Analysis Today

Author(s): Edward T. Cone

Source: The Musical Quarterly, Vol. 46, No. 2, Special Issue: Problems of Modern Music. The
Princeton Seminar in Advanced Musical Studies (Apr., 1960), pp. 172-188
Published by: Oxford University Press
Stable URL: http://www.jstor.org/stable/740369 .
Accessed: 24/06/2014 21:24

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.

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

http://www.jstor.org

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
 ANALYSIS TODAY

By EDWARD T. CONE

THE analysis of music-especially of traditionalmusic--is one of

the most respected of theoreticaldisciplines,but the respect in
which it is held would do it a disserviceif it preventedthe periodic
re-evaluationof the subject. What is analysis,or what ought it to be?
What are its purposes? To what extent are traditionalconcepts and
methodsapplicable to new music? What are the relationsof analysisto
performanceand to criticism?My title refersto a discussion,fromthe
pointof view of today,of thesequestions;it is in no way meant to imply
that I have a new systemto promulgate,or that I have made startling
discoveriesabout new music.
I

Rather than presentingat the outseta naked definitionof the term

under consideration,let us begin by lookingat a familiarexample. The
firstfew measuresof Tristan have performedmany servicesotherthan
theiroriginalone of openinga music-drama;let them serveyet another
and open the argumenthere.
Ex. I

This chordal sequence can be accurately enough described as a

minor triad on A, a French sixth on F, and a pimary seventhon E;
but such a description,revealing nothing of the relationshipsamong
the three chords, involves no analysis whatsoever. If, however, one
5 67
refersto the passage as I -II 4-V he has performedan elementary
31

172

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
 Analysis Today 173

analyticalact: he has related each of the chordsto a tonic, and hence

to one another. He has made a discovery,or at least a preliminary
hypothesisto be testedby its fruitfulness
in leading to furtherdiscovery.
But the analysis as such ceases with the choice of the tonic; once this
has been made, the assignmentof degree numbersto the chordsis pure
description.If, on the otherhand, one pointsout that the second chord
stands in a quasi-dominantrelationto the third,he is doing more than
simply assigning names or numbers: he is again discovering and
explainingrelationships.
Ex. 2

I *""* U.

Turning now to the actual score,the analystmightbegin a program

note thus: "The risingleap of the 'cellos from A to F is succeeded
by a chromaticdescent,followedin turnby ..." He need not continue;
this is pure description.But when he points out that Example 1 repre-
sents the chordal skeletonof Example 2, he is once more on the right
track.He can go stillfurtherby showingthat all the appoggiaturashave
half-stepresolutions,and that the motif so created is augmented in
the motion of the bass, and paralleled in the alto, in such a way that
the chordal progressionof measures2-3 becomes an amplificationof the
melodic half-stepof measure 1.

Ex. .
- 9r'-..-~
3

The fact that in the above diagram no such analogy has been pointed
out in the half-stepsE-D$ and A-An is in itselfan importantthough
negative part of the analysis, since it implies by omission that these
if relevantat all, are incidentaland subordinate.
progressions,

Going one step further,one might claim that, from a serial point
of view, the opening sixth is imitatedin the thirdE-G# (see Ex. 4).
This is the point at which analysisproper passes over into what I call

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
174 The Musical Quarterly
Ex. 4

