MEYER, Leonard B. - Explaining Music - Essays and explorations.pdf

•

I

ESSAYS AND EXPLORATIONS x

x

x

x

x

x

x

x

x

x

x

10.00

XllLAINING ~1 ESSAYS AND EXPLO RATIO NS

x

x

x

x

x

x

LEONARD B. MEYER ' l'l1c cc11rral cc>11cern <)f chis l><><>I' is r11c critici. 111 of 111t1~ic-n1ai11I,· . l)llt r1c>t cxclu~i\•cl\r . \ \ ' cstc.:r11 t dist'<>\·er rhc secret <>f rhc singt1lar-t<> ... c\plai11 11<)\\ t!1c J>:ttter11s 1>ccnli.1r t<> a c<>1111J<>sitio11 arc c<>11111rcl1c11Jeti I>)', a11tl atfet·t, the listener. ldic>))·ncraric rel.1• ri<>11sl1ips ca11 l1c cxpl.1i11cd (>nl~· in t crn1s c>f gc11cral J>ri11ci11lc.. l3ecat1sc SLtch pri11ci1>lcs, as f<>r111t1larccl in existing 111usic rhc<>r)·, often arc inat>O. CS llC\\' CXJ>lanat<>f_\ 11) J><>thcsc. fr<>n1 rit11c t<> ri111c. Con\jc<1uc11tl)'. tl1c boo!~ is tl1ccJrcrical as \\ell as critical. I 11 cx1>lai1li11g tile f<>r111al a11(l r1rc1grc. si,·e rclctcit>11slups <>f a s~1cti1ic 1)assagc. criticis111 i11Llicaccs he>\\" J>l1rases sl111lild be . haJ)Cll a11t1 srrt1crurc artict1lat cd- in sl1orr l1c>\\' rhc passage or ,,.t,rl~ sl1ould be l)erf<>r111cd. l~llLt), alrl1ougl1 tl1c criteria f<>r 11crfor111a11cc are i1111)licit tl1rc,ughoL1t the l1<><>l~, pcrfurn1a11cc iii rl1e special suuject of Ll1c scconcl cl1a11rcr. 1 hc rhir
.,

-----____ ,,...., __ _ • •.··. x 01,.ations

~I

.

--

_;_'._ •

•

I

""'*'

-

-~

'

"

1

j

- ; _

~ '

j

I

2

2 j .....

I

'

- . . .

- -

-.......

ii

i

---1;- . . . ._.-·-...

........... - £1

, -

-............-J j

-I - ---

- - - - ·

~ -·

by Leonard B. Meyer

UNIVERSITY OF ,C ALIFORNIA PR ESS BE ,R KELEY,

LOS

AN G ELES ,

LONDON

197 3

Material corn direitos autorais

U1dversity of Califor1zit1Press Berkeley and Los Angeles, Califor1zia U1z:iversity of Califor11ia Press, Ltd. London, E1i,gla,1z,d Copyright © 1973, by The R egents of t 1':1e Uriwersity of Califorma ISBN: 0-120-02216- 5 Libra-r;1of Congress Catalog Card N1it1-1ber: 73- 187749 Prir1ted i11 t!Je U11ited States of A111erica Designed by .Dave .C0111stock

Material corn d1re1tos autorais

The Ernest Bloch Professorship of 1\1usic and the Ernest Bloch Lectures were established at the University of California in 1962 in order to bring distinguished figures in music to the Berkeley campus from time to time. Made possible by tl1e Jacob and R osa Stern Musical Fund, the professorship was founded in memory Er.nest Bloch (s 880-1959), Professor of Music at Berkeley from 1940 to 1959. THE ERNEST BLOCH PROFESSORS

1964 1965- 66 I 966-6 7

RALPH KIRKPATRICK

1968-69 I 97 I

GERALD ABRAHAl\11

1972

EDWARD T. CONE

\VINTON DEAN ROGER SESSIONS

LEON ARD B. MEYER

5f9T- 8HF-4HS5 Mat rial com dtr 1tos autora1s

To 11zy

brotl~er

Dan

Material corn d1roitos autorais

·re .·ace

- ---

--·-·...--:·-· - ·...........--- ·-....

This book is cou-cer11ed wit.h tl1e criticis1n of n1usic. As I intend the tern1, criticism seeks to explain how the structure and process of a particular ·composition are related to the competent listener 1s comp1·ehension o.f it. In other words, tl1e role of t he music critic is similar to tl1at of t.he literary critic. A11d jt1st as the literary critic need t1ot explicitly consider questions of ,ralne-exhibit the greamess of x·mg Lear, or demonstrate the brilliance of Middle111.t1;rch but is content for these to r·emain iu1plidt in his anal)rsis of the ways in wl1ich plot and. character, setting and diction shape our understa11dit1g of and response to literature, so tl1e n1usic critic 11eed not eA11ressly

inquire about the excellence o.f the comp.ositions he cl1ooses to discuss. Values are, of cotuse, al'''ays latent in "''ltat the critic does in ]us account of the \.vays musical relationships affect the listenerjs u11derstanding of ~1t1d response to particular n1usical worl{s, and also i11 iris choice of " 0.rks to analyze. But the critic does not, I think,, begin "vith aesthetic principles and arrive at critical judgments. Quite the 01)posite. H e t)egins with l1is own responses-his cognicive-affecti,,e sense of '''herher a composition is convincing and exciting, i11triguit1g ar1d e11tertaiitlng. TI1e11 11e atternpts to find rational grot1nds for his j u.dg-ment. Ai!oreover, since l\e dea!s for the most part with worlcs by acknowledged niasters, it seems a bit precentious for the critic to talce as bis 1nain rasl\: a11 examination of the virtues of compositions by Bach or Beethove11, Haydr1 or I-IandeI. To do so is probably also <..-ircular, arid perl1aps some\J\rhat disit1genuous. For the works of such master-s are in some se11se the initial basis for his stylistic stan

PREFACE

.nor come to f)taise 111ascerpieces, b t1t to explicate and illuminate t hem: to

ur1tlerstand and explain, that is, how the various sores of tonal .relationships in a particular con1position are underst:ood and enjoyed b)T experienced sensitive liste11ers. Every ex:planation 1nusr be it1 terms of some ge11eral principle. A nd to a co11siderahle extent, the reasons and arguments t1sed by the critic come from the t heory of mt1sic. Becat1se ge11uine rr1usic rheorisrs, such as Heinrich Schenker, a.re few ai1d far between-moSt so-called tl1eorists are teacl1ers of the grarnmar and syntax of mlisic-music theory is ar present rather rudirr1enrary. Consequently, the critic n1ust at cin1es assurne the role of theorist. I..Je must engage in the fort11ula.tion of general hypotheses and. principles '-Vhich relate eve11t.s \\rithi11 co1npositions to 011e another. It1 the course of this bool{, I lia,,e not infrequently had to assun1e su,ch a role, and1 as a result, tl1e book is theoretical as 'vell as critical. The relationshi,ps arnong eve11ts '~rit:l1in musical con1posicions-e\re11 seemingly si111ple 011es--are freqltently surprisil1gly co1nplex •1nd st1btle. The analyses explaining them. are,, accordingly, often con1pl.icated an.d. 1nvoived.

I have not sougl1t to sir11plify the difficult, or to gloss over intricate interactions \vith plal1sibie generalities and vague poetic appeals. Ratl1er I ha\r,e tried co n1alv of 110 other "\.V a}r of doing justice to c:he \:\ro11der of n1t1Sic and the miracle of hun1ai1 intellige11ce '\.:vl1ich 111al<:es and con1pre-

hends it~ Tl1is brings me to a quescion traditio11all)r anS\~,ered in prefaces: namely, ro wl1om is tl1is book addressed? Franl{Jy, l ' ln not sure. Sornetirnes in those mo1ne11ts of doubt and depression~ that co1ne Vl hen one is. near to finishing, I have felt that it is addressed n1ostly to its at1tl1or-thougl1 I t1ope riot. I wort1d lik:e to believe that a wide range of readers seriou.s1)7 interested in n1usic, in aestl1etics, and in cricicis1n v.rill .find tl'ie proble1ns I ha\re grappled ' vith i1ot 01tly in1portant a.nd challe11gjng, but even an1using. W1~iting the preface to a

boo1{ is al\\rays both a sad occasion a11d a. ha.ppy 011e. Sad, because rl1e ft1n is i11 the search, 11ot tl1e solution; in the pro'ble1n, not: tl1e propositio11; it1 tl1e forn1t1lation, r1ot tlte fi11ishJng. Sad, too, bec~11se th.e at1tl1or li:11ows better thar1 a:nyo11e else tl1at his boolc is goi11g to press witl1 all its imperfe,ccions on its l1ead. Specific--ally, I an1 acutel)' a\\rare of all tl1at h.as be,et1 left U ' ndo11e: of tl1eoretical problems not sol,red st1ccess-


•

PREFACE

fully. of particular

Xl

pQ~ages

not adequately explain.edt and of examples not

used wnose analysis would have illustrated yet another fascinatin.g aspect of r11usical strl1ct11re. I can only console myself '~'ith the thot1ght that the only perfect bool{ is a book. i'l1iag-ined- not 011e actually \vritte1L For to \Vrite is to .fur \Vhat is in flux and to pigeo11hole the parti{..-ular. BL1t, as I argue in Chapter I, there is no alternati\1e. To e,"
depressed and caln1ed r11e ·when I was uptight about deli,rering 11\1blic lec1:llres. I am also indebted to Helen FarnS'\vorth. who nc>t only inade snre that I got the nlate,cials and eqtupment I needed, bt1t re.ad a.t1d corrected. po1~rions ·of this manuscript. Jane Willrinson prepared tl1e musical examples for publica•

tion.

Tl1e first part of this bool{ the Essays-are b,asically the same as fot1r of ·tl1e lectures given at Berl{eley. ~The second l}art- Ex.ploration&-which began as the fifth lectur·e · has obviously been much expand.ed. But eve11 thus expanded"it is a kll1d of co111pro1nise. Some two yea.rs ago my colleague. Profess-Or Ed\\•ard E. Lowinsky, trrged 1ne to write a book abo\1t melody. I thought about it and began \vo,rk on such. a book, hue found tha,t the rask was beyond my capability. The seco11d part of this book, which attemp~c; to explore some aspects of tonal melody, however, is the result of Professor Lo,,rinsky's suggestion. It inay be misraken in part: or in. parcel. But whatever its v'alue-ar1d, r1eedless to say, I be.lieve it to 'be interesting and inventive- it is offered to him as a sort of separate Festschrift, .a token of respect and affection. Parts of this book have been read an.d cr.iticiz.ed by a number of scholars. All ha\;'e my heartfelt thanks. Professors J\1Ionroe Beardsley and Fore!>'t


••

PREF1\ CE

:t\11

Hansent \Vho sa~r an early draft of Chapter I togettier \vith pa,rt of tl1e chapter ,,.. .11ich closes the boolc, made several important suggesrio11s. Professor· \iVillia111 Thon1S(>t1's cricicisn1s of Chapter III were specially careful and cogent. Professor Barbara H. Smith read the first part of the boo1c. In addition to her specific con1rnents and corrections, I h:r\1e benefite·d e11ort11ously 'botl1 fron1 her writings <.l11d f1-01n rhe general disct1ssions vve 11ave had abtJut 'vor.Irs of art and related matters. Professor Janet Lev}' rea,d mosr of this book, and, again, it is not 1nerel)r l1er percepti,re and specific criricis1ns that merit gra:tefitl aclcno,vledgn1ent, bt1t tl1e seriot1s and sympathetic talks \Ve h.ave had abor1t parcic11lar cornpositions, and abot1t the tbeor)' a.nd history" of •

lllUSlC.

Like n1ost teachers, I am specially i11debted to rn)r srode11ts, '\vhose probing qt1escions and ent.httsiastic skepticism have been a source of concint1u1g dcligl1t a11d constant learni11g. Thougl1 I ca1mot name the111 all, or1e in particular 1n t1st he n1encioned. Eugene N arm.ottr, '''ho t".IO\V teaches at the University of Pennsylva11i~ has been \Vorking on the analysis of to11al 111elo·dy, coo. The last pare of ch.is bo-ol( O\Ves a \rer)r co11siderable debt to our man;r cliscu~io11s.

V\tl1ile I was worlcing on rhis book-and particttlarly in the difficult beginning stages-Zita Cogan asst11neci 1nan)1 admittistrativc chores which ·w ere ·p roperly rnir1e. Her de·votio11 arid co11cern n1erit rnucl1 n1orc than perfunctory aclcno'-vledgment. F'inally, what can I say ( rl1at has 11ot been said tiine and rune a.gain by otl1er authors in si111ilar circt1mstances) ahouc Lee's help? Tl1at she was

patient, untierstanding a11d forebearing? S~1e vvas all ·of the~e. That sl1e saw relativel)r little of Califor·nia becatise I vvanted to "''O·rl{ ~ind. that, on our ren1rn to Cl1icago, we neglected ol1r friends beca11se I \\ as tr)rit1g to finish this boolvhat is her r e\vard? '011ly the ci1sto1nar}r connubial t l1anks-and the prospect of proofreading! 1

L.B.. Nl .

Material corn dirc1tos autorais

. . . . on tents Preface PART ONE-ESSAYS I On the N atlJre and Limits of Critical Analysis JI Critical Analysis and Perfor111ance: The T heme of Mozart's A-Major Piano Sonata Ill Confor1nmt Relationships l V Hierarchic Structures

PART TWO EXP! ,ORATIONS: Implicarion in T onaJ lvlelody v Introduction VI

Definitions and Methodology

Vil 1Welodic .S tructures Conjunct Patterns Disjunct Ptlttenzs Syrmnetrical Patterns

Contextual Discrepancy Archetypal Schemata Prolrmgatio11s

VIII A Su111111ary_ Exam'/!_le

•

ix

1

109

!..!.j_

!.1!. ~ ~

!:J..J. 196 213

m

~

Index of Subjects and N a111es

Index of Music

Mat rial com dtr 1tos autorats

ne

Material corn
c---. .

..L,.L

TERI

. . _.,_-----·~:-·--.~---·~r--~:-·---

On the Natzt1~e and Li1nits o Critical Analysis I.

Experienced naively- w it hout any ps)rchological predispositions or cul-

tural preco:nceptions '\Vhatsoe,1er- tl1e world, as Willi~1m James ob.served, is a buzzi11g, boonling confusio11 of discrete, ur1related ser1Se in1pressio1is. It is ''full ·of sound and fury...." One may, of course, tr}r to experience existence i11 this way : unr11ediated ·by concepts1 classes, or relationships. And a nu.mber of artists and vl1'iters for example, Joh11 Cage, No·r man 0. Bro·wn, and Alain Robbe-Grillet-have urged such mindless innocence upon us. So, too, have son1e n1en1bers of tt1e h.ippy-drt1g subculture. In tb.e Electric KoolAid Acid T est, Ton1 Wolfe gives us this vie'-'rpoir1t t1nadt1lterated: TI1at bab}" se-es the \vorld with a complere11ess that you and I \\ril1 never know again. His doors o.f perception ha,re not )'et been closed. He still experiences the nmn1ent l1e lives in. T he ir1e\•itable b1ulsl1it t1aSt1't constipated lris ce,r ebral cortex )~et~ H e still sees the \vorid as ic r·eally is, while we sit .here, left with only a din1 historical version of it inanu.facrured for • us by '\Vords and o-ffi.cial bullshit, a11d so f <)rtl1 and so o·n. • Bnt Wolfe's world of scatological 1~01nanricism, remindi11g us of 'i\lo.rds-

worth's Ode, cannot be: understood. lt has 11eitl1er process nor form, xneaning nor value. Like the "''orld of Benji in Faulkner's The Sozind and the Fi1:ry, ''It is a tale told by an ic.liot . .. signifying nothing.'' It merely exists. To understand the \:vorld~ we must abstract fro.m the ineffable unique11ess of stimuli by selecting and grouping, classifying and a:i1alyzing. '¥e mt1st


4

EXPLAINING MUSIC

attend to some features of an ol,ject, f)erson.t or l)rocess rather tl1an ot~1ers discingtJislu11g (from so111e particular point of vie·\v) the essential fro1n the accidental, the intrii1sic f ro111 tl1e i11cide11tal. The intelligible present is nor an isolated instant in rin1e, bttt, as "\/\'hitehe-a:d p11r it: ''Wllat we perceive as rhe pr·esent is the vi·vid fringe of 111em01J' ti11ged witl1 anticipation." A meaningful, a hl11nanly' viable \:Vorld n1tlst be order·ed and patter11ed into relatior1ships of some sort. Tlus is the case not ortly in everyday existence, but in the arts and scie1lces as well, Ti1e order tl1us discerned-whether i11 natl1re, culture, or artis not, hovvever, arbitrary or .fi.cticiotis. The processes a.i1d forms., pattert1S and principles disco\rered by scie11tists, sot,"ial scientists, at1d .h.u111a1'lists are derive.cl, directly or indirectly, fro1n exiStent events that are really there in the world. TJiey are not arlJitrary fig111e11ts of stll)jecti e ir11agir1~rcion. Tl1e similarities bet\veen events, th.e ,orclerlines.~ of 1.,rocesses, and the hierarchic structtrring· of relationships ~rre just as r:eal as the differences bet\veen e\re11ts, the features \\' hich are disord.ered, ~tnd tl1e a'bse11ce of relario11Ship. Tl1e critic does not, like God, bring order oi1t of cl1aos. !lather, lil
aesthetic ex_pe.rience its special savor and i11describable affectiv·e quality. In a strict sense, this charge ca.n not, I rhinlc, be refuted. A specific inusical e..~erience "'' hich combines the percepti-0n o.f n1usical e\rents ,~1irh. tbe subjective peculiarities of an individual hurnan psyche at a specific 1no111e11t in its hisrory, is unique. Criticism cannot fully k11ow or explain that experience. Nor is it co11cerned to do so. f"or criticisrn endea\ro.rs to understand and explain tl1e relationships among and between mt1sical events, not tl1e re-

sponses of individual listeners. '"fh<>se "\\' C must lea,re to the shamans of the r11iddJe-class t:he psychoa11aiysts. However, though the individual's J)nrticular e,;li.rperience is uniqtie a11d perl1aps unla1owable, th.e perceptions '"' t1ich shape that ex:perience .are not so. Wl1enever it goes be7ro11d the J11ere se11sing of i11corning stimtili listening is necessarily analycical-abstracting, classif}ring, ar1d organiz;ing musical stimuli into patterns, proce~-es, and relatio,nships. ru soo11 as one perceives tll.e ro11e of, say, an oboe, is a"vare of octave identit)r, or gror1ps to11es i·nto motives and motiv,es into phrases, one has .abstracted- has ignored a myriad 1


ON THE NATURE AND Ll:tvltTS OF Cnrl"lCAL ANALYSIS

5

tile series of srin1uli. A\\rareness virtually co·mpels conceptualization. 011r perception of the splendor of a sunset or o·f the subtle

of attribt1res present

i11

nu.ances of a lovely theme are inseparable f ro1n our l<:nowledge of the event

as being a sunset or a theme. In his novel, End of the R oad, John Barth l1as put this n1uc.b better than I can. His hero, Jacob Horner·, a writer, e."{claims: ''Articulation! There, by Joe1 'vas my absolure, if I co11ld be said. to ha"re one. . . . To n1rn experience into speecl1-that is, to classif)r to conceptualize, to grammarize, to syntactify it-is always a b,etrayal of experiei1ce, a falsification of it; but o·n ly so berxayed can it ·be deatt with at allt and 011ly in so dealing with it did I ever feel a man alive and kicking." And just as the at'ti.st in presenting a reality in words, visual materials, or musical tones in this sense dist.ores l1is and our experience of existence, so criticis·m in its turn 11ecessarily falsifies tl1e experience of the art work. But on1) so 'betra.yed, to paraphrase Bartl1, C
confusion which everywhere surrounds us. On the other hand, particularly in an academic context s11cl1 as this, it is \vell to remember that we tend to teacl1 and study those aspecrs of e:A.-perience that n1ost easily lend t hemselves to abstraction and syntactification. There ar·e, ho\vever~ other, r11ore elusive forces sh.aping a11y rich human e,,·perience. For insw1ce, pace and timing (how lon.g a particular sort of e\1er1t sl1ould contint1e, and how differe·n t sorts of events should f ollo\.v one another) are,, I thinlc, of central importa11ce i;n both music and litera:tute. But I kno"v' of no adequate study of tl1ese as-pects of temporal experience. Thust even thougl1 \Ve reject Macbeth's picture of tl1e world, we sr1ould tal{e f-Ian1let's caution seriously~ ''There a.re more things it1 heave11 and earth, Horatio,/ Than are dreamt of in y our philosophy.'' It is also ofte11 objected that criticism or analysis is coldly intellectual and. itlhunl3.llly detacl1ed, fragmenting what is really one and conceptualizing what should be felt. With regard to the first, oo.e ca:n 0111}' ans,ver that good


