Beyond Analysis

Author(s): Edward T. Cone

Source: Perspectives of New Music, Vol. 6, No. 1 (Autumn - Winter, 1967), pp. 33-51
Published by: Perspectives of New Music
Stable URL: http://www.jstor.org/stable/832404
Accessed: 22-02-2016 03:02 UTC
REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/832404?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
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.

Perspectives of New Music is collaborating with JSTOR to digitize, preserve and extend access to Perspectives of New Music.

http://www.jstor.org

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

ANALYSIS*

BEYOND
EDWARD

T. CONE

E x AMPLES 1-3 present the beginnings of three hypothetical compositions.If theysound both oddly familiarand familiarlyodd, thatis because theywere derivedby the simple application of a mirrorto three
well-knownsources: Schoenberg'sKlavierstiick,
Op. 33a, and the first
and thirdmovementsof Webern's Variationenfiir
Klavier,Op. 27. Hence
he will findit a
if the reader wishes to complete these constructions,
even
mechanical
task.
and
straightforward
=

120

Ex. 1
S=40

Ex. 2
d=8o
fAff

tlF

rit.

oI

=f

p!

Ex. 3
* This

formas a lectureat the SummerInstitute

paper was presentedin slightlydifferent
ofCompositionalStudiesof the AmericanSocietyofUniversity
Composersat the Berkshire
Music Center,August1967.

33 *

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

PERSPECTIVES

OF NEW

MUSIC

The possibilityofsuch derived compositionswas suggestedto me by

a famouspassage fromSchoenberg'sessay "Composition with Twelve
Tones":
The unityofmusicalspace demandsan absoluteand unitary
In this
perception.

space,... thereis no absolutedown,no rightor left,forwardor backward. Every musical configuration,everymovementof tones has to
be comprehendedprimarilyas a mutual relationofsounds,ofoscillatoryvibrations,appearingat different
places and times.To the imaginativeand creativefaculty,relationsin the materialsphereare as independentfromdirectionsor planes as materialobjects are, in their
sphere,to our perceptivefaculties.'
No doubt I have taken this passage more literallythan its author
intended.So faras I know,Schoenbergnevertriedto demonstratethat
the strictmirrorinversionof a twelve-tonecompositionmustbe as valid
as the original-but thismightindeed be one conclusionthat could be
drawn fromthe quoted passage. It is also-and thisis my real startingpoint-a conclusionthat mightbe drawn fromreadingmuch,and perhaps most,accepted twelve-toneanalysistoday.
My researchinto this question has been by no means exhaustive;
furthermore,
although I feel confidentthat the analytic essays I have
studiedconstitutea representativesample, I have no way of proving
this.The onlyfairway of presentingmy case, then,is to listthe actual
examplesI have used and the resultsI have obtained.
To beginwith,the master'sanalysesofhis own worksin the essayjust
cited would apply equally well ifthe compositionsin questionwerereplaced by mirrorinversionsof themselves.One need only make the obvious adjustments:substituteforthe original formof the set its inversion,forany transpositionits complement,and so on, and the analysis
can easily be made to read accurately. Only the referencesto instrumentation(which appear by way of descriptionratherthan analysis)
mightcease to be relevant.
One may immediatelycounterthat what Schoenbergwas presenting
was not analysisbut an explanationof a method-and a veryprimitive
explanation at that. One could not expect him to have developed the
sophisticatedand powerfultools of analysisat our disposal today.Very
well, then, look at as varied a compilationas the following:Milton
Babbitt'sthreeclassic statements,"Some Aspectsof Twelve-ToneComposition,"2"Set Structure as a Compositional Determinant,"3and
"Twelve-Tone Invariants as Compositional Determinants";4 Ernst
1Arnold
Schoenberg:StyleandIdea,New York,PhilosophicalLibrary,1950,p. 113.
2 The
Score,No. 12 (June1955), pp. 53-61.
3JournalofMusic Theory,Vol. v, No. 2 (April 1961), 72-94.

4Problems
Music,ed. Paul HenryLang,New York,W W. Norton,1960,pp. 108-21.
ofModern

. 34 *

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

BEYOND

ANALYSIS

Krenek's analysis of his own Lamentatio

and Sestinain "Extents and
Limits of Serial Technique";5 the entire second issue of Die Reihe,
devotedto Webern; and, despite theirpromisingtitles,George Rochberg's"The HarmonicTendencyof the Hexachord"6and his "Webern's
Search forHarmonicIdentity."7In none of theforegoing
would the line
of argumenthave to be changed ifthe entirebody oftwelve-tonecomintoitsexact inversion,forin every
positionwere magicallytransformed
case the only pitchrelationshipsdiscussedare thosethat remaininvariant underinversion.Even such extendedmonographsas JosephRufer's
Compositionwith Twelve Toness and George Perle's Serial Compositionand

exhibit only a few unsystematicexceptionsto this general

Atonality9
principle.One furtherexample that is especially indicative is Allen
Forte's analysisof the SchoenbergFantasy Op. 47 in his Contemporary
Tone-Structures,
10forit is the onlyanalysisin the book that foregoessome
sortof Schenker-likelinear reduction.In demonstratingthe continuity
ofthe Fantasyit reliesentirelyon connectionsbetweenrow-statements,
all of which would work equally well for the mirrorinversionof the
composition.
As might be expected, PERSPECTIVESOF NEW MUSIC offersan unusu-