L
_

prescription:the insistenceupon the validityof relationshipsnot sup-

ported by the text. In the above case, for example, the orchestration
implies the wrong-headednessof the suggestion,since the opening
interval,played by the 'cellos alone, is heard as a unit, whereas the
E-G$ is divided disparatelybetween 'cellos and oboe.
Analysis,then, existsprecariouslybetween descriptionand prescrip-
tion, and it is reason forconcernthat the lattertwo are not always easy
to recognize. Description is currenttoday in the form of twelve-tone
counting-necessary,no doubt, as preliminaryto furtherinvestigation,
but involvingno musical discrimination on the
whatsoever.Prescription,
other hand, is obvious in the absurd irrelevanciesof Werker'sanalyses
of Bach but is equally inherentin some of Schenker'smore dogmatic
pronouncementsand in those of his followers.
It should be clear at thispoint that true analysisworksthroughand
for the ear. The greatestanalysts (like Schenkerat his best) are those
withthe keenestears; theirinsightsrevealhow a piece of musicshould be
heard, which in turn implieshow it should be played. An analysisis a
directionfor a performance.
In order to explain how a given musical event should be heard,
one must show why it occurs: what preceding events have made it
necessary or appropriate, towards what later events its function is
to lead. The compositionmust be revealed as an organic temporal
unity, to be sure, but as a unity perceptibleonly gradually as one
moment flows to the next, each contributingboth to the forward
motion and to the total effect.What is often referredto as musical
logic comprisesjust these relationshipsof each event to its predecessors
and to its successors,as well as to the whole. The job of analysisis to
uncover them explicitly,but they are implicitlyrevealed in everygood
performance.Description,restrictedto detailingwhat happens, fails to
explain why. Prescriptionoffersits own explanation, referringto an
externallyimposedscheme ratherthan to the actual courseof the music.
One more familiarexample may clarifythisview of logical- or, as
I preferto call them,teleological- relationships.

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
 Analysis Today 175
The recapitulationof the PrestissimofromBeethoven'sSonata Op.
109 burstsin upon the developmentin such a way thatthe II$ (V of V)
is followedimmediatelyby I. From a narrowlydescriptivepoint of view
one could call thisan ellipsis,pointingout that the normallyexpectedV
Ex. 5

i Le I

Ag I` ~I =f

has been omitted.Looking ahead, however,one will find that the first
phrase of the recapitulationends on V, and its consequent on I. The
puzzling II4, then, only temporarilyand apparentlyresolved by what
immediatelyfollows it, actually points ahead in such a way that the
whole passage is bound togetherin a cadential II-V-I. The propulsion
thus generatedis given an extra spurtby the compressedII-V-I at the
end of the consequent,and the forwardmotion is renewed with fresh
energyby the elisionthat sets the next period going.

II I

I need hardlymentionthe obvious effectsof such an analysison the

performanceof this passage. Whateverdoubts one had as to the proper
placing of the main accent in thesephraseswhen theyfirstappeared can
if need be, in the
now be resolved; the expositioncan be reinterpreted,
new light of the recapitulation.
II
It should be apparent at this point that analysis- and hence per-

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
176 The Musical Quarterly
formanceas it has been discussedabove - cannot apply to certaintypes
of compositionin vogue today. When chance plays the major role in
the writingof a work, as in Cage's Music for Piano 21-52, logic as
defined above can take only an accidental part. The same is true of
music written according to a strictlypredeterminedconstructivistic
scheme, such as Boulez's Structures.In neithercase can any musical
eventbe linkedorganicallywiththosethat precedeand thosethatfollow;
it can be explained only by referring to an externalstructure- in the
one case the laws of chance and in the otherthe predetermined plan. The
connectionsare mechanisticratherthan teleological: no event has any
purpose- each is thereonly because it has to be there.In a word, this
music is composed prescriptively, and the only possible or appropriate
analyticmethodis to determinethe originalprescriptive plan. This is not
analysis but cryptanalysis - the discovery of the key accordingto which
a cipheror code was constructed.(If we are lucky,the composeror one
of his initiateswill spare us a lot of hard workby supplyingus with the
key.)
A third categorythat does not permit analysis is representedby
Stockhausen'sKlavierstiickXI, where improvisationis given such free
rein that it actuallycreatesthe formof the workanew at each perform-
ance. Thus KlavierstiickXI does not exist as a single compositionand
cannot fruitfully be treatedas one. Each new renditioncan be discussed
on its own merits,to be sure; but the relationshipof all such versions
to the abstractidea of the piece as a whole, and the decision as to the
estheticvalue of such an experiment- these problemscan be argued
endlessly.At any rate theyare far afieldfromthe practicalconsiderations
that are our concernhere. (It need hardlybe pointedout that improvi-
sation as traditionallyapplied to the frameworkof a Baroque concerto,
for example, had purposesquite different. A cadenza servednot only to
show off the soloist's virtuositybut also to punctuate an important
cadence; the soloist'selaborationof a previouslystatedorchestralmelody
clarifiedthe dualisminherentin the form.The qualityof a givenrealiza-
tion depended on its appropriatenessto the compositionalsituation; the
performancedid not, as in many present-dayexamples, create the
situation.)
III
The analysis of music of the periods closelyprecedingour own -
the 18th and 19th centuries- has almost always assumed the applica-
bilityof certain familiarnorms: tonally conditionedmelody and har-
mony,periodicrhythmicstructureon a regularmetricalbasis. Naturally