6

EXPLAINING MUSIC

criticism separates "''here separation is \-VUrat1ted b)' the musical structure and unires \\rherever the musical organizatio11 pern1its. Because mtisic is hierarchic~tones combining to for111 motives, lnori\*es phrases, a11d so onwhat is separated on one level beco.r11es u·11ifiecl on tl1e t1ext. Only cl1e 1nusic of transcendentalism- th:e music of Cage, Earl Bro"vr1., Pousset1r, etc.-is there complete homogeneity an.d nondifferentiation. A nd such: mt1sic C
u1

be .analyzed, onl.Y d.escribe-d. V\7it.b regard to tl1e assertion t11at criticism co11ceptl1afrz.es '~hac should be felt and is, therefore, son1el10"'' i11ht1n1an, t \Vo observations seem perru1e11t. First, t11e ·c har.ge restS upon a doubtful dichotont)' : namely, that which separates mind and body, and inc:ellect fro1n affect. Ottr emocio·nal respon.scs to the world are invariably Linked to cognitive f>atterr1ings. Conce11rualizacion precedes and qt1alifies affective exrperience. Tlitning to V\Tillia1n Jar11es again: he reminds us tl1at a grizzly bear securely co:nfined behind bars elicitS on.e response perl1aps 0 11e of amused ernpatl1y.; tl1e sa1n.e 'bear escaped a1.1d running toward us, ,q-uite another anxio·cis a11tipathy, And the differer1ce li·es in our conceptual u11derstanding of tl1e s.irua·cio11. Second, there are r easonable grounds for believing that the musical processes and strtlctures explicitly conceptualized i11 criticism (Lfe those \Vhich e\ oke affectiv•e: responses in sensiti,,e and. ex~1erienced listeners. To concli1de tl-1e first part of this cl1apter, I dispute vel'ten1ently the 1

.r1orio11 tl1at an intellectu.al response to '\.VO'rks of art, ai1d to rl1e "vorld in gert-

eraJ. is inhl1n1an or t111desirable. Quite the opposite. Tl1e artsf philosopl1y, and history, as \vell as the scie11ces and social sciences, are \raluable and rele\'ant be.cause they are entertaining. Not in tl1e sense of rl1e Ed Sl11Ji,ra11 sl1ow -it diverts,. But in tl1e sense tl1at T. S. Eliot l1ad .ir1 mind when he said

tltat p·oetry is si1perior arnusetnent. For to e11tertain ideas-to see pattern a.nd strllcture in rh.e w orld ar1d to be entertai11ed by ide(tS is both tile n1ost 11wn,an and the rnost h\tm~111e CO·t1dit'io11 t<1 which rr1an can aspir,e. 2•

Criticism (or critical analysis) rnust l)e distinguished from style analysis. For these disciplines, thot1gh co1nplementary, involve .different vievvpaints, mecl1ods, and goals. Critical ru12'.lyrsis seeks to understax1d and explain what is idiosyncratic about a particular c<>mposition: ho' v is this piece differe·n t fr.o n1 all other pieces-even those in tl1e sa1ne sryle and of the same genre? It is concerned with the implications of this specific motive or process, rhe function and structure of


ON nm N.ATURE AND LIMITS OF CRITICAL ANALYSIS

7

this specific hannonic progressioll,; tl1e relatio11ship between this particular slow introduction and the Allegro whicl1 follows it, the reaso,n why t!1ere is a sforzando on this note or why this the.me is interrupted at this particular point. In short, criticism tries to disco.v er tbe secret of th.e singular-to explain in ~\rhat 'wys the patterns and processes peculiar to a particular work are related to one anotl1er a11d. to tl1e hierarcl1ic strticttrre of whic'h they for1n a part. Style analysis, on the other hand, is normativ·e. It is co11cerned ~rith dis~ CO\rering and describing those attributes of a compositio11 \Vl1ich are common to a group of worlcs-USll<-illy ones '\Vhich are similar in style, form, o:r genre. It aslcs, for instance, about the characteristic features of late BarG"C}Ue 1.11usic --its typical textures, l1armonic procedures, and formal organization; or it inquires into the features common to diverse 1novements u1 sonata form or differ,e nt types of operas. Style analysis:, in its pure form1 ignores tl1e idiosyncraci.c in favo·r of generalization a.n.d. typology. Consequently statistical methods are as a ntle :m.ore appropriate in St)rle analysis tlian in. criticism. For sty le analysis, a 11articulax con1position js an it1stan.ce of a technique, a form, or a genre.

ln describing and classifying typical :processes and schen1ata1 style analysis: discloses and d·efines those probabilities-tho.se rhyt l1mict rnelodic, harmonic, and textural relationships-which are chara.eterisric. of the music ·of a pa.rricuJar period, a form, or a genre. 1 H ere style analysis shades i11to what is commonly called music theory. For l1armony and counterpoint too, are normative and probabilistic. To ·cake an obvious example, a progression from the dominant to the tonic .is normative ax1d probable in the harmonic practice of the eighteenth century, as is indicated by the fact that it is called an authentic cadence. A motion from the dominant to the submediant is less probable, and is said to be deceptive. lnd.eed, w hat has traditionally been called music theory is by and large the translation of tl1e normative pra.crice of some style 'p eriod into a set of syntactical rules for ~rriring exercises in that style. .But a significant disrinccion rn·ust, I believe, be made bet'\vee11 the essentially inductive norms of scyrie anal)rsis and an authentic theory of music. Briefly, a real theory of music endea\"Ors, where possible.. to disco·ver the 1

' iVe nee-0 not consciously classify it1 o.r der ro understand wha.t is n.o rmari.ve:. \ Ve c:a:n and do lear.n tl1e nom1s of a stylec-get to k11ow its t)rpical proced ures and schem-a - through listening a.n d performing, jnst as t\.'e do ~~hen we learrt a l3nguage or any other kind of culroru behavior.


•

s

E.~PLAINING

.?.-!USJC

principles governing the forn1acio11 of the t)rpical procedures and s<::hemata, described in st}rle ana1)rsis. Let rr1e give a sir11ple e,xa.rnple of \Vhat I ha,re in mind. In books on counterpoint there is g'enerall)' a rt1le '\Vluch srates that after a skip, a r11elody sl1ouJd r101mally 111ove by step it1 tlte opposite direction. This rule is a generalization fr.on1 sixteenth-century practice. It describes a practice, 'but it does 11ot explain \\.1 hy the practice 111akes setlse or v;,rhy it was de,reloped in the first place. Suppose a student \vei·e to ask·; <~V\lh.y does cllis melody of Palestrina descend u1steplvise fashion after an u1:>ward skip of a sixth?,>·w e ·migl1t a11swer: <''Veil, that was the rule," or <'That's what composers writing at the time usually did.,' But this is clearly circular reasoning, si11ce Palestrina1s 111usic \Vas pa.rr of the data usec.i ir1 deriving t:l1e rule. So

the student- probably he is ru1 ·z1,1uiergradi1,ite-presses us further, asking: ''VVhy did Palestrina follo\V this practice?,' rlere we V\rould l1ave to anS\';\ er him \vith a general law ·o f some sort. \¥ e 111.ight~ for instance, cite the Gestalt la\.V of cornpleteness, wh.ich asserts that tl1e ht11nan mii1d, searcl1ing 1

for stable shapes1 wants patterns to! be as complete as possible. A slcip is a lcind of incompleteness; the liste11er is av;rare Qf the gap bet\\:een the first pitch and the seco11d, a11d ~'-t,ra11ts'> tl1e gap to be filled ir1~ \\ ltlcl1 step\\rise motio.n in the opposite direction does. This ~'la\:v) of r11elody is presut11ably 11ot St}rle~bound, but applies to the n1usic of Beetl1over1 or that of south India, as t~ell as the music of the sixteenth cenrury- though. \vhat represe11ts n sati.r; fa.c tory fillin.g of a gap '"'rill depe11d upon tl1e repe.rtor-y of tones prevalent in cl1e rnodes of a stvle. 2 " Oi1e might, of course, arce111pc to ge:r1eralize still furtl1er, aski11g wh}I che mind searcl1es for stable sl1apes. And or1e rnight explain t:l\at becat1se l1uIliL'Ul behavior is not for t he m.ost part geneticall'r deter.nlinecl, men n1t1s t envis-age tl1e co11seqt1ences of cl1oices ir1: order to kno~1 110 , ;v to act it1 the present; and they can envisage a11d choose only in. tern1s of f)atterns and pr
1

the e\rentS describe(i.3 To put ti1e nla.tter in another

\Va.y :

\Ve

endea\ior to

e The tnelodic .st;rle of contemporary poit1tiilists is not an exception to ct1is ''la\v1" because the relatior1:ships among intervallic events in, say, a piece by \;\lebern are not processi\re but f.or n1al. 3 As lvtario Bttn,ge has pointe(i our: "Every !'J'Sterr1 tmd ever')' .evl!'nt ca71 be acco1mted for (described, explained or predicted, as the case TPMY be ) prima-ril:t in t eN1ts of its own levels and tl;e adj:oi1Jing levels . .. , Fo.r exeJnf)le, r11ost historical events


ON THE NATURE AND LtJ;..:J.rrs OF CRJTICAL ANALYSIS

9

go beyond descripti~.te or statistical t1orms to the ~imple~'t explanation "-'hich takes the form of a gener·aJ principle. The goal of music theory is to discover soch principles. Ir is not, however, the goal of cricic'll analysis to do so. Critical a.nal}rsis uses the laws foro1ulated b)• n1usic theory-and, as \\re shall see, the norn1ative categories o:f style analysis-i11 order to explair1 ho'v a11d why tlie particular events within a specific composition are related to one another. Theoryr gi~tes us the general pru1ciples governing, say, tl1e processes of melodic implication and cl0St1re, wlille criticism is concerned witli the ways in which those principles are a.ctt1alized-or perhaps e,raded-in the case o:f a specific motive, tl1e111e, or section ir1 a particultrr worlt. A descriptionot no matter ho~, detailed and elaborate, is n.ot an explanation. A catalogue of successive pitch·es (e.g., observing that "the melody begins on D , skips up to

Bb,.

and the11 n1oves do\.v·n ·t o A '-Vhich is

trilled ... .'').,the labeling of chords (as C n1a)or, F niajor, etc.), or a listing ·of dominant instrumental timbres n.one of these constitute a critical analysis of a composition. I do not intend to st1ggesr th.at descripti,re discourse is noc relevanteven necessary-in t.'ritical a:r1a1ysis. To be so, 11owever, it musr be used in conjunction. with a viable theory abo·u t how the variotlS pa.ra111eters of n1usic - melody, rhythm, cin1bre, harn1ony, and the like-function; that is, ·h ow they give rise to pattern_s and rel.ations}ups. For instance, give·n an hypothesis abou.t the psychological f unctioni1ig of structural melodic gaps, it is mean-

ingful to observe tl1at "the melody begins on D, skips up to Bb ...') llltd so on. Sin1ilarly given an l1)rpothesis about tl1e role and function of ornaments in melodic-rl1ythmic processes, it becomes significant to describe a note as bearing a trill-um1ally gi\ring son1e reason wl1y it does so. But in the absence of a background of theory \.Vhich relates events ro or1e another, description, even when disguised in a cloak of obscure tecl1tucal ja·rgon, ex-plains t1othing. 3· Elren wl1en nor explicitly stated,. ger1eral l1ypotheses are invariably i111plicit in critical ailalysis. Of'tert ~ucli l1}rpotl1eses are of a co1nrnor1-setl:Se can be accoumed for without resortir1g to physics and chemistry, btlt chejr cannot be properly understood without so111e behavioral science.'~ 0 The Metnph)~ 'ics~ Epistemology and fviethodology of L.evelsl11 in Wh:yte, \¥ilson, and. Wilson, eds., Hierarchic Struct1:1res (New York: Else•.rie.r, 1969) 1 p. 2+

Material corn d1ro1tos autorais

EXPLAINING ?v!USlC

IO

variety. For example, near tl1e er1d of the second the1ne of tl1e Finale of I\l.iozart's S)'mphony No. 39 a11 interru.ption, follo'"'e
•

Ff•

p •

• -

p <

~· '

•

p t

•

•

•

•

Ff.

In~rruption


ON T HE NATURE AXO I.J.\JITS OF CRITICAL A::-\AL YSI

I I

Had the consequent })hrase hee11 regt1lar it \\'Oltld have bee11 as f.ol]o\vs:

•

•

• •

•

... .,,

,-

Exan1plc

-

2

Tl1at is, the major n1ode \VOt11d ha,re ren1ained a11d the melody and harmo.n y of the cadence ':\'Ot1ld l1~1ve been:

Exa111ple 3

And this cade11ce exactly is presented at tl1e end of tl1e clig:ressit)Il (Ex.an1plc 4). Now tl1js i11terruption migl1t b·c explained by poi11ting out that the second theme employs the sa1ne n1orivic material as tl1e first, and both are antecedent-consequent phrase structure. . iVloreover, tl1e form and process of the second theme are particularly .Pate11t and predictable. Tt1erefore, had the conseque11t ,p hrase closed in the expected ~ray, tl1e \.i;rhoic theme 'vou1d l1ave seemed obvioL1s and a11ticlim.actic. Or1e couldi of course, for111ulate a general law covering the case: v.that is coo }Jredjce,able is u11il1teresti.ng a11d is1 as a rule, avoid.ed. But tl1is is scarcel)r i1ecessary. I11deed, because con11non sense lets us take the proposition for granted, as a kind of cultural do1ine, the explicit state1nent sce1ns artificial and pretentious. This is nor to Sl1ggest that one sl1ould not try to build a n1ore refu1ed and co·n1prel1ensi,,c tl1e<)ry of n1ttsic. Bt1t because specific n1usic-al e ents are the result of nonrecurring co11cate11atiot1s of co.nditions and variables, 110 set of general laws can adeq.t1ately explain tl1e particular relationships c111bodied ii1 an actual con1position. l n other \\rords, 110 matter ho\v refined and inclusive the laws of music theory bcco1nc, their use in the CA1Jlanarion of

Mater al corr d1yc tos autora s

EXP LAl Nll G MUSIC

Jl

--

Fl.

•

~

•

f

'

'

;Ji_

~

t •~

,~

,,_

'

'

-··••

·,,

~I• '

'

sJ

~

..Jf" ',sf

<(#-

'

-'

!

''"'f l;2

i

'

sf

I A. . '

- -

'

1',.--; i ..

-

•

'

.

'

•

- •

' W4 5W ,,

-

_.,..

'1

'1

' '

I

~,~ s. '', ' -

~

-

•

- -

•

-

' ::. .,..t

~

'

'·.

== ,-

'

':al

~

1·

'

t

sf

~

-

'mi'

,

~

~

"

'

'

.

• -' I ,

-

'

tor. - ' "'::' . ~ !

- i_

'

-~-

'

a

• ''

..,

-

• .._.;

".llf.•

'- c

•

-

'

f ,.. _.

_

f

,

'

i

4t

' '

.

'

iOf

I

-

' '

'

'

''

.

.,'

l

I

•

._t

I!

~

tf

Ml-

i

-~ -

j

J

$J

:

-

~ - ~lr, L . .,

~

-

1

sJ

...

~

'

'

7'

i• i '

!!!! - ~~

-

:

-

:,_.l . :..·

~ •

-

I

-'

'

'ti:-"-

'

•

'•

-

'

:Jiiiii:~

i:fJ ,

-""

'

~

"..

f

.sf

'

.:

2=::

'

-'

•1 .J

., •

~

'

~

I,

:Ii

.

-

.

'

f

-

j~

'

.,.--

'-

'

..

~

-

C f

~

-;;.

'

~c

'

..

... '"' - le :

~---

I::

f

-

sf

- -~ -'

-

'

•

'

• 1

''

,

If

•

-

Jj

•

'

f

''

~

sf lf -

~

•-'

.l

'

lli

T-

'

'.f

---"'

I

JC

'

' '

I

'

(,,:·

'

'

~

sJ

...

.

f

I

'

•

'l

''

et

.~

-'

'

:!

~

'

':

•1

$.f

Cf '

"

..

'

' '

If

..

'

:J

-

-

'

l: '

i=;D '

,_ '

- -

~

-

'

;ti·

.

-~ ~

-...

