridge, Mass., 1968).
20 See for example: Minsky, M. and Papert, S.,Artificial Intelligence (Condon Lectures,
Oregon State System of Higher Education: Eugene, Oregon, 1973).
21 See for example: Kosslyn, S. M., Murphy, G. L., Bemesderfer, M. E., and Feinstein,
K. J., 'Category and Continuum in Mental Comparisons', Journal of Experimental
Psychology: General 106 (1977), 341-375.
22 Kuhn, T. S., The Structure of Scientific Revolutions, 2nd edition enlarged (The
University of Chicago Press, Chicago, 1970).
23 Darwin, C., Charles Darwin's Diary of the Voyage of H.M.S. 'Beagle' (edited from the
MS by Nora Barlow) (Cambridge University Press, Cambridge, 1934).
24 Darwin, C., Journal of Researches into the Geology and Natural History of the
322 HOWARD E. GRUBER

Various Countries Visited by H.M.S. Beagle (Colburn, London, 1839).

2S Darwin, C., Journal of Researches into the Natural History and Geology of the
Countries Visited during the Voyage of H.M.S. Beagle round the World, 2nd edition
(John Murray, London, 1845).
26 Krause, E., Erasmus Darwin, with a preliminary notice by Charles Darwin (Appleton,
New York, 1880). [The preliminary notice is a 127 pp. biography. - HEG].
27 Darwin, C., The Autobiography of Charles Darwin, 1809-1882. With original omis-
sions restored (edited and annotated by his grand-daughter Nora Barlow (Collins, Lon-
don, 1958).
28 These papers are reprinted in The Collected Papers of Charles Darwin, ed. by Paul H.
Barrett, 2 vols. (University of Chicago Press, Chicago, 1977).
29 Darwin, c., The Formation of Vegetable Mould Through the Action of Worms with
Observations on Their Habits (John Murray, London, 1904).
30 Darwin, C., On the Origin of Species by Means of Natural Selection (John Murray,
London, 1859).
31 Darwin, C., The Descent of Man and Selection in Relation to Sex (John Murray,
London, 1871).
32 Gruber, H. E. and Gruber, V., 'The Eye of Reason: Darwin's Development during
the Beagle Voyage', Isis 53 (1962), 186-200; Ghiselin, M. T., The Triumph of the
Darwinian Method (University of California Press, Berkeley, 1969).
33 Miller, G. A., Galanter, E., and Pribram, K. H., Plans and the Structure of Behavior
(Holt, Rinehart and Winston, New York, 1960).
34 Ibid., p. 16.
3S Holton, G., Thematic Origins of Scientific Thought: Kepler to Einstein (Harvard
University Press, Cambridge, Mass., 1973). See also, Holton, G., 'On the Role of Themata
in Scientific Thought', Science 188 (1975), 328-334.
36 Galton, F., Inquiries into Human Faculty and Its Development (Dent, London, n.d.;
first published 1883).
37 Darwin, C., The Life and Letters of Charles Darwin (ed. by his son, Francis Darwin),
3 vols. (John Murray, London, 1887), Vol. 1, pp. 238-239.
38 Galton, op. cit., p. 61.
39 Piaget, J. and Inhelder, B., Mental Imagery in the Child (Basic Books, New York,
1971).
40 Darwin, C., On the Structure and Distribution of Coral Reefs (Ward, Lock, and
Bowden, London, 1890; first published 1842), p. 21, p. 24, and p. 53.
41 Gruber, H. E., 'The Fortunes of a Basic Darwinian Idea: Chance', in The Roots of
American Psychology: Historical Influences and Implications for the Future, ed. by
R. W. Rieber and K. Salzinger Annals of the New York Academy of Sciences 291
(1977); and Gruber, H. E., 'Darwin's "Tree of Nature" and Other Images of Wide Scope',
in Aesthetics in Science, ed. by J. Wechsler (MIT Press, Cambridge, Mass., 1978).
42 Gruber, 'The Fortunes of a Basic Darwinian Idea ... " p. 236.
43 Gruber, 'Darwin's "Tree of Nature" ... '.
44 Burns, C., 'A Case Study of Mary Wollstonecraft: The Development of Her Feminist
Thought', Proceedings of the Sixth Annual Symposium of the Jean Piaget Society
(Philadelphia, 1976).
 BIOGRAPHICAL NOTES

JOSEPH AGASSI is Professor of Philosophy at Boston University and Tel-Aviv Univer-

sity. He is a Fellow of the American Association for the Advancement of Science, a
Correspondent Member of the Academie Intemationale de Philosophie des Sciences,
and a former Senior Fellow of the Alexander von Humboldt Stiftung and the Zentrum
fUr Interdisziplinare Forschung, Bielefeld. He has published about 200 items in the learned
press, including Towards an Historiography of Science, Theory and History, Beiheft 2
(1963, fascimile reprint, Wesleyan U.P., 1967); The Continuing Revolution: A History of
Physics from the Greeks to Einstein (McGraw-Hill, N.Y., 1968); Faraday as a Natural
Philosopher (Chicago U.P., 1971); Science in F1ux (Reidel, Dordrecht and Boston,
1975); (with Y. Fried), Paranoia: A Study in Diagnosis (Reidel, Dordrecht and Boston,
1976); Towards a Rational Philosophical Anthropology (Nijhoff, The Hague, 1977);
Radiation Theory (forthcoming); Science and Society (forthcoming); The Philosophy
of Technology (forthcoming).