ally richharvestof appositeexamples.These include David Lewin's "A

Theory of SegmentalAssociationin Twelve-ToneMusic";11 John M.
Perkin's"Dallapiccola's Art of Canon";12 Babbitt's "Remarks on the
Recent Stravinsky";13
Perle's "An Approachto Simultaneityin TwelveTone Music";14 Peter Westergaard's"Toward a Twelve-TonePolyphony";15and about a half-dozenof the "Younger Composers" series.
Especially interestingis another essay of Babbitt's, "Twelve-Tone
RhythmicStructureand the ElectronicMedium,"16 which developsa
methodof derivinga rhythmicrow fromthe intervalsof the basic set.
the original
Perhaps here one can find a criterionfor distinguishing
compositionfromitsinversion.But no: since the directionwe choosefor
countingnotesor forcalculatingintervalsis a matterofpure convention,
an invertedset can always be made to yield the same rhythmicrow as
its original (i.e., by countingintervalsdownrather than up fromthe
origin).
5Ibid.,pp. 72-94.
6J.M.T, Vol. III, No. 2 (Nov. 1959), 208-30.
7Ibid.,Vol. vI, No. 1 (Spring1962), 109-22.
8Translatedby HumphreySearle,New York,The Macmillan Co., 1954.
9
ofCaliforniaPress,1962.
Berkeley,University
10New York,Teachers
College,Columbia University,
1955,pp. 110-27.
11Vol. 1, No. 1 (Fall
1962), pp. 89-116.
12
Vol. 1, No. 2 (Spring1963), pp. 95-106.
13Vol. 2, No. 2 (Spring-Summer1964), pp. 35-55.
14Vol. 3, No. 1 (Fall-Winter
1964), pp. 91-101.
15Vol. 4, No. 2 (Spring-Summer
1966), pp. 90-112.
16Vol. 1, No. 1 (Fall 1962), pp. 49-79.

* 35

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

PERSPECTIVES

OF NEW

MUSIC

Allen Forte's"Context and Continuityin an Atonal Work""17

shows,
by its treatmentof Schoenberg'sOp. 19, that my suggestedtransformationneed not be limitedto twelve-toneworks.From thisessay (as well
as fromappropriatesectionsof Perle'sbook) one mightgo much further
and conclude that,barringpurelyinstrumental
a new comdifficulties,
can
be
fit
constructed
to
contextual
position
always
any purely
analysis
merelyby invertingthe original--regardlessof its styleand technique.
(It should perhaps be pointed out here that the aforementionedinstrumentalobstacles to literal inversionare not so formidableas one
mightthink.Much twelve-tonemusic is conceived in a texturethat,
even when not strictly
polyphonic,nevertheless
dependson an equalizationof voices and registers.When the analysesreferto instrumentation
theyusually do so to point out identitiesand contraststhat can easily
be maintainedunderinversion.)
So far I have said nothingabout the possibilityof another kind of
which,ifacsystematictransformation,
namely,completeretrogression,
as
cepted,would in turnimplythe availabilityof retrograde-inversion
well. AlthoughSchoenberginsiststhat,just as thereis theoreticallyno
"absolute down," thereis no absolute "forwardor backward,"thereare
neverthelessoccasions (as oftenwhen a row is divided among two or
morevoices) when an exact reversalwould failto producea correctsetform.The reversionof a twelve-tonepiece, then,cannot always be depended on to produceanother"correct"twelve-tonepiece. On theother
hand, thereare certainlymany examplesthatcan be reversedwithimpunity,especiallyifone is not doctrinaireand allows the reversalofapas an alternativemethodto thereversaloftimeproximateattack-points
values. And slightmodificationsof the rules governingnote-counting
(such as the optionof countinga note on its lastappearance in a given
context)would open the door to universalretrogression.
To be sure,the distinctionbetweenforwardand backward oughtto
be made froma wider point of view than that of pure note-counting.
Schoenberghimself,later in the above-quoted essay,implies that,regardlessof theory,practice may require such a distinction.His statementthat "One could perhaps toleratea slightdigressionfromthisorder [of the basic set] .. . in the later part of a work, when the set had

already become familiarto the ear,"18suggeststhat a composermust,

sometimesat least,take into account the orderin whichmusical events
take place. But thisrule is vague and by no means self-evident;
besides,
thereare manycompositionsto whichit does not apply,sincetheynever
depart fromthe originalset except in canonical ways. And when these
methodicaldeparturesare used, Schoenberg'srule is frequentlydisre17 Vol.

1, No. 2 (Spring1963),pp. 72-82.

18
Op.cit.,p. 117.

* 36 *

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

BEYOND

ANALYSIS

garded.We have his own example,in the FantasyOp. 49, of a compositionthat beginsby developinga singlehexachord,statingthe definitive
set onlywhen the piece is well under way. And Milton Babbitt'sCompositionforFour Instrumentsreservesits definitivestatementforthe
end, aftera systematictreatmentof derivedsets.
One may neverthelessfeelintuitivelythat somethingis wrong: that
in music, whatever its technique, should have as little
retrogression
generalvalidityas in literatureor in cinema.And certainlycompositions
statement,
planned accordingto traditionalrhetoric-e.g. introduction,
development,climax,restatement,peroration-hardly admit of intelligible reversal.Yet it is just theseelementsof formin the music ofthose
composers,such as Schoenbergand Berg,who relied on older models,
that a later generationhas foundold-fashionedand is tryingto purge
fromitsown music.Accordingly,
it isjust theseelementsthatare ignored
in many analysestoday.
If we search the above-citedessays,we findverylittlehelp in decidingjust whythosecompositionslackinga textmove in the directionthat
theydo, or-a relatedquestion-why theyend just when theydo. The
analyses,withfewexceptions,demonstrateconnections-how one section is related to another-rather than progressions--howone section
followsfromanother. Such relationshipsas repetition,similarity,contrast,common-tonelinkage,and the like,are as independentoftemporal
as of pitch direction.Similarly,discussionsof harmonyconcernthemselveswiththe derivationof simultaneities,
but hardlywiththejustificationof the motionfromone to another;criteriaformelodicconstruction
are nevermentioned.Thus, forpurelyinstrumentalcompositionslacking passages wherethe exigenciesof strictnote-countingdeterminethe
directionof events,forwardand backward indeed seem to be indistinguishable. Webern's fondnessfor the palindrome, which celebrates
musical reversibility,
may be an indicationthat his own thoughtwas
movingin thisdirection.
(In thispaper I have not consideredthe systematictransformations
effectedby equating the chromaticscale withthe circleoffifths.
I leave
to othersthe exerciseof determiningto what extentthe cited analyses
would remainapplicable to versionsso derived.)
So far,none of the transformations
I have discussedhas affectedthe
internalstructure
ofthecompositions
in question.Now, however,I should
liketo suggestthepossibilityofoperationsofthiskind.One ofthepoints
that emergesfroma recentcolloquy among Babbitt,Perle, and Lewin
on the SchoenbergViolin Concerto19is that,althoughit may be imprecise to treattransposition
as analogous to tonal modulation(as Perle at
19 Lewin, op.cit.Perle, "Babbitt, Lewin, and Schoenberg: A Critique," PERSPECTIVES OF NEW
MusIC,Vol. 2, No. 1 (Fall-Winter1963),pp. 120-27; followedby Babbitt'sreply,pp. 127-32.