'

-

Example4

particular music:al e'\re11ts \:\"ilJ lta\1 e to depen

ON THE NATURE AND L'IMITS OF Cl:UTICAL ANALYSIS .

if it is to be

useful~

can specify only a pan of v.rhat is to be presettted in a performance, so mnsic theory,. can formulate o.n l}r some of tl1e l1ypotheses

needed i11 the analysis of a particular composition. To put the niatter tlle other way around; n score whicl1 contained all of the inforrm.rion communicated by a particular perfo.rmance· e\ ery 11uance of duration, pitcl"4 dynan1ics, timbre, etc. would not only be unreadable. but ¥11ould take years to '\~te down and months to decipher. SimilarlyT a theory which c.overed everjr possible interaction of all possible variables \vould be useless bec-ause it wonld lack precisely wha.t a11y theory m t1St have- namely, generaliry·. For the performer, the 001nposer's score constitutes a n1-0re or less definite set of directions. '"rhich suggests a parricula.r interpretation; and in parallel fashion,, the explicit f or1nulation.s of music theory suggest p-ossibf e ,expla:oa.tions of particular musical events to the critic. Ju.st as t l1e per·forn1er's acrualization of a sco1·e is controlled in part by· the srylistic tradition. of performance practice, so the critic's use of music th·e ory depends i11 part upon. the con1rn,011...sense tradition. of bjs ct1lni:rc. If perform.ance traditions may {')e considered as a kind of unwritte11 nota.tion, then common sense may be regarded as unformulated the.ory. Because this point is of crucial irr1portance for criticis1n. I should like to emphasize it by stating it in another 'vay. In. the Gl;ort i1z the i\IJ..acbine, Arthur Ko-estler points out that e\rery skill- for ou.r purposes: coinposit.ion, perfonnance, and listening as well- has a fixed aspect and a variable one: 1

Tf1e former .is detennined by its c~non, the 'rules of the ga111et~ which lend it its charaeteristic pattern-whether th.e gn1r1e Le; n1nking a spider's \Veb., constructing a bird's nest, ice-skating, or play'ing ch.ess. But the rules permit a. certain va.riet)r b)' al.ter11ati,re cl1oices: the v.reb cru1 be suspei1ded fr-0n1 three or f oux poin·ts of atttachment, the nest can be adjusted to the angle -o.f the fork i11 the branch,, the che.ss-piayrer has a v~ctt: cb·oice among pexn1issible rnoves. These cl1oices, havi11g bee11 left ·open by the rules depe11d upon tl1e lie of the land. the local environment in \Vhich the holon operares---.;they are a matter of st:rnteg}.r, guided by feedbacks. Put in a different -,.vayt the fixed code of rules deter·1nines the permissible moves. :flexible ~'trnt(;.'g}r dctertn.itles tl1e c'hoice of the actual moves among the pernlissible ones. The larger the num'ber of alternative choices, the more :complex arscl flexible tl1e skill.4 1

TJ'e Ghost in tl>e Aifacbhw ( Ne\V York: M.acmillnn, 1968)., p. 105 • .A ' holon." is a n1ore or less siepar-able entity o.r even-c t l1at forins pa.rt of l:l hierarchic structure. For instance, a rno.cive Vi'Ould be a holon on a low level; a tl1cme "''ould be one on a 4


UPLAL.~lNG

MUSIC

.In mus.ic, psycl1ologic~1l constants sucl1 as the pri11cifJles of })attern organization, the syntax of pa.rricular styles, a[ld typical scl1en1ata such as triadic holons constitJJte tl1e ·ritles of t/Je ga111e. ~I-heir actualization as specific n1,usic.-al e\renrs is the realization of what I{oestler calls fiexible strateg; 1• For any given n1usical repertory t:l1e 'tr1Jles" determine tl'1.e lcir1ds of patterns that can be e11.lploy"ed in a cor1'.lposition. They are the province of style criticisrn. Strategies> '~hicl1 are variable and nonrecurri11.g, gi·ve rise ro particl1Iar insr.111ces of some general type or class. The cask of critic(]:l ru.l31ysis is to explain \-Vb.y a general rl1le \;vas ~ictualized in tl1e Vla)r it \¥as. For instance, in the exa111ple fror11 Mozart's S}rmphony No. 39, t~1e rules of tl1e ga111e re]l us t11at cl1e theme belor1gs to the characieristic })1.l tter11 called "a11tecede111:consequent pllrase. 'T he strategy to he explained is rf1e partict11ar realiz_·atio11 o·f this nom1acive patter11. Becal1se rules do nc>t deter.mine srrategies, co111mo.n sense reasons are necessary to explain specific mtl.sical events. T l1ey bridge the gap between. rule and s·n·aregy. And be<..~use common-sense reasons are necessarily ad /,oc, criticism is, a11d vvill al vva)'S be, ai1 art- 11o·t a science, The reasons used. to explain a particular n1t1sical event \'rill, tl1en, he of two di:tfere·nt sorts: rule reaso11s, derived from. style ana1) sis 1tnd n1usic theory, vvluch will te11d to be consta11t, and strategy r·easc>11S vlf1icl1 \vill be of the ad !:Joe, con1mon-sens·e variety. Because they deper1d upon partict1lar circu11:istances, strategy reasons are ge11erally e:clecric. Sor11eti1nes they 'viU 1

1

'

1

I

be dni\vn from estal>lislied disciplines such as acotistics or psycholog;r; at other tin1es t he)' \Vill be b~sed upo11 con1mon ser1se. Rule reasons, too·, at least for

tl1e present, ~rill from ti1ne to cin1e he eclecric. This, because music theory is still rt1dime11tary and style anaJ ysis 0111}' so·n1e\vl1at le:>-s so. Not onl}r \vill criricisrn rend to be eclectic, but so1ne aspects of 111t1Sic may for a time simply be inexplicable. Fortt1na:tely, howe\rer, explanation need t1ot be exl1aust:ive at1d absolutely certain in order to be illu1nit1:1ting. \Vere complete information a.nd i11t~ontro\1ertibl e tl1eo0r a i1rereqt1i.
Even rhough critical ai1al.)'Ses are seldorn comprehe11sive, all coo often tl1ey v\rill seem t1ndt1l)' arduot1s and protracted. Tl1is is because rl1ere is

in.variabl}T a disparity bet'\\1 ee11 th.e s1>eed and ease "''itl1 \vhich 'tnusic is experienced .a nd t1nclerstood, a.J1d the length and co1111Jlexity of ·tJ1e discussion bigl1er level. Esse11tlally tl1e san1e poi11c is n1ade by I--! erllert A. Sit11011 it1 Tl:1e Sciencer of tl1e Artificial (Can1bridge: 1\.l.I.T. Press, 1969)t Chapter 2~ parti'-'tlla.rly pp. l3- 31.


ON TRE NATURE A.i.~D LI~'1ITS OF CRITICAL ANALYSIS

needed t:o expla1n wl1y and t1ovv it is experienced and understood'" A simple melody of., say, sineen measures, \¥.hich rallain precisely v\rl1at llappe11ed on th:e ridehow and ;vl1y t11e rider sltlfted his weigl1t, turned the wl1eel, changed gears, and modified his speed in or,der to follow 'the specific course taken~ It is this sort of particular inosical event-series \vhich the critic attetnpts to explain. Considering that even a si111ple n1elody is at least as con1plex an event as a sl1ort bike ride it is scarcely surprising tl1at explru1arions in criticisr11 are usually loriger and :more involved tl1an one mjght v\rish. Jones)s statement calls attenti()O to another impo1~ta.nt consideration. Just as one can ride a bicycle tvithot1c knowing ho\v a hic;rcle really works, so experienced listeners can respond setisiri\rely to music \Vithot1t knowing anything about what mal\:es music worl<: \VithOll.t kno""ring aboltt the theory or history O'f n1t1si.c. Beca11se it involves attencling ro ar1d comprehe11ding tonal relationships, t1ndersrandir1g 111usic is, I have argued, necessarily cognitive and analytical. Bur it does not fol]ow fron1 this that t1ndersra11diI1g depends 3

5 °The

Stabilit'y of the

Bi~ycle.''

P/;;i;ysics Today, XXIJI, 4 ( April, 1970), 34·


16

EXP LAINING MUSIC

upo11 li:11o~rledge of theory or it1forn1atioc1 about n1eans and techniques. \!Ve can perceive and comprel1e11d action.s and relationslliJ)S-mt1sic~ll as well as o.o nmusical-"vithol.it the explicit conceptualization ne.cessary for explnna.. •

t iOll.

Zealous listeners are sornetimes }1eard to protest tl1at tl1ey <'love>' 1nt1sic, but don.' t understand it. Th.is is, of course, absurd People seldom lilce '\that they do not understand. Quite the Of)pos:ite. Because it tl1reate11s a deep need for psychic sect1rity, men generally detest and reject ·w hat seems inco111prehensible. '\Nit11e.ss the hostility whicl1 conterr1porary n1usic so often excites in audiences a.ccu.stomed to the synrax an.d Structure of tonal music. Wl1at listeners n1e:an "'1hen tl1ey say tliat the)' d.0 11'r understand mtlSic is that tliey can1t read it, nan1e synt~1ctic proce.sses, classify forn1~l procedt1res, or othenvise explain how music \vorks. Lilce many theatergoers in ancient Greece or E liza·b ethan England, tt1ey are illit·erate; but they are by no n1eans ignorant. 6 Luckily, u11derst~lnding does not depend upo11 literacy or upon th.eorerical }{t1owledge. If it did, the al1diences for the plays of Aescl1yTlus and Shakespeare, or for the n1usic of Bach and Bra11n1s, would t1.t1ve been very small i11deed. U nderstanding a_nd enjoying a Bacl1 f\1gl1e or a Brahms sonata does not it1volve knowing about-conceptualizing cadences contrapuntal devices, bridge pas..~ges, and the like, any more tha11 being ei1tertain.ed by H anilet 1

involves kno,~ving about syntactic ft1nccions, prosodic .d evices, or dra:tnacic .means. Understanding m11sic, to paraphrase wl1at Bertrand Rt1ssell has said of language~

is not a n1att:er of kno\vi11g the tecl1nical terms of n1usic theory, but of habits co:rrectly acq11ired in oneself anti. riglitly presu.med in others. Listening to 111t1sic ir1telligendy is more like kt10\;\1ing ho\v to r1de a bicycle than knowing wh)r ,a bicycle is ridable. e Al1d th.is. may be crue of creative arcistS as \VeU. aq o f QUdierlces. For iru.'t:ance, Albert B. Lo.rd>s account of the singing of epics in Yugosla:via indlca.ces that con-

eepru.al,ization and classificatory knowledge ar e not necessal!)' for t'he composirion of oral poetry: "lvlan \vithout \.· vritin.g thinks in terms of sound groups anrl not in words, and tb.e t'4vo d.o not necessat"ily coincide. When asked urhat a word is, he will reply that he does riot kno\111, or be "~rill give a sot1nd group which n1ay var)• in lengtf1 from what \Ve call a word to an ·entire line .o.f poetry, or even an entire song.. . . Wlien a singer is pr:essed then to 53y \Vhat a line is, he, \.Vllose chlef claim to fa1ne is that l1e traffics in lines of poett}', will he e11cirely baffled b),. the <.\\lestion; or l\e will say that sinc e h.e l1as been dicntiug ru1d has seen his Utte.ran,c es being writte11 do,vn, he has disoo\fered what a line ist altho ugh he did not kno\.v it as st1ch before,. because he h,a d never gone 't o schoo.J!; T he Singer of Tales ( Can1bridge: Ha~rvard Uruversit}~ Press, 1900) , p. 15.


ONT.HE NATURE AND Lll\
This is not to contend that edt1caci.on cannot en!1ance u.nderstandin·g ani;i

hence appreciation ai1d enjoyn1ent. B}r calling attention to patterns and relationships which migl1t other\vi.~ have bee11 missed, it refines the a.u.ral imagination and increases the sensiti\riry of the cog1uri\ e ear. And to ·this enterprise, critical analysis can certainly make an i1npor,ta11t contribution. But education is not its primary goal. The primary goal of criticism. is explMiation for its own salce. &cause music fascir1ates, e.xcites, ar1d moves us, 've '\.\'ant to explain, if only in1perfectly1 in '\Vt1at \:vays the eve11ts wicliin a particular composition are related to one another and ho\\" suc~1 relationsl1ips shape musical experienc-e. Though kno\vledge abou.t rhe theory ai1d ltlstory of music are n<>t a prerequisite for sensiti\
zation, an.d "vhether an,d ho"v these are actualized; ho'v the event is strucrured hierarchically, and in wl.i.at: 'vays the several levels of the hierarchy are related to one another.

The answers to q·uestions such as these are not al,vays obvious at first. Repeated playing and listening may be reqtiired. Becat1se tl1e several parameters do not 11ecessarily move in congrue11t fas,hion (with the r:est1lt th.a t harmony, melody, rhythn1, and so on may each yield a different pattern of organization), it will at rirnes be helpful to analyze ch.e parameters separately in order to study tl1eir interrelationships. Often it i.~ iJluJninating to ''nor.nialize•, a passage- re\\'rite it in a simpler archetypal forn1-in order to understand how the composer has modified a traditional schema. Always it is important to discover which tones or l1armonies are strucn1rally essential and w:hich are ornamet1tal. \Nhen employing sttcl1 t:ecl111iques..._wluch a.re t

7

P:ersonal K'r1owkdge (Chicago: Universit;1 of Chicago Press, 1958); p. 360.


18

.E XPLAINING

~l.U'SlC

not modes of explanation, b11t methor disclosing l10'\Pv a n1usical event functions-the critic's ''ear,' 1 llis inusicaliry, n1ust guide analysis. [' t .m.u.s t accept or reject a line~r abstraction, ai1 harmonic reducrion1 or a rhytl1n1ic analysis. His ear keeps tl1e critic honest. vVirhout its control, theory or St}rle anal)rsis rends to become a Procruscea11 bed to v\rhich the praccice of composers is made to conform. Because its reasons ar e often .ad /'Joe and its explanatio11s eclectic, criticist11 IllaY ~it tin1es see1n some,~1 hat it11J)f0''isatory. But this does 11ot inean that it is:arbitrary or illogical. Different sorts of argurne11ts f rotn a \rariet)r of sources 111ay be e111ployed, but they 111ust be applied ol>jectively: rt1les a11d tecl11uql1es, argun1enrs and evidence n1t1st be used in the sart1e \vay in each anal)rsis; an(l,, chougl1 not systematized, re~so11s n1ust l)e consistent witl1 one an.ott1er . Criticism must ob,riot1sl~,. be musically ~1ersua,sive, b t1t tllis is not enough. For "' hat finally convinces is aural cogency combined with logical coherence. :Becat1se it must be scrupulous in reasoningl but flexible in strategy criticisn'l might ' veil be called t he delicate cliscipli1ie.

5· L-0okcd at from a11od1er point -0f ·\•ie\v, criticis1n attempts to underStand and explain the choices n1ade by a con1poser in a particular \\'orl<:. In order to do so, the critic 111uSt be aware of the options available to the co111poser at each pc_>int ill the co1ni)ositio11, and l1e must be able to. cscinriare (in a ge11eral \Vay) 'vhat rhe probable co..r1sequences of aJt,ernati,1e decisions '\vo11ld be. The critic inust have not only a. viable theoretical fran1e'\:vork, but equally

important a se11sicive f eelit1g for the s't yle. Sryle analysis is tl1erefore necessary to ar1d releva11t for criticis111. A particular melody, harmonic progression, or formal procedure is almost al\1lays understood in terms of the normative t)rpe or sch,ema ,o f \<\*'hich it is an exe1npli1icacion. As Nlorris Cohen points out, ''T he absolutely uniq11e, rha.t which has no eleri1ent in cominon \:v.ith anything else, .is indescrib·able, since all description .a nd all anal)1Sis are in r.ern1s of p redicates, class concepts~ or r epeatable relations.,, 8 For i11Stance, a listene.r who has IlOt learned. tllrough 0.11.tural e.1rperience, not 11ecessarily through .c la.c;sroom instructio11-the St}rlistic syi1rax of r.on~tl harn1ony "\= vill t1ot be able to appreciate t11e deviant deL1y of a decepti,re cadence. Sirnilarly, \vritl10 .u t a se11s-e of ti1e t1orn1acive procedures of a classical rondo "1':ith its more or less tegular r eturns of the mair1 a

8

T he .J\4ea11i1zg of H

i 11r1all

.H ittOr'J' ( LaSalle, Illinois: Open Court, 1947)1 p. 84.


ON 1'BE NATURE AND LIMITS OF CRITICAL ANALYSIS

theme, n11uch of the delightful play pres,ent iii, say, tl1e Finale of Haydn's Sy1nphon}r No. to~ wo11ld be nlissed. Style analysis is necessa,ry for criticisn1 not only because particulars are invariably und.erstood in the light of classes and 11orms1 but because such typalogies suggest ho"', th.e passage or event being considered v.rill probably <'work.'' i\ nd,. as noted earlier, an.alysis of a11 eve11t mu.st begin with some son: of hypothesis about irs funcrio.n. Pola11yi's observation is akin to \.Villiam Dray's co.ntention (n1ade in co1mectiot1 \vitf1 tl1e study o·f l1istory) that "explaining what a thing is . . . is just not the same e·nterprise at all as explain.ing why it . , . happened." 9 A patt of a composition without p11tent and closed melodic shapes and charllcterized by rapidly shifting ha._rmonic motion 1 is more underst:a·nda.ble vvhen conceived of as a ''developme11t section,' just as an historical period marked b:}r disturba11ce and rurmoil can be better con1prel1ended as a ''revolution.' 1 Moreover, to classify an event is to call attention to the Wa)' it fun.ctions and to pro1note a heuristic attitude with. respect to events \Vithi11 it.. Tl1e classification of the IV1ozart theme discusse-0 earlier as an. antecedent~con,sequent type led us to look for the missing nornurive cadence and, in this case, to find it. 1 To underStand a comp-0ser s choices is to en\risage the psychologicalstyliscic alter11arives ope11 to him at a particular point in a composition. 'F or th.is reaso·tl, particularly in tl1e sl1ort run., onr guesses a.·bout implicacio11s and concinuarions n1ay often be partly or wl1olly t1ustake11. Ends are ge11erally more accurately envisaged tha11 tneans. A11d cl1e predictable rout«! which suggests itself ro the critic will not as a ruJe be tl1e one cl1osen by the co,mposer. T11cir inve11tion is botl1 n1ore subtle and more adve11turous tliarl ours -""rhich is wh.y they, and not \Ve, are creators.. That our guesses may l'>e mistaken d.oes n"Ot, 110\vever, gainsa}r the it11port a11ce of consideri11g possible alternari,res. For 011r ·understanding of what tl1e con1poser actually did is significantly depende11t upon our understanding of tvhar he tnigbt have done. (Fron1 an aesthetic IJ-Oir1t of vie''' ' this is crucial. For it n1akes clear that musical enjoyn1ent Lies as much, if not n.lore, in the act of traveling as in the fact of arriving. V\'l1a.t d,elighrs aitd n10\res us, as \.\re listen to a compo,sicio11, are the cl1ru:1ging landscapes, the n1rns in the roacJ reveali11g unexpected vistas, an.d the surprise of delectable detours encountered en route to goals of relative repose.) 'Explairiing Whar i11 {:.C{istory,,, in Pa·t tick Gardiner, ed.,, T heories of H istory (Glencoe. Illinois! The Free Press, 1962 },. pp. 4.03- 408. 9u


EXPLAINING .i\iUSIC

20

But eve11 i11 the Iorxg rtin., our n1ost co11fidenr sur1T1ises abo.tl.t rot1res and goals 111ay pro\re \i\rro11g. Th.is is because, gi\re11 rhe parcit-ular style witl1in \:vhic.h he works. the composer is a free agent. He invents and sl1apes l1is u1itial musical st1'bstan.c e-his th.emes, ha.rrn.oiuc progressior1s, textures,. and t:l1e like. These ~la.ve implications for st1bseque11t eve11ts. But tl1ey do not determine t hose events. T his for c'1\ro n1ain .reaso.ns. First of all, the im11licarions--the possible consequences of a 11111sical e\rent, of a n1ocive> phrase1 or even section-are aJ,vays pltual. A n1usical e\rent implies a nt1mber of alternative actualizations. \~'hat tl1e cor11poser does is to disco,rer t~1e j)Ossibilities implicit ir1 his ow11 .n1usic.-al ideas. 111 Stravinsky's '\\'Ords, ''Step by step, lir1k b)r lir1k, it ' "rill be granted [the composer ] to discover the "''ork. ,. 10 The quality of his compositions depends botl1 upo11 his ability to discern or, if ,you \vill, to invent sucl1 i1npliDirions, and t1pon his arristic jt1dgn1ent in seleccir:1g inreresti11g and fruitful ones for his cor11posicion. Determinism is ~'l rnistake11 n_otion applied to worlts of art not only be,cnuse implications are plura~ but alst) because, \vithin the style he e·mploys, the composer .may at any parrici1lar poli1r in a piece be absolurely arbitrary. That is, he ma)!T it1vent and t1se a 111usical idea or relatio1iship \Vhich has r1otl1ing to do witl1- was in 110 wa)r ii11plied by or de1Je11der1t upo11- preceding events in th,e piece. Though he is free at any point in a ~rork to do as l1e likes,, a .re.spor1sible co1nposer V\' ill SLtbsequent ly take st1ch tl11: arbitrary a.ct i11to a:ccou11t. TI1at is, the relacio11ship berweer1 at1teceder1t eve11ts at1d the arbitrary one "i.vill, take11 together, have co11Seq\.1et1ces later in the con1posi1 tio11. F'or instance, the interrltption t)f the conse(1uent p hrase in ·1ozart s Symphony No. 39 ca1111ot (as f tU" as I car1 see) be inferred from a11ything that preceded it. It is simply a decisio11- rhough rhe critic cru1 suggest, as I tried to do, \Vh.y it is not an ur1reasonable one. Tl1e interr11ption is acce:prable for a nl1mber of r,easons. One of these is that it is subsequently seen to have important consequ·ences~particularly in tl1e de,re]oprr1e11t s:cctjon. All of this suggests tl1at the notion of {'inevitability'' in inr1sic inust not be talr11me.rcials. Ratl1er a piec·e of 1nusic n1USt see111 i1z retrospect to have fitteti togetl1er- to have been "right." A goocl con1posiStra.r1vinsky. I)oe:tics of A!ftl.'J~ic, trar1s. b:j' A.. Kn.odel a11d l. Dahl (Ca mbridge: Harvard Univer.si:r:y Press 1947) , p. 50. 10 Igor

1


ON TBE N ATURE AND LIMITS OF CRmCA.L ANALYStS

21

tion makes us feel the u11certainty of the ir11probable, C'len. \vlule convincing us of its propriety. It confro11ts us "''itl1 tl1e capriciotlS .and c.ons us into believ• • mg it was nec.essary.