LUIGI BELLONI is Professor of History of Medicine at the University of Milan. Among

his published books are Documenti Bassiani (Milan, 1956); Storia della medicina a
Milano (Milan, 1958-62); Opere scelte di Marcello Malpighi (Turin, 1967); Essays on
the History of Italian Neurology (ed.) (Milan, 1963); editor of many medieval medical
texts.

VINCENZO CAPPELLETTI is Professor of History of Science at the University of

Rome. He is Director of the Istituto dell'Enciciopedia Italiana (Rome), President of
the Domus Galilaeana (Pisa), Director of the International School of History of Science
(,Ettore Majorana' Centre for Scientific Culture, Erice), Scientific Adviser to the Italian
Prime Minister. Author of many items especially in the field of history of biology and
psychology. Among his published books are Entelechia. Saggi sulle dottrine biologiche
del secolo decimonono (Florence, Sansoni, 1967); Freud. struttura della metapsicologia
(Bari, Laterza, 1973); La scienza tra storia e societa (Rome, Studium, 1978).

GUIDO CIMINO is Assistant Professor of History of Science at the University of Rome

and a member of the editorial staff of the Enciclopedia del novecento, Istituto della
Enciciopedia Italiana, Rome, since 1975. He works especially in the field of history of
biology and psychology and among his publications is Problemi e momenti del pensiero
neurologico nel XIX secolo (Rome, forthcoming).

SALVO D'AGOSTINO is Associate Professor of History of Physics at the University

of Rome. His historical researches have concentrated on nineteenth century electro-
dynamics, and he has published articles on Maxwell, Lorentz, Hertz, in Scientia, Physis
and Historical Studies in the Physical Sciences.

323
M. D. Grmek, R. S. Cohen, and G. Cimino (eds.), On Scientific Discovery, 323-325.
Copyright © 1980 by D. Reidel Publishing Company.
324 BIOGRAPHICAL NOTES

GABRIEL GOHAU is a Teacher of Biology and Geology at the Lycee Janson-de-Sailly

(Paris). He is the author of Biologie et biologistes (Magnard, Paris, 1978), and of some
papers on pedagogy of science.

MIRKO D. GRMEK is Professor of History of Medicine and Biological Sciences at the

Ecole des Hautes Etudes, IV Section (Sorbonne), in Paris. He is Secretary of the Inter-
national Academy of the History of Sciences, Scientific Director of an International
Encyclopaedia of Sciences and Editor of a new review History and Philosophy of the
Life Sciences. He has been Editor of Archives intemationales d'histoire des sciences.
Among his published books are Santorio (Zagreb, 1952); On Ageing and Old Age (The
Hague, 1958); Introduction to Medicine (Zagreb, 1961); Mille ans de chirurgie en
occident (Paris, 1966); Raisonnement experimental chez Claude Bernard (Paris, 1973).

HOWARD E. GRUBER is Professor of Psychology and Director of the Institute for

Cognitive Studies at Rutgers University, Newark. His book, Darwin on Man: a Psy-
chological Study of Scientific Creativity (E.P. Dutton, New York, 1974) received the
Phi Beta Kappa Award for books in science, and other awards. He is author and editor,
with Jacques Voneche, of The Essential Piaget (Basic Books, New York, 1977).

JOHN NORTH is Professor of the History of Philosophy and the Exact Sciences at the
Rijksuniversiteit Groningen (The Netherlands). He is Editor of the Archives lntema-
tionales d'Histoire des Sciences, the journal of the Academie Intemationale d'Histoire
des Sciences, of which he is a member. His publications include The Measure of the
Universe (Oxford U.P., 1965) and Richard of Wallingford (Oxford U.P., 3 vols, 1976).

MARCELLO PERA is Associate Professor of Philosophy of Science at the University of

Pisa. He is the author of lnduzione e metodo scientifico (Editrice Tecnico Scientifica, Pisa,
1978) and of several papers on the logic of scientific discovery. He has edited and intro-
duced a collection of essays of H. Feigl,Induzione e empirismo (Armando, Rome, 1979).

DEREK DE SOLLA PRICE is Avalon Professor of the History of Science at Yale Univer-
sity. In 1971, when the International Council for Science Policy Studies was founded, he
became its first president. In 1976 he was the recipient of the Leonardo da Vinci medal,
the major award of the Society for the History of Technology. He has published nearly
200 scientific papers and six books, including: Science since Babylon, enlarged edition,
1975; Little Science, Big Science, 1964; Gears from the Greeks: The Antikythera
Mechanism - a Calendar Computer from ca. 80 B. c., 1975; and was editor of Science,
Technology and Society: A Cross·disciplinary Perspective, 1977.