. 37 *

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

OF NEW

PERSPECTIVES

MUSIC

createthe
one pointseemsto tryto do), transpositions
can nevertheless
froman originally
effect
ofa moreor lesswidedeparture
statedquasiinregister,
butalsoandespecially
harmonic
area-notjustbydifferences
of
of
common-tone
by
relationships
amongsegments twoor moreforms
hexatheset.The numberofsuchcommontones,e.g.betweenthefirst
chordoftheoriginalstatement
and thatofa giventransposition,
might
be a measureof the "harmonic"distanceof the transposition;
and
measuresofthiskindmightthenforma basisfor"harmonic"progressionthrougha piece.To returnnowto a composition
to whichI have
alreadydoneviolence,and whichI intendto manhandlestillfurther,
letus see howthisconceptappliesto Schoenberg's
Op. 33a, and howit
can be usedto composean alternative
to Schoenberg'sdevelopment
an alternative
shouldbe an
that,accordingto theacceptedprinciples,
fortheoriginal.Here are theset-forms
adequatesubstitute
Schoenberg
R and RI rightto left):
uses(withP and I readingleftto right,
I

Po:Bb F
Io:Eb

G
IF B

P7:F

Ab Db D

P2:C
12F

Bb
a

D
E
G

iB
E
C,
E I

F# C

D# G

Bb Gb F

G# D# F
A D C

F# A

A
A

Bb E
G C

Ab G

C# G# A# D

F# E

17:Bb Eb Ab A

Ab D

C#

Eb A

Db C

:Ro

A :RIo
F# :R2

F# E

"R12
:R7
:RI7

Firsttrichordsarrangedin fifths:

P7

12

Io

Db Ab E'b Bb F
17

Po

P2

Common tonesin firsthexachordsof Po and 12: Bb F A F#

Common tones betweenend of developmentand beginningofrecapitulation:C F Bb Eb Ab
Of the above forms,the expositionemploysonly the To; the developmentuses theT2 and T7; the recapitulationreturnsto theoriginalforms.
Now, it can be shownthat,in makinghis firsttransposition(T2), the
composerhas exploitedtwo relationships,both indicated in the above
chart: the common tones of the firsthexachordsof Po and 12, and the
seriesof fifths
implied by the firsttrichordsof Po and Io and explicitly
statedin m. 25. These fifths,
extended,thenhelp to make theconnection
* 38 *

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

BEYOND

ANALYSIS

betweenT2 and T7. But in thispiece everystatementof a P-form(or R)

is complementedby its combinatorialI-form(or RI). The same common-toneand fifth
then,could equally well have been exrelationships,
the
in
the
reverse
directionby usingT-2 (i.e. To10),
ploitedby
composer
followedby T_7(i.e. T5). Thus:
Eb Bb A
Plo:'Ab
Ilo:Db Gb Cb C

C# F

Gb C

:Rio

Bb Ab E

D# A

:RIo

P5 :Eb

Bb F

F# G# C

Db G

:R5

15 :Ab

Db Gb G

Bb E

:RI5

Eb B

Firsttrichordsarrangedin fifths:

Ilo

Io

II

II
Ab...iEb

15

Plo

Cb Gb Db

P5

Bb F

Po

Common tonesin firsthexachordsof Io and P10: Ab Eb G E

Common tonesbetweenend of developmentand beginningof recapitulation:F Bb Eb Ab Db
I have writtenout a hypotheticalnew developmentsectionalong these
lines (Ex. 4).
Schoenberg'sdevelopmentinitiallyexploitsthe Eb-Bb-F fifths;
mine,
the Ab-Eb-Bb. For his common-tonesbetween Ro and I2, I have substitutedthose between RIo and P10. The restof my developmentcan
easily be followedby comparisonwith the original.At the recapitulation,Schoenbergmakesa connectionfromthelast tetrachordsofR7 and
RI7 to the firstof Po and 1o. My version,leading fromR5 and RI5, preservesthe same number of common tones between the tetrachordsof
the developmentand thoseof the recapitulation-five;fourof them,includingthe importantconnectivesBb and Eb (the firstnotesof the recapitulation)are the same as in the original.(In fact,of the eighttones
theend ofmydevelopment,all but one are thesame as those
constituting
ofthe original.)
Beforegoingfurther,I must insistthat my attemptin none of these
has been to improveon, or even to equal, the original.I am
rewritings
merelytryingto show that the analytical methodsused by the essays
citedofferno criteriafordecidingin each case betweenthe twoversions.

. 39*

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

OF NEW

PERSPECTIVES

7.

m. 27

8va-

Sf
i Ww

Idk
'

MUSIC

>i.

vv

-----------8va

go in

Al N

P6,"

---

8va8v

,,?

...

Ii

'".fro!so,"

,.,A

ff

8va-A-F

ta
'

Ex. 4

It is now time fora brieflook at those analysesthat do tryto offer

criteriafordistinguishingup fromdown, forwardfrombackward. My
admittedlyincompletesurveydisclosed several worthnotingfortheir
in thisregard.Claudio Spies's discussionof Stravinsky's
Abraham
efforts
andIsaac,20like Edward Laufer's account of Sessions'Montezuma,21
can
and on other associationsbetween
call on the demands of text-setting
musicand words;but deprivedof these,Spies's analysisof the "Huxley"
Variations22has to fallback on such conceptsas thoseofantecedentand
consequentphrases-usefullyevocative,perhaps,but undefinedin this
context.In the same spirit,Ren6 Leibowitz,in his Qu'est-ce
la Musiquede
makes
between
Webern's
Sons,23
vague
analogies
phrase-construcDouze
a la MusiquedeDouze Sons,24in its
tion and Beethoven's.His Introduction
longanalysisofSchoenberg'sVariationsforOrchestraOp. 31, pointsout
the traditionalmodels the composerused to give shape and temporal
Vol. 3, No. 2 (Spring-Summer1965), pp. 104-26.
Vol. 4, No. 1 (Fall-Winter1965),pp. 95-108.
22Ibid.,pp. 62-74.
23EditionsDynamo,Liege, PierreAelberts,1948.
20
21

PNM,

PNM,

24 Paris, L'Arche, 1949.

40

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

BEYOND

ANALYSIS

directionto his large-scaledesignsbut evokes no furthercriteriaofthe

kind we are seeking-save the quotation of the BACHmotif.Even the
discussionsof orchestration
emphasizesymmetries,
parallelisms,similarities,and contraststhat,as I have already suggested,can easily be retained underinversion.
PeterWestergaardgivesus a glimmerof hope in his attempttojustify
themeterofthesecondmovementofWebern'sPiano VariationsOp. 27.25
He suggeststhatheretheinvariableappearance ofthelowerofeach pair
ofthree-notechordsat the beginningof the measuresin whichtheyapmeter.But even ifone decidesthatBoulez
pear emphasizesthe two-four
is wrongin maintainingthat Webern'smetersare purelyconventional
and notmeantto be observedin performance,26
one mustpointout that
the placementof low chordsin the othertwo movementsgivesus no indication whatsoeverof theirmeters.One would also question whether
the regularappearance of the higherof each pair of chords on strong
beats mightnot equally well establishthemeter.In fact,thismovement,
the inversionof which is such a trivialoperation that it can almostbe
performedat sight,offersa simple and complete demonstrationofthe
problem I am raising. (It should be noted that Westergaard,in his
mentionof "the Haydnesque wit" ofthe two-quarterrestjust beforethe
the originaldirectionofthis
end, does give us one reason forpreferring
movementto its reversal.But one mightwonderwhy,ifsuch a gesture
so clearly-and so wittily-marksthe end, the entiresectionis thenrepeated.And mightnotthesame gesturewittilyserveas an introduction?)
Of the remainingcritiquesthat I have considered,mostofthosethat
make a structuraldistinctionbetween soprano and bass-to put the
problemoftotalinversionin its simplestform-and concernthemselves
withprogression-todo the same withreversal-do so by means of linear and harmonicoutlinesvaguely derived fromSchenker'smethods.
Attemptsof this kind may be seen in two articleson Sessions: one by
AndrewImbrie27and the otherby Edward Laufer.28Richard Swiftalso
movesin thisdirectionin his account ofJ. K. Randall's Demonstrations.29
But what righthas one to call on such devicesin thiscontext?In tonal
music,the motionof the bass can be derivedfromsome expansion30of
the tonicchord; that ofthe soprano,by passing-motion
withinthe scale.
But what does eithertonicor passing-notemean whenthereare no previouslyor permanentlydefinedchords,and no functionallyoperative
25"Webernand 'Total
Organization,'" PNM, Vol. 1, No. 2 (Spring1963),pp. 107-20.
2me cahier,1948,p. 67.
26"Propositions";
Polyphonie
27"Roger Sessions:In Honor of His
Sixty-Fifth
Birthday,"PNM, Vol. 1, No. 1 (Fall 1962),
pp. 117-47.
28Op.cit.
29
PNM, Vol. 2, No. 2 (Spring-Summer
1964), pp. 77-86.
30I.e., by theelaborationoftheintervalbetweenrootand fifth.

. 41 *

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

PERSPECTIVES

OF NEW

MUSIC

scales? Can this music really be approached through attitudesand

habitsderivedfromlisteningto tonal music?And would a tentativeand
qualified assent to that question commit us to an acceptance of the
tonal analogies Spies findsin the late Stravinsky,31
or to approval of
Martin Boykan'sstillbolder tonal approach to the same composer32not to speak of Hindemith'srigidapplicationofhis own tonal principles
in his well-knownanalysis (or mis-analysis)of a passage fromour old
in thesame
friendOp. 33a?33Or should we put all such interpretations
categoryas the explanation of the French word-sequencePas de lieu
Rhonequenousas makingsense in spokenEnglish?
be takenat
Again,shouldeven the presenceof clear triadicreferences
tonal face-value?Leibowitzrecognizesthe possibilityof theircreatinga
"tonalite vague,incertaine,"especially in the worksof Berg, although
ultimatelyit is the "logique du maniementseriel" that must provide
justificationforall that happens.34Rufer,on the otherhand, seemsto
believe that such tonal impressionsare more illusorythan real, and at
any rate are uselessforour purposes:
Thus triads of tonal structurecan appear too, as, forinstance,the
in
"Ode to Napoleon" shows. But these, like all chord-structures
twelve-notemusic,are ofpurelylocal importanceand do not produce
harmonicprogressions
whichhave the effectof creatingform,as happens in tonal music; forthe relationshipto the key-noteis missing.35
Who is right?
The factthatone can raise such questionsshowsthatwe have arrived
at a crucial point in the historyof Westernmusic. Up until now there
has been no ambiguitybetween up and down-at least not since the
in effectfromthe fifth;therehas been no quesfourthwas distinguished
tion of choice betweenforwardand backward since the appearance of
the melodic cadence-and, later and a fortiori,
the harmoniccadence;
therehas been no transpositional
relationshipthatcould notbe explained
by referenceto some sort of tonic. But these aspects of composition,
hithertoaccepted as basic, are apparentlyunaccounted forby twelvetone theory.
If one acceptsthisconclusion,one can adopt one ofthreeattitudestoward it. One can welcomeit wholeheartedly,
agreeingthat therereally
is no basis forchoice among my hypotheticalversionsbeyond the con31 "Some Notes on Stravinsky's
Requiem Settings,"PNM,Vol. 4, No. 2 (Spring-Summer
1967),
pp. 98-123.
"
32 'Neoclassicism'and Late Stravinsky,"
PNM, Vol. 1, No. 2 (Spring1963), pp. 155-69.

33 The CraftofMusical Composition,Book I, New York, Associated Music Publishers, Inc.,

1937,pp. 217-19.

34Introduction,
pp. 282-85.
35 Op.cit.,p. 126.

. 42

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

BEYOND

ANALYSIS

venienceof acceptingwhat is already givenand the comfortoffamiliarity.But that only throwsthe problem back where it really belongsin
the firstplace: on the shouldersof the composer.How did he make his
decisionsin thesematters?
This leads us to a second point ofview: that twelve-tonetheoryis as
yet incomplete,and that the superiorityof one versionofa composition
over another depends on purelyformalfactorsas yet unanalyzed but
neverthelesseventuallyanalyzable, analogous to the laws of linearand
harmonicprogressionin tonal music,possiblysimilarto those but not
necessarilyso. A compositionis successfulinsofaras its composer has
made his impliedchoicesamong conceivablealternativesin accordance
withhis intuitive,or, better,his partlyrational understandingof these
presumedlaws.
Finally,one can accept the primacyofthecomposer'sconcretechoices
but insistthat,farfrombeing made in obedience to laws knownor unknown,they are so fundamentalto the composer'sconceptionof his
workas to belong,so to speak, among its basic assumptions.They are
determinedby what may be called absolutedecisions,
i.e. decisionsfor
whichno adequate analyticalreasonscan ever be adduced.
If many of us at firstglance opt forthe second point of view, it is
because the success of theoristsof tonality,notably Schenker and his
has givenus hope thatall thesecretsofcontemporary
followers,
compositionawaitanalogoustypesofexplication.But a moresophisticated
generationof theorists-asexemplifiedby Milton Babbittand Michael Kassler
-has been pointingout what a flimsysystematicbasis even Schenker's
splendid constructionrests on.36 In tryingto establish tonal theory
more firmly,they dismissSchenker'sappeals to Nature, the Human
Spirit,and the Overtone Series,in favorof a strictlylogical systemderived from-a limitednumberof axioms and rulesofinference.For these
axiomstheyoffer--naturally-noproofwhatsoever.But ifwe acceptthis
approach, we must admit the possibilityof equally consistentsystems
that we mightcall anti-tonal.By regularand easilydefinablemodifications of the axioms and rules of inferencesuch systemscould lead to
or invertedretcompositionsthat are the totalinversions,
retrogressions,
of conventionaltonal compositions.Other transformations
rogressions,
too, are possible.Deprived of all natural bases, what appeals could the
conventionalsystemmake againstsuch rivalssave thoseof convenience,
here to note thatin
tradition,custom,and familiarity?
(It is instructive
36See Milton Babbitt,"The Structureand Functionof Musical
Music
Theory: I," College
v, Fall 1965,pp. 49-60; Michael Kassler,"A TrinityofEssays,"a dissertation
for
Synposium
the Ph. D. in the Departmentof Music, PrincetonUniversity,
1967. The essaydealingwith
the twelve-tone
systemwas publishedin PNM,Vol. 5, No. 2 (Spring-Summer1967), pp. 180, as "Towarda TheoryThat is the Twelve-Note-Class
System."

. 43 *

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

PERSPECTIVES

OF NEW MUSIC

theearliercase we couldperform
thehypothetical
on indioperations
vidualworks,
fortheoperations
themselves
constitute
"rulesofinference"
ofthesystem.
Sincethisis nottrueoftonalmusic,theoperations
must
be appliedto thesystem
as a whole,notto individual
works-a possibilof harmonic
ityadumbratedin the case of inversionby proponents
dualismfromZarlinoto Riemann.)
We can perhapsrecognizehereone motivethathas drivenso many
theorists
to findsomekindofsupportin theexistenceoftheovertone
and
withthemeventhoughwe cannotfollow
series, we can sympathize
them.Theyseemtoconsidertheroleoftheseriesas somehowanalogous
tothatofgravity
inarchitecture:
a rawfactofphysics
thatmustbe taken
intoaccountincreating
viablestructures.
Buttheanalogycanbe turned
isa successinsofar
as itdefeats
Moreagainstthem:everybuilding
gravity.
the
vault
and
the
cantilever
attest
the
of
over, gothic
futility arguingthat

gooddesignis necessarilybased on thevisualexploitationofphysicalprinciples.True,the overtoneseriesdoes indeed make a distinctionbetween

up and down withinthe individual tone,since overtonesare, afterall,
above the fundamental.Furthermore,one must take account of the
series in the physical constructionand practical use of instruments.
Neitherofthesefacts,however,justifiesthe claim that the auditive
structureofmusic,whethertonal or not,necessarilydepends on the compositionof the series.In fact,only today, throughelectronicmeans, is it
becomingpossibleto integrate,in a systematicand thoroughgoing
way,
overtonespectra,whethernatural or artificial,intomusicalstructures.37
Ironically,thesame media now offerforthefirsttimethetheoreticalpossibilityof invertingthe audible spectra.Such completetone-colorinversion would at last deprivethe individual tone itselfof the possibilityof
distinguishing
up fromdown!
If now, in spite of the discouragingexample of the tonal system,we
stillinsiston seekingsome basis formakingdistinctions
thatwe stillfeel
to be somehowessential,let us turnto the thirdalternative:that there
is, and can be no analyticalgroundforconcretemusical choices,i.e. no
groundwithinthe internalstructureof the music itself;yet that these
choicesare crucial in determiningmusical values, i.e. salientcharacteristics that afforda basis for distinction,comparison,and judgment.
(Critical listeners,as well as composers,must also make such choices,
althoughin a slightlydifferent
sense; forall judgmentsare based on imbetween
actual
and possiblecompositions,
and hence
plicitcomparisons
on a choice among concretevalues. Indeed, itwas fromthispointofview
thatwe initiallyapproachedtheproblem.)To put thepositionsuccinctly:
37See, forexample,J. K. Randall: "Three Lecturesto Scientists,"PNM, Vol. 5, No. 2
(Spring-Summer1967), pp. 124-40. The thirdof these,"Operationson Waveforms,"
deals
withthispossibilty.

. 44 *

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

BEYOND

ANALYSIS

concrete
musicalvaluesdependonabsolutedecisions.Remember that by absolute

I do not mean arbitrary:theremay be, as we shall see, good reasonsfor

makingone choice and not another.By absoluteI mean independentof
purelyanalyticalconsiderationsand unsusceptibleof purelyanalytical
justification.
Let me tryto clarifythispoint by referring
to anotherart,thistime
an
artist
is
monochromatic
a
painting. Suppose
painting
picture,or
a
formal
element
of
the
making
simply
drawing.Every
pictorialstructurewill thendepend on line and light-value,not on color-relationships.
But how does the painterdecide what color to use? He mightrule out
certaincolorsas incapable ofsustaininghis design-yellow mightbe too
light,forinstance;but he would stillhave a wide range of choice.If the
decisionis not a purelycapriciousone, it mustbe based on reasons;but
thesereasonscannotbe analytical,sincethe internalstructureofthepicturewill be the same in any case. The reasonsmustthereforebe external to the structure.The picture may be intended for a room with
a given color-scheme.The artistmay feel that a warm or a cool color
mightbe more appropriateto the subject of the picture.He may even
feelthatone color has a vague expressivevalue consonantwiththe subject. Or he may simplyrevelin the sensuousqualityofone color.
Let us take anotherexample,one somewhatanalogousto theproblem
with which our discussionbegan. How does an artist(or an observer)
decide which way a pictureshould hang-which way is right-sideup?
Good designseemsoftento be independentofwhetheror notitis inverted
-an assumptionsupportedbythehabit,commonamongpainters,oftestingtheircompositions
byviewingthemupside-down,as wellas bythefreof
in
mistakes
the hangingof abstractions.(We seem to measure
quency
balance with referenceto a verticalaxis, possiblybecause of our own
rotationis seldoma live
physiologicalorientation,so that ninety-degree
Carl
Pickhardt
has
aboption-although
experimentedwith free-form
stractionsthat can be hung at anyangle.) In the case of a representational picturethe answerto our questionis obvious-unless the artistis
Chagall (or, apparently,sometimesMatisse, whose Le Bateau hung
upside-downin the Museum of Modern Art in New York City from
Oct. 18 to Dec. 4, 1961).38But we arriveat thisanswerby a reference
outsidethe picture-to the depicted subject.Indeed, fromthe pointof
view of pure design,the orientationofa picturemustoftenbe based on
an absolutedecision-one made withreference
to representational
rather
than to structuralvalues. Apparentargumentsfromdesignwill in such
cases merelyconceal externalreferences.
For example,to the claim that
a landscape musthang as it does because the lighterarea looksbetterat
38 Norris and Ross McWhirter, GuinnessBook of WorldRecords,Rev. and enlarged edition,

New York,BantamBooks,1966,p. 157.

* 45

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

PERSPECTIVES

OF NEW

MUSIC

forthispreference
is
thetop,one can counterthattheonlyjustification
thatthisisthewaylandscapeslookinnature,and onecan pointtomany
in whichthelighter
areasarebelow.How,in fact,doesone
abstractions
How doestheartisthimofan abstraction?
determine
theorientation
the
from
selfmakethatdecision?In theabsenceofanyclearindication
which
it
is
based
The
on
must
be
reasons
the
decision
absolute.
design,
theartistsayssimply,
willbe external
whether
to thepictorial
structure,
effect
of
"Thisis thewayI likeit,"or morespecifically,
"The expressive
thepicturewouldbe harmedifit wereinverted."
We havearrivedhereat an important
valuesin any
point.Expressive
art-if theyexistat all-depend on concrete
values.Theycannotarise
fromanalyticalvaluesalone.How couldthey?Unlessonewishesto explainwhatit couldpossiblymeanfora workofartto "expressitself,"
thenone mustagree that expression,
by its verydefinition,
implies
a relationship
betweentheworkofartand something
else;whileanastructure.
Thisis in no
lyticalvaluesare derivablepurelyfrominternal
thatstructure
has nothingto do withexpression.
waymeantto suggest
in a verballanguagedependson bothsemantics
Justas communication
and syntax,
mustinvolvebothconcreteand anaso artistic
expression
thestructure
couldconveyno meslyticalvalues.Withouttheformer,
thelatter,
themessagewouldbe limitedto theequivalent
sage;without
ofprimitive
substantives
and exclamations.
Thus theexpressive
power
ofan abstract
canvascannotstemfromitsdesignalone;it mustdepend
in parton somecovertrepresentational
or otherassociative
element(as,
forexample,theillusionof"mass"or "movement").
The foregoing
thatthosewhowishto makespecialclaimsfor
suggests
theroleoftheovertone
seriesin tonalmusic,or forwhatcan be much
moreeasilydefended,
theprimacyofthefifth,
a morefruitful
analogy
in architecture
thanthatofgravity
be
ofrepresentation
that
in
might
For
whereas
law
of
is
a
that
all
nature
controls
construcpainting.
gravity
tioneventhoughitmaybe apparently
refuted
totheeye,representation
is merelya reference
to naturethatcan be utilizedor not according
to the purposesof the artist.Similarly,even if one holds thatthe
in tonalharmony
ofthefifth
derivesfroma naturallaw,one
supremacy
mustadmitthata greatdeal ofmusicignores
it;henceitmustbe a law
in a different
senseofthewordthanthelaw ofgravity.
Yetit couldstill
be a law to thisextent:thatin all musicthatexploits
ina tonal
thefifth
to fundamental,
whether
due to defisense,thespecialrelationoffifth
nitethough
ill-defined
rootsinphysical
and anatomical
orsimply
nature,
to thegrowing
forceofconventional
habitoverseveralcenturies,
inevitheorientation
ofthemusic,i.e. itsdirection
bothin
tablydetermines
determines
theorientation
of
pitchand in time-justas representation
a picture.
S46