If the goal of criticism is to· u11derstand a11d ·explain the mt:is:ical decisions made by composers, then ideas about music expressed by the composer himself, or by critics and theorists close to his time, should. be partic:t1larly rele~ vant for tlte present-da}r critic. They are, but \Vith important qualifications. Such u1formation is ju.st as relevant, but IlOt 0.Ile \~hit 1nore so, as Statements ma·de by other protag·onists in history~by Icings and pl1ilosopl1ers, genera.ls and social reformers, tradesmen and theologian5'-about their beliefs, m.otives., and goals. They may be· reliable and perti11e11t, or tl1ey may be biased, in.c-0m1llete, at1d misleading. Just as 'the policical or social historian evaluates th,e asserted. beliefs and, vie"\\1S of a11 historical figure in the light of his actual actions (and vice verse) , so t~1e critic n1ust evaluate r_he statements of composers or theorists in the light of tl1e co1nposicional practice ro \Vhich. the}'

refer. Such doct1mentatio11 may pr·o \ride fruitful avenues for criticism and analysis. Thus, as Le·onard Ratner has sho'\i\rt11 the views of eighteenth-century composers and tl1eor.ists about so1tata fo.rm help us to understand the practic.e of the period. On the otl1er hand ideas and theories from t he past must, when contradicted by the practice of tl1e pasc as we see it, be rejected o.r modified by prese11t-da:y theory and criticism,. In Donald Jay Gr,out's \vords: ''the correspondertce of theory and practice is no, more exact for medieval modal 111elodies than for a.ny otl1er rype ·o f acn1al mt1sic in a.ny 11eriod.,, 11 Most often, ho\vever;, differences between th.e views of composers 1

and theorists of the past and critics at1d t heorists ·o f the present are ones of etnphasis. For example, n1-0st critics toda;r \Vould explain Bach's Brandenburg

Concertos or tl1e \¥ell-Tempered Clavier largely in rero1s of syntactic processes and fonnal organization, rather than in t·erx.ns of t he doctrine of affecew York:: \iV,, iV. Norton, 1¢0) , ?· 53· Lawrence Gusl1ee makes a similar point '\vhe11, h1 a reviev.1 of a cranslacion of Gtlurius' P-ractica 1\t!usicae, he 'vrite:s: "But what are ,,,.e to make of the facr that most of Ga.furm.~· .foi1.rrh book described proportions whicl1 ai-e not e11countered in the '-\' titte11 xnusi:c of the time? Was Gifurius inroxicared ·by the spirit of systetnacization, mating the arithmetic thoocyT of proportions inherited from Classical A11tiqwry with tl1e principles of me1:isw:-al n1usic developed during the 13th and 14th centuries? . . . But a.put fru1n se:r1sitizing ~wde11ts to t:l1e i1nportance of .rhythmic .r·elationsl1ips and o.f nnderst:~nding the original fom1 of notation. of a Vl'·O rk-!>11Ch sensiti,Tit)' \vould also arise frnm a study of the works cl1emse1ves I cannot sec ni.any insights accruing from the study of Gafurius." Journal of A1'u.sic Theory, XIV, 1 (Spring, 1970 ), 12?-?30. 11 A History of M i llie (


EXPLAINING MUSIC

22

tio11s, a f(Jrm of explanation '"'hich \\t 011ld probably have b,een f~''·Ored ir1 Bach's O'tVn time. Finally a some\vhat general obse1~vation. Because tl1eyr have a veryr speciaJ relationsl1ip t<), and technical li1terest in, tl1eir art, composers are often inaccura.te reporters of vvh~tt tl"iey do and not u.nbiased jt1dges of the vvorlc of otl1ers. T he precedi11g discussion calls atte11tio11 to tl1e fat't that there is a significa11t difference betwee11 the co11cer11 of the critic or theorist who artempts to t1se present. day lcnowleclge about 1nan to explair1 tl1e art of mt1sic, and the concern of the hiscoria11 of tl1eory or of criticis1n \1lho seelts to account for the seque·nces of ti1eories and critical viewpoints about music. The distinction, analogot1s to tl1at betvveen a scie11tist an.d a 11istoria11 of science, l1as 11ot alv\rays been recognized by musicologists. For them theor)r in par·ticular has n1eant explai1ling "''l1at: past treatises ha,re said about m11Sic-~\:1~tll'j1 n1usic close to che time the treatise \Vas "vritten. Sucl1 studies a.re undoubtedly

important; bur the;r are essentially historical, not theoretical. The task of the tl1eorist is differe11t. Using as his· prim.a ry (iata the rnttsic itself, together \Vitl1 his O\Vtl st}rlistic e>i.-perience a11d wl1ate\•er can be ascer·cained about perforn1a11ce tradition1 11e atternpts to construct hypotl1eses abot1t the ways mt1sical e,1ents-n1elodic, rl1ythmic, l1armornc, textural, etc. -are related to 011e another. In so doing, he nta}' refin.e exisci11g 'hypothesest devise quite 11ew ones, or borro"v concepts and :methods fr·O·lll other disciplines sucl1 as linguiscics, psycholog)', or s_. stetns analysis. His theory 1na'}t be corroborated by treatises \11ritten a:t the riine tha.t the mt1sic he is co11cer11ed \Vith was composed, and this \vill constitt1te sup11orting e\ri.dence. B·u t the

absenc·e of historical corrobo1·ario11 \vi11 not

r1ecessarily~

prove tl1e newly

formulated t heory ,, mistal

ON THE NATURE AND Lli\f:ITS OF CRITICAL AN ALYSIS

2.;

does not mean knowing \vhat actuall}:r \Vent on in tl1e composer~s mind

wl1en he wrote a p.articular work. Probably neither he,. 11or we, will ever kno·w his 1nencal p1·oc..-esses as tl1ey actually occurr·ed. Rven ~"hen a. composer was conscious of n1aking a~ decision- \vl1er1 his l1abit of craft was not immediately adeqt1~1te to the problem at ha11d

cl1oice n1ay ha.ve been largel).r

intuitive. After considerable thought ·trial and error experimentation, and just plain daydreanll:ng, the right sol ucior1 111ay 'ba\re appeared, as it were, out of the blue often \vhen least expected. The result seerns clear and ''logical," but the ro1Jtc follo¥ 7 ed it1 reaclill1g it n1ay \Vell have beer1 veiled ru1d circuitous. In o·rher ~.rords, just as there is a differenc·e between tl1e logical steps throt1gh ivhich a sc.ienrific argument is presented and tl1e act of scientific discov~ery ~rluch n1a}r have been the res11lt of u.ncol1scious pr.ocesses,, owing somethitlg to trai11ing, to dis1)ositiot1, to tl1e cu.r rent state of tl1e discipline, ·and frequentl)" quite a bit to chance; so t:l1ere is a difference benv·een the coherence ai1d consiste11cy~ of a completed composition and the composer's creati\te p:roeesses \vhich depe11d upo11 a, combii1atior1 of tra.illirig, traditio11,. perso11ality., and, again, plain luck. 1 ~ In s11ort, tl1e critic attempts to understand not the history of the decisions \vhich resulted in a coinposition, but tl1e '< logical'' alter·natives pr·esented to the composer gi\1 en the strucn1re of a particular set of mtisical circurrlStances. He is, to paraphrase Aristotle,

cor1cerned

~rith

wl1at migl1t be called rl1e poetry of creative cl1oice, not itS

history. This point is especiall)r rele,,,rant be~use there has of late been a salutary interest on the part of historians as \;\t ell as theorists ir1 the skercl1es, dr:afts, and autographs of co111p-0sers-tl1at: is, i11 l1ow they \Vent a'bout writir1g 1nus,ic. An.d \vhile sketches and r1otebooks may rJe rei.1 ealing, they should. not be confrased \Yith. the finished \VOrk.1a Eqt1ally imr>ortant is the fact t.h at the cl1anges n1ade by con1posers fro m sl
1a Se:e th.e
Beetho,1en)s i 1Er0:iC11u S}~mpt1ony.


EXPLAINING MUSIC

cedures, not to explai11 their compositional tlun king.

v\'e can expla:i11 ' vhy

a composer cllanged a passage as he did put ~1 sforzando on tl1is n·ote, modified a n1elody in this way, <.>r altered a. rnodulatory schen1e in a. particular fashion

only if we have a cheory about the nature of niusical relationships.

6. I should like to close b)r briefly considering the limits .and 11azards of critical analjrsis. Clearly just as rl1e pu11ish111e11t n1ust according to a famous Oriental potenrater fit tl1e crime, so an analytic metho·d or theory n1ust be appropriat e to the style of tl1e co1nposition being studied. It is pointless to analyze, say, a \Vork by Nlacl1aut or one by Bot1lez it1 te:rn1s of techniqt1es developed by H eirrrich Scl1e11l{er for tll.e analysis of tonal m11sic. In rl1is respect critical r11ec:hods .n1tlst be exclusi,,e. Given t.his important li1nitarion, hovvever, critical analysis should in general be as inclt1sive as possible. That is, nil the n1ethods, tl1eories., and techniques ~,hich are relevant to a11d '~rill illt1minate the co11rposition being consid.ered sl1ot1Jd be br·ougi1t to bear. In anal)rzing a Beetl1oven sonata, for instru1ce, a nt1m.ber of tecl1niques and tl1e-0recical ~rpproaches ar e app1·01Jriate: conventional harmonic analysis, morivic derivation (judiciol1sly er11ployed ), tl1e 1nethods developed by Scl1er1lcer, tl1eories ha\?ir1g to d·o w itl1 the ·etl1os and cha.racter of inusic, and. so 011. lfl hich analytic n1eans are used "'rill depend 11ot o nl)r t1pon wl1at lcii1d .of relaci.011.s hip is beir1.g co11sidered , but a.lso upon the ll:ierarchic le\1el being a11a1yrzed. F'or n1star1c.e, con:\'e11tional harmonic analysis is ar}propriate to the sn1dy of the lowest le\rel of harn1onic progression, \\rhile the techniques of Schet1ker are rele,rant to the 1Jndersta:nding of higher levels-the middlegrot111d a1'1d backgrot1 nd. But no matter ho'\' i11clusi\1 e and derailed a c.rirical ar1:al)tSis is, it is • seldon1 exhaW>"tive, and it is 11ever definitive. It is seldo1n exhaustive because n1ost pieces of serious mt1sic are complex. Conseql1ently it .is aln1ost a1,~1ays p-0ssible to discover relationsl1ips not previously obser~vecl. The c,ritical analysis o.f a particular '\Vorlr is .never defi11itive because the tl1eor}1 of rnusic and tl1ose of related disciplines st1ch as ps1rchology are likely to cha11ge. i\nd heca,use it is partly depen.dent upon sucl1 t'l1eories, a11al}1tic criticism '\Vill probably cl1ange too. Tl1is should not, however, be cause for despair. For thot1gh criticis1n is x1either exhaustive nor de.fi11itive, it does not folio"'' (an}' more th.a n it does il1 the fields of chemistry or biology t 1)sycl1ology or philosophy) tl1at \ralid and valuable insights are in1possible-that criticism can explain nothing


ON TBE NATURE A.ND Lu\.':lttS OF carrICAI, ANALYSIS

about music

~nd

our experience of it. On the 0th.er l1and, the inchoate state

of n1~1sic theory ~ttld St}rle analysjs, together with t11e need for specially sensiri~re judgme.nrs it1 the explanation of particular instances (strategies),

make criticisn111ot merely a delicate, 'but a hazardous discipline. The desire for certitude a11d per111anence is both :deep and a,biding. Conseqt1ently style an,alysis, chronological studies, and paleography tend to be 1r1ore attractive to rnost me111bers of the aC'ademic establishment r:l1an theory, criticism, a.c1d history. For, h1sofar as style a11alysis n1erely d.escribes and classifies, and l1istory 111erely authenticates or arra11ges data in chronological order, their observario11s •1nd res:ttlrs ap1Jear to be certain and secure. They \\rill stan,d. the test of time, except, ·of course, w hen ne,vly discovered info.rrnation n1akes it necessary to revise exisr.ing nor111s or cltronologies-a possibility which diminishes as the accumulation of data grows. Theories and criric.-al analysis, on tl1e other hand, are fallible; c.iebatable at1d pro""'.isional; a,11d so a.re those histories \vl1ich attempt to explain "'1l1y a sel'ies of events happened as it did. Theories are rejected or revised, lristoi~ies are rewritten, and criticisms are not ,d efiniti\~e. Disheartened and perhaps disr11ayed by the speculati,re uncertainties of tl1eory, criticis111 and1 one shoulci ad.d,, history as disrinct frorn chronicle, too many humanists, 1>articularly those in music, have ·te11ded to follow t he well-worn 1)atl1 of safe scl1olaxship. But to choose prospective certainty over

present insight is bol:h 1nistalcen and misguided. It js nristaken be.cause the search for fuial, definitive a,!Th"\Vers is ar1 lJna:ttaina.ble goal for ·chose disci1}1:ines concerned wit,h understanding and explanation. For, since the furure is open and influential, it can change our un·derstandi11g botl1 of past cotnpositio115 and of past historical eventS. It is misguid.ed-:paradoxically so· because the enduring mo11t11nenrs of scholarship, '\Vhich have shaped inen's 11unds and beliefs, far from being cautious a11d circun1spect, have been those wl1ich illuminated a relatio11ship, a work of art, or a past epoch thro.ugh a bold, encompassing hypothesis. Thou,gh in all probabiliryt tl1ey will s·ubsequently be revised, or even rejected, su·ch. \vorks and theories e11dure because they are exciting and semi11al: they lead to ne'v discoveries and further formulations., and thereby continu.e to affect language, rhought, a.11d behavior.


ERII -

-

I

I

.

! . ·~------· --

Critical Analysis and Pe ormance: The Theme o J\1ozart's A-Major Piano Sonata PRELUDE

Thougl1 tl1e patterning of tonal mtisic '\.vill be the specific concern of this book, one or t'\¥~o general observarioiis need to be made a.t tl1e outset. The .first is th.at the existence of borl1 sirnilarit)'• a11d difference betwee.n n1usical events is a necessary con<.t ition for ptltterning of any sort. Si11ce t}1e concept of partern ir1cludes rl1e 11otion of regt1larity· and ord.e.rliness; patterning is incon1patible with complete l1eterogeneit)' . \!\That is less obviotis; b·ut equafly important, is that t<>tal uniformity also tends co preclu.de patterning. A note or a harmo11)r repeated without a.n.y change in tiy11amics; durarion, or timbre is not a pattern. Like the ticlcing of a clock it is 11ot disordered, rather it is unordered. E\r·en a motive (\vltlcl1 is its.elf a pattern) 'vill not, if repeated exactly, creare irnplic-ati\re patterning on the next hierarchic le,reL What ordering there is will be formal, not processive. Sirnilarly, the pattern.s produced by con1pletely t111iforn1 changes '\eviJl be tlnstl:tisfactor}r. For UlSta11ce, ·cl1rorr1atic and \vf)ole-tone scales or exa.c tly uniform rnelodic and har moi:tic sequences, thougl1 tl1e)' give rise to \rer}r regular lo,~-level relationships. es.tablisl1 no points of relative stability and clost1re. The series

of events is so undifferentiated tl1at it is potentia11y e11dless and, in this sense, u11orde.red and aborti,1 e. If a repeated e\r.e nt is et1tirely formal, tl1e11 a ttnifortn series of changes is exclusi,1ely processi,·e (see Chapter IV) . Viable patterns fall some,vhere benveer1 tl1ese ex:rren1es. They involve t)oth form and process.


CRITICAL ANALYSIS AND PERFOR..t'VlANCE

2.

7

In general, the more clearly and co.mpletely stiaped a patter11 is ·-'t he more patent its order on a. 11umber of hierarchic levels--tl1e more specific and compellir1g irs irnplicatio11s \;\1i]l be. l\ si11gle tone, for inSta11ce, is not a patterr1. Eve11 within a particular musical style, it might be followed by an indefinitely large nu111ber of a]remati\re contit111ations t11a11y of \vl1ich would to all intents and purposes be equally p·robable. As, tones are added, durati-On.al relationships established, and other parameters specified, tl1e relational ordering- the pattern- becor1-ies ap1>arenti a.nd unplicarions are defined. In other wc>r{is, as an event t111folds in time, the prolJal>ility of' some modes of continuation and closure it1creases relative to others. Delaj1S-digressio11s, extensiorlS, ar1d tl1e like-occur towar·d tl1e ·end of patter11s botl1 beca,use they are most effective "''hen specific goals are in ,rie'\v 1 an.d because the clearer and more probable the mode of continua:tio.n, the gre.a ter the n eed for tl1e heightened it1terest provided by delay .o .r deviation. Mu.~cal p~tterns arise \vithin a11d are gov·erned by the gramn1ar of a specific style. In ti1e case of t11e t11t1sic co11sidereci in tl1is b.oolc, the style is t hat of Western tonal music fro·m about 1650 to, rot1ghly, i910. J t1st as literary criticis1n broadl}r spcalVhicl1 are less so, '"''l1icl1 triads are relatlvelyr stable a11d '\vhicl1 1

tend to be mobile and on-going, and so on.

I must also assume that the concepts of rhyth111ic strt1cture and S)Tllt'
1

1

1 Indeed 1 dela.}7 entails tlle notion of goa!s-i11-vie:\\>'. 2 See The RJ;1ytl:nnic S tr1.£tti1re of 1'\tl1JSic ( Chit.~go : Universit)1 of Chicago 196.0).

Press~


EXPLAL~TNG

MUSIC

of all paran1eters of 1nusic- rnelodic, h~lr1nonic, d)rnamic-as '\Vell as dura-

tional relatio11sltips. Nlore specifically . End-accented groups ia111bs ("' ·- ) and anapests ( v w ._ , -are ge11erally spealru1g n1ore stable and closed than eitl1er rni
·:

')

1

3

Hu111an J(norwledge.: Its Scope
p. 343. For tltis .reason, as occ:ed in Chapter l ( p. 7)t sratistical methods a.r e appro1>riate for style ana.l}1 sis, bat only indirectly (throng+1. c·he relevance of tyle analysis) for • • • crtt1c1Srn. 4 lbid., p. 343.


CRITICAL ANALYSIS k""m PERFORMANCE

upo11 probability esrin1ates.

011

the one hand, the history of music seen1S to

show that a change in the freque11cy \\11th \vhicl1, say, .a partict1lar chord progression occurs in the literarure of mt1sic does affect the listener's sense of harn1onic probability. On tl1e other l1ar1c.i> \Ve are not simply products ·of Pavlovian conditioning, as an esse11tially statistical th.eory of pattern comprehension \vould irnply. Since I have considered this point else~rhere, 5 let n1e only restate n1y judgme11t tl1at n11plicative probabilities, thougl1 it1fluenced by frequency at rin1es, arise out of and are sigr1ificar1tly tl1e result ·of fundan1ental modes of human perception. a11d cognition- dispositions and proclivities having to do "vitl1 the vvays i11 \vl1ich the l111ma11 nervous system process and patterns 9ense-experien.ce. r.

In Chapter I, I a.rgued that anal}1s:is is something \Vhicl1 11appe11s whenever one attends intelligently to th.e '';r.orld. W11ene'\rer srin1ttli ar·e grot1ped, ordered, and related into coherent pattertlS a11d processes, analysis has tak ei:1 place. The performance of a piece of music is, therefore, the actualization of an analytic act-ev en though st1cl1 analysis ma)r l1a\re been inttticive and unsjrstematic. For '\\1l1ac a performer does is to make the relationships and patterns potential in the co·mposer's score clear to· the mind and ear of tl1e expei-ienced listener. C-0n,rersely, as Edward Cone l1as pointed out ''Active listening is, after al~ a kin.cl of ,rjcario·tIS performance . . .'' 0 Just as analysis is i111plic:it in what tl1e performer does, so every critical an~alysis is a x11ore or less precise indicatio11 of how tl1e work being analyzed sho1lld be performed. Bj" explaining the processive and formal relatio.nships of a con1position, analysis suggests ho"'' I)hrases, progressio11S rhythms,. and l1igher~level structures should be sha.p ed and articulated by the performer. At times such relatio11ships n1ay be equivocal either because the patterning itself is so or because several different grottpings are implied simt1ltaneously. In such cases alter11ative a11alyses will be possible. But su.ch alterna.tive interpretations, and the performances to vvhich they give rise, Virill 'be complementary ra,t her than contra
M1uic, the Arts a'lzd Ideas ( Chicago: U n.iversit;r ·O f Chicag·o Pres.s, 1.967 ), pp. r91

201 ·26o.

e Afu.sical Fort1i and Perfo'Y'1flance ( .ew Yorlr: \;\l. W . No:non, 1968}, p.