GERARD RADNITZKY is Professor of Philosophy of Science at the University of

Trier, corresponding member of the Academie Internationale de Philo sophie des Sci-
ences; author of Contemporary Schools of Metascien'ce (Chicago, Gateway Editions,
3rd ed., 1973), Preconceptions in Research (Literary Services and Production, London,
1974), Epistemologia e politica di ricerca (Armando Armando, Rome, 1978) and more
than 60 papers, among them 'Justifying a Theory versus Giving Good Reasons for
Preferring a Theory', in Radnitzky, G. and Andersson, G. (eds.), The Structure and
Development of Science (Reidel, Dordrecht, 1979), pp. 213-256.
 BIOGRAPHICAL NOTES 325

JACQUES J. ROGER is Professor of the History of Science at the Sorbonne (University

of Paris I) and Director of the Centre International de Synthese. Among his publications
are Urspnmg der Formen und Entstehung der Lebewesen (Marburg, 1962), a critical
edition of Buffon's Epoques de la nature (Paris, 1962); Les sciences de la vie dans la
pensee tran~aise du XV/lIe siecie (2d ed., Paris, 1971), Un autre Button (Paris, 1978)
and several articles in the history of science and history of ideas published in American,
English, French and Italian reviews.

VITTORIO SOMENZI is Professor of Philosophy of Science at the University of Rome.

Editor of the volumes La filosofia degli automi (Boringhieri, Turin, 1965), La fisica
della mente (Boringhieri, Turin, 1969), L 'evoluzionismo (Loescher, Turin, 1971) and
L 'etologia (Loescher, Turin, 1979), he is the author of La scienza nel suo sviluppo
storico (E.R.I., Turin, 1960) and of papers in the fields of the history and philosophy
of physics, biology and cybernetics.

RICHARD TOELLNER is Professor of the History of Medicine at the University of

MUnster and Director of its Institute for the Theory and History of Medicine. Charac-
teristic for the work of the Institute is the consideration of the connection between the
history of science and philosophy of science and their application to the problems of
medicine as a science in the past and today. The emphasis of Professor Toellner's re-
search has been on the early modern period (l6th-18th centuries) and on this he has
contributed to journals of the history of medicine and of science. He is the author of,
among other works, Albrecht von Haller. tiber die Einheit im Denken des letzten Univer-
salgelehrten (Steiner, Wiesbaden, 1971). He is an officer in numerous societies in his
field, including the WolfenbUttier Arbeitskreis fur Renaissanceforschung, the Deutsche
Gesellschaft zur Erforschung des 18. Jahrhunderts, the Lessing-Akademie WolfenbUttel
and Gesellschaft fur Wissenschaftsgeschichte.
 INDEX OF NAMES

Agassi, J. 38,58,97,99,162,163,323 Belloni, L. 274,275, 278, 323

Agostino, S. d' 292, 293, 323 Bemesderfer, M. E. 321
Albert, H. 61,65, 68, 69, 97,99 Benassi, E. 276
Alembert, J. L. R. d' 255 Bergmann, G. 49
Alibert, J. L. 265 Bergson, H. 36, 163
Ampere, A. M. 126,153,154,282-283, Berkeley, G. 122
284,289 Bernard, C. 2,6,12,14,15,18,19-20,
Andersson, G. 60, 61, 65, 97, 99, 22, 25, 27-28, 31, 39, 141, 142,
100,101,102,324 164,195-196,200,205
Andral, G. 262 Bernoulli, R. 255
Antiseri, D. 33,38,99,100 Bessel, F. W. 155
Anzieu, D. 38 Best, C. H. 23
Apel, K.-O. 50,69,99 Beveridge, W. J. B. 39
Aquinas, T. 122,135 Beverwick, J. van [called Beverovicius]
Arago, F. 154 256
Archimedes 23, 175, 216, 247, 256 Bilharz, T. 267, 277, 278
Aristotle 13, 22, 106, 134, 216, 229, Bizzozero, G. 262,270,273,278
230,243,244,256 Blanche, R. 192, 209
Ashby, W. R. 176 Bloch, M. 295,308,320
Astolfi, J. P. 201 Bohr, N. 77,162
Athena 19,199 Boirel, R. 39
Auenbrugger, L. 261 Bonomo, G. C. 270
Ayer, A. J. 60 Boring, E. G. 37,222,226
Bourgeois, C. 197, 209
Bachelard, G. 31, 38, 193, 204, 207, Bouvard, A. 155
209,210 Boyle, R. 235
Bacon, F. [Lord Chancellor of England] Bozzolo, C. 270,272,273
21, 30, 95, 108, 109, 110, 123, 142, Brecht, B. 242, 255
146,159,161,163,243 Brentano, F. 222, 226
Bacq, Z. M. 15, 38 Breuer, J. 213,219-220
Baer, K. E. von 231,233 Bridgman, P. W. 113
Baldini, M. 23,38 Brouncker, W. 231
Balsamo-Crivelli, G. 277, 278 Brown,J. 168,176
Banting, F. G. 23 Browne, Peter [Bishop of Cork and Ross]
Bassi, A. 265,278 122
Bayes, T. 156, 163 Briicke, E. W. 219
Bechler, Z. 162,163 Bruner, J. S. 37
Beethoven, L. von 306 Brunhold, C. 193, 194,201,209
Bellarmino, Cardinal Robert 104-108, Bruno, G. 104
110-112,113 Bucellati, L. 266,269
Bellone, E. 38 Buck, R. D. 164
327
328 INDEX OF NAMES