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
 Analysis Today 177

such standards cannot be applied uncriticallyto the music of our own

century,but on the other hand they should not be dismissedwithout
examination.I contend that, in a more generalizedform,they are still
useful. Regardless of vocabulary, linear and chordal progressionsstill
show strikinganalogies to older tonal procedures,analogies that are in
turnreinforcedby rhythmic structure.Only in thoserare cases wherethe
music triesto denythe principleof progression(as in the examples cited
in the immediatelypreceding section) are such analogies completely
lacking.
This pointof view is moregenerallyaccepted withregardto harmony
than to melody,perhaps because harmonic analysis is the more firmly
entrencheddiscipline.Afterall, formanymusicianstheoryis synonymous
with harmony,melodybeing supposedlya freecreativeelement,neither
in its compositionnor in its perceptionsubjected to rule. (They forget,
of course,that the object of the studyof counterpointis primarilythe
construction,and only secondarily the combination, of melodies.)
Whereas Hindemith'senlargementof traditionalharmonyto encompass
present-dayvocabularies is generallyknown and often applauded, his
attemptto finda melodic framework,actually a much less questionable
procedure,is oftenignored.
Anotherreason forshunningmelodic analysisis that it is not always
easy or even advisable to abstractthe purelylinear elementfroma pro-
gression.Wagner, in such motifsas the Wandererand the Magic Sleep,
is writingpassages in which the melodic aspect is an incidentalresultof
the chordal motion. A little later, Debussy offersexamples (like the
openingof Refletsdans l'eau) in which a linear phrase is dissolvedinto
an atmosphericallydispersed harmony that implies without actually
statingthe expectedmelodic resolution.Hyper-impressionistic pages, like
partsof the Night-SoundsfromBart6k'sOut-of-DoorsSuite,fragmentize
the melodyto such an extentthat the progressiveelementis heard to be
the increase and decrease of densityas the motifsfollow one upon the
other,ratherthan the specificallylinear aspect, which is here reduced
to a minimum.Nevertheless,whereverthere are successivedifferentia-
tionsin pitchthereis melodyof some kind,and whereverthereis melody
the ear will tryto hear it in the simplestpossibleway.
This is not meant to implythat we must expect to findbehind con-
temporarymelodic lines the simple stepwise diatonic frameworkthat
Schenkerhas pointed out in Classical examples. But the ear will natur-
it in pitch.The adjacent pitches
ally connecteach tone withthoserinearest
may be diatonic or theymay be chromatic;theymay be actually adja-

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
 178 The Musical Quarterly
centor displacedby one or moreoctaves; theymay be presentby implica-
tion only. In some cases motivicassociationsor peculiar scale-formations
may enforcethe acceptance of a largermodule- as in the simple case
of bugle-calls,the adjacent tonesof which are a thirdor a fourthapart.
(In the case of microtonalmusic, smaller modules may be in effect,
althoughit is doubtfulto what extenteven present-dayears can accept
them.) In everycase the ear will do the best it can with the available
intervals.It is the dutyof the analystto show the patternof connections
by which an educated ear - his own - makessense of the total melodic
flow.
Even less than in traditionalmelodies must one assume that there
is one uniquely correctway of hearing. Rather, the best analysisis the
one that recognizesvarious levels functioningsimultaneously,as when
a tone resolvesonce in the immediatecontextbut turnsout to have a
differentgoal in the long run. Two very brief examples may help to
clarifythis point of view.