11.


EXPLAINING l\>iUSIC

dei>end upo.n th.e taste, ten1peran1ent, ar1d training of the critic or· (he performer, perhaps tl'1e best tl1e least an1biguous~'Vva y to deal with these 111atters is to consider an inter·pretacion \vh.ich I tl1ink is defi1utely \.vrong, a11d t1~y to sl10\v "'''hy it is so.

Ir1 tl1e Peters edirio11 of iVl.Oztrrt,s l'i.1110 So natas, edireci by Louis Kohler and Adolf Ruchardt, the Theme of the first n1oven1ent of tl1e A-Maj·or Sonata is phrased as sho'\vn.i11 Exa1nple 5.

E.xarr1ple 5 According co the n1ost recent autl1oritative etiitiot1 of rhe sonatas, done by Nathan Broder~ Nloza.rt pl1rased the Then1e as sho\vn in Exru11ple 6.

'

.

. •

'

Exai:nple 6

In n1y judgment the: phrasing ii1 the Peters edicion. is dovvnrigl1t \vrong. The question is: -i;.vl1at is wro.ng witl1 it? What reaso11S can one give for preferring l\1ozarc's phrasi11g? 011e n1ig·ht content 011eself by argt1ing from at1tl1oriC)r, saying: t'\ Vell,


31

ClUTICAL ANALYS1S AND PERFOltl\..1ANCE

that's the -uray the ct>mposer wrote it.' Bu.t such an argument \Vould nor. help us in making phrasing decisions in B.ach. s music-wt1ere such m.arks are 1

1

few ru1d far betwee11. Nor does it tell us wl1y tvlozart's 111arking 1nakes t11usical sense, 'vhile Kohler's and Ruthard.t's phrasing does 11ot. At fust blr1sh th,e difference bet\Veen the. t\\10 versions seems slight. After

a11, che notes,, durations, a11d barmo11ies remain tl1e same. Bt1t tl1e rhythms of • the nvo are .different. Corisider the low-level rl1ytl1n1ic analysis of i\>lourt's phrasing (Exan1ple 7A) a11d th.at ir1 the Peters edition (Lxa111ple 7B) .:

A.

B.

Example 7

In N!ozart's version, tl1e lowest level ( i) consists o:f a serie-s of trochees, e..xcept for the fin;al groi1p "'' hich is pivoted and forms tl con1pound event. The groups stay v.rithin t:l1e 111easure, as indicated, e\ en on the second rhytl1mic level (2). The phrasing in t11e Peters Eciition,, 011 the other ha11d .makes the weal{ eighth-notes, pa.rricularly those a.t the end of each measure, function as t1pbeacs. T he unirs are pri111arily jan1bic-tho11gh con1plexJy so. The second level ( 2) , too, is rather irregular at1d lacl!s S)rrilmetry; particularly, one is in d.o t1bt abot1t l1ow to g.rc)Lll' tl1e tliree ia.mbs in tl1e rniddle of the phrase. 1-iere a general reservatio11 about Kohler's and Rurhardt,s abet'rant phrasi11g should be voiced. Its asyrnl'r1etric irregularity seems at odds v\ritl1 the basic simplicity of th.e u1ne :and the har mony. M.oreo,rer, variations are to be ba.i;ed upon this rl1eme, and sinc.e complexity ''till probably con1e later in the movement, its propriety is at least doubtft1l i1ere. Someho~v. too, the rather strong upbeats of the aberrant version lack refinement and nuance. Next to the Sl1br]e cool of Moza.rr:'s phrasi11g, they seem blatan.r and gratU·itoosly emphatic. Tl1e aherra..nt versio11 '

31

EXPLAINING MVSIC

the short notes te11d to be l1eard as upbeats to the longer ones '''hich f ollo\v~ rather tl1an as "''ea.k afterbeats. For instance, a series of notes of equal duration in duple meter tend to he heard as trochees:

This effect \Vill be t)articu1arJ.y clea! if a slight stress is placed upon eacl1 main beat. If evenr we~]c beat is no"'r shorteneti .relative to its accent, th.e \\'eak: beats '\\1ill begin to group \i\rith the 11otes \\rhicll follo~r ratl1er rl1an \v•ith those which precede; ,J

And if these triplets are n11ade into dotted eighths and sixteenths1 the, effect is e\ren in.ore striking:

For tl1is reason, cl1e ·eigl1t:h-nores .follo\ving qt1arter-notes in a % meter tend to ftmctic)n as upbeats (a1u1cruses) . '"f11is is also the case in ~'1
beats. Th.e same is trt1e of rl1e \~·eak be.ats on A and B in tl'le third n1easure. The a111biguity of the E-the i1ossibility of he:aru1g it either as the 'veal< beat ·O f a trocliee {· , .. J ;*1'I ) or as the tip,. . beat of a11 .ianlb ( -j· ... ,J:, )-is easily tested. For instance, if rhe E at the e11cl of n1easure .1 is follo\.\1ed b'' conjt1nct n1otio11, as it1 Example 8, then the final eigl1tb-11ote of the 1neasure is perceived as an unequivo-cal an9crus:is. Pitch proxin1ity allows tl1e potential mobility of the final eightl1-nore ro be acrt1a1ized. Notice that duration.al

,!J

~

J

. ,,-:-- "'f 25

•1

[

,, w _

'

~

,~ "· ... ""9.'1.

- Wi l11___ "c•

·_ L•

Example 8

relationships ac1d harmony a,r e tl1e san1e as it1 Mozart's Theme. Tl1e eigl1thaote E in .i\tlozarr's Tl1eme acrs as che 'veak part of a trochee .rather than as an


•

CRITICAL ANALYSIS AND

PERFOR~'1AN CE

33

upbear because pitch disjunction (the skip of a fourth ) tends to prevent r hythmic grouping across t:he barline and because the seq\1ential repetition of tl1e inelodic patter11 in measiu:e 2 1nakes it clear tl1at t11e first measure is a separat:e, integral musical event. The po·ceuria1 mobility of tl1e last eighth.-11ote of 1neasure 1 is also sl1-0\vn in the fact that its function is easily i11flucnced by ptior org11nization. Thus if the theme is made to begin with ·an upbeat, E, as in Example 9, tl1en the Eat the end of measure 1 and the D at the e11d of measure 1 '\.Vill tend to group witl1 the notes '''hicl1 f ollo~r t hen1, as ao.acrt1ses.. This ic; the case because 011ce a particular ·01ode of organization., such as a rhythmic grouping,

is esr;ablished, it tends to be continued 11s a '\fa)~ of patter1ung until c-0unce.r vailing e,1ents take place. A11d for this reason, the V\1eak beats in measure 3 (of Example 9) \vere also probably felt to be anacruses.

p

\~ \a?::: ,--:-, t.. •:r-.__ '1"' .._, V

.,._ v Ji

_g

~

__ J S

W"'"'-

·-•-• : ·t

~

f

Exa:r11ple 9

But an e'\ren more modest change '\vill show that the E at the end. of measure .1 is a. potential upbeat. For if the sL~tee·n.th-note D in the first measur·e, and t.l1e C# in the second, a.re deleted"' as in Example 1 o, the groups tend to be heard as e11d-accented-as ia111bic. Obsen1e th.a.t thot1gh these sixteenth-notes seen1 to be me.rely decorative, they perform an imp·o rtant

, •

E.~.at11ple 10


34

EXPLAINING MUSIC

n1osical function. Conseque11tly conscie11tiot1s a11alysis vvill consider each noce and duration as well as every 11uanc.e of d.y11an1ics and phrasing to be syntactically significant l111less shown to be ocl1erwise. Th·e si.~teenth-.note D is crucial becatise it ma1
--

-

illf"" -

I i --·' C. R""""'h--.&, . -.'!:I~"' 1£

. . . . 21

- w. . _..,, "

£1

-

.

-.

.... - . . D. -,;.."'d..u.. -

.it,

, , ,

Exa111ple

....,

t, · 7

~,

...

"..,-, •

- -

1i

an anapest ( i 1B), or as a.11 a111pl1ibracl1 ( 1 i C). V\' hat cha1iges in ea.c h case is the way in vvhich ihe '\rea:k beats are grouped '\vit:h the. acce11t. But if the first note is dotted ~t11d the secotlti is ma.(le into a si.-xteentl1 (as in Exarr1ple r 1D ) i then tl1e final '\>veal{ beat is tied to the preceding accent and can11ot

function as an anacrusis. Now tl1is is ,pre.cisely \Vl'l.}lt the sixteenth-note, D , in i\ilozart's Theme do-es: it prevents the follovving weak bear, tl1e C# f1'\om f unctio.n.ing as an anacrusis to the E. Once this trocl'iaic grotlping is established- and reinforced in the second meast1re- it acts as ~i 111odel for subseque11t grot1ps u11til ,meast1re 4. As indicated in Example 7A, tlle \Veale third beat (the tvlO si"iiree11tl1s) i11 n1easure 4 acrs as a pivot li11l{it1g tl1e t\\'O C#,s. Tl1e sixteenths a.re heard \\-1ith tl1c preceding quarter-note Cl because of the previotlSl}' established trochaic group.ing, an.d the)' act as t1pbeat to tl'ie follovvin.g C.~ both because rl1ey are sepa~1ted frorrt th·e prece,ding c~ by a skip and beC~lUSe tt1ey n-1ove to the second C$ by more r·a pid, conjunct n1ocion. vVere tl1e phrasing of the Peters edition, sho\vn i11 Exarnp.le 7B follo\ved, no pivoting


CRITICAL ANALYSIS A!.t~D PERFORMANCE

35

""'ould take place because the pre\riousiy established iambic pattertt, far from li111ring the c~ and E of n1east1re. 4, \VOuld separate them. The result, as \Ve will see, \VOuld be an asymn1etrical, a~rk\~rard plU11Se strucrt1re. Like the a11teced:e11t phrase, the consequent begins '°vith crocl1aic groups. Bur, as indicate,d in Example I 2, the organization is ·modified in 01easure 7. TI1e eighth-note B serves as a pivot linking A an.d C#, so that it is al1n0.St an upbeat- thougl1 an internal or1e. T he finru eigl1t-note, D, howev.e r., is a

-- -v .,,,. -

22

1t

'

._

.

.

-

• ,-

;I

ii

rl

•

,._ a

~l E

-~ z •

,_, _,

~~

~

Ex~1nple

-

.~ § ~I

_.. t

- -1CJ-

:"

""" -- · - -

·,

rz

strong, una1nbiguous anacrtISis. Th.is anacrustic ft1nctio11 is einphasized by the skip i11 the base f rotil A do,vn to the D \Vhicl1 pre,ce,des the conjunct motion across t:he barline, by the t11icl{ening of t,he chord sonority, and, above all, by tl1e sforzando '\cvhicl1 the weak bear bears. l11deed,, the sforza.ndo marking makes sense because it 111al{es clear that tlle fi11al eightl1-r1ote of the % tneasure has at last b:ecome ·an upbe:at. T his perhaps needs some clarification. It is often thot1ght that d:ynamic 111ai·kings-sforzttndo', forte-pimzo, cres·cetzdo, and the like~essencially act to define the character and mood of a composition. S1rrely they do so. But they have an even more important syt1tactic f11nctio11;. they• serve to articulate and sl1ape patter11s a.r1d formal st1-uctures. Whe;n it comes 011 a weak beat, as it does u1n1easw:e 7 of l\iiozart's The1ne a sfarzando serves as a role to· malce it clear that the "irea.k beat is an anacrusis. 7 For ~'la.mple, in the 1

In mcetered ro.n al music,. a sforzando does 11ot as a rr1le make a. weal{ beat into

au accent'ed one. The added stress serves ro :lrticulsce a grouping, t)ut does not change the 1'l'letric functions of the bears. W'l'len a: sforzemdo co1lles on an accent, :it ofter1 malces the accent funcrior1 as th:e begirutir\g of a r l1)rtl1111ic pnttern at1d ties sub-sequent weak beats t0 the accent. This is 'vhat the sfo1':Uflzdi in meastlres r :r and 12 of l\11ozan's T hct11e do {see Exan:iple 17). These a.tld oth~er r11atters l1aving to do with the articuJn.cion. of rhyth11uc patterns are (iisc11ssed in Grosvenor vV. Cooper and Leonard B. £\1'.eyci-, T he Rliytl:J-mic Str1J.ctzire of iW11sic ( Cl1.icago: U rliversity of Cl1icago P:ress1

1960) 1• Chapters I-111.


EXPLl\JNING 1"1USIC

36 Afl&yo

D

~·-~

. r;r

..

"'

•

·• r

I

I

. ..

.

'

.

.

•

~· ... ~~~~ --~·· °'' r-=tt= ' ·~-- '~ -·· . . .· ~ ' If •If • -

a.

.,, : ',

•

' 'f

1

• i°'.l • : ' ,-. ~ .

-

-

I

I

.-

..,·

""

:t

'

~-

f

...

. , .... .

.

~

•

I

~ ,

r

E.~ample

c3


CRITICAL ANALYSIS AN D PER.FORl\UNCE

37

f0:urth tnovemenr of Beedlo,ren's Fifth Sympho11y, the a11acrustic grou,pings are to a considerable degree depe11dent t1pon the series of sforzandi (Example 1 3). Observing, incidencl.y, t hat orchestration too. helps to make the syntactic patt:errung clear. Recurnin,g to i\tlozart's The1ne: the decisive upbeat at tl1e end of measure 7 is the realization of a potential ' vhich "''as latent, but not actualized, in the preceding \veak beats, When it finally arrives, tl1e clear 311acrt1sis is experienced as a kind of achieveme11t -a psychic satisfaction. Were the groupings unambiguously anacrustic fr<>n1 the out.Set, as c.-alled for by the p.hrasing in the Peters edition, the emphatic upbeat at the end of ineasure 7 '\vould be anticlimactic, a11d tl1e sforza1zdo \c\•hich enforces it would be red\111da11t- \vitho·u t syn.tactic significance. T l1ese rer11arks suggest that the perfortner n:1ust. take considera.b le care not to str~s the \\'eak beats i11 the first six n1easures-·. emph,a.~

or by sligbtl}r lengthe11ing th·en-1. For if he does so, th.ey will tend t o he perceived as actual, rather than potential, upbeats. If the E. at tl1e e11d ·o.f the first measure is understood as a potential upbeat, then the interval fro111 E ro B across tlie barline mt1st l>e an tinrealized perfect fourth. That is,, if the structure of t h.e first two rhytl11nic le\rels is trochaic,. then (as sho'\.vn in the analysis in E.xan1ple 14) the melodic motion witltl11 each mea~'ttr.e "'rill be heard as a rising third, and t he. r11ocion between measures ,vifl be a conjunct d.esce11c fro1n C# to A,. Co11sequently the either by d)rnamic

Exn.111ple 14 fo·urths .. from E to B and fro·m D to A-are 'llOt perceived as active syntactic connections. The relationship is potential. le rernains to be actualrred. I-Iowever, not u11til the first variation are che fo1ml1s hear,d as synta.ccically explicit events. T !:Jere they bec-0me aural a·<..'tualicies. A.nd wl1en they do so, the upbeats latent in the first ro.east1res of the theme are also actuali1..ed. Once again the phrasing in the P eters edition tnisconstr11es and distorts the musical tnea.ning of the tl1erne. The fourtl1 relationslup 1nust be kept potential in the Theme if its a.cmaliza,t ion in the first \rariation is to be interesting and significant (Example 15) .


EXPl.1\INlNG i\-1.USIC

-

- - - - - -· -

If,

lllli

I

~,,

These observations call ~lttenrion to ~t i11ore ge11eral at1d perliaps n1ore serious shortco111ing of the aberra11t f)hrasi11g. Namely) to· e111pha:size the connection betvveen \veak eighth-ot)tes an.d the aocenterl qt1arrer-notes, \vhicl1 folio\"', is to direct perception to low-level, foreground events at tl1e expense of h1gher-1e,,el sm1c.'tural relacior1ships. Thus. if the E is heard as directl)r coni1ected to the B, the importance of the laiger lllotions fro.n1 Ctt to B to A, 011 accents and from E to D to C# on weak beats tends to be obscured. This differen.c e is indicated, perhaps ir1 a sotne\vl1at exaggerated way, i11 the a11al)"Ses u1 Ex11n1ple 16.. Note1 incidentally, tl1at the structural irnpor'ta11ce of seco11-A.a.r

-----~--

Put.rs

£...~ample 16

the C# and tl1e B i11 111ea.sure 4 is the result not 011ly of their harrnonizatior1 as a ! :; semicade11ce, but also llf the fact that the primary melodic line (Cl-B-A) and the secondary rr1el(>dic line (E-D ) co11\r,erge on these pitches. T l1e structure of Nlozart's T her11e, tl1ot1g.l1 h.ierarcluc..-ally c-0n1plex, is at the same time exceptionally clear, at least when ,perforrr1ed as Nloza:rt pl1rased it. This scrucrure is sho,,rn i11 E.x3n1ple 17, I11 rneasures r and 2 rhe rhythmic groups •ire as "ve i1ave see~ trochaic on rl1e first t-vvo le,1 els. On the third


·CRll"lCAL ANALYSJ.S AND PERF01Uv1ANCE

t-

39

A .

f

r

0

I

_& .

,_....,....,.• ---.~-•~~"'•'--1,.~,----oa·-------iu M J... ..J! __.,..._ _-

JJ

, '•~z ; 9;wd

\,Jr' >fr ;

I

:a,11:S7 ..L_ ll

&

"<

' '

· f'

'

z

t

•

5

I

~,

I

... _k+ , • . ac, '111

\ u-

.-

-

,.,,,,-y , ..,.,.,)', "*W'~ ·-

rt , _,.... ·

"c' 2' +

'91'

t

-1

I

FF 7

I

<"-_-___________________,,

-------~.;: - -_..... . _-. . .;:ill

TJt ._

I

1

J l Jlf,

·J j ..

t . .••, ,.••• ,• .*'''•Mr ,pM,; .· P'" , V r' • * -

~1

.J

,-

•~1-w™~ --- ,,,.µt--• 1 ...u:wz--._. : z~~

J JJJ J _.

• • • :--

-

...

a

LY

a

a

3 kl

5il&F

8

' §

I 1P

I

# ...!~:-~~--~,~~~c!tl~;j(•~ - ~lcl~"~-~~-~~·~1~__;~~1:__~~~-------·-------' -------':.-------, & • m' ,,,,••,.. .. .. . ,_.. - . ..... ... - •. m" m• -

#i

I

'

_r, r-•--~ ' ..__ _"""' 1 ~

J

2 \

;i

~

;_

-

J - - - - • ~ ._ •

~

....