Biihler, K. 64,100 Darmstaedter, L. 188

Buffon, G. 205, 206 Darwin, C. 5, 12, 29, 146, 148, 162,
Bunge, M. 39, 67, 80, 97, 100, 135, 164,167,174, 175, 2l3, 296,297,
140,151,162,164,289,293 300, 305, 306-307, 308-309, 312-
Burns, C. 320,322 320,321,322
Butler, S. 167, 169, 307 Darwin, E. 313
Buzon,C.de 21,39 Day, J. P. 151, 152, 154, 162, 164
de Graaf, R. 229, 230-231, 233, 234
Cajetan, Cardinal 122 Debus, A. G. 255
Campbell, D. T. 168-169, 170-172, Dement, W. C. 175,176
174-175,176 Descartes, R. 95, 109, 110, 134, 142,
Campbell, L. 139 161, 164, 170, 215, 229, 235,
Canguilhcm, G. 30,39 241-242, 246-254, 255, 256, 257,
Cannon, W. B. 14, 39 258
Cappelletti, V. 323 Devoto, L. 274
Carmichael, R. D. 39 Diderot, D. 255
Carnap, R. 151-152, 162, 163, 164, Dockx, 1. 100
209 Dowdeswell, W. H. 209
Castaldi, L. 276 Drake, S. 304-305, 321
Cauchy, A. 131 du Bois-Reymond, E. 215,219,220,225
Caws, P. J. 39 Dubini, A. 262-267,268,269,274,276
Cazzaniga, A. 261, 274 Duhem, P. 16, 79, 83, 92, 94,223,226
Celsius, A. 205 Duncker, K. 37,39
Cesalpino, A. 24
Christina, Queen of Sweden 246 Eales, N. B. 184
Cimino, G. 323 Eccles, J. 14,97, 101, 168, 176, 177
Claparede, E. 37, 39 Eddington, A. 223
Clark, R. 305 Ehrenfels, C. 221,222,226
Cohen, I. B. 240, 254 Ehrlich, P. 35
Cohen, M. R. 145, 146, 164 Einstein, A. 47, 86, 88, 89, 99, 107,
Cohen, R. S. 100, 101,164,165 113, 211-212, 213, 214, 216, 217,
Cole, F. J. 184 220, 221-222, 223, 224, 225, 285,
Collip, J. B. 23 293,305
Colombo, R. 24 Elkana, Y. 149, 163
Colomiatti, F.-V. 271 Enriques, F. 39
Comte, A. 115 Erikson, E. 308,321
Concato, L. 270,271 Euclid 99
Copernicus, N. l3, 107, 108, 184, 187, Euler, L. 138
241 Evans, R. I. 176
Copleston, Edward [Bishop of LlandaffJ
137 Fabre, J. H. 35
Corneliani, G. 262 Fabricius, R., ab Aquapendente 24,243
Coulomb, C. A. de 124, l31, 138, l39 Fabrizio, G., d'Acquapendente, See
Craik, K. J. W. 169, 176 Fabricius, H., ab Aquapendente
Cremonini, C. 242 Falloppio, G. 230,233
Faraday, M. 124-125, 126, 127, 129,
Dante 106 131,283
 INDEX OF NAMES 329