Ex. 7

. I u(~q

I
The firstis the opening of Schoenberg's Klavierstiick Op. 33a.
Chordal ratherthan melodic in conception,its linear structureis never-
thelessclear. Despite the octave displacements,a line can be traced in
the uppermostvoice fromthe F$ in the firstmeasure to the B in the
third. (Notice, however,that at one point two adjacent tones are pre-
sented simultaneouslyinstead of successively.) At the same time, the
original Bb leads, through various voices but always at the original
octave-level,to the same tone of resolution.At this point the entrance
of the F, repeatingthe climacticF of the second measure,begins a new
motion that is carried forwardthroughthe succeedingphrase.

Ex. 8

~P ma aS

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
 Analysis Today 179

~'t ~i~i

The second passage is from the second of Sessions's piano pieces

From My Diary.' Here both the F in the firstmeasure and the Gb in
the third are associated with upper and lower chromaticneighboring
tones.But what of the cadential motif?Why is the patternaltered?And
why is the linear descentfromthe Cb in the second measure brokenat
this point?There are several possibleanswers,all of which are probably
relevant.Firstof all, the most prominentbass-notein each of the four
measures- as indicatedby its repetitionand by its quarter-stem - is an
F, which can be heard as a resolution, at another level, of the hanging
Gb - a resolutionconfirmedby a direct Gb-F in the bass. But at the
same time,thereseems to be an implied E fillingthe space betweenthe
Gb and the D in its own voice - a tone suggestedby the originalassocia-
tion of E withGb, and by the prominentwhole-stepmotionin the melo-
dic descent.In thiscase the line graduallyincreasesitspace as it descends.

ES?9~,~,~?\ i Lf~?-~_;2~~

-----'
1
But if it seems far-fetchedto introducean unstated,understoodelement,
one can hear the skip Gb-D as a way of emphasizingthe cadence, and
point out that the motif of neighboringtones aims each time more
directlytowards its resolution: the firsttime the neighborsfollow the
principal;the second timetheyprecedeit; and the last timethe principal
takes the place of one of its own neighbors.Finally,it should be noted
that the next phrase takes offfromthe dangling Gb in a subtle motivic
referenceto the beginning.
Ex. io T'r

It is of course impossible to do justice here to the role of such

details in the total melodic structure,but on examinationone will find
1 Copyright 1947 by Edward B. Marks Corporation. Reproduced by permission.

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
180 The Musical Quarterly
the same kind of connectionat work in the large. Note, for example,
how much of the firstthemeof the Schoenbergpiano piece is controlled
by the high F already mentioned- whetherin its original octave or
in another- and by its associationwith the adjacent E. It is again this
F, in its highestregister,that preparesfor the recapitulation;and it is
the E that, returningfirstwith the tranquil second theme,later closes
the motionin a loweroctave in thefinalmeasure.In sum,modernmelody
can not get rid of stepwise motion, because that is the way we hear
melody; but it can and does expand (or on occasion contract) the
distance, both temporal and spatial, between successive steps. From
this point of view even Webern is found to be no pointillist,but a
draughtsmanof subtle and fragilelines.

The role of harmonyin the music of our century,although more

extensivelyexplored, is perhaps more difficult,complicated as it is by
many factors,such as the frequentexploitationof the static, sensuous
effectof the chord in additionto or even at the expenseof its progressive
functions.As a result, one can no longer assume the easily defined
functionality of obviouslytonal music. Chords can no longerbe precisely
named, nor can theiridentitybe maintainedin differing contexts.But it
is important to realize that, even in stubbornlynon-triadic music, the
conceptof the chord remains,by analogy at least. The composercan set
up arbitrarysimultaneitiesthat, by their commanding position or by
repetition,are accepted as the controllingsonorities- the chords
-