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

BEYOND

ANALYSIS

Such a view of tonalityis by no means inconsistentwith the recent

attemptsto explain the systemaxiomatically.It merelyinsiststhatsuch
explanationscan neveradequately deal withthe problemoforientation.
If tonal music carrieswith it its built-inorientation,then it is built in
absolutely,not analytically.It rests,not on the internalconsistencyof
the system,but on some connectionbetweenthe axiomsand rulesof inferenceon the one hand, and the externalworldon the other-whether
thatworldis representedby acoustics,psychology,
or history.
physiology,
The orientationis, so to speak,semanticratherthan syntactic.
One who accepts the analogy implied in the last sentencemay be
and admit the relevanceof tonal orientationto the
willingto go further
problemofmusicalexpression.If the effectofthe fifthin tonal musicis,
to some extentat least,independentof contextand externalto pure design,then elementsof musical forminferablefromthe role of the fifth
could serve as vehiclesof some of the associative
(e.g. tonal
'cadences)
elementsnecessary
to expression(e.g. theassociationofa perfectcadence
with fulfillment
or satisfaction).It is temptingto say of such instances
that the structurealone is the vehicle of the expression;and fromthis
erroritbecomeseasy to generalizeto theextentofbasingall musicalexpressionon puresyntax.That is because tonalmusicmarriesthesemantic
and syntacticaspects so closelythat it is difficult
to conceive of the semantic elementin isolation.One should reallyspeak here,not ofsyntacticand semantic,or of analyticand concrete,but offusedvalues; for
in the best tonal music the two aspectsof tonalityare indeed indissoluble. But recentmusichas suggestednew possibilities.
Justas representational implications(such as thoseofmass and motion)can impartsome
meaning to a pictorialabstraction,so tonal referencescan functionin
non-tonalmusic, not so much syntacticallyas associatively,bringing
withthemimplicationsoftheorientationaland expressivevaluesinhering
in tonal contexts.At the same time thesereferences,
arisingas theydo
fromsyntacticaloriginsin tonal music,must,iftheyare to be successfullyemployed,satisfywhateversyntacticalexpectationsof thisnature
theyarouse. Such referencesmay vary from,say, a bald statementof
consonanttriadsto a generalizedadaptation of melodic-harmonic
relationshipsand phrase-structure.
Thus music whose syntaxis primarilytwelve-tonemay nevertheless
legitimatelycall upon implicit tonal functionsto clarifyits concrete
values-so long as the functions,once summoned,are permitted,so to
A completeexplicationofthis
speak,to fulfilltheirtonal responsibilities.
music mustthentake thesetonal allusionsinto account-whether they
are overt,as is oftenthe case with Berg, or concealed, as in much of
Schoenberg.(Note, forexample, in Op. 33a, the V-I effectcreated by
the bass connectionBb-Eb fromthe developmentinto the recapituS47

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

PERSPECTIVES

OF NEW

MUSIC

lation-an effectsignally,and perhapsdisastrously,

lackingin my
version.)
whether
or notto utilize
can chooseforthemselves
Todaycomposers
had
Forcenturies,
ofcourse,theindividualcomposer
tonalreferences.
no suchoption.The decisionwas alreadymadeforhim,justas thedewas alreadymadeforthepainter.
cisionas to theuse ofrepresentation
chooseconcrete
well
as
as individuals,
Nationsand historical
periods,
absolutedecisions;hencewe can speakofnationaland
valuesthrough
as wellas individualstyles.That is whatstyleis: thetotality
historical
a givenbodyofworkas a whole.
oftheconcretevaluescharacterizing
deterThe stylistic
decisionofa groupmayseemto be so completely
not
should
or
that
it
minedbyevolution,
environment,culture,
properly
in thesamewayas an indibe calleda decisionat all; yetit functions
forit resultsin one modeofactionthatrulesoutall alvidualdecision,
thesedeciternative
modes.Perhapsbecauseoftheirdeterministic
origin,
thanhispersonalchoices,
sionsare evenmorebindingon theindividual
carrieswithitcerwhichmayvaryfromworkto work.Thusiftonality
are boundto leavetheirmark
thentheseassociations
tainassociations,
ofthe
as thepresence
on all musicofthetonalperiodjustas inevitably
affect
the
is
bound
to
content
of
human
figure
depicted
realistically
the
the
end
of
the
nineteenth
from
Renaissance
to
up
century.
painting
Tonal functions
are, to be sure,not the onlysourceof associative
values.Once one admitstherelevanceofthesevaluesat all, one finds
theminvolvedin almosteveryarea of concretemusicalchoice.And
oftonality,
we findthatmanycononcewe leavethespecific
problems
in many
cretevalueshave been equallyat the disposalof composers
Butall thesevaluespresuppose
absolute
styles,
usingdiversetechniques.
decisions;so, althoughthe tonal composermay neverhave had to
was running
in therightdirecor nothiscomposition
wonderwhether
was constantly
confronted
tion,evenhe,likehispresent-day
successor,
bychoicesthatcouldneverbe madeon analyticalgroundsalone.How
did he determine
ofmostcompositions
tempo?The internalstructure
limits
within
which
must
be sought,butthese
certain
a
imposes
tempo
limitsareoftenverybroadindeed.Wecan all thinkofcompositions
that
wouldstillmakeperfect
musicalsenseiftakenat a tempotwiceas fast
ortwiceas slowas thatindicated;whythenshouldtheindicatedtempo
have precedence?Because the composerchose it? But whydid he
chooseit?
is anotherexample.How wouldthestructure
oftheChopin
Register
ifthepiecewerewritC minorPrelude,or ofhisFuneralMarch,suffer
tena fifth
higher-oreven an octavehigher?Yet such transpositions
wouldmanifestly
ofthepieces,and hardlyforthebetter.
altertheeffect
(RogerSessionsreportsthathe once succeededin turningScriabin's
S48