Fawcett, H. 162 Gohau,G.210,324

Fechner, G. T. 215,218,220,222,226, Gola, D. 261
283 Golgi, C. 262
Feigl, H. 324 Granit, R. 39
Feinstein, K. J. 321 Grassi, B. 262, 268-270, 271, 273
Feldhaus, F. M. 188 Graziadei, B. 270,279
Fenoglio, l. 270 Green, G. 124
Feyerabend, P. K. 24, 39, 58, 59, Greenacre, P. 306,321
161-62,67,100,165 Gregory, R. L. 169,176
Feynman, R. P. 159-160, 164 Griesinger, W. 267,276,277
Figuier, L. 20 Griffith, B. C. 179
Fisher, R. A. 29-30,39 Grinfield, E. W. 137
Fitzer, W. 245 Grmek, M. D. 13, 22, 39, 55, 100,
Florkin, M. 30, 39 176,195,209,324
Fogel, L. J. 176 Gruber, H. E. 12, 37, 39, 40, 175, 176,
Forbes, J. D. 125 177,320,321,322,324
Fourastie, J. 39 Gruber, V. 322
Fourier, J. B. 124, 131-132, 139 Gruby, D. 265
Frada, G. 279 Griinbaum, A. 59,97,100
Frank, P. 67 Guilford, J. P. 321
Franklin, B. 158, 164 Gutenberg, J. 208
Freud, S. 211,213,214,220,221,225,
300,306,308,321 Habermas, J. 99
Fried, Y. 323 Hadamard,1. 37,40, 104, 176
Friedlander, S. 306,321 Haller, A. von 242,255,256
Frisch, K. von 14 Ham, J. 232,233
Hamilton, W. 125
Galanter, E. 315,322 Hanson, N. R. 24, 29, 40, 140, 157,
Galen 13, 24, 244 163, 164
Galileo 15, 22, 27, 57, 86, 104-107, Harding, S. G. 40
110-111, 134, 137, 184, 187,218, Harre, R. 100, 101
235,241,242,243,304-305 Hartsoeker, N. 232
Gallois, J. 231 Harvey, W. 12, 21, 22, 23-24, 32-33,
Galton, F. 316,322 37, 229-230, 231, 233, 234, 241-
Gambarini, A. 262 245, 248-254, 255, 256, 258, 259
Garden, G. 232 Hattori, K. 180,188
Garnett, W. 139 Heimann, P. M. 125,138
Garrison, F. H. 188 Heisenberg, W. 67,88,97, 100
Gauss, K. F. 283-284,286 Helmholtz, H. von 2, 6, 14, 219, 282
Gerlach, W. 241, 254 Hempel, C. G. 145, 163, 164, 209
Geymonat, L. 15,39 Henson, V. 19-20
Ghiselin, M. T. 314,322 Hertz, H. 217,287,288,323
Giere, R. N. 39 Hesse, M. B. 40
Gillieron, C. 297,320 Hippocrates 218,229,272
Gingerich, O. 39 Holmes, F. L. 12,40
Giordano, A. 273,279 Holton, G. 315,322
Goethe, J. W. von 213 Hooke, R. 30, 116-117, 136, 137,236
330 INDEX OF NAMES

Hooker, T. 137 Kosslyn, S. M. 321

Horne, J. van 230,233,234 Kotarbmski, T. 44,49
Horton, M. 40 Koyre, A. 103,305
Host, V. 201 Kramer, H. 306,321
Hiibner, K. 60 Krause, E. 322
Humboldt, A. von 312 Kuhn, T. 12, 19, 23, 24,40,59, 110,
Hume, D. 28, 52, 95, 170 149, 162, 200, 209, 281, 312, 321
Huygens, C. 121-122, 130, 137, 162,
218, 232 Laennec, R. T. H. 261,262
Lakatos, I. 12, 40, 85, 88, 100, 101,
Inhelder, B. 40, 296, 316, 321, 322 114,142,149,161,164,282
Ivanov, G. M. 40 Lamarck, J. B. 148,167
Laplace, P.-S. de 103-104, 131, 215
James, W. 172 Largeault, J. 40
James I 243 Larmor, J. 138
Jammer, M. 97,100 Lava, G. 279
Jaynes, J. 176 Lavoisier, A. 187,206
Jenner, E. 22 Lazerges, G. 209
Jevons, W. S. 28, 121-122, 137, 172 Leeuwenhoek, A. van 35, 229, 231,
Jewkes, J. 40 232,233,234,235,236
Joan of Arc 104 Leibniz, G. W. 232,242, 255
Jones, Ernest 219,220,226 Leonardo 308
Joyce, G. C. 137 Leroi, J. A. 266
Jung, C. G. 40, 211,213-214,222,225 Leroy, G. 204,209
Lesky, E. 255
Kant, I. 74, 95, 96, 97, 99, 100, 104, Lessona, M. 278
109,150,215,241,242,254 Leuckart, K. G. F. R. 268, 277
Kedrov, B. M. 40 Leverrier, U. J. 155
Kelvin, Lord (William Thomson) 116, Liebig, J. von 28
123-129, 130-132, 134, 135, 136, Limoges, C. 296,320
138, 139, 140, 162, 286 Lining, J. 158
Kepler, J. 21, 57, 105, 162, 241, 242, Linnaeus, C. 174,205
255 Lobachevski, N. I. 38
Kerckring, T. 230,231 Locke,J. 122,297,305
Kerenyi, C. 40 Loewi, O. 14,15,40, 175
King, William [Archbishop of Dublin) Lomonosov, M. V. 38
122, 137 Lorentz, H. A. 323
Kirchhoff, G. 289 Lorenz, K. 25, 40, 168, 169, 171,
Klein, M. 30 176,293
Knudsen, O. 125,138 Ludwig, C. F. W. 35
Koch, R. 268, 272 Luther, M. 308
Kohler, W. 172, 175, 176, 222-223, Lyell, C. 146
226
Koestler, A. 40 Mach, E. 2, 172, 176, 214-221, 223,
Kohlrausch, R. 285,287,292 225,226
Koningsveld, H. 163,164 Mach, L. 226
Kordig, C. R. 40, 163, 164 Machamer, P. K. 40
 INDEX OF NAMES 331