againstwhichothertonescan functionin the mannerof traditionalnon-

harmonic tones. Bart6k's ImprovisationsOp. 20 show how by such a
techniquequite complicatedsonoritiescan be used to harmonizesimple
modal folk-tunes.In the followingexample fromthe last of Sessions's
Diary pieces the metricalpositionand the half-stepresolutionssuggest
that the firstchord is an appoggiaturato the second; thissuppositionis
confirmedby the appearance of the root-likeD in the bass, and by the
clinchingrepetitionsthat ensue.
Ex. II

i 2'I

--

In fact, only where the contrapuntalaspect becomes so strongthat

everyelementof each sonorityis heard primarilyas a point in a moving

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
 AnalysisToday 181
line, or at the otherextreme,wherethe textureis completelypointillistic,
is the chordal concept seriouslychallenged. In such cases one further
assumptionof traditionalharmonythat must then be questioned is the
primacyof the bass. Contrapuntallyor coloristically,of course, it will
have gained in importance,but at the expenseof its role in definingthe
harmony.A beautiful example of this process already at work over a
centuryago is shownin the openingof Liszt's Vallie d'Obermann,where
the melodic action of the bass clouds the harmony.Not untilthe return
of the theme adds a new bass underneaththe originalone is the situa-
tion made clear. A furtherstep in this directionis taken by Mahler,
who by his polyphonicallyopposed chords pointsthe way towardspoly-
tonalityin the magical cowbell passage in the firstmovementof his
Sixth Symphony.A more thoroughgoingexample is Stravinsky'sSym-
phonies pour instruments a vent,a more trulypolytonalwork than any
of Milhaud's often-citedSaudades, which in fact presentonly extended
and elaborated harmoniesover a single real bass.
There are otherforcesat workunderminingthe primacyof the lowest
voice. Impressionisticparallelism,which reducesits role to that of color-
istic doubling,is too well known to require citation.Less frequent,but
possiblymoreimportantin the lightof later developments,is the masking
of the true harmonicbass by a decorativevoice below it, a technique
seen clearlyin the repetitionof the opening of La Fille aux cheveux de
lin. Another device, common to the Impressionistsand Mahler, is the
ostinato.From one point of view the persistentvoice is emphasized,but
at the same time it is removedfromthe sphere of action. In Debussy,
as later in Stravinsky,the ostinatoresultsin harmonicstasis; in Mahler
thereis a constanttensionbetweenthe harmonyimplied by the motion-
less bass and those outlinedby the movingvoices and chords above it.
In both cases the functionalrole of the bass is called into question.
So far no specificreferencehas been made to the problemof tonality.
Except in comparativelyrare cases, such as passages in Le Sacre du
printemps,where an almost completelystatictone or chord of reference
is set up, tonalityis creatednot by harmonyalone, nor even by harmony
and melody,but by theirrelationshipwith the rhythmicstructure:in a
word, by the phenomenonof the cadence. A discussionof certainrhyth-
mic aspects,then,can no longerbe postponed.

IV
Much of the vitalityof the music of the Classical period derives
fromthe constantinterplayof meterand rhythm,the formerdetermined

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

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

by regularbeats and measuresand the latterby constantly varyingmotifs

and phrases. This tensionbetweenthe abstractand the concretebegins
to break down duringthe 19thcentury,when phrasearticulationis often
eitherslavishlytied to the meteror else so completelyliberatedthat the
sense of the meteris almost lost. The retentionof the measure in much
Impressionisticmusic is purely conventional,and it is no wonder that
later composershave abandoned the effortto keep an abstractpattern
when it would conflictwiththe actual rhythm.For thisreasonthe regu-
larityof the meterin such composersas Webernmustbe carefullyexam-
ined. Is it to be felt as a constantlypresent control? Is it a pure
convention?Is it, as some would have us believe, an evidence of the
composer'snumerologicalsuperstitions?
The answers to such questions must always be given with specific
referenceto the textinvolved.When, as in the case of Example 11, the
motifsets up a clear cross-rhythm, the explanation is relativelyeasy.
Webern's Piano Variations,on the other hand, presentthe problemin
an acute form.What has happened here, I think,is that the composer
has called on a complex set of interrelationships of rhythmic,metric,
dynamic, and textural factors to compensate for the tenuityof melodic
and harmonicinterest.In the firsttwelvemeasuresof the last movement,
for example, I findat least seven differenttime-divisions simultaneously
functioning.These are set up by the meter (3/2), a possiblecross-meter
(5/4), the rhythmof the two-note motifs,the rhythi of the phrases,
the tone-row,the dynamicalternations,and the linearpattern(Ex. 12).*