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

BEYOND

ANALYSIS

"Black Mass" Sonata [No. 9] into a White Mass by playingit an octave

higher,and in turningthe "White Mass" [No. 10] intoa Black Mass by
reversingthe process.)
Even instrumentation
depends to a large extenton absolutedecisions.
This is especiallyeasy to demonstratewith regard to monochromatic
media, where the problemsof interrelationships
among colors hardly
arise. Beethoven,forpractical purposes,was willing to transcribehis
GrosseFugeforpiano four-hands.Brahmsdid the same forhis twostring
sextets.They are funto play that way-but it is hard to get anyoneto
listen.Why? What crucial analytical values, presentin the stringversion,are lostin the transcriptions?
Decisionsin thesemattersmustbe made by all composers,regardless
of styleand technique. Each one of them determinescertain concrete
values that,moreover,are associativevalues; and whetherwe likeit or
not, theseassociationsare bound to inherein the music itself.Tempo is
inevitablymeasuredby unconsciouscomparisonswith rates of human
action; registerrelatesitselfto our conceptsofheight,weight,and mass;
tone-colorbringswithit obviousconnotationsofall kinds,fromourtendency to identifymelody with the human voice to resemblancesofthe
sortthat so delightedthe littleStravinskyin the "dubious" noisesproduced by the red-hairedpeasant.39Many otherareas in whichassociativevalues are unavoidablyimpliedwillcome to mind: absolute dynamics,melodicdirection,rhythmicand metricpatterns.Again,whetherthe
associationsare in some sense "natural" or whetherbased on generationsofconditioning,
theycannotbe escapedbyanyonemusicallytrained
in the Westerntradition.
To be sure,choicesin theseareas are influencedbystructure-butthey
controlstructureas well. Insofaras theycharacterizeeven the primitive
gesturesof the composer'sinitial ideas, and hence precede the musical
design itself,they are independentand necessarilydefyanalysis.The
designmusttake shape in accordancewiththeirdirections.
If one acceptsthe possibilityand the relevanceofmusical expression,
one may indeed feelthat one's decisionshere are governed,consciously
or unconsciously,by the expressivepotentialitiesof the associationsinherentin one's concretechoices. Or one may insistthat the decisions
are, in everysenseofthe word,absolute.What I suspect,but am unable
to prove,is that any concretechoice made on the basis of pure personal
functionsin the same way as one made withexpressiveintent
preference
and thatthe two may indeed be equivalent.To put the case at itsmost
trivial: Why is my compositionsuperiorto its inversion?Because its
melodydescends.Why should the melodydescendratherthan ascend?
Because I likeit thatway! But whydo I likeitthatway?Because I wantit
An Autobiography,
New York,Simonand Schuster,1936,pp. 3-4.
39Igor Stravinsky,

* 49 *

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

PERSPECTIVES

OF NEW

MUSIC

to have the effectthat can be produced by descentbut not by ascent.

Or-because I want it to expresswhateverit is thatdescentcan express
and ascentcannot express.
It should now be obvious that what I have been calling concrete
choicesare in many cases not choicesat all, in the senseof representing
the exerciseof a live option.The absolute decisionsof a composer-for
thismelody,in thistempo,in thisregister,
forthisinstrument-areseldom
the resultof the consciousdismissalpf otheralternatives,even though
any voluntaryaction impliesthe rejectionof everyotheractionpossible
on thatoccasion.The composerdecideswhat piece he is goingto write,
not all the pieces he is not going to write; what I have been calling
choicesare reallythe assumptionsbasic to his conceptof thatpiece. Yet
thereare certainoccasions,especiallyfrequentin connectionwiththe
developmentof previouslystated ideas, that do seem to offerseveral
workable alternatives.As I triedto show by means of a change in the
to advance analyticalreasons
developmentofOp. 33a, it is oftendifficult
to justifyone's choice at such a point; we may perhaps now be willing
to admit the example as evidence that the domain of a composer'sabsolute decisionsembraces even the internalstructureof a twelve-tone
piece. As a finaltask,I shall tryto show the same principleat workin
tonal composition.
For obvious reasons,it would rarelybe possibleto invertsuccessfully
theharmonicdirectionofa tonaldevelopmentas I triedtodo withSchoenberg's.But one fieldof choice presentsitselfwitha high degreeofregularity:the opportunityof changingmode. Once the conventionof the
it became a matterof the composer's
tiercede Picardiewas overthrown,
choice whethera piece in minorended in major or minor; later on, in
the nineteenthcentury,it became increasinglycommon for worksin
major to end in minor.In many cases it seems impossibleto findadequate analytical reasons forthe ending actually adopted. Think over
Schubert'sMoment
MusicalNo. 3 in F minor.Can you adduce any analyticalevidence forthe inevitabilityof its conclusion?Could you not
rewritethe coda so that it ended convincinglyin minor?Compare the
MusicalNo. 4 withChopin's Etude in E minorOp. 25
C#minorMoment
No. 3 and hisNocturnein C minorOp. 48 No. 1. All threeofthemmove
to the tonicmajor in the middle section,so that all have, so to speak,a
motiveforending in major. Only the Etude does so; the Nocturneremains in minor;while the Moment
Musical,aftera shortreferenceto the
major section,returnsto minor.And what of Chopin's Nocturnein B
major Op. 32 No. 1? Is thereany necessaryreason forits conclusion?
(And just what is this conclusion,by the way? Some editionsend in
minor,some in major. Historical evidence seems to favorone overthe
other.Would you be willing,on analytical grounds,to decide which?)
S50

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions

BEYOND

ANALYSIS

If you deny that theseromanticexamples are in the frontrankoftonal

thenworkon Beethoven'sStringQuartetin F minorOp. 95.
structures,
You can perhapsjustifycertainaspectsof the coda by analysis,but that
is notthesame as provingitsinevitability.
Can you,on internalevidence,
show thatjust this coda, in this tempo,above all in major, is the only
possibleendingforthisquartet?
Whether Schubert,Chopin, and Beethoven-or, to returnto our
originalproblem,Schoenbergand Webern-made theirdecisionson expressivegroundsor whethertheywrotetheircompositionstheway they
did simplybecause theylikedthemthatway,mypointis thesame: their
reasonsare beyondanalysis.And ifwe as criticallistenersconcludethat
the composerswere right,it should not disturbus to findthat our own
reasonsare oftenbeyondanalysis,and that,when we tryto explainthe
of a compositionover any alternativeversion,sometimesall
superiority
we can say is, "It soundsbetter."
A greatdeal ofcurrentwritingon music seemsto implythatnothing
about composition,or nothingimportantabout composition,is beyond
analysis. But surelythe single most importantthing anyone can say
about any compositionis beyond analysis: namely,"I like it." It is especiallydisturbingto findthatmanyyoungcomposers,who presumably
writeabout the musicof othersthe way theythinkabout theirown,are
eitherinsensitiveto non-analyticalvalues or-as I thinkmore likelyafraidto admittheirimportance.As a resulttheyoftenseem to be writing,not about actual compositions,but about abstractionsderivedfrom
compositions.Now, I recognizethe greatdebtwe all owe to increasingly
rigorousmethodsof analysis,and I am fullyawake to the dangersof
criticism;yet I findmyselfcompletelyon the side ofthe
impressionistic
youngcomposer-a ratherwell-knownone-who, when asked whyhe
wroteas he did, replied,"I like the tunes."

51

This content downloaded from 194.27.18.18 on Mon, 22 Feb 2016 03:02:13 UTC
All use subject to JSTOR Terms and Conditions