McMullin, E. 88 Navier, C. L. 131

Magendie, F. 37 Neumann, C. 293
Maggi, L. 268, 278 Newell, A. 177, 297, 321
Magini, G. A. 242 Newton, I. 21, 27,47,57,62,66,86,
Mahdihassan, S. 176 88, 89, 93, 107, 108, 116-122,
Maier, H. 37 123, 129, 130, 132, 134, 136, 137,
Maillet, B. de 205 139, 151, 162, 187,241,291,308,
Malpighi, M. 231,233 317
Malthus, T. R. 319 Nicolle, C. 14, 36,40
Manuel, F. E. 308,321 Normand, L.-A. 271
Manzoni, A. 261 North,J. D. 317,324
Mao Tse Tung 307
Margenau, H. 289, 293 Oersted, H. C. 154
Maxwell, J. C. 116, 123, 125-132,134, Ohm, G. 138,289
135,136,138,139,212,217,281, Olby, R. 14,41
285,286-293,323 Oldenburg, H. 116, 117, 136
Mayer, J. R. 2 Olschki, L. 255
Medawar, P. B. 40, 145, 146, 160-161, Olson, R. 139
164,177 Orsi, F. 269
Meinong, A. 222 Owens, A. J. 176
Mendel, G. 29-30,80
Mersenne, M. 249, 258 Pagel, W. 12,41, 255
Merton, R. K. 180, 188 Pagliani, L. 272
Metzger, H. 135, 140 Pallas Athena, See Athena
Meyerson, E. 28,40, 222 Panizza, B. 262
Michelson, A. A. 223, 290, 293 Papert, S. 321
Mikulinski, S. R. 40 Parona, C. 268-269,270,271
Mill, 1. S. 116, 123, 127, 129, 130, Parona,E. 269,270,271,273,278,279
132-135,140,151,164 Pasteur, L. 22, 35, 200, 268
Miller, D. 59,60,61,99,100 Paulesco, N. 22
Miller, G. A. 315,322 Pavesi, P. 278
Minsky, M. 321 Peirce, C. S. 141,157
Mirandola, Pico della 105 Pera, M. 162,164,324
Monod,l. 14,40 Perroncito, E. 264, 270, 271, 272,
Moore-Russell, M. E. 297, 305, 320, 273,278,279
321 Pettenkofer, M. 272
Morelli, C. 267, 277 Petzoldt, J. 225
Morgagni, G. B. 261 Pfeifer, H. 100
Morin, E. 35,40 Pfluger, E. F. W. 20,41,198
Morley, E. 223,290,293 Piaget, J. 25, 26, 41, 175, 177, 204,
Muller, J. 219, 223 296,314,316,321,322
Murphy, G. L. 321 Planck, M. 150,164
Murray, I. 22,40 Plato 209,218
Musgrave, A. 12,40, 85, 98, 100, 101 Pledge, H. T. 188
Plempius [PlempJ, V. F. 258
Nagel, E. 40, 145, 146, 164, 209 Poe, E.A. 209,210
Napoleon 103 Poincare, H. 170, 172, 173, 177
332 INDEX OF NAMES

Poisson, S. 131 Sawers, D. 40

Polanyi, M. 110 Scarpa, A. 262
Pontieri, G. 279 Schaff, A. 41
Popper, K. R. 12, 23, 24, 28, 30, 32, Schaffner, K. F. 41, 281, 289-291,
33, 34, 36, 41, 46, 49, 51, 53-55, 292, 293
58-59, 61, 62, 64, 68, 69-70, 71, Schatzman, E. 208,210
73-74, 76, 78, 80, 86,87, 88,91, Scheffler, I. 10, 11,41
95,96,97,98, 101,112,141,142, Scherer, J. J. 19
143-144, 145, 146, 147-148, 150, Schilpp, A. 41, 100, 101, 165, 176,
155-156, 161, 162, 163, 164, 165, 211,224
168, 169-171,177,198,209 Schlegel, P. M. 245,256
Porep, R. 41 Schleiden, M. 30
Postmann, L. 39 Schwann, T. 30
Pratt, J. H. 23,41 Selye, H. 14,41
Pribram, K. H. 177,315,322 Serveto, M. 21
Price, D. de S. 188, 303 Shaw, G. B. 104
Priessnitz, V. 265 Shaw, J. C. 177
Pruner, F. 267 Siebold, K. T. E. von 267
Ptolemy, c. 187 Silvestri, F. 269
Simon, H. A. 173-174, 177
Rabelais, F. 191-192, 202, 203, 209 Simonton, D. K. 180, 188
Radnitzky, G. 96, 97, 98,99,100,101, Small, H. G. 179
10~ 161, 16~ 165,324 Socrates 37
Raiberti, G. 265-266 Solomon, M. 321
Rainoff, T. J. 180, 188 Somenzi, V. 325
Ranke, L. von 9 Sommerfeld, A. 288, 293
Rasori, G. 261 Sonsino, P. 267, 277
Razzell, P. 22,41 Sormani, N.-M. 262
Redi, F. 35, 234 Sorokin, P. A. 180, 188
Reichenbach, H. 163, 165 Souriau, P. 172
Rey, A. 16,41 Spinner, H. 49
Riolan, J., the Younger 33, 233, 252, Stahl, G. E. 206
253 Steinberg, M. 321
Robinson, G. S. 60 Stendhal 9, 41
Roger, J. 210,237,324 Steno 230,233,234
Rossi, P. 41 Stenson, N., See Steno
Rothschuh, K. E. 242, 255,257 Stent, G. 41
Rousseau, J. J. 95 Stuewer, R. 292
Rumbaugh, D. M. 177 Stevin, S. 217,222
Rupke, N. A. 29,41 Stillerman, R. 40
Russell, B. 170, 305 Strachey, J. 225
Russo, F. 41,101 Suarez, F. 122
Rutherford, E. 162 Suppes, P. 209
Sviedrys, R. 293
Salamun, K. 101 Szent-Gyorgyi, A. 14,41
Salmon, w. 157,163,165
Sanga1li, G. 267,276, 277,278 Tarski, A. 64,97,209
Sarton, G. 9 Taton, R. 41
 INDEX OF NAMES 333

Terrell, G. 40, 177 Weizsiicker, C. F. von 67

Thom,R.26,41,174 Wertheimer, M. 37, 40, 41, 177, 221,
Thompson, S. P. 139 222-223, 226
Thomson, William, See Kelvin, Lord Westcott, M. R. 41
(William Thomson) Westfall, R. S. 39,116,136
Tichy, P. 60,61 Whateley, R. 123,129,132,137
Toellner, R. 254,255,325 Whewell, W. 123, 154
Tomita, T. 180,188 Whittaker, E. T. 116,138,292
Toulmin, S. 59, 101, 177 Wiederkehr, K. H. 292, 293
Toynbee, A. 179 Williams, L. P. 12,138
Triker, B. 1. K. 209 Williams, N. 189
Wisdom, J. 162, 165
Valobra, N. 278 Wittgenstein, L. 58,59
Vesalius, A. 13,24,230,233 Wollstonecraft, M. 320
Vesling,l. 33 Worden, F. G. 41
Virchow, R. 28, 37 Worrall, J. 85
Volante, G. 273 Wright, G. H. von 42, 57, 151, 162,
Voneche, 1. 1. 321,324 165
Wucherer, O. 267,277
Waerden, B. L. van der 30,41 Wundt, W. 215,222
Walden, P. 189
Wallace, A. R. 307 Yaroshevski, M. G. 40,42
Wallace, D. 320 Young,J.Z.169,177
Walsh,M.J.176 Yuasa, M. 180,188
Wartofsky, M. 100,165
Watkins, J. W. N. 50, 97, 101, 102 Zahar, E. 85,93
Watson, James D. 14,41 Zaidel, E. 177
Weber, W. 126,215,281-293 Zeiger, J. 101, 102
Wegener, A. 206 Zeus 19
Weiling, F. 30,41
 BOSTON STUDIES IN THE PHILOSOPHY OF SCIENCE

Editors:
ROBERT S. COHEN and MARX W. WARTOFSKY
(Boston University)

1. Marx W. Wartofsky (ed.), Proceedings of the Boston Colloquium for the Philosophy
of Science 1961-1962. 1963.
2. Robert S. Cohen and Marx W. Wartofsky (eds.), In Honor of Philipp Frank. 1965.
3. Robert S. Cohen and Marx W. Wartofsky (eds.), Proceedings of the Boston Collo-
quium for the Philosophy of Science 1964-1966. In Memory of Norwood Russell
Hanson. 1967.
4. Robert S. Cohen and Marx W. Wartofsky (eds.), Proceedings of the Boston Collo-
quium for the Philosophy of Science 1966-1968. 1969.
5. Robert S. Cohen and Marx W. Wartofsky (eds.), Proceedings of the Boston Collo-
quium for the Philosophy of Science 1966-1968. 1969.
6. Robert S. Cohen and Raymond J. Seeger (eds.), Ernst Mach: Physicist and Philos-
opher. 1970.
7. Milic Capek, Bergson and Modern Physics. 1971.
8. Roger C. Buck and Robert S. Cohen (eds.), PSA 1970. In Memory of Rudolf
Carnap. 1971.
9. A. A. Zinov'ev, Foundations of the Logical Theory of Scientific Knowledge
(Complex Logic). (Revised and enlarged English edition with an appendix by
G. A. Smirnov, E. A. Sidorenka, A. M. Fedina, and L. A. Bobrova.) 1973.
10. Ladislav Tondl, Scientific Procedures. 1973.
11. R. J. Seeger and Robert S. Cohen (eds.), Philosophical Foundations of Science. 1974.
12. Adolf Griinbaum, Philosophical Problems of Space and Time. (Second, enlarged
edition.) 1973.
13. Robert S. Cohen and Marx W. Wartofsky (eds.), Logical and Epistemological
Studies in Contemporary Physics. 1973.
14. Robert S. Cohen and Marx W. Wartofsky (eds.), Methodological and Historical
Essays in the Natural and Social Sciences. Proceedings of the Boston Colloquium
for the Philosophy of Science 1969·1972. 1974.
15. Robert S. Cohen, J. J. Stachel and Marx W. Wartofsky (eds.), For Dirk Struik.
Scientific, Historical and Political Essays in Honor of Dirk Struik. 1974.
16. Norman Geschwind, Selected Papers on Language and the Brain. 1974.
18. Peter Mittelstaedt, Philosophical Problems of Modern Physics. 1976.
19. Henry Mehlberg, Time, Causality, and the Quantum Theory (2 vols.). 1980.
20. Kenneth F. Schaffner and Robert S. Cohen (eds.), Proceedings of the 1972 Biennial
Meeting, Philosophy of Science Association. 1974.
21. R. S. Cohen and 1. 1. Stachel (eds.), Selected Papers of Leon Rosenfeld. 1978.
22. Milic Capek (ed.), The Concepts of Space and Time. Their Structure and Their
Development. 1976.
23. Marjorie Grene, The Understanding of Nature. Essays in the Philosophy of Biology.
1974.
24. Don Ihde, Technics and Praxis. A Philosophy of Technology. 1978.
25. Jaakko Hintikka and Unto Remes, The Method of Analysis. Its Geometrical Origin
and Its General Significance. 1974.
26. John Emery Murdoch and Edith Dudley Sylla, The Cultural Context of Medieval
Learning. 1975.
27. Marjorie Grene and Everett Mendelsohn (eds.), Topics in the Philosophy of Biology.
1976.
28. Joseph Agassi, Science in Flux. 1975.
29. Jerzy 1. Wiatr (ed.), Polish Essays in the Methodology of the Social Sciences. 1979.
32. R. S. Cohen, C. A. Hooker, A. C. Michalos, and J. W. van Evra (eds.), PSA 1974:
Proceedings of the 1974 Biennial Meeting of the Philosophy of Science Association.
1976.
33. Gerald Holton and William Blanpied (eds.), Science and Its Public: The Changing
Relationship. 1976.
34. Mirko D. Grmek (ed.), On Scientific Discovery. 1980.
35. Stefan Amsterdamski, Between Experience and Metaphysics. Philosophical Problems
of the Evolu tion of Science. 1975.
36. Mihailo Markovic and Gajo Petrovic (eds.), Praxis. Yugoslav Essays in the Philosophy
and Methodology of the Social Sciences. 1979.
37. Hermann von Helmholtz: Epistemological Writings. The Paul Hertz/Moritz Schlick
Centenary Edition of 1921 with Notes and Commentary by the Editors. (Newly
translated by Malcolm F. Lowe. Edited, with an Introduction and Bibliography,
by Robert S. Cohen and Yehuda Elkana.) 1977.
38. R. M. Martin, Pragmatics, Truth, and Language. 1979.
39. R. S. Cohen, P. K. Feyerabend, and M. W. Wartofsky (eds.), Essays in Memory
of Imre Lakatos. 1976.
42. Humberto R. Maturana and Francisco J. Varela, Autopoiesis and Cognition.
The Realization of the Living. 1980.
43. A. Kasher (ed.), Language in Focus: Foundations, Methods and Systems. Essays
Dedicated to Yehoshua Bar-Hillel. 1976.
46. Peter L. Kapitza, Experiment, Theory, Practice. 1980.
47. Maria L. Dalla Chiara (ed.),Italian Studies in the Philosophy of Science. 1980.
48. Marx W. Wartofsky, Models: Representation and the Scientific Understanding.
1979.
50. Yehuda Fried and Joseph Agassi, Paranoia: A Study in Diagnosis. 1976.
51. Kurt H. Wolff, Surrender and Catch: Experience and Inquiry Today. 1976.
52. Karel Kosik, Dialectics of the Concrete. 1976.
53. Nelson Goodman, The Structure of Appearance. (Third edition.) 1977.
54. Herbert A. Simon, Models ofDiscovery and Other Topics in the Methods of Science.
1977.
55. Morris Lazerowitz, The Language of Philosophy. Freud and Wittgenstein. 1977.
56. Thomas Nickles (ed.), Scientific Discovery, Logic, and Rationality. 1980.
57. Joseph Margolis, Persons and Minds. The Prospects of Nonreductive Materialism.
1977.
58. Gerard Radnitzky and Gunnar Andersson (eds.), Progress and Rationality in Science.
1978.
59. Gerard Radnitzky and Gunnar Andersson (eds.), The Structure and Development
of Science. 1979.
60. Thomas Nickles (ed.), Scientific Discovery: Case Studies. 1980.
61. Maurice A. Finocchiaro, Galileo and the Art of Reasoning. 1980.