The reallyimportantquestionto ask in all such cases- and even in

cases wherethe composerhas deliberatelytriedto get rid of all traditional
metricalmeasurement--is, can we locate the structuraldownbeat? If
we can, then we can proceed with analytic concepts in some way
analogous to those of the traditionalrhythmand meter,phrase and
cadence. If not, some completelynew rhythmictheorymustbe devised.
Some musicians,like Stockhausen,are tryingto do this,but I have as
yetseen no satisfactoryone emerge.

By structuraldownbeat, of course, I do not mean the arbitrary

accentuationof the firstbeat of everymeasure; I mean ratherphenomena
like the articulationby which the cadential chord of a phrase is identi-
fied,the weightby which the second phrase of a period is feltas resolv-
ing the first,the release of tensionwith which the tonic of a recapitula-
2 Copyright 1937 by Universal Edition, A.G., Vienna. Reprinted by permission
of Associated Music Publishers, Inc., sole agents for the United States.

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
 AnalysisToday 183

RuhigflieBenddJ=
e so
2 3

7, f

' 6
_ j ... .,,

4 5 6

7 8 9

7f

10 rit .. . . ... . . . tempo

11 "1

, .
^_"

tion enters.(In the Webern example,I hear the downbeat as the Eb at

the beginningof measure 12; and I considerit no accidentthat it occurs
at the beginningof a measure,precededby a ritardando.)
It is just here that the importanceof rhythmto the establishmentof
tonalityemerges,for the cadence is the point in the phrase at which
rhythmicemphasisand harmonicfunctioncoincide. It would be partly
true to say that the cadence creates tonality,but it would be equally
trueto say that tonalitycreatesthe cadence. Where the cadence exists,it
is impossibleto hear musicas completelyatonal,even thoughone may be
unable to definethe key in conventionalterms.

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
184 The Musical Quarterly
We know the signs by which a cadence can be recognizedin tradi-
tionallytonal music: its position at the end of a phrase, the melodic
resolution,the change of harmony.The actual downbeatmay not always
exactlycoincide with the cadential point, but such unusual cases arise
most often when the phrase is rhythmically prolonged (the feminine
ending) or when it pointsahead so clearlythat the next phrase acts as
a huge cadence to the first(as when an introductory sectionis followed
by a main theme). In any case, keysare definedby the appearances of
strong,cadential downbeats- whetherclearlyon the tonic as in most
Classical examples,or on deceptiveresolutions,as notablyin the Prelude
to Tristan.

The extent to which analogous principlesgovern the structureof

contemporarymusic is surprising.A few examples will show them at
work.

The openingof the second movementof Bart6k'sFifthQuartet may

prove puzzling untilit is heard as an upbeat. The firstdownbeat comes
on the D in measure5, clinchedby an even strongercadence on the same
tone (now supportedby its fifth)in measure 10. The digressionthat
followssuggeststhe key of C, but this tonalityis not confirmedby the
cadence, which, when it arrivesin measure 20, is again clearly on D.
The firstpage of Sessions's Second Sonata for Piano is much less
triadic; yetwhen the downbeatcomes in measure 11, the harmonyof Bb
is clearlyestablished.Not onlythe V-I impliedby the progressionof fifths
in the bass, but the melodic resolutionto D, accented by the downward
leap, pointstowardsthistonal center,whichis confirmed by what follows.
In the second movement, no such clear downbeat is presented,but the
two importantfemininecadences of measures 177 and 190 both suggest
an unstatedresolutionto E. The importantdownbeat of measure 191,
comingas it then does on F, is in the natureof a neighboringharmony;
and not untilmuch later,at measure 213, does the expectedE occur, its
extensionas a pedal forten measurescompensatingforits long postpone-
ment.The last fewmeasuresof the Lento act as an upbeat releasedin the
returnof Bb in the opening of the finale.But this in turn,aftera long
battle with conflictingelements,gives way at the last to the key of C,
on which a downbeat is firmlyestablishedin the final chord.

Stravinskyis sometimesreferredto as a "downbeat composer," by

which I suppose is meant that he oftenemphasizesthe beginningsrather
than the endings of his phrases. This resultsin a weakening of the
cadential sense,it is true,the phrasesso accented being as it were huge

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
 Analysis Today 185
feminineendings to their own opening chords. A typical example is
the opening of the Sirinade en la. The harmonic progressionwould
be describedin traditionaltermsas VI 6-I in A minor; actually the

F of the firstchord is heard as hardly more than an appoggiatura re-

solvingto the E of the second. This would appear to be no progression
at all, in which case the phrase should be a huge diminuendo.Yet we
cannot be too sure: in a similarsituationat the beginningof the third
movementof the Symphonyof Psalms, the composer,by changingthe
mode and the orchestrationat the cadential word Dominum, creates
a clear accent even thoughthe chord has remainedessentiallythe same
(C) throughoutthe phrase.
In any event, whatever we may decide about the reading of his
phrase-accentin detail, Stravinskyis perfectlycapable of producinga
big structuraldownbeatat preciselythe pointwhereit is required.I need
only point to the huge deceptive cadence that opens the Symphony
in Three Movements, the dominant G of the introductionresolving
finallyupward to the A of the ostinatotheme (rehearsal number 7);
or to the way in which the Interludeacts as an upbeat to the C major
of the finale.

More controversialis the attemptto find traces of tonal form in

avowedly atonal compositions;yet I do not see how music like Schoen-
berg's,withits usuallyclear cadentialstructure,can fail to arouse certain
traditionalassociationsand responses.The previouslycited Klavierstiick
Op. 33a begins with six chords,of which the second throughthe fifth
are very easily-although not necessarily-heard as forminga progres-
sion referringto E minor.This in itselfis nothing,but when the opening
phrase is heard as an upbeat resolvedin the thirdmeasure, and when
the resolvingsonorityis recognizedas a seventhon E, a tonal analogy
is set up. The firstsection of the piece concludes even more unmistak-
ably on E, withthe added emphasisof a ritaridando;and the themethat
followsin measure 14 gives the effectof a sudden shiftof key. In the
recapitulation,the ritardandoof measure 34 again calls attentionto the
following downbeat, where the E appears in the upper voice, but
supportedin the bass by A-in the manner of a deceptivecadence on
IV. It remainsforthe finalcadence to confirmthe E, which is so strong
that it is not dislodged by the dissonant tones with which it is here
surrounded.

Several objectionscan be made to the above account: that it picks

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
186 The Musical Quarterly
out isolated points withoutreferenceto the movementbetween them,
that the "cadences" on E are a resultof the fact that the row ends on
that note, that such analysis is irrelevantto music in this style.
To the firstcount I plead guilty.I have indeed picked out isolated
points,because theseseemed to me to be the important"full-cadences"
of the piece. (Important "half-cadences"occur at measures9, 24, and
32.) The movementbetween them cannot, I grant, be explained in
simple tonal terms.At some points,linear or contrapuntalmotiondomi-
nates-in which case the melodicprinciplessuggestedabove will indicate
the logic of the chosen cadences. At other points the sonoritiesthem-
selves dominate-and thesecan of coursebe shown as derivedfromthe
opening chords.As a resultthe entirepiece can be heard as a develop-
ment of its original cadential progression-thatis, as analogous to a
traditionalstructure.

I agree that the cadences are partiallydue to the use of the row.
Depending on one's point of view, this effectis a virtue or a vice of
Schoenberg'stwelve-tonetechnique. It may even have been one of the
points persuading him to turn towards the system,away from freer
atonal methods.In no case can the argumentinvalidatethe actual mus-
ical result.

To the charge of irrelevancy,I answerthat one who cannot indeed

hear such cadential phenomenain thismusic mustjudge the analysisto
and inapplicable.But one who does hear themmustadmit
be prescriptive
to that extentthe validityof the approach. He may counterthat one
ought not to hear the music in this way; but he is then criticizingthe
music, not the analyticalmethod.Unwanted cadential effectswould be
as great a flaw in atonal music as the chance appearance of a human
figurein a non-representational
painting.

V
The last point suggeststhat thereis a relationbetween analysisand
criticism.It is not a simple one. Analysiscan often reveal flaws in a
work, it is true- oftenbut not always. If it were dependable in this
regard,we should be able to decide definitively between the disputed
C. and CX in the last movementof Beethoven'sSonata Op. 109 (meas-
ure 55) or whetherthe famousAL in Schoenberg'sOp. 33a is indeed an
Ab (measure 22). But unfortunately such cases all too oftenwork both
ways: the C- that fromone pointof view preparesforthe advent of D
two measureslater mighthave been avoided in order not to anticipate

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions
 Analysis Today 187

it; by the same token,although the A, seems more logical in the row-
hand), it may somewhat
structure(in spite of the At lacking in the left-
spoil the freshness of the Ab-Eb fifth that comes soon after. The ear
must be the ultimatejudge of such subtleties,but insofar as analysis
trainsand sharpensthe ear it makes its contributionto the finaldecision.
It would be temptingto go furtherand state that analysis can
demonstratethe qualityof a work,but thisrequiresa faithin rationality
that I am unable to summon. Judgmentof final excellence must be
fundamentallyintuitive.If analysis leads one to condemn a work he
neverthelesscontinuesto hear as good, he must conclude that there is
somethingwrongeitherwithhis ear or withhis method.Since he cannot
dispense with the only pair of ears he has, upon whose evidence the
examinationshould have been based in the firstplace, he mustblame his
method. He must then find a new one based on his own hearing,one
thatwill substantiate,not contradict,his musicaljudgment.He may then
claim that analysishas establishedthe excellenceof the workin question,
but he will be wrong; his own judgment will have established the
analysis.
One positivepoint emergeshere, and it is a crucial one. The good
compositionwill always reveal, on close study,the methods of analysis
needed forits own comprehension.This means that a good composition
manifestsits own structuralprinciples,but it means more than that. In
a wider context,it is an example of the propositionthat a work of art
ought to implythe standardsby which it demands to be judged. Most
criticismtoday tacitlyaccepts the truthof thisstatementand sets about
discoveringthe standardsimplied by a given work and testinghow well
it lives up to them. For investigationof this kind, analysis is naturally
of primaryimportance.
Criticismshould take a furtherstep, however,and the best criticism
does. It should questionthe value of the standards.A work that sets no
clear standarddeniesor defiesthe possibilityof evaluation; one that does
set its standardfailsor succeeds insofaras it measuresup to it; one that
measuresup completelyis at least flawless- but its value cannot exceed
the value of its own standard. It is this final step that is completely
beyond the confinesof analysis.
The music of Webern is a prominentcase in point. No seriouscritic
denies the perfectionof his formsand the complete consistencyof his
style. Its paucity of normal melodic and harmonic interestshas been
mentionedabove, but in connectionwith other values that, replacing

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

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

these, uniquely characterizehis manner. What is seldom questioned is

the significanceof the styleitself- of the restrictive
standard (for it is a
restrictive one) that Webern set for his own music. Are the limitstoo
narrow to permit accomplishmentat the very highestlevel? Only a
decision of this point can determineone's final evaluation of the com-
poser. It is a decisionthat depends on one's beliefsabout the limitsand
aims of art in general and is thus not exclusivelymusical, although it
mustat the same time be peculiarlymusical. It mustbe made on faith,
and it must be accepted or rejectedin the same spirit.

This content downloaded from 141.135.81.223 on Tue, 24 Jun 2014 21:24:29 PM

All use subject to JSTOR Terms and Conditions