Вы находитесь на странице: 1из 24

Extents and Limits of Serial Techniques

Author(s): Ernst Krenek

Source: The Musical Quarterly, Vol. 46, No. 2, Special Issue: Problems of Modern Music. The
Princeton Seminar in Advanced Musical Studies (Apr., 1960), pp. 210-232
Published by: Oxford University Press
Stable URL: http://www.jstor.org/stable/740372 .
Accessed: 13/07/2013 16:12

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Oxford University Press is collaborating with JSTOR to digitize, preserve and extend access to The Musical


This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions



T HE propensityof presentmusical theoryforterminology

belonging mathematicsand physicsis characteristicof a styleof
thinkingessentiallydifferentfrom earlier ways of viewing the subject
matter.Although some of this language sounds merelypretentious,it
has neverthelessadded usefultermsto musical discussion.One of these
is the conceptof "parameter."It was introducedintorecentmusictheory
by Dr. Meyer-Eppler,of the Instituteof CommunicationTheory at the
Universityof Bonn, who was associated with the work of the electronic
laboratoryof the West German Radio at Cologne. It is borrowedfrom
mathematics,whereit means "a variable enteringinto the mathematical
form of any distributionsuch that the possible values of the variable
correspondto differentdistributions."'
Serial organizationof a certainnumber of parametersof a musical
process causes a certain number of other parametersto be left uncon-
trolled.A detailed studyof the relationshipsof these two areas was the
purposeof the seminar.The titledid not, as was surmisedby some, hint
at a discriminationbetween accomplishmentsand shortcomings of serial

Serial music was definedas a method of compositionthat has been

developedas a sequel of the twelve-tonetechniqueinauguratedby Arnold
Schoenbergaround 1923. While the serial conceptin that techniquewas
embodied in the twelve-toneseries,i.e. an orderingof the pitchesto be
adhered to throughoutthe course of the composition,the new idea of
I American College Dictionary, New York, 1948, p. 879.


This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
Extentsand Limits of Serial Techniques 211

serialismencompassesall aspects (or "parameters") of the musical pro-

cess,such as timbre,dynamics,articulation,and above all, time,i.e. dura-
tionof the individualsoundingelementsand theirmutual relationships in
time,subordinatingall theseaspectsto premeditated serial In
thisview the twelve-tonetechniqueappears to be a special, or limiting,
case of serial music,similarto an interpretationof Newtonianmechanics
as a limitingexpression of the Special Theory Relativity,whichin turn
has been explainedas a limitingexpressionof that General Theory.

Anton Webern and Olivier Messiaen were mentionedas the best-
knowngeneratorsof the new way of serial thinking,the formerbecause
of the extraordinary impacthis workhas exercisedduringthe last twenty
yearsor so, the latterabove all throughhis experimentswith "rhythmic
rows" (or "modes") and his immediateinfluenceon such composersas
Boulez and Stockhausen.The discussionthen turnedto the significance
and consequenceof the gradual expansionof the musical area that was
subjected to premeditatedorganization.It was recognized that serial
orderingof the factorof time (i.e. premeditatedfixationof points of
entranceand durationof the individual musical elements) caused fun-
damental changesin thestructure, and mean-
appearance, perceptibility,
ing of music. Thereforethe largerpart of the investigationwas devoted
to the methodsof organizingseriallythe parameterof time.The discourse
was mainlybased on myown workin the serialstylebecause myintimate
knowledge of this work allowed succinct presentationof the relevant
details, whereas the few available analyses of other composers' serial
worksare frequentlyambiguousand far fromenlightening.


By rotationwe understanda procedurein which the elementsof a

givenseriessystematicallyand progressively change theirrelativepositions
according to a plan whichin is
itself seriallyconceivedin thatthe changes
occur in regularphases.l

I applied this principlefor the firsttime in a large choral work,

2 In his book, Die Komposition mit zw6ilfTiinen, Berlin, 1952, p. 113 ff. and
passim, Josef Rufer points out that Arnold Schoenberg occasionally let neighboring
tones of his rows exchange places, or groups of tones change their positions within
the row. Rufer's discourse and the examples quoted show that this was done
sporadically and mainly in order to create a musical context that would not have
been served as well by adhering to the premeditated succession of pitches.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
212 The Musical Quarterly
LamentatioJeremiaeProphetae,3writtenin 1940 and 1941. The twelve-
tone seriesof thisworkreads thus:

Ex. I --:--?---?-------
Each of its two constituentsix-tone groups is progressivelymodified
by making the firsttone the last:
Ex. 2


'I *.~~;~= ~;;r ~,

The patternsthus obtained may be called "diatonic" since they con-

tain the same six tones.The rosterof patternsis doubled by transposing
all those of the leftcolumn of Ex. 2 to begin on F, all thoseof the right
column to begin on B.

g~I). i ~ 0 1'



3 B~ienreiter-VVerlag,

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
Extents and Limits of Serial Techniques 213
These new patternsare "chromatic" because they eventuallyinclude
all twelvetones.The rotationtakingplace was inspiredby the construc-
tion of the Greek modal scales and theirtranspositioninto one "char-
acteristic"octave. The purposeof the operationwas not so much to make
the serial designstricter,
but ratherto relax it, insofaras the wide variety
of available six-tone patternsmade it possible to remain within the
frameof referenceof the twelve-toneserial techniquewithoutconstantly
having to use complete twelve-tonerows. Thus it became possible to
give various areas of the compositiondistinctiveharmonic flavors.At
that time no attemptwas made to organize seriallythe selection and
successionof the rotationalpatterns.
A more consistentand systematicapplicationof the principleof rota-
tion may be found in my orchestralwork, Circle, Chain and Mirror,4
writtenin 1956 and 1957 forthe Basel Kammerorchester. The tone-row
of this work reads as follows:

3 4 5 6 7 8 9 12
I 2 3 4 5 6 7 8 9 0II 12

In the course of the compositiontwenty-four derivativeformsof this

row are employed.The principleof derivationmay easilybe apprehended
by comparingthe originalrow with its firstthree derivativeforms(the
tones in theiroriginalsuccessionare numberedfrom1 to 12):
Ex. 5

I 3 2 5 4 7 6 9 8 II 10 12

3 5 2 7 4 9 6 II 8 Z 0o

3 5 7 Z 9 4 II 6 12 8 10

The rotation taking place here consistsin forminga retrogradesuc-

cession of each pair of two adjacent tones. Aftereleven such operations
one arrivesat the complete retrogradeformof the original statement.
The twelvefollowingderivatesrepresentthe retrogradeformsof the first
twelve,and the twenty-fifth transformation is identicalwiththe original.
The same procedure was applied to the invertedform of the original
series (see Ex. 6). This arrangementsuggestedthe "circle" part of the
titleof the work.

SOriginal German title: Kette, Kreis und Spiegel. Biirenreiter-Verlag,Kassel.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
214 The Musical Quarterly
The sequence in which the forty-eight rowsthusobtained were used
in the work was determinedby the decisionto have each originalform
followedby the second of the two formsof the inversionwhich would
have fortheirfirsttonesthe last tone of the precedingoriginal,while this
inversionin turnwould be followedby an originalformbeginningwith
the last tone of the precedinginversion.This interlocking
meant by the term "chain" in the title.The sequence of rows obtained
through this operation may be partiallyseen in the followingtable
(O=original, I-inversion, R=retrograde, RI-retrograde inversion):

_ 12

4 ....._ __

........ . ..
4 ...
....1 2
2 22

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
Extents and Limits of Serial Techniques 215
The symmetry resultingfromthisorganizationis obvious: The sequence
1, 6, 10, 4, 12 in lines 1 to 9 of the O column is identical with the
sequence 1, 6, 10, 4, 12 in lines 13 to 21 of the I column. The same
relationobtains as regardsthe sequences 12, 4, 10, 6, 2 in lines 4 to 12
of RI and 16 to 24 of R. The positionsof 18 between 01 and 6 and
of 08 betweenI1 and 6 are equally symmetricaland correspondto the
positions of the 8s in R and RI between 2 and 6 of RI and R
Ex. 6

I 8 3 10 6 5 12 2 II 4 9 7

~- "---
I 3 8 6 10 12 5 II 2 9 4 7

)C . .
_b ..
3 I 6 8 12 I0 II 5 9 2 7 4

3 6 I 12 8 11 10 9 5 7 2 4

The term"mirror"finallyrefersto the fact that the musical config-

uration that opens the work and is expressedin termsof the row 01
returnsin invertedformwhen the serial "conveyorbelt" produces the
formI 1, in retrogradeinvertedformwhen the row RI 1 appears (not
shown in the above table), and at the veryend of the work in termsof
the form R 1. The remainingareas of the music are not any longer
occupied by thematicstatement,development,recapitulation,and the
like. Whatevermorphologicalkinshipmay be detectedbetweenadjacent
sectionsis a resultof similaritiesof intervallicshapes that may occur in
neighboringformsof the tone-row,the vicinityof which, however,is a
consequenceof the premeditatedserial arrangementoutlined above and
not dictatedby requirementsof a so-called musical nature.

In thiscompositionno otherparameterbeside the successionof tones

was seriallyordered.In this respectit belongs to the provinceof "class-
ical" twelve-tonemusic. It transcendsthat provincein that it allows its
structureto arisefromtheserialarrangementof the rotationalderivatives
of its tone-row.

The principleof rotation,which, as may be seen here, I discovered

and utilizedforreasonsnot relevantto the evolutionof pan-parametrical
significancewhen I became
organization,turnedout to be of far-reaching

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
216 The Musical Quarterly
interestedin that kind of organization.The point is that the notion of
invariancyinherentby definitionto the concept of the series,if applied
to all parameters,leads to a uniformity of configurations that eliminates
the last traces of unpredictability,or surprise.But unpredictability ap-
pears to be not only especially of
characteristic so-called "atonal" music,
but desirable,or necessary,in any workof art. That the composerswho
have made the most consistentattemptsat "total determinacy"are
aware of this need transpiresfrom this utterance of Pierre Boulez:
"L'inattendu, encore: il n'y a de creation que dans l'impr6visible
devenant n~cessit6."5

Combinationof the various configurations that resultfromrotational

procedure with constant (non-rotating) serial elementsmeans that the
principleof order that governsone set is applied to another,unrelated
set (as if one, for instance, would order the numbers from 1 to 5
alphabetically: fivefourone threetwo). Since thisis one of the defini-
tionsof randomness,we meet here forthe firsttimethe factorof chance,
which has attainedhigh significancein recentdevelopments.


Accordingto GyorgyLigeti's analysis8of Pierre Boulez's Structures

for two pianos,' the composer has interpretedthe transpositionsof his
twelve-tonerow to various pitch levels as a formof rotationand has
transplantedthe resultsto the parameterof time in order to obtain an
analogous sequence of derivativeformsof his timeseries.
5 "The unexpected, again: there is no creation except in the unforseeable
becoming necessary" (Revue musicale, April 1952, p. 119, as quoted in Die
Reihe, No. 4, Vienna, 1958, p. 71). It is interestingthat this statement almost
verbatim sums up Carl Bricken's brilliant argument about "inevitability and the
unexpected" in his analysis of Beethoven's Quartet Op. 18, No. 3 (Some Analytical
Approaches to Musical Criticism, in Proceedings of the Music Teachers National
Association for 1936, Oberlin; 1937, p. 262 ff.). In Bricken's discourse the "inevi-
table" is, of course, represented by those musical processes that appear to be most
likely to occur within the frameworkof tonal harmony so that they constitute a
predictable, "normal" set of events. The "unexpected," then, consists of the devia-
tions from the norm introduced by the genius of the individual composer. In the
case of serial music the inevitable is what serial premeditation ordains. The unex-
pected, however, is not a result of the composer's kicking against the self-imposed
limitations,but of the built-in surprise mechanism, as we shall see later on. In my
article Is the Twelve-Tone Technique on the Decline? (in The Musical Quarterly,
Oct. 1953, p. 523 ff.) I indicated that Boulez in his Second Piano Sonata probably
applied the principle of rotation.
6 Die Reihe, No. 4, p. 38 ff.

Universal Edition, Vienna.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
Extents and Limits of Serial Techniques 217
The elementsof the tone seriesare numberedfrom 1 to 12:
Ex. 7t
I 2 3 4 5 6 7 8 9 10 1 12

To this a seriesof time values corresponds,expressedin termsof

I 2 3 4 5 6 7 8 9 0 II 12
If we transposethe tone row, for instance,a major third higher,the
originalorderof the tones is changed into:
Ex. 8
5 6 8 9 12 10 4 II 7 2 3 I

the time serieswould take on the shae:

5 6
8 %)J
4 .).. 12 10 4 i- 7 2 3 I
InJ-J. U,. , of the tone-
fact, the whole work consistsof manifoldcombinations
and time-setsthus obtained.

Karlheinz Stockhausen'sworkdescribedalternatingly as Komposition

1953 No. 2 and ElektronischeStudie 18 is based on a six-toneseries
which accordingto the composer'sown elaborate analysis9is an expres-
sion of thisseriesof ratiosof frequencies:
12 4 8 5 5
5 5 5 12 4
Expressed in vibrationnumbers,or cycles,persecond,thefirstseriesreads:
1920 800 1000 625 1500 1200
12 : 5 8 : 5 5 : 4
4 : 5 5 : 12
In notesit reads approximately:

Ex. 9
Five more seriesare derivedby makingthe consecutivetonesof the first
series points of departurefor new series identicallybuilt (a procedure
somewhatreminiscentof my Lamentatio rotation):
800 333 417 260 625 500
1000 417 521 325 781 625
625 260 325 203 488 390
1500 625 781 488 1170 937
1200 500 625 390 937 750
8Universal Edition, Vienna. Recorded by the Deutsche Grammophon Gesell-
STechnische Hausmitteilungen des Nordwestdeutschen Rundfunks, Vol. VI,
No. 1/2, Cologne, 1954, Item 10, p. 46 ff.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
218 The Musical Quarterly
A second set of six seriesis obtained by making the second line of the
firstset the top line of the new set,then the third,and so on.
All parametersare seriallyordered in termsof some variantsof the
numericalsequence 1 to 6. For instance,the combinationsof the above
frequenciesfollow from the series 4 5 3 6 2 1 in that the firsttone-
combination("Tongemisch") has fourtones,the second five,and so on.
There are foursuch "Gemische" in "sequence 1" (a "sequence" being a
grouping of consecutiveelements), and four "sequences" in the first
"structure,"which is the next highercompound, "horizontal" or "ver-
tical." (It does not become quite clear on what groundsone or the other
of thesetwo dimensionswas chosen.) There are six dynamiclevelswhich
are assigned to the various frequenciesin proportionto their relative
positionsin the groupsand columnsof the entiresystem.The seriesthat
ordersthe successionof dynamiclevels withinthis frameof referenceis
3 4 2 1 6 5. Finally,the time factoris determinedby relatingthe dura-
tionsof the individualsoundingelementsto the pitch levels and degrees
of loudnessof those elementsas orderedby the previousrules.The gov-
erningseriesin this parameteris 2 4 6 3 5 1.
The details of this organizationare far more complex than what we
are able to indicate here in an abridged sketch.Unfortunately the pre-
sentationby the author is not always felicitous,so that some of the
intricaciesof his work remain obscure. At any rate, the characterof his
reasoningseemsto reveal a desireto derivethe rulesof serialorganization
fromthe natureof the chosen materialand its intervallictexture.In this
respect Stockhausen differssomewhat from Boulez, who has a rather
mechanisticapproach in assigningnumericalvalues to the various mag-
nitudes manipulated in his work. While this procedureof Boulez's has
been criticizedas produceda fascinating
it has nevertheless
piece of music. On the otherhand, Stockhausen'sStudie, althoughmuch
shorterthan the Structures,suffersfromconsiderablemonotonyof har-
monic flavor,which is due to the prevalenceof augmentedtriadsin the
originalseries (see Ex. 9). The extraordinary subtletiesof combinations
of dynamicshadings,timevalues, echo effects,and the like cannot over-
come thisinitialhandicap.
The objection was raised that music here becomes the victimof an
abstractnumbersgame which is contraryto the nature of music. While
thereundoubtedlyis room formore than one definitionof the nature of
music, we did not extend our inquiryinto this field.The numbersused
1o Ligeti, loc. cit., p. 41.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
Extents and Limits of Serial Techniques 219
in the orderingof the parametersof serial music are almost always de-
rivedfromproportionsand measurementsof the basic musical substance.
Of course,thesenumbersdetach themselvesfromthe objects withwhich
theywere associatedand take on a lifeof theirown in the variousopera-
tionsperformed.The resultsof theseoperationsare, however,retranslated
into musical termsand applied to the soundingmaterial.In thisrelation
of numberand realityone may see a vague analogy to the connectionof
contemporarymathematicsand physics.

In myoratorioforvoicesand electronicsounds,Spiritusintelligentiae,
sanctus,"thereis a sectionwithoutvoices (so to speak an "instrumental"
interlude). The material of this sectionis a temperedscale of thirteen
tones.From the continuumof thisscale, groupsof toneswere selectedto
form alternatinglydisjunct and conjunct heptachords of equal and
symmetrical structure(see leftside of Diagram 1). A seven-tonepattern
(seven-tonerow) meanders throughthis systemof pitches constantly
retainingits principleof progress: from any tone on which it startsit
goes up to the thirdand fourth,thenback to the second, up to the sixth,
back to the fifth,and it stops on the seventhtone of the networkof
pitches.Since the patternalwaysprogressesconjunctly(which means that
the firsttone of its next appearance is identicalwith the last of the pre-
ceding) while the pitch systemis based on the alternationof conjunct
and disjunctshapes, the internalintervallicconfiguration of the pattern
is always different, although its general outline remains the same (see
rightside of Diagram 1). Afterthirteenappearances the patternlands
again on the tone fromwhich it started,and the "rotation"has come to
an end.
The interludein question may technicallybe called a double canon.
One of the two elementssubject to imitationis a tone-lineconsistingof
the chain of the thirteenpossiblevariantsof the seven-tonepatternjust
described,the otheris an analogous line presentingthe chain of the in-
verted formsof the pattern.The firsttone-lineis so designed that it
begins on the centraltone of the entiregamut (330 cycles), risesto its
highestlevel (4754 c) in the firstthirdof itslength,returnsto the center
in the second third,and descends to the lowest level (26 c) in its last
portion.The second line beginson the lowestpointwhen the firstreaches
its apex, risesto crossthe firstline whereit passes on its descentthe cen-
tral tone, goes up to its own high point which it reaches approximately
when the firstline ends, and returnsto the center.
11Recorded by the Deutsche Grammophon Gesellschaft (LP 16134 Hi-Fi).

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
220 The MusicalQuarterly

Read frombottomup
Heavy linesindicateoctaves
Chain of
conjunct Progressof the
heptachords seven-tonepattern

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
Extents and Limits of Serial Techniques 221
The canonic imitationswere obtained by rerecordingthe original
material at a higherand a lower speed, in which procedure the pitch
level of the originaltape was automaticallyraised or loweredin the same
proportion.These imitationswere so synchronizedwiththe originallines
that the slowed-downversionof the ascendingbranch of the firsttone-
line would reach itshighestpoint (proportionately lowerthan thesummit
of the original) when the originalline had returnedto the center.It was
followedby the slowed-downimitationof the descendingbranch of the
second line. The above-centerarcs of both lines were imitatedin accel-
erated versionsreaching their (proportionatelyhigher) apices shortly
afteror beforethose of the original lines. Finally, a veryhighlyaccel-
erated imitationof the below-centerbranchesof both lines was inserted
shortlybeforethe end of the section.
To determinethe timevalues of thesingleelementsthe whole expanse
of the piece was viewed as one unit. Through measuringthe linear dis-
tances of the importantpointsof articulation- entrancesof imitations,
turningpointsand such - a seriesof elevenspans was established,a sort
of macro-rhythm articulatingthe over-all structure.It was reduced in
scale to a micro-rhythm In order to determinethe durations of the
individual tones in each tone-line.Since each line takes approximately
three quarters of the entirelengthof the piece and each line contains
ninety-onetones (seven times thirteen), the micro-rhythm of eleven
values has to be repeatedeighttimes,leavingthreetonesfreeat the end.
This conceptdeterminedthe ratio by which the macro-rhythm had to be
reduced. Since the rhythmicseries thus establishedhas eleven terms
whereas the tone-serieshas onlyseven tones,it followsthat the last four
termsof the firsttimeserieswill apply to the firstfourtonesof the second
tone series, and so forth,so that here again mechanical repetitionis
avoided while uniformity in a highersense is maintained. (See Diagram
It may be statedthatwhateveroccursin thispiece at any givenpoint
is premeditatedand therefore technicallypredictable.However,whilethe
preparation and the layout the materialas well as the operationsper-
formedthereinare the consequence of serial premeditation,the audible
resultsof these procedureswere not visualized as the purpose of the
procedures.Seen from this angle, the resultsare incidental. They are
also practicallyunpredictablebecause the simultaneousprogressof highly
complex rhythmicpatternsat various relativespeeds togetherwith the
correspondingtranspositionsof equally complex pitch patternscreates
situationsthat defyprecisevisualization.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions

a" Fc


, t


L'ne End
, . L

M&re- I ? | ? s
? ? , _

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
Extents and Limits of Serial Techniques 223

The Sestina is one of the poetic formsdeveloped by the

poets of the twelfthcentury,its original specimen being ascribed to
Arnaut Daniel. It may well be called a serial formof poetry,and its
essentialformativeprincipleis rotation.
The poem consistsof six stanzas of six blank verseseach. It hingesupon six
keywordswhich appear at the endings of the individual lines. If in the first
stanza the order of these words is 1 2 3 4 5 6, the words will appear in the
second stanza in the order 6 1 5 2 4 3. The principle of rotationwhich is
applied here consistsin switchingthe positionof everytwo keywordsequidistant
from the center of the series, proceeding from the end toward the middle.
Accordingto the same principle,the positionsof the keywordsin the subsequent
stanzas are 3 6 4 1 2 5; 5 3 2 6 1 4; 4 5 1 3 6 2; 2 4 6 5 3 1. The process
ends here, since the next rotationwould produce the original series. The six
stanzas are followedby a Tornada of threelines in which the keywords,one of
each pair in the middle and the other at the end of the line, appear in the
order 2 5, 4 3, 6 1.
The contentof the Sestina which I wrote (in German) as textforthe present
compositionis a contemplationof the implicationsof the idea governingthe
of the work.13
musical construction

The firsttwo stanzas may sufficeto indicate the characterand form

of the poem:
1. VergangenKlang und Klage, sanfterStrom.
Die Schwingungder Sekunde wird zum Mass.
Was in Geschichtelebt, war's nur ein Zufall?
Verfall,Verhall, zerronneneGestalt?
Die Stunde zeitigtWandel, wendetZeit.
Das Vorgeschrittne ordnetsich der Zahl.
2. In Schrittenvorgeordnetdurch die Zahl
gestaltetsich Gedanke,doch zum Strom
wird strengeTeilung,uhr-genaueZeit.
Ist es vermessen,solches Mass von Mass
dem Leben aufzuzwingen, der Gestalt?
Der Zwang zerrinnt,erzeugtden neuen Zufall.14

12Birenreiter-Verlag,Kassel. Epic Records, LC 3509.

13 Quoted frommy noteson the jacket of the recordcited in note 12.
14In a nearly literal translationwhich reproduces the positions of the key words:
Bygone are sound and mourning,tender stream.
Vibration of the second becomes the measure.
What lives in history,
was it only chance?
Decline, fading sound, vanished shape?
The hour causes change, turns the time.
What looks ahead subordinates itself to number.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
224 The Musical Quarterly

The music of my Sestina is based on a twelve-tonerow divided into

two groupsof six toneseach:
Ex. Io
, 1 a 2
4 3 6 I 6 4 6 s (2)

The figuresindicate the size of the intervalsmeasured in half-steps.

These tone-rowsare rotatedaccordingto the principleof the sestinaso
that the second A- and B-groupsread:

Ex. rI ;I ,'-I" "

,.- ,

6 4 4 3 5 4 z 2 (1)
The thirdline is:
Ex. 12
S 4 2 44 I 3 I 4 6 4

and so forth.The tones are always placed so that theywill not exceed
the ambitusof the originalrow and the intervals(indicatedby the num-
bersbelow the staff) are so measuredup or down that theirmagnitudes
will not exceed the figure6. Obviouslythe sequence of these intervallic
magnitudesconstantlychanges as a resultof the rotationof the tones
prescribedby the sestina pattern,but these changes are of a different

The durationsof the tonesof the whole compositionare derivedfrom

these magnitudesin the followingmanner: each intervallicmagnitude
correspondsto a time segmentwhich containsas many basic time units
as the intervalfigureindicates.Consequentlythe firsttime segmenthas
four units,the second three,etc. Each segmenthas as many tones as it
has units (4, 3, etc.). The durationof the individualtonesis determined
by a subdivisionbased on the same serial sequence of magnitudes.If the
firstsegmentcontainsfour units and fourtones,its subdivisionis based
on the firstfour values of the original series: 4 3 1 6. The sum of
these being 14, the subdivisionunit within the firstsegmentis 4/14.
The durations of the individual tones within the firstsegment are
determinedby multiplying4/14 consecutivelyby 4, 3, 1, and 6. The
In stages preordained by number
thought takes shape, but a stream
is (the result of) strict division, of clocklike, precise time.
Is it presuming to force such an extent of measure
on life, on shape?
Force vanishes, brings forth new chance.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
Extents and Limits of Serial Techniques 225

durations,then, are 16/14, 12/14, and 24/14, or 8/7, 6/7, 2/7, and
12/7 of the basic value.

Actually the determinationof the durationsis due to much more

complicatedcomputationbecause it is influencedby serial organization
of other parameters.In order to achieve higherrhythmicdiversity,the
concept of "internalspeed" was introduced.It is derived fromthe as-
sumptionthat in every group of six tones one to five tones might be
sounded an octave higherso that the magnitudeof the affectedintervals
would be augmentedby twelve. The successionof "internalspeeds" is
derived fromthe position of the tones in group B (see Ex. 10). The
lowest (A) is designatedas 1, the highest(F) as 6. The initialrow of
internalspeeds is therefore5 1 4 3 6 2. The firstsegment,then,has the
internalspeed 5 so that 12 is added to fiveout of six subdivisionnum-
bers. Thus these numbersread 16 15 13 18 14 instead of 4 3 1 6 2.
The followingnumber- 1 - remainsunaltered.The sum of the num-
bers attached to the firstsegmentis therefore62, instead of 14. Conse-
quentlythe durationsof the individualtoneswill be considerablyshorter
than if the "internalspeed" were, for instance,1 or 2.

To facilitatecomputationseach basic unit is assumed to contain ten

micro-units.We arriveat the subdivisionof the firstsegmentby dividing
40 (fourtimesten) by 62. The resultis 0.645. This numberis multiplied
consecutivelyby 4, 3, 1, 6. The resultsare 2.58, 1.935, 0.645, 3.87. If
the work had been realized by electronicmeans on tape, these values
could be produced with utmost accuracy. Since it was conceived for
conventionalmannersof rendition,the time values had to be adjusted
as follows: 2.5, 2, 0.5, 4. If the smallestnumericalunit is expressedby
therhythmic shape of the firstfourtonesis ~
S9/16. __,

"Density" is the next parameter to be determinedserially.There

are six degreesof densitywhose successionis determinedby the position
of the pitchesin group A (Ex. 10). Again the lowest (C) is called 1,
the highest (G#) 6. Consequentlythe initial seriesof densitiesis 6 3 5
4 1 2. In "density1" the two tone-groupsA and B run offsimultaneously
in a sortof two-partsettingin whichthe durationof the individualtones
is determinedby the mechanism described above. In "density2" the
firstand second time segmentsof group A run concurrentlywith the
firstsegmentof group B. In "density3" two segmentsof each group are
developed simultaneously, and so forth,untilin "density6" six segments
of each group,i.e. twelveall together,run offat the same time.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
226 The Musical Quarterly
Anotherparameteris the location of the tones withinthe gamut of
six octaves designatedas the ambitusof the work.The serial statement
adopted for this area reads that the tones of each segmentshould run
throughas many octaves as thereare tones.The directionof the motion
is determinedby the direction of the correspondinginterval in the
original series. Since many segmentscontain less than six tones, they
coverless than six octaves and thereforecould extendover variousbands
of the complete ambitus. This, too, is regulatedby special serial state-
ments.Needless to say that all theseserial organismsare subject to rota-
tion accordingto the sestinapattern,whichis the supremelaw governing
everymove of everyvariable withinthe whole composition.
The structurallayout is designedto combine each "rotated" version
of any six-tonegroup witheveryother.Thus the musicof the firststanza
is based on the firststatementof the A-group,in each consecutiveline
of the poem combined with one of the formsof the B-group rotated
fromB 1 to B 6. The second stanza has A 2, combinedagain with all
six B-groups,but now in a different sequence, accordingto the sestina
pattern: 6, B B B
1, 5, 2, 4, B B 3.

Paralleling the arrangementof the key words in the tornada, the

tone series assigned to it reads 2 5 4 3 6 1. The music of the tornada
consistsof six sections,the firstfourand the last of which are givenover
to the instruments alone. While the tone row of the tornada undergoes
the now familiarsix sestina transformations, the densityincreasesfrom
1 to 6 so that in the firstsection of the tornada onlyone each of the A-
and B-rows are employed,while in the last sectionsix of each, that is
twelve,or all available formsare used simultaneously.
The parameterof "externalspeed" has six stepsalso, the lowestbeing
M 90, the highest ) 180. The formeris associated with the
highestdegree of density,the latterwith the lowest.
Example 13 shows the firstten sixteenths(micro-units)which form
the firstbasic time unit of the Sestina. On the leftside one may see the
distributionof the tones of the A- and B-groupsover the twelve layers
(density6) of simultaneously progressing timesegments,each tone enter-
ing at the point assignedto it by the time mechanismexplained above.
The tones occupy theirplaces fromtop to bottomlayerin theirorderof
successionin the row. The "internalspeed" for the A-layers(top six)
is 5, forthe B-layers (bottomsix) 1 (no acceleration). Encirclednum-
bers indicate the number of tones allotted to the particularsegments.
Arrowsindicate the directionof the tone lines. The figuresabove the

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
Extentsand Limits of Serial Techniques 227

top staffgive the durationsof the firstfourtonesin , as computedon

p. 225. The rightside of the example shows how these tones are repre-
sented in the actual score, and a few connectinglines were drawn to
demonstratewheresome particulartonesmay be found.
Ex. '3
0.5 4
A )2.5
.. ..
_ .----,--:.=--,
A, '- - 14
1 \'


,-= ,
?~ ~ 1,-
: i ,Gu.
B~ 6




It is easy to see that the parameterof timbrelies beyond the limits

of the presentserial arrangement.If this parametertoo were organized
seriallyand this procedurewould, for instance,require the firsttone of
the top layer (G.) to be played by the trumpet,it would obviouslybe

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
228 The Musical Quarterly
at variance with the octave registerdemanded by the serial regulation
of spacing,since the trumpetcannot play the G$ in question.


Other parametersmay be affectedin the same way. If the succession

of tonesis determinedby serial regulation(as is the case in the classical
twelve-tonetechnique) and, in addition to this, the timing of the
entranceinto the musical processof thesetonesis also predetermined by
serial calculation (as, for example, in the case of the Sestina), it is no
longer possible to decide freely(that is, by "inspiration") which tones
should sound simultaneouslyat any given point. In otherwords,the so-
called harmonicaspect of the piece will be entirelythe resultof opera-
tions performedon premisesthat have nothingto do with concepts of
"harmony,"be it on the assumptionof tonalityor atonalityor anything
else. Whatever happens at any given point is a product of the pre-
conceivedserialorganization,but by the same tokenit is a chance occur-
rence because it is as such not anticipatedby the mind that inventedthe
mechanismand set it in motion.

Generallyand traditionally"inspiration"is held in great respectas

the most distinguished source of the creativeprocessin art. It should be
rememberedthat inspirationby definitionis closely related to chance,
forit is the verythingthat cannot be controlled,manufactured,or pre-
meditatedin any way. It is what falls into the mind (according to the
German term Einfall), unsolicited,unprepared, unrehearsed,coming
fromnowhere.This obviouslyanswersthe definitionof chance as "the
absence of any known reason why an event should turn out one way
ratherthan another."'5Actuallythe composerhas come to distrusthis
inspirationbecause it is not reallyas innocentas it was supposed to be,
but ratherconditionedby a tremendousbody of recollection,tradition,
training,and experience.In orderto avoid the dictationsof such ghosts,
he prefersto set up an impersonalmechanismwhichwill furnish,accord-
ing to premeditatedpatterns,unpredictablesituations.Ligeti character-
izes this state of affairsverywell: "We stand in frontof a row of slot
machines ["Automaten"] and we can choose freelyinto which one we
want to drop our coin, but at the same time we are forced to choose
one of them. One constructshis own prisonaccordingto his wishesand
is afterwardsequally freelyactive within those walls - that is: not
entirelyfree,but not totallyconstrainedeither.Thus automation does
not functionas the opposite of free decision: ratherfreeselectionand
15 The American College Dictionary, New York and London, 1948, p. 200.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
Extents and Limits of Serial Techniques 229
mechanizationare united in the process of selectingthe mechanism.""'
In other words,the creativeact takes place in an area in which it has
so far been entirelyunsuspected,namely in settingup the serial state-
ments (selectingthe slot machines). What happens afterwardsis pre-
determinedby the selection of the mechanism,but not premeditated
exceptas an unconsciousresultof the predetermined operations.The un-
expectedhappens by necessity.The surpriseis built in.


A later serial work of mine is a set of six piano pieces, called Sechs
Vermessene.This German title is a play on words, since vermessenin
German means "completelymeasured" as well as "presuming,"a pun
that cannot be reproduced in English. While the time mechanism is
similarto that of the Sestina, the constructiondiffersfromit in that for
the firstthree pieces a systemof fivelayersis set up in which the first
has "density 1" (i.e. one tone at a time), the next has two tones
together,the third three,the fourthfour, and the fifthsix tones. The
time measurementsforthe variouslayersare a resultof summingup the
intervalmagnitudesinvolved in the consecutivetone combinations.For
example, the tone seriesof this compositionbeing:

. .... _ 1_ _ , .. ,

the firstcombinationof tonesin "density2" is:

Lx. 15

The numerical values derived from this progressionare 3 (a minor

thirdfromG to Bb) and 1 (a half-stepfromE to F). Consequentlythe
firsttimesegmentof the firstlayerhas threeunits,the firstof the second
has four (3 + 1). As the densityof the layersincreases,the numberof
simultaneouslysoundingintervalsand thus the numericalvalues of their
sums become higher.Thereforethe timesegmentsbecome longer,which
means that the chords, or tone-clusters, with increasingthicknessare
spaced fartherapart, while the singletones of the firstlayerfolloweach
other more rapidly. Computationsof this kind form the basis of the
whole composition.
As explained before,phenomenain the parameterof harmonymust
be accepted as resultsof the operationsin the sectorsof pitchsuccession
and time. In the fourthof the piano pieces an attemptwas made to
Loc. cit., p. 38 (translated from the German by this writer).

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
230 The Musical Quarterly

begin with a selectionof sound elements.From the tone row we devel-

oped twelve sets of four elementseach (consistingof one, or of two,
three,or four tones played simultaneously)plus two six-tonechords.
These fiftyelementswere numberedfrom 1 to 50 and theirsuccession
was determinedby progressing along thisseriesby the distancesindicated
in the numericalvalues of the intervalsof the basic row:
seriesof elements: 1 2 3 4.5 6 7 8 9 10 11 12 13 14 15...
intervalsof the tone row: 3 2 5 4
selectedelements: 1 4 6 11 15 ...

In the fifthpiece the five degrees of thicknesses(see above) are

distributedover fivelayerswhich progressat various speeds so that the
timemeasurementsof the slowestlayerare reducedto 1/2 in the second,
to 1/3 in the third,to 1/4 in the fourth,and to 1/6 in the fastestlayer.


In the field of serial music one may observe a tendencytowards

using series of magnitudesthat progressively vary according to some
serial orderingof their own. The speed levels of the Sestina are an
example. Anothertimeseriesof thisnaturewas establishedforthe voice
line of this work. It is based on the successionof 1 2 3 5 7 and 10 )
for the accented syllables.The openingsuccessionis 2 3 10 5 7 1 and
the followingforms are obtained throughthe sestina rotation. Since
each line of the poem has only fiveaccented syllables,interesting situa-
tions of overlaps occur.
It may be seen that the serieshere applied is a modificationof the
so-called Fibonacci seriesin which each termis equal to the sum of the
two precedingterms: 1 2 3 5 8 13 21 34 55 etc.1'Luigi Nono has used
the firstsix termsof this seriesas factorswith which he multipliesthe
basic time values of his II Canto sospeso in order to obtain the actual
durationsof the individualtones.18I have used the termsof the Fibonacci
seriesfrom2 to 21 to determinethe speed zones in a recentorchestral
compositionentitledQuaestio temporis(A Questionof Time). This work
is based on a twelve-tonerow that containsall eleven intervalsin this
order (measured in half-steps):
The entireexpanse of the compositionis thoughtof as consistingof 66

1' Cf. Matila Ghyka, The Geometry

of Art and Life, New York, 1946, p. 13 f.
' Cf. Karlheinz Stockhausen's analysis of the work in Darmstiidter Beitrage
zur neuen Musik, Mainz, 1958, p. 70.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
Extentsand Limits of Serial Techniques 231
time units (the sum of the above figures),which formeleven sections
of varyinglengthsaccording to the magnitudesof the basic series.To
thesesectionssix different
speeds are assigned:
M -= 20, 30, 50, 80, 130, and 210

It appears thatdensityis a functionof speed and thicknessof texture.

If the lattermay be called the verticalcomponentof densitybecause it
depends on how many layersare in operationat the same time, speed
is the horizontalcomponentof densitysince the tones followeach other
more closelythe fasterthe tempo of the music is. If both parameters
approach maximumvalues, a degree of saturationis reached at which
accurate computationsof time points and durationsbecome irrelevant.
When in the finalsectionof Quaestio twelvelayers (maximum vertical
density) progressat a speed of J = 210 per minute,the tones come
so close togetherthat nearly everysixteenthis sounded, frequentlyby
several tonessimultaneously. The velocityof the music causes 14 to
run offper second. At thisrate even the successionof pitchesis not any
longerof greatsignificance.It seemssufficientto determineby experiment
withina limitedarea the average numberof time unitsneeded for run-
ning throughthe twelve-toneseries.The resultsof thisstatisticalexami-
nation are then used in order to fill this area of highestdensitywith
actual musical sounds.


One of the parametersthat obviouslycannot be controlledby pre-

meditationwhen those so far discussedare subjected to serial ordering
is the expressive,or communicative,aspect of music. If a serial composer
were concernedwith this problem,he would have to set up a seriesof
"moods," or "ideas," or somethingof thissort,to begin with,and then
let the otherparametersfall in line. It so happens that serial composers
are not thinkingin such terms.
In a more pessimisticattitudethan he now seems to entertain,the
German composer and philosopher,T. W. Adorno, has criticizedthe
recentdevelopmentsof serial music'9because in these the (according to
him) deep-rootedand essentialanalogy and affinity of music and speech
is abandoned. While it may be true that music fromthe time of plain-
chant has been orientedtowards speech-likearticulation,diction, and
over-allstructure,and while especiallythe exploitsof Expressionismand
19Das Altern der neuen Musik, in Der Monat, May 1955.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions
232 The Musical Quarterly

atonalitypoint to a veryclose associationwith the free articulationof

prose,we have to face the fact that under the influenceof the construc-
tive rigorthatwas the veryconsequenceof Expressionistic roamingserial
music has turned away fromits rhetoricalpast. Since whatevermusic
seemsto communicateis not so much the supposedcontentof the audible
matteras it is the product of the listener'sreactiontouched offby his
auditory experience,there is no reason to assume that the nature of
serial music excludes the possibilityof interpreting it as a medium of
some sortof communication.The interestit may evoke is similarto that
elicited by the process of life, to which serial music is related in the
paradox of the chaotic appearance of totallyand systematically traceable
causality. may mean as much or as littleas lifeitself.

This content downloaded from on Sat, 13 Jul 2013 16:12:29 PM

All use subject to JSTOR Terms and Conditions