_

~·

Ofk.1t;SlQ1t

.f • '

't '"~t11.•0Y1 tr'' ,,, cVY , -4·Wt0J:-~;Kat ·•-t~XA\'tilf1«• 1e~rttt~_!:S~''" ,·· 'ft 13 ··v , ••t!!· ~, t.'"' __ ,, Mfr· • t .t\m,¥ (';..... ¥ I ' ;er v -. ., ,~ s1's 1 ' " )( 1 ' >' T....... , -\;,\,_;:i)m11(..._ ......~---.1;

j.

l""__

.:._,___ __.....:==- =- :.,_--Ji ' 2 ' -

.....:'<~ -

,

1

''

=7 5_

{J I

I

\

y

zj _' '

2

~

, 8' -. ¥ )

--

..,

rt

"- -

l

•

2 I

, _

level, measure 1 is ir1itially heard as accented, bur is u11d-e rstood as being weak "vhen grouped in retrospect '"'ith the seco11d .measure. T'hac is, botl1 measures function as anacn1stic events leading to measures 3 and + 1nus the organrt.ation of the 'vhole antecedent phrase is~ as indicate·d over E.'i:ample 17, t11-m:1~n or, i11 eightl\-11ote values, 6+ 6+ xz . Such patterns, wl1ose pr·oportioos are 1 + I +z,. are ca.lied bar-for111s. Notic·e ·t hat measur·es 3 and 4 are


EXPLAINING l\!USIC

a mi1tlarure version of this lru1d o.f structure. The first crocl'\ee of rn·easure 3 whi.cl1 is on. A, is inlti:ally hearcl as accer1recl; ic con1bi11es ~rith the follo,ving trochee on B, and bot1h .ft1nction as a11acn1ses to me~sure 4 ~vh.ich is a single pivoted grot1p. In other words, the t\\10 halves <)f measl1re 3 (p ancl p') are to ineasure 4 (q) , as tl1e first t\''70 meast1res (m a11d m') ar e ro tl1e third and fourth measures (n) . Nleasures 3 and 4 are a bar-form with tl1e proportio1is 3+3 + 6. If this elegant parallelism of s·trl1cture is to be projected in perforn1ance, the prevailing rhychn1 n1ust be trochaic. If it is iambic, ineasure 4 \:vill fail to f.orm a single pivoted grot1p, and the bar-fo r111 it1 1ne:asl1res 3 and 4 vviJt be obsc11red. Tl1is is what happens \vhen the Tl1en1e is played as Kohler and Ruthardt h~lve phras·ed it. Tl1e l1ierarcl1ic organization of tl1e phrase, as \.\tell as its formal pai1111elis1n, is destroyed-as eiU1 be seen by cor111Jaring a diagram of Mozart's groupings m(6) - m'(6) ~ n(12) p ( 3}-p' ( 3) -q ( 6) with the aberrant one (Exarnple 1 8}:

+

f

3

&EU

+3+3

;d

+

E.x.nmple 18

Like the antecedent pl'rrase,. the coriseque11t is also a bar-forn1, 111-n 1'-n', on level 3. But no miniature ba:r-forrn arises in measures 7 and 8,. ~lS \Vas th:e case i11 1neasures J ai1d 4, because at tl1is point tl1e pre,vale11t o·ocllaic rl1ytlun is decisively broken, and the upbeat grouping, potential in the preceding m.easures., is acruali:z.:ed. 2.

TI1·e influence of tempo and d.y 11amics must be considered in any disctissio11 of perforrnance. T err1po is ge11erally tl1ougl1c to be i1111Jortat1t because it alfecrs tl1e cl1aracter a11d nlood of the music. A fas"t tempo is gay or energetic, a slow one i.c; sad or cor1te111plativc. Tl'1e perforn1er prest1ma.bly ser1ses the etho·s of a co:t11posirioo a11d then decides upon an appropriat e tempo. Though this seems circt1lar-for cl1aractet is n1ade depe11den,t upo·11 tempo and choice of ten11)0 u.pon i111agined cl1aracter- there is no: do11bt


CRI' rICAL ANALYSIS AND PERFORJ\1.ANCE

so·meching to it. But tempo and dyna.rnics (including foregrot1nd accentuation) are e\
greater the tende11cy to emphasize metric accencs-rl1at is, to play downbea.ts lol1der tl1an \>Veak beats. And stress on ac.cents tends to create unambiguously beginn.ing-accented rh,ytl1111ic groups. On the other hand, if the tempo is too slo~', the \veak cigJ1tl1-notes Vlill alt11ost certairtly be perc.eived as upbeats. This is so because, thougl1 grou.ping is mostly. . a matter of relative durari.on, absolute dtiracio11 ca11 aJs-0 be influential. As the absolute amot1r1t of t1me bet,veen the sou11ding .o f an accented quan:er-11ote and ~n1 unaccented eighcl~ 1

note increases~ t he tenden(..,7 of the eighth-note to function as an anacrusis grows. It is possible, 110,,rever, to play tl-1e theme quite slo·\\rl)' , yet still phrase it correctly. To do so, great care nrust he talceI1 in ·t he articulation of dynamics. Tl1e weak beats n1ust recei,re 110 stress. For just as the sforza1zdo on the D in measure 7 emphasized its upbeat cltaracter, so even a slight stress 011 an equivocal weak beat ·will ten.d to n1al
•

Example 19


EXPLAL.~lNG

There is also a danger, if the te111po is too slo'\\r, cl1at

MUSlC

order to project rhe correct rl1y~l1mic groups \r1Jdt1e stress "\ivill be placed. ot1 tl1e begi11r1i11g of each measu~re, witl1 tl1e ,result th.at the rhe111e 'vill see1n too obvious and assertive. In short, tl1ougl1 ~ifozart s Therr1e is synraccically sirnple ar1ci strucrurally clear, a good 1>erforn1ance reqt1ires great tec~lniC'al cor1trol a11d i11ter1>retative finesse. Tempo and dyr1amics 1nust so com1)leme11t one a11other that a delicate ar11bigtJit)r of rhyrhm is felt by tl1e lister1er. l\11d all tl1is is lost, if the i11

1

phrasing in the Peters edition is follo1'ved. If a slovv 'te1npo te11ds to prodt1ce iarnbic groupings, tl1en ~1 performer pl1rasing as in the Peters edition might play the Theme n)ore slo,vly rha11 one; follo\,1ing l\i(oz:~lrt's n1arki11gs. Tl1e hypotllesis is diffic·ult to .confirm directly because the.re is no 'vay of lcnowing \vhich piaJusts, if any., use this edition i11 performa11ce. Bt1t tl1e 11ypotl1e.sis ca11 be checl
. .-

.

.J

'

..

.c:,c:

-

•

,. ' """"° -

~

•

'

~

'

.,.. '

,~ .

•

~ .

I

J· -

:::::> ,•

'

.. I •

· ::::::-·

r·~--1-

~

I

·•

.

l --

-r.

•

..

I

t'f

'

, -

·)

-

•

......:: - ;

,

,_

-

-

.

...::::. '

l

[-f::. u.

~-~

_,_.. ....,• .

-:,.

•

~.·

~ ••

-

'

:

...

-

~

*-. ~

.

•

-

'

-i

.

,.. J

.b4t

I

-

:s

---..·--

.0

-

-.

• '

.

'.,

:•

.-<::::'..

!hf'· ~'f

Example 20 Indeed, tl1e Pe'ters edition i1Luasing may be talcen fror11 Reger a rhought \vrucl1 suggests the desirability of usitl.g' primary sources. The only recordu1g of this \vork I 11ave been able to find is by Edt1ard van Beinun1 condueting the Amsterdam Concertgebou"·~l Or.chestra. Despite the 111etr·o nome marking in the score, rhe perforn1ar1ce tempo i., ;. = 28! Loolring at Reger's f)hrasir1g, I lvondcred -vvh)'" an excelle11t 111usicia.nan.d Ileger· was that, \Vhate,1er you n1ay tl1in1{ of his compositions----sl1ould have made wha.t I considered to be an unn1t1sica.l mistake. Or u as T perhaps 1 "'rrong? T hen I had one of tl1ose happy ''insp.irations ' for 'vhich one th.a.nl

CRITICAL ANAL1tSIS

AL~D

PERFORMANCE

43

Riemarm, had a. theor y· that all music vlas essentially anacrustic--e,ren tholxgh

the upbeats n1ight be suppressed. I checked for a {)l)SSible connection, a11d there in old, reliable Groive's Dic·rio1za1•::;f, I learned th~it: ''. . . in 1890 young Reger 'vent to [Rie1narul] as a pi1pil, fo llowing 'him the next year to Wiesbaden and soon becomi11g a teacher in the same conservatoire as his master!, s But the gifts of Goddesses gen.er·aily have tl1eir price-tl1eir onco1nfo.rtable side-as this did for m'e.. For tl1e m,oral of tlie story of R eger's aberrant phrasing ' vould seem to be: ''Cultivate a taste for speculative theory, but season it \V.ith a sou~on of saline sl

CHAPTER III

...._..--Con ormant Relationships ----·~·--------.....,..-·· --

1.

By conforma11t relationsl'lips I mean simply those in which one (more or less) identifiable, discrere musical event is relatecl. to anot l1er such ei.renr by sin:1ilarity. Consider, for instance, tl1e follovving f(>lli; tut1e (Example 2 1).

• •• ~\a_. -*•'

. _- *' a

JU!J

't,.. -' f f .,..,,

-

\a; ~ -===--....!""'~-,

a, ~

•

j --

\I I

"JM :' ~ [,,.

§

'

'

'

Exa1np1e 21 The first measure forms a clearly ide11tifia'ble n1otive or e.ve11t (rn), 't\1hich is t"l1ytl1mically" closed and inter\raJlicaJly coher ent.1 The second meaS\1re (m') repeats this 1noci...,Te at a lovver pitcl1 level. T l1e conforma11t relat:io11ship be~1een the e\.re11ts is u11111istal{able becat1se t \VO of rl1e prim.ar)r patter11f or ming para1neters-pircJ1 and rltytltn1- form a ~ucc~'Sion of similar eve11ts a11cl because fe\V ott1er paran1eters (for instance, timbre, dynat11ics, register, etc.) are varied. i\iieasure 4 ( tiY') is related to tn easure t, but less patent])' so. For i11 this case rhythm conforms, bt1t pitch does not- except for the f iil Ct that tl1e first 11ote, G, is the same. The last t\Vo quarter-notes of 1 1 hst is. both r1ores ca11 be u11dersroo,d as belonging to a single harn1on}' ·


CONFORM.ANT RELATIONSHIPS

45

n1easure 7 and the first beats of n1easure 8 are also related to the opening motive by conformance (m''') .. This time, though pitch relationships are identical,. rhythm is \raried. For e'\ren though the durational relationships are the s:nne- JJ J-t.he quart.er-11ore.s no'v have an ariacrustic function,. as indicated in the analysis under tl1e example. T hat measures 5--8 are a varied repetition of me~sures r- 4, and .m:easures r 3- 16 an ~~act repetition of measures 5- 8, is '\vorth 11ocing because it calls att:ention to two facrs of some importance. First; it indicates. that, like many other musical relationships, conforn'lant ones are .oftet1 hierarchically structured. Tl1at is, not 01uy n1ay lo,v-le\rel ever1tS like motives be related through conformance, b11t the1nes a11d sections may be so too, Indeed, strophic forms, such as a theme and variations, depe11d for their coherence primarily upon tf1e per:eeption of C·onfom1ant relatiotiships. Secondly, it should be observed that though measures 13- 16 are an cxaccio11 repecitio11 of 5-8 frcJ1rl a s.trx,1,ctural po·int of view, they are quite different from a functional poi1zt of v-iew. Because they come after a less stable section (111arked B in ri1e exa111ple), tl1ey 11ow function to pr-01note closure- they constitute an arrival and at the Sarne time a return. This change can easily be tested b)' playir1g O\rer tl1e first half of t11e tune. No matter how often it is played it \vill 11ot end-even 'thoug'h the notes of A' are literally the same as those of A''- unless the 1niddle part of th.e rune precedes A''. Not a.II conforn1anr relationhips are 11s obvious as tl1ose tl1us far 11-ientioned. For instance, measures 10 and. 12 of the folk tut1e 1night be analyzed as varied retrogrades of measur,es 2 and I respectively, ~~th the interval of a tl1ird filled in. But tl1e .relationsllip, if percei,,ed at all, is subli1ninal: we are not really a'\V"d.re of it. Tl1eir latent si111ila.ri'ty is tnas1

.EX_pLAlNlNG l\>1USIC

A.

Exa111ple

21

r1ot on ly~

l'1as rhe functio.n of the retrogrades (m·r atld 1nll") changed---tl1ey are no"'' part of th.e relatively stable portion of the

Notice, 110,ve,rer, that

tu11e-bu.t so have the mor·phological lengtl1s out of "vl1ich the rune is constr11cted. For though it can be Sl1bclivided, the first cott1plete e\rent is no"'" nvo measures long-and so, .o f course, is tile second eve11t. This change in 1norphological length is sl10,vn by the fact that in compariso11. \vith. rl1e first four n1easures of Exan1ple 2 r, these four meas tires s·eem incomplete-not even a pluase. More music is required: as illustrated ir1 part B of Example 21, ttt least fottr n1easures are needed for the phrase to reach closure. Our a'\va.re11ess of co1uormance depends tIDt 011ly upon the i)roxmnty bet\veen tl1e varia1Jt a11d its n1odel, but also, as suggested earlier, upon tl1e d.egree of sinulacigr betwe·en theni. That is, ·t:he more all the pararnerers arc dt1plicated in model a:nd varianti the stronger the conforn1a11t relatio11Ship. This .is specially the case '\Vith the prirnary pattern-forn1i11g f>arameters of pitcl1, dr1racion, and ha.ikmony. For instance, a motive can be changed in register, d_ynamics, ten11:>0, and in.)t:ru111entatiot1 a11d still be recognizabl)r th.e same. But if rirne a.11d. pircl1 relatiOt'I?hips a.re sigr1ificantly altered., eicl1er se1)arately or together, conformant relationships tend ro be n1asked. To tal{e an unco1np1icated exan1ple: tl1ougl1 ti1·e final tTto'(rerrient of •.1ozart's ,Clarinet Qt1inret (.Exrunple 23B) begin.c;' ith rile same n1elodic {2 3B1 ) and e\ren rl1yrhmic structure as the folk tune \Ve h.ave l)eerJ discussing, the conforn1a11t relacio115hips between measures 1 and 2 of the ~lozart worlc are less patent than

I

'

i . '-'--==:.,_ ,_,__~!'l('t..-.--''

•


C-ONFORZlt1ANT llELATIONSHlPS

47

those of the .folk tune simpl}' because the first n vo beats .o f the second measure are varied. Or perhaps not so ''simply.'' For the morion to the F' not only somewhat disguises the n1otivic repetition, b1rt it creates implications not pr·esent in the basic descending step-motion com.n1on to both tunes. As indicated in part A of Example 2 3, the motion from D to F# as part .o f a triad implies tl'le possibility of the .high ;..A,.. as a goal. 2 At least parrly for tlus re:ason., the middle p-m of 1Vlozart's Tben1e (Ex. 19B) ex1)loits tl1e upper pe:rfect founl1 of the A-octave (en1phasiilng G#)-as '\Vell as the lower perfect fiftl1-and the melody closes on a high A in n1easure 16, while tl1e folk t11ne is bou11d, so to speak1 ro the lower fiftl1. Had tl1e 111ocivic re1)eticion been disguised as sho\vn in Exarnple 23C, we \Vould ha,,e bee11 more aware of t11e confor1nru1t relationship between the first t'\VO measures-not only becat1se 1n.otive and var·ia11t are more .alike, but also bec..--ause no alternative implications \VOt1.ld have dr·awn our atte11rion a'\vay f ron1 the similarity benvee11 tl1e r11easures. And had the motive been varie·d as in 2 3C, tl1e higl1 A \vhicl1 closes rl-1e first part of the melody would have been less. ' ca!Jed for'' a11d .migl1t ha:\re come as a sligl1t surprise. Often secondary parameters~ SllCh as dynamics, register, and tirr1bre~ help ·to disguise w.hat might other'vvise ha\1 e be·en a rather obvious confor~ mant relationship. This occm·s, for u1stance, at the beginning of the exposition section cJf the :first moven1ent of I-Iayd11 s ''Surpriseti Syinphonjr. The melodic-temporal pattern O·f ·measure 20 is con.tinued in some'\vbat \raried form in measures 1 ·1 and 22. This. can be seen fron1 tl1e anal)'sls over Example 24. But this similarity is virtu.all}r masked by the abrupt change in dynamics, o.rcheStntion, regisrert ar1d low-level ornamental motion (i.e., the sixte·enth-notes), and, most important of ill, tJ1e ct1a11ge in rhythmic • groupmg. 4

1

Fo.r a disc-ussion of intplication~ ee Part II of this book. 111 the a1lal)rses, an arrow is used as a sign of implication. \ ¥ }1e11 the irnpJied event is separated fro111 the generating pattern. the arrow is brols:en: generating patte.r11 > goal. 2


EXPLAL"{l NG MUSIC

One of the reasons f o.r this c'maskjng effect>' is, l suspect, t hat the similarit)' between events is not merely one o·f syntax and fo1m, bt1t one of character·o f ethos....~d in this example, the lyrical-playful character of the first measures is so different from the assertive cl1aracter ·of later r11e~sures that \-ve are only partially a'vare of their n1ori\1ic similarity. In addition to temporal proximity a.nd si11UJarity of structure and character, our perception of conformanr relationships depends upo11 t he in.divi.duality of the model event. O ther things being equal, the inore striking its 111elodic, rhythmic, or harmonic profile, the n1.o re e~sily it v\ri.ll be recog11ized when varied and in diverse ,co11texts. T he phrase ''other clungs being e.qual,, mustt ho\.vever, be uiker1 seriotlSly. Fl)r we cai1 I}erceive conforn1a11t rela,t ionships only if \V·e ca.n reo1en1ber the model to wluch sttbsequenr e'vents are t o be related. And to a cons:ider~ble extent 111e1nory rnt1sical, linguistic, or other-depends upon the presence of a. constant and co11Sistent syntax and of 1·elacively regular l)atterns vvhicl1 often become the basis for arcl1etypal schen1a.ra. The linear descer1ding t11otior1 of falling thirds tl1e t11otion w hich begins both the folk nine and i\1101,.an's theme-is l)oth simple and regular. But in neither of these rnelodi~~ is tlle scl1e111a give1'1 a particularly individual profile. E.ssencially the same sche111a, though in the nlinor l11ode, is useti as the motto "'· hic}1 opens Beethoven's Fifth Sympho11y. There. howe,rer, tb.e famot1s upbeat rl1ythn1 gives t he pattern a very strilcing idet1ticy (Exan1ple 25A ).

.. /&""' .· ·..""""""' -

1'1. · '

~'

.

' t :

- . /411cyo

rr_• _ wd'

-,

-

_ ---

.,.__ __, -

'

~m'~ - - ~~~m Example 25

As the mlISic unfolds, \\1 e hear the n1otto t1n.dergo a myriad of trarlSforxnacions. It acts as tl1e basis for t11e rema.ir1der of rhe first ther11e group (Exa1nple 15B), leads to the secood k ey area \:vl1ich. it an11ou11ce.~ (Example


CONFOR1vIANT RELATIONSHIPS '

49

'

15C), and then acts as an acco,mpaniment motive for the lyrical portion of the second theme. In rhe first mo,rement it is ubiq,wr.ous: in the rl1eme groups, the develop111ent secrio11, and the codas. The special rhythm and repeated notes of the motto ar:e so vivid that there is no doubt about its recu.r rence in the Scherzo (Exarr1ple 25D )-tl1ough Beethove11 is careful to mark the return by prece.ding it \vitl1 a ritarda1ido, a.nd a ferniata . Nor is there a.ny difficulty in recognizing the reperirion of the Scherzo-version of the motto when it occurs just before the recapitulation in the Finale of the Sympl1ony. The preceding disct:LSb'ion might be sun,marized in a formula such as: regularity' of

Strength of pe,rceiv"!d conformance

=

individuality similarity of pattet""n (sc11e1nata.) · of profile · patterning_ variety of i11tervening temporal distance events

• bet\veen eventS

In other \vords., the gre-ater .the variety of intervening events and. the greater the separation in time between t\VO comparable events, ·rhe more pat,ent the shape of the model must be if ~i conforn1ant relatio11Sl1ip is to be perceived. Or, to put the matter the 0th.er "''a)r around: the more regular and in.dividua] tl1e patt er11 (and, of course, t he n1ore alilce ,events are in interval, rhythm, etc.); the greater can be the ten1poral separation bet'\veen model and variar1t and the greater the \rariery of interve11i11g 1nocives, ' vith the conformant relationship still recognizable, Perhaps it is partly for this reaso,n that, as symphonic movements tend to gro\V in scoi)e and length du.ring the 1nneteenth cenn1ry, tl1ere is a correlatit.re tende11cy for con1posers to inve.ot motive.5 and melodies tv·hlch f ulfill these conditions. As we listen to c:l1e :first 01ove1n.e nr of Beetl:i:oven's Fiftli: Symphony, the varied repetitions--the 511ccessive modifications of th·e motto are heard, generally speaking, as a more or less co11ci11n;ous proc,ess of transformarion. 'W hen the varied version of the motto oc<..'Urs in tl1.e Scherto .a nd the Final e, however, it is heard not as having '

n"PLAINING MUSIC

50

·their most signlfic·ant function-and i11 works of any con1plexiry it is an iinport~lnt one. Thoi1gh ret~urn contribt1tes co the impression of unity 'vhen it er11phasizes closure, neitl1er t1nity n.or closure depet1d upo11 the f)resence of 1·eturn. vVere the rerurn of n1elodic patter1ung a necess~r}r co11dition fo.r l.tnity and closttte, tliere could be

i10

rn'o-part melodies st1ch ~s the one

show·n in Ex:a111ple 2 6 8

. Is

'

'

'Y

.

not to mention co·uncless Bach chorales, many classic1l rl1e111es, and ''God rounded bi11ary structures (~A:: BA: ) lil{e the foll< tune in Ex.arnple 21, .return is a neces.sa,ry cot1ditio11 for closlu·e, bt1t TlOt a

Save tl1e Queen.''

I11

sufficient one. Thar ir is nor a sufficient condition can be seen by repeati11g phrase A. instead of J,Aa.', after the B section of the tt1ne. Tl1at return i~ a necessary condition for clo~me in such tu.nes is indicated by tl1e fact that, despite their similarities, it \\rill not do to taclc t}1e final phrase of the Mozart tl1en1e o·n to tl1e B sectio11 of the folk tu11e -as is done iI1 Example 27.

'

'

.

,

Ex.arnple :c 7

The result is incongruous partlyr because, a.5 11oted in. co1mection ''rirh the discussio11 of tb.e f\4oza.rt, that t l1eme 1'1as implications '\vl1ich shot1ld have been actualized in tlte B porcio11. Partly. l1o"''ever, the feeling of distortion arises because experience leads tlS to expect: tl1~1t the relative u1stability of tl1e B phrase \Vill be follo\ved by ti1e repose of tt1e fa111ili~1r. We aie prepared for redu11d~1nc)", r1ot for 110\1elty. It .is as though \ •V e took a trip a11di on rett1rning home, discovered. tl1ar a l1ew house has been bt1ilt vvl1ere Ottrs had fo1·rr1erly stood. R.epetition, as disti11guisl1ed f ro..i:n return, gives rise botl1 to forn1al a11d co processive confor111a.nt: relacior1slups. It does so becallSe of the natt1re of luerarchic srrucntres. To see ho"v repetition vvorlcs, let us begin "'Tith ·\.v ha.r


CONFORl\Ui~

RELATl ONSHlPS

51

at fir.st seems to be a rather str1111ge iror1y: the mo.re alike rn·o successive

eve11ts a.re, the more separate-dL~ti11ctl)' articulated-each appears to be. The .first two measures of Mozart;s then1e are related, 1tS 've have seen, by co11for111ar1ce (Example zSA) . Th.eir articulatio11 as discr:ete entities is clear, but it would have been more ·marked lud tl1e seco11d beei1 an exact: repetition of the :first (Exan1ple 2.SB). The re..11Son for this difference is that in Mozart's theme the first nvo n1easures in,rol\re a mel
A

c.

D

:Ei3:e::li: J

' .. Wt•= w[' ,;.. ,,,,,• .Md ~I

-

7

\-- St

cl...

¥ -

SF

,

w

~ri ....

t \ _,,., ;~ . -, "-\,!111."' v" ;

-

I

-

-

,- -Yi •

,?....

~-"'

\.._;:r=-~'a!(.. _• -' so •·""'_Cl'._ ·· _ _..__..,_'i,..,,+

£.'\'.ample z 8 lower-le,-rer structure (Rxi1r11ples 28C anti. D ) . Now tl1e \\~eak beat

i11

't he

sequential version ( 2 8C, level 2) becon1es a latent up-beat-as indicated by the inverted brackets in the analysis. 'N o such latency is pr'esent, ho,vever, in E.Yample 28D: each n1easure is a distinct, separate e\rent. Three points n.eed to be niade in connection "''ith this part of our dic;cussion of repetition,. The first is that thot1gh repetitio11 n1ay be ex.act f,rom a for1nal i1oint of view, it is ne,1er so psychologically- for the obvious reaso·n that bei1lg a repetition in itSelf qualifies an.cl changes the event which is reiterated. And it is a curious fact tl1at i1nmediate repetition tends to emphasize the differences bet\ve,e n lil~e events, ""rhile remote repetition-th.a t is, retu.rn-te11ds to .call attention to their si111ilarities. The second point is that the more exact the repetition of a.11 e\1enr is, the n1ore strongly we e>..rpect change -we feel that further repetition is unlikely. Yet this statement clearly requires qualificarjon.; .a nd tins bri1igs us: to th.e ttli.rd point. T11e implicative eff'e ct of repetition -depends tl~mn conrext. For u1sta11ce, if a reiterated pat-


52

EXPLAINING MUSIC

ter11 is understood to be pa1·r of an ostin.aro or of a ground basst '"'e do not necessaril}r expect change. S.imil.arly~ repetition in a coda or of a cadential figure repeated as an echo, has quite a differ,enc e:ffect from repetition vvhic·h is understood to he part of sotnc on-going })rocess. An)' parameter-melody, rl1ythm, harmony, rext11re, timbre, or dynamics- n1ay co:11tribt1te to tl1e articulation of rnusical structure. In Exan1ple 24, for instance, an abrupt change in dynamics, tin1bre, register, and rhytl1m marks tl1e 'begiru1i11g of cl1e second F)art of rh·e first tl1e1ne group. And closurCr created b)r some co111bination of parameters, can articulate structure in the absence of repecitior1. Tl'le er1d of the conse<1t1ent pl1rase in the Mozart theme (Exampie 19) is vet)• clear, even though it is not foilo"ved by repetition~3 Ne,rertheless, one of the n1ost effective w~iys of empl1a.sizing that an event is ended, is to begin it again. Re:petitio11 ca.n arcict1late form.al relationships not orlly st1cl1 as that bet\veen tl1c antecede11t at1d conseqt1ent phrase in Exampl:e 29 but also on higher levels of structure.

F o-r instance, the O\rer-all strt1ctt1re of a theme a11d variatio1lS is as a rule decisivel}r articulated rl.ot only because the tl1eme and each of the variations are t1sually closed slLapes, but because tl1e element .o f repetition is so pronounce.d. A11d w!1en con1posers sot1ghtt as chey did d.uring the nit1etee11tl1 century, to make tl1eir variation n10\l'e.r11el1ts n1ore proccssi,1 e (at1d less ft>rmal) in character, tl1ey tended ro mask the existence of repetition. It is perhaps partJy because repetition serves as a cue that a preceding part ha..~ ended, that transitions from the first to tl1e second then1e grot1ps in n1ovements in sonata form often begi11 with a repetition of tl1e openit~g melodic rnaterial. Ha.ydn;s ''Clock'' S1f'mpho·n)? begins ~1 itt1 the tu11e quoted in Exarnple 30A. After sorr1e twenty-five nleasr1res, a strong cade11ce on tl1e domi11ant is follo"''ed by a slightlyr varied repetitio11 of the opening tune (Ex.an1ple 30B). T l1e repetition, together with tl1e }1.re.cedir1g clost1re a11d tl1e cliange 3 This suggests that, generally speaking., the more decisive the pa1"t!.01etric cloSltte,

the )e~-s the need for repetition

i1.1

the defin.itio11 of structu.re.


CONFORM.ANT RELATIONSHIPS

53

in orchestration, d;rnamics, and texture indicate tha:r a major section ·of the . . .1S completed·. an d a ne'W~ part is "beginrung. .. exposroon >

,.

(J)

•

:t

JI'

,, •

r._

.~ ~

•

Exa111ple 30

Observe that in this case Haydn does 11ot, as lie son1ecimes does~ l<:eep us in doubt about the fu11ction of the repetition. The new harn1onization (\T of II to n, ratl1er t:han I ro IV) 'iivith its emp.hasis on th.e minor mode and the n1elodic modincarion (C instead of D it1 111e
On tl1e hierarchic lev·el · where repetition is immediate, it tends to separate events. Bur on the ·next levei.- vi.lhere similar events are grouped together as part of some larger t1nit-r:epetition ter1ds to create cohere11ce. This is parcicttlarly the case when similar events are part of an in1plic-arive, goaldirecte
usually· t:he next higl1er 011es.6 Processi,re conform.ant relationships, Sl1ch as t hose descril1ed in connection '..vith the first movement of Beetl1ove11 s Fifrl1 ·symphon}r (pp. 48-49) also play a significa11t role in the perception of musical structure. B·y ''leading>' the listener's eat and 1nind througl1 a series of gradual modifications, 1

Sornetin1es Hayd11 will, however, use this S£)rt of sign in a deceptive '"1 ay-he '\vill sig1ial that l1e is. going to 1r1od.ulate, only to bring t:is back ro the tonic and the 4

opc:tling tl1eme. See, for instance, the Fin~e of t.he '4 Jvfilitat)'., S}"mphony, in '\>vhich a.t measi1re 10 a Shift to the submediant (VI) suggests that the transition is beginning, b:ut no real change of key occurs, and the first the.n1e is repeated at .measure 42. It would, I believ·e:, be ITIOSt re\varding to analybe the procedures of the ciassiw

period as a. ''syst.em of signs." 6 This aspect 0£ 111usical relationsl1ips: is discussed in Chapter lV.


EXPL1\1NING MUSIC

54

the successive tt~nsformation of a

contributes to-but does n-ot in itself create-a sense of cor1ci11uity and col1ere11ce. 0 11ce agai11 ar1 ir1teresting paradox is u1vo1,,,ed: the existet1ce of inoti ic constanC)' of · aried repei11o ti\1e

tition-allows t11e lisr,e·ner to arter1d to higher-level processes. Becau.t)e lo-xvJevel eve11ts are .relacivel}~ regtilar ar1d persiSt:e11t> tl1ey c.u1 be, if i1oc ig nored, ar Least 111ore or less taken for gr.anted, so that atte11tion can be directed to tl1e larger patterning of the n1usical str11ctt1re-ro long-range harn1oxuc, melodic, and rhytl'u11ic reiatio11sl1ips. 'f o use an analogy from visual experience: the relative consta11cy of size and shape of the sto11es it1 a· lnosaic facilitate Ollr perception of its o er-all ,pattern. Tlris perhaps explains in fJart \\' h)r bridge p:~~ages a11d developn1e11t secti-011s i11 so11aca-for1n rnoven1enrs tencl to be n1oti\rically stab]e: morivic constancy allo\VS th e liste11er to focus u pon and con1prehend tl1e har11lonic and textl:tral processes v\rhich are ce11r:ral. O ne car1 n1odul,1te \Viti1ot1t repeiting motives, bu.t tlien the n1ind has to cope ·\\rith much more simultaneous info·rmation ar1d attention tc11ds to be di,rerted fro1n the larger sm1ctures ordering musical events. T l-1e motivic const111cy of classical development sections .is 1zot, as some l1ave suggestedt t.he result of th.e Ct) n1poser's desire ro ''exl1aust" his materials-to prese11t au rl1e perrnutations a11d. con1binations l1e can discover. For t his is seldom done. (Any third-rate con1poser can invent trariancs on the first theme of, sn,y, Beethoven's Eighth Syn1phon}r, \Vhich Beethove11 never uses.) R~1ther mocivic co11srancy is used because it allo,~.rs for the perceptio11 of larger processes, an1.i becatise the use of pt1rt of a previously established i,,vl1ole (a n1oti,1e caken from a theme) is implicati~1 e of t.he .renirn of tl1e total patteri1. These obser,racions ca11 be state,d in more ge11eral f orrn: tl1e greater the am.ount of ctW1ge- in both rate tind degree-i11 one parat11eter, rt1,e sn1aller n1ust be the ch.anges in othe.r parameters if pattern.U1g is to be perceived. If all 1)arameters are varied sin1ulta11eously and indeper1dently of ori.e anotl1er, cl1e result is not necessarily a more complex and interesting patter~ bt1t of ten none at all-a confused t1odgepodge of sounds. T i1e amOltnt of silnultaneous variation possible also depends t1porl the nan1re of the patcerr1s themselves; tl1e more pacer1tly stn1ctured and arcl1etypal 011e aspect of a, f)1ltten1 (for instance, its melodic sl1ape) , the n-~ore ot.l1er para111eters (e.g .. rl1ytl11n, har1nonyt etc. ) c:t1n be varied \Vitl1ou.t destroying che impr~~ion ·Of cot1formance. Because rl1e ar11ot1nt of i:t1foro1acion '\>vhicl1 tl\e llu111an. rr1it1d can comprehend at one time is li111itecl, ri1e n1ore information <)ne para1ueter c'lrries


CONFORMAN"f ltELA'flONSHIPS

SS

th.e 1nore redundant others mu~'t be if music·al relationships are t<> l)e perceived. Tllis proposition evidently a1)plies to n1usical styles as "vell as ro indi\ridual ·Compositions. For instance, a highly cornplex ar1d su.btle 1nelodicrhythmic style, like that o·f the music of ·s outl1 I11dia, generally nu11i1nizes (or does witl1011t) co111plex luir1nonic pr·ocesses st1cl1 as ha,re characterized Western m·usic since the Re11aissanc·e. Even more modest style differences may be distinguishecl, in _part, it1 tern1s of which pa.ra1ncters tet1d to be varied m-0st. Compare,. for instai:1ce, the begi11nmg phr~(jes of two funeral marcl1es (Example 3 r ) . In the first, fron1 Beetho,ren's Third Syn1pho11y, ~i l1igh degree of melodic-rhythmic variety is coupled ~·itl1 :l1ar1non.ic .restra.int -01tly one change of ham1ony in fol1r me~ures~ T ;h e second phrase, fro·m Schun1am1's Piano Quintet in Eb, '\Vor1{s tl1e other· '\\1 a.y arour1d: 111ininl..'l.l n1elodic-rl1yth-

mic '1~ariety is accompanied by considerable harmonic change._seven changes ii1 four 1neasurcs.

I- . - - -

ii!!!!

...

•

-

..

·-

-

..

~·

911!:

-

•

-~

R-can1pl e 31

This discussion suggests one reaso11 wh;~ then1acic ttansforination be~ came an impt)rta.nt concern o.f nineteenth-century c<>mposers: Liszt, Schu .. nla11n, w~1gner, Brahms, Franck, and Strauss- to nam.e but a .few. l\t1otivic co11st'ancy "'as r1eces.sary if t!-1e expressive possibilities of instn1me11cal timbre, register·, and foreground harmonic color 'vere to tJ-e realized. On higher levels, an increase in the scope and 1·apidity of l1annonic change could be


EXPL.c\IN1NG MUSIC

effective only if n1elodic cl1a11ge empl1asized co11ti11uity ratl1er tl11tn contrast. Bur this is only part of tl'1e story~. As so often in hist:ot')r, a particular rrer1cl or pr.oclivity is tl1e resttlt ·of a conc:atenacion of causes. Another ret1Sot1 for the prevalence of tl1ematic transformation \Vas that the ideal of persona.!, disrincrive self-expression (coupled with a tend.e11c31 to \Vrite 1011ger rnoverr1ents) led to the invencio.n an.cl tise of individ.ualized and cl1aracteristic tl1emes and motives. Because the singularit}" of sucl1 materials would ·have been ioco11grt1ous \\rltl1 co11vel1tion.al, transitional a11d pass-age-\vork figures st1cl1 as scales a11d brol suggest 110\V pervasi\re these ideas vvere.7 Art, like natttre, is made tlp of grad2H1l trttnsitions, "''hicf1 li11k to.getl1er the remotest classe:s and tl1e mosr dissi111ilar species. (852, iralics 11une) i1an1re, Lr\ the huma11 soul, a11d it1 art, t he extren1es, opposites, and. high points a.re b-0un.ci (>ne to a11otl1er b)r a conti11uous series of various varieties of beiT1g. ( 5z )

111

Arr., lil{e nature, '\Veds related. or cot1rradicto11r .fo.rrns. (852) [A.rt] is impelled r:o'\vard an un predicted and u11predictable fintd goal i11 pe7pett-tnl ·t:ransfor?1-u1ti<>1zs. ( 854, ira.lics rru11e)

6The fact, rioted by Gerald Abraham, tfle.t nit1et:eentl\·Cet1tt1r y ·corr1posers. tinlilre eigl1tee·11tl1-century ones, '\ve.re literary n1en n11d ofren belonged co rl1e intellig~i1tsia is of signal importance for rhe hisrory of recent ' iVestern nrosic. See A Hif.11dred :I" ears of JW:usic (CE:Ucago: A1dine Pttblisfring, 1964), p. 2of.

1 TJ1e e..'l:ce.1,1)ts a.re taken from Olive:r Strunk, So'l1:ree Readings it1 1\t!11sic History (Ne'IV York: V\1• '1\!. Norton, r950) ; page nttrnber-s are given in paren tl1eses after ·each • quotatlon.


OONFORMANT RELATIO~SHTPS

57

The poetic solutio11 O·f i11strumental music contained in tbe progran1 seems co us ra.ther o.11e of the various s·teps for,,rard \v hich the art bas still to take, a necessary result of the developnzem of o·ur ti?ne. (859; italics mine)

The particular point or argument in\. ol''·ed in these statements is n-ot important. vVhat is crucial is that tl1e ge11eral conceptual frarne'\v.ork must at the. very least have influenced Liszt's u11conscions attin1de toward his art. An.d it is difficult, in vie\v of th:eir pre\ralence, ro doubt that tl1ese ideas affected other compos·ers also. 1

During tl1e nineteenth century, formal conf£,rn1ant relacionship~ ways vitaJ in the artict11a.tion of n1t1sical :form-beca1ne increasingly impor.m11t in the minds of com.posers. Here Beethoven1s infiu.ence \Vas strongly felt-particularly the exan1ple of tl1e Ninth Sympl1ony. Formal conforrr1a11ce was extended in order to relate n1ovementS to one another. Be.rlic>Z.'s invention of the £dee fixe, the cyclic pti.nciple employed by Francl<, D 'Indy, Fatu:e, and other French co·mposers, and the use of tl1emari.c reminiscen.c e in the inusic of Scl1uniann1 Brahms, and Brt1clcner are all instances of this tendency. Wagner's case .is some"''hat special ~ 'his use of processi\1 e conformance is obvious, «b ut tl1e ''·ren1r11'' of leitmotifs throngl1out the Rffl.g, tho11gh a.rcicularing strncrure to sorne extent, is also used to refer to ideas or characters in the narrative ·Of the operas. The preoccupatio11 ~ritl1 conformant rela.tionships- form.al as " 'ell as processive-contint1ed into the nve:ntieth cen,tury. It is clear in the '\-Vork of tonal and nontonal composers alike. Bartolc•s Sixth String Quartet may serve as an example. But the ultimate and ''logical'' consequence of tl1e concern wir.h c·on.formance is f ot111d in the twelve-tone method of composition, wh.ere the total pitcl1 structure is derived fro111 a si.t1gle twelve-tone .row. Initially, in the music of Schon.berg and Berg, the row e,rave rise to botl1 processi"~e and formal conformant relationships. But subsequently, in the music of Webern and his follo,vers, processive conforn1ance became less and less in1p-0rtant: empl-iasis was placed upon inte.rvallic reCl1rrence rather tl1:a11 1notivic resernl'>lar1ce. Formal conformance the nonproce.ssive ordering of rows, subsets of ro·'\\~s, an.d their presentacio11 in variot1s permutations and combinations-has become the chief con.c ern o.f recent serialism. That this mode of musical understat1din.g is essentiall)' formal, rather than proces* si''e, is ii1dic-ated 'by the analy•ses made by serial co1np<>sers of one anotl1ers' music. It tends to be in tern1s of row manipulation~ not in terms of function, implication, and sy11tactic strucmre that these \Vorks are discussed.


EXPLA!NlNG l\itfSIC

A nun1ber of reasons for the increased use l) f f<>rrnal conf·ormance benve.en 111ovements. suggest ther11sel\res. Fro.m a n:rusical-perceptual poir1t of view, as multimovement works (particularly sy111phonies) gr·e'\v in size and cornplexity; it became both increasingly importar1t a11d increasingly" difficult to reme111ber tl1e basic tl1en1atic ideas. Tl"Le return of a pattern, bet\1.reen moven1ents as well as vvithin them, both reduces the n11n1ber ·o.f different ideas ·to be ren1embered ar1d rei11forces tl1e rnemor),. of those ~ilready presented. In the area of cul.rural histor}r the conceptual n<>tions mencioned above we.re influential: the prevale11ce of ideas of in11er necessity.~ di,'llecticai conflict, and resolution, and tl1e germi11acio11 of a, large worlc from a single rnorivic celJ irl c,t lcind O·f miniatttre evoltition. St1cl1 gern1ination, it \Vas thought, would not onl) create audible col1'C;~r
These lectures are it\tendecl to shovv the path that has led to this music, and ro n1ak.e jr clear tf:1at it liaa to h(f!Ue this 11nt:t/.ral or.1-tca17'le. (3 J, ·italics nline)

:Ntuch later I discovered d1at all this tnent. ( 51., itaHcs mine)

\ V as

a pnrt of the necessary dervelop-

AU tl1ese fugues are based on one single th.eme, "vhicl1 is consta1ztly tra1'J.Sf omted (34. italics mit1e)

T o d·evelop everything else f1·on1 one prir1ciple idea! That's the stror1gest tinit)' . . . ( 35)

Co1nposers cried to create unity .in the acco1npac1.i1nec\t, t,o ·\i;1orl! rl1etn aricall)r, to derive every thir1g from or1e chi11g, and so to produce the ·tigl1test- 11iax.iu111-t111ity. Ar1d no"v ever ytlung is clerived f ron1 this chosen .succ·essior1, of t\velve 11ote~ and. thematic tech11ic.1\.1e "vorks as before, on tJliS basis. Bt1c- the great advaru;:age is tf1at I car1 treat the-

a Anto.r1 \:\'eb.er.o, T IJe Pa:tiJ to tiJe N e'll.1 .Nltuic (Bnrt1 Ma'''r, Pa. : Theodore ' Pl:iesse.r, 1963 }; page r1ur.nbers a.re give11 ii:1 parenthese.i; nfrer each <:Jtlot.•ition. ~


59

C'ONFOIUYIANT RELATIONSHIPS

matic teclu1ique n1uc}1 more freely. For it11it:l is co'.f11pletel.y the miderJy.i1tg series. (40, italics mi11e)

.by

2.

Recogn.izi11g the presence of confonnant relationslups in music, theo.rists becat11e intereste.d i11 the te,clmique and significance of thematic transforniacior1. Notable among tl1ese \vas Rudolph Reci, \vhose book. T/Je T/:>e'l11atic Process in ~~lusic,u seeks to demo11stra:te that in rhe worl{s of t11e ''masters1 ' ' all thematic ideas are deri\red from a si11gle gernlinal motive and, fu.rther, cliat the succession of transforn1acions is a process '\vlucl1 imparts mea11i11g and unit)r to Vl1l()le co111p-0sitions as "veil as to single n1ovementS. I shall discuss one of R eti's anal}1ses not only in order to i11dicate son1e of the pr,oblems .and pitfalls faciJ.1g this sore of er1tcrp·rise, but because this 'vill lead to a consideration of a nt1n1ber of ftrndamental iss1.1es-n1ethoclological1 theoretical, and pi1ilosophical. There is, I think,, no qu.escion that the 01)et1mg pluase of Brahms'.s Second Symphony (Exan1ple 32, r11easures 1- 5) presents n1otives which ~re co11cint1ally varied and ttansformed througl1out tl1e movet11ent-and are even used i11 other rnoven1ents of the 'vork. This is specially true of the neigllbor-note (D-CJ-D) figure of the first meast1re '\\1hich I shall call tlie

,,.. . !

.... .

.L

Example J.2 9

(r ew York: Macmillan, 1951 ) .


60

EXPLAlNING MUSIC 11

''n1otto. Anil 11ndoubtedly tl1e cornposer consciously v.rrought t l1e relationships. Reci, hO\\rever, '\vants to do more th~1n contend that Brahn1s used conformant relationships to create co,J1erence; ir is l1is conviction that every important theme in the first ffi()Ve111ent is derived f1·on1 the openi11g n1easures. Reti begins by atternpring t o perSl1ade tls tl1at tl1e second tune of tl1e first ke)r area (Example 32, measu.res 44f.) fu11criotlS as a litlk or comn1on ter111 uniting rhe opening phrases and the 1}rrical melod)' \Vhich begir1s the second group (Exan1ple J.6) . That rl1e opening n1easures lead to the second tune cannot be disputed. This is a11 t1nequivocal instance of processive confon11a11ce. \Ve. /,ear the trat1sform~1cior1 of the n.eighbor-11ote figure and .e<>11nect its varied staten1ent in n1easures 42- 43 1'vith the ne'v th.e1l1e (m. 44£.)

b.ecause of ten1poral proxi1nity and pitcl1 ic.iencity. H ·o wevet, doubts 'b egin to arise Vll1en Reti asl{s us to agree tl1at rneasures 50 and sr of the nevv n1ne (Example 33B, 1T1oti\1e c') are derive.cl fron1 the end of the first phrase of the openi11g the1ne (E.~an1p.le 33A, n1otive c). L eaving aside the fact that the first (c) phr.ase does 11o t end 011 D as R eti's an.alysis suggests, but continues to an A (see Exani.ple 31, 111easure 5 or its transposed equivalent in Exan1p1e 33C), the rhythmic-melodic· ft1nctions of tr1e tones are significantly different in each phiase.

·\ z

Ell

-1

&

,

Exa1tlple 33 (after Reti1 p. So)

In the ope·n ing phrase tl1e first D (r11easure 3) is a weak beat, w hile tl1e second (in 1neaStire 5) is accented; tl1e second tune j·uSt the opposite is r.he case.. A.n d the rh}rtl1mic position of the E is sitnilarl)r altered. l\1.orcover, in the opening phr-ase the Fjt: (meaSl1re 4} is a changing-note (rel~ited to the first rn.easure by inversion) "vitt1our tnarked mel<:>dJc direction,; but in the second • tune it is strongly directedt botl1 because ir is prece(le
u1

The second nine of t:l1e first l

61

CONFORMANT RELATIONSBIPS

cionship to the opening n1otto becomes eve:o more apparent (Example 34)~ forms the basis for t}le transitio11 to tl1e lyric 111elody \vhich begins the second lcejr area.. Again the processive ·COt1form.ant relationship is at1dibly clear.

t

,

"'

\

•

•

•

nn

I

I\

•• '

: I

'

Exa.mple 34

But it does not f ollovv f ro1n th.e processi\re cl1aracte.r of the transition th.at this lyrical melody (the second rl1en1e) is related to an;rtl1i11g wlucl1 has gorie before. Reti p.resencs an e.'l:ample wlucl1 ostensibly shows the relationship hernree11 the second theme an.d the second tune of the first gi:oup (Exampie 35):

8.

Exarnple 35 (after Reti, p. 8-0} and he comn1ents; {'This mt1st strilce t1s most forcibly . . . The composer states his firs.t theme, '\\rruc11 is follo·wed by an int·ermediary theme that reiterates the substance of the first. But tl1is interr11ediary tl1err1e is in tum, and at the same time, a foreteller of the seco11d rlteme . .. If we single qut certain

1zotes1 tl1Je first tbeTne co1rzes to tJ;e fore; if we single out others, the sec·ond the7ne appears.'' 10 10

Ibid., pp. 80-8 1.


61

EXPl.AINrNG MUSIC

R ather tl1an striking r11e ' '111osr f orcibl)' ," it strikes 11:1.e .as being sornewhat fo.rced. To begin wirl1, Reri (lid not pick the melod)' of the second theme which is played. by the celli (Exarr1p1e 36), but ratl1er the parallel

•

line pla;red by cl1e violas .a tl1ird belo':\' the celli. Tl1e reaso11 for this is, I suspect, rl1at llad he used the melody irself, no '11alf-sce11 rnocion (A-G~-A) could ha\re been extracted, \Vith tl1e result rc1at tl1e si'n1ilarit)r bet,;veen the begitu1:it1g of tl1e second tune arid that of the seconci tl1en1e \¥Ould l1a,re seemed eve11 111ore te11l1oi1s. Bt1t even as it stands, ti1e st1ggestion of similarity is t111cor1vit1cing 'becatise t11e tonal-melodic f:i.111crions of the co111pared pitches is so differ·e11t. Fo1· insrancei tl1e of tl1e seco11d tu11e (35B) is tl1e fiftl1 of a triad, "''t1ile ·chat of the se.cor1d therne is tl1e rhird- a·nci in this respect Reti wot11d have been better off with tl1e cello \rersion of tt'l·e cl1en1e, for it begins on tl1e fiftl1. The G~ 'vhich perforrr1s an in1porta11t rnelodic fu11ction ii1 tl1e second tune of the :first group (leadi11g the 111elodic lit1e dovvr1 tl1rot1gh Fl to E a.11d D ) is an t1ni1nportant passing tone in the versio11 ·of tl1e second theme presented by Reti. Fir1aily, as }1€ did \:\1itl1 rl1e er1d of the ~tc'' n1otive (see Exatnple 33), Reti sin11)ly le·aves ot1t i;;vharev.er seems incompatible v;rith his argurnent-fo·r exan1ple, thot1gh. the last r\ of tl1e second t11ne (35B) is a sta'ble structural to11e, the A at the e11'les co G,# and tlletl to G· and

F#.11

My resen ·ations about Reti's analyses a1·e both methodological and theoretical. ivlethodologically, if one can picl< and choose-selectir1g those ' 'oices or pitcl1es wllich St1pport one's i1ypothesis, a11d disregru-di11g those \vhich do 11ot (the srnall 11otes i11 lleci 1s an~ilyses)-then almost ar1y rnelodyl' cru:1 be related to any ot her· i.;\1 h.etl1er \Vithir1 or 'bec\vee11 worl{S. In Tovey's \vords: •tof a.11 the pastimes of n1usical a11al)rsis, tl1e easiest is tl1e ide11rify1ng Tr1ere is, ho\.\7 ever, at1 e.xplicit the beg.i1mic1g o·f the second l
cor111eccion bet\voo11 th.e transition secrion and not on.e of rnain n1elodic idea.st btlt of accomin tl1e fi1"St and second. violi11s \vhich begir1s i11 accompanies the the111e in iTieasures 82 ac1d 83


OONFORMANT RELATIONSHIPS

of melodic :figures. An t1ncontrolled in'"1ag.ination- that is to say, an uniinagin.~1tive 111ind
The prolongations are followed b·y a slcip of a fourth, after '\vhich both m.elodies descend in step\vise fashio.n -thot1gll, as noted earlier, tlle)r l1a\1e differe11t }}()ints of closure. \\That is needed is a set of rules, l10\;\1ever informal, for distin.guishlng structural from ornamental tones in an. objective and consistent \\ ay. Every critic \Vho \\. ants ro illumir1ate tl1e larger struc,rure of a con1position-the rela.tionsh.ip and interacrio11 a11101ig foreground., niiddlegrot1nti, and bacl<1

1

12 Donald

F . Toveyl .i\ll1tsica:l T ext-tlres ( Lo11don: Oxford University Press, I 941 ),

P· 48.


EXPLAINING MUSlC

ground- 111ust face this problem. lt co11fro11ts tl1e follo·wcrs of Schenker, and "'rill pL1g11e n1e later ii1 this book. Bl1t eve11 if a reaso11able set .of r1.1les were devised, arbitrary selectio11 \\1ould not be con1pletely precluded . For, as argt1ed i11 Chaf)ter It a gttp 11ecessarily exists betv\1eet1 a11y generi1l set of rules a.nd their application to particular inStances. Consequently judgm.ent a11d self-criticisn1 111ust re111ai11 tl1e lllcimate controls. \~lithout t11en1, there is a strong te1nptation to allow favored h. potheses to influence a11al}rtiC c:hoices. This temptation is one to which, I thin\{, R.e ti frequently St1cct11nbs. I-I.e does so beci:iuse in llis vie'r111ing the the1.ues frorn one itie11·tical mt1sica.l st1bstance.,, 14 It is 11ot 111y purpose, at this poi11t, to dispute Reti,s co11cei)tion of the n11u1re ·o f mtlsical uniry. Tl'1at must \vait a bit. llather I ,~rant to suggesr that l1is positi-011 ' rinuall)r co11·1pels l1in1 to discover tt1e ki11ds of relationsl1ips 'he has h}rpothesized. For if the valL1e of a \llOrk depe11d, {as it does al111ost by definition) u_pot1 unity, and if unity in tur11 d·epet1ds tlpo11 the ''variation of one identical 111usical thought~') 10 tl1e11, if a11 acl{t10\¥ledg·e(t n1asterpiece is being ~

anal)~~ed,

the hy~iothesized t:i1e1nacic proce.'>s ll1ttst "villy -nilly be uncovered. An(f Reti seen1s to feel t hat tl1e 111ore i.11sta11ces o·f transformatio1i, tl1e better (rriore unified) the col11})osirion. R eti's positio11 leacls hin1 to ernplo}r \lirl1at E. H. G·o111bricht criticizing tl1e writiligs ·o f culuual liistoria.ns, l1as called tl'1e ' 4exegeric'' 111ethod: ''. . . the r:tlethod, tl1at is) tl1at bases its interpretation 011 the detection of that kind .o f 'likeness: that leads the iJ1terpreter of the scriptures t<) link

t:he passage of the Je\\tS tl1rougl1 ct1e Red Sea with the Baptisn1 of Cl1rist. . , . The assu1npcio11 is al\:va;1s that so111e essential structural similarit)' tn\1St be detected \vl1ich per1nits tl1e interpreter to st1b·sume cl1e \1 ariol1s as1)ects of ct1lttire untier a single formula.'' 16 Ai1d just as tlie cttltltral historians who emplO}' this method te11d toward a kin,d of I'"'1'.egeiianism,11 so Reti see1111s becattse, like rhe rest ·o f us, he finds pleasure in discovering l1idden relationsliips-ir1 "solving the pt.1Z7Je." 1
Partly he succuinbs,, I

l1Sf)e<;t,

1


CONFORMANT RELATIONSBIPS

65

inclined toward a. dialectical conception of 1nusic. ot only does a tl1e111e move ''by transforr;zation toward 11 goat,>' 1 bu.t a process of thesis, antithesis, ancl S)rnthesis seems implied: . . . the real plan of this 'vhole becorr1es appare11t \\'heil the first Rhapsody ['Opus 79, by Brahms] is t1ot only cor11plemented but almost contt-adicted, and so ":resolved,'' b}r the sr.rucrure a.nd id.ea. of the ,. seco11d.1t) Th:us this Finale then1e is indeed a synthesis of all tl1e thematic i111pulses of tt1e :,ymphon)' ·20

As Gombrich indicates, an in1porta11t assi.1111ption underlying this method is that manifest differences n1ask ai1 underl)ring, but n1ore significant similarity. According to Reci, L'ln reality [ tl1e first and secor1d tl1e1nes of a so11ata-form mover1.1enr J are co11rrasting ot1 the surface bu.t ide11tical in sulJstai1ce. ' 21 This is the crucial loopl1ole! For there is r10 vv~1y of conftrting this a.:na.lytic 1n.e thod. If it is objected that a relationship is no-r at:-idible or that the notes selected as sigillfiea11t complet.ely cha11ge tl1e p:rocess an.d organizatio11 ·Of a tl1eme, the ansvver is: Of course! it is part o·f the co1nposer's pla11 to disguise similarities. \1Vl1en this occurs, cabalistic circwty has taken t11e place of seriot1s cricicism.22 But it 'vould be foll)' to let a. distaste for exegetic excess blind us ro the real significance o-f conformant relacionsliips. F<)r th.ey create foregrotmd coherence u1l1ile at tl1e san1e time allc>wi11g atte11tior1 to be clirected to higherlevel syntactic processes, contribute to the forn1al articulation of musical structure on all hierarchic levels, a11d 4lllso pro,r.ide the satisfression of .cl
•

Reti, Tl3e'lnatic Process of Ai!u.sic, p. r 39. 19 Ibid., P· i:45. ZO- /bid., p. i64. Tl1e \Vo.r k referred to is Bral1nlS' Second S)rmphony. 21 lbi~, P· 5. 18

22 Reti believes that tl1e sinlilariry bet,ve,er1 die cones he h.as selected as «strucroral" • and the rnodves fro111 " 'l1icb the)' are prcsm11abl)r derived <.."an be /Jel'l'rd. In n1any cases this m.ay be true, b:ectl:t.ise if selection is 1nore or less arbitrary, any 111otivc of some con1plexity can be made to resemble a model n1otive. l\lloroover, as has fr~c1uently been note,d, "believing is seeing,, (or in th.is case "hearing") . For illS:tanc·et a tile"\'\' familj• arriv·es in our comn1unity 1 and ""'e ren1ark upon the resemblance benveen paten.ts and ehildrea-only ro learn thnt tlte children are a

66

EXPLAINING M USI:C

S:lf)" condition fo1· n1usica.i understanding. And. in. tl1i5 sense at least t11ey appear to be 011e basis for n1usical t 111.ity. Tl1ey are not in 111y judgment ho\vever, a sufficient cause for u11ity. TI1ough this co11tention can be justifiedi it is diffi.ci1lt to prove in a rigorous \'~. tay. For unity is i1ot an objecci.,1e, subsrai1r.ive ·e11tit)' like a n1ocive, a perfect cadence, or a change ir1 texrure or d ynt1111jcs. Ratl1er it: is a psychological effe.c t-an impre.~ion of propriety, integrity, and cor11pleter1ess. A11d it is, I think, as difficult ro specif;r rl1e sources of this i1Trp1·ession in n1t1sic as it is to say \Vhat rr1akes a \\roman attractive. V\lhen it's there~ you lmow it. B11t the special con1binacio11 of ca\1ses that procluces the effect in. any particular case defies precise definitio.n. Neverthele.ss, it seems safe to say that all the kit1ds of relatio11sl1ips present i11 a composition- 11rocessive, tectonic, etheric (i.e. pertainir1g co ethos) as well as conforn1at1t ones contribute to t:l1e in1pressio11 of ·unity. The 1111portance of a particular kind of relatio11st1ip will vary not only fro1n style co style hllt from one ki11d. of \Vorlc to anotl1er. For insta:nce1 ethetic relationships are a necessary condition of tlnity in the mttsic of India, \Vl1ere mood or rasa is of central in1portance; in the rnusic of conten1porary serialism, on the .o ther l1and, co1tformant relationships are the IJrime; basis for C()mJJOsitio11al ir1tegrit)r. In tonal rnusic, processi,,e and l1ierarchic relationshiJ)S ai·e the most important in creatil1g a ser1se of unity in stn1crures such ~1s sonara-forrn movernent~ w hile confo.rr11ant relacionsJ1i11s pla)r a vital part i11 ur1ifyir1g strophic for ms Sltcl1 as tl1e1ne a11cl variations.. A11 i1npressio11 of unity thl1s d:ei>ends to son1e exte11t tipon rl1e ki11cl of relationstup wl1icl1 do1nit1ates the orcleri11g of a particular hierarchic level. Despite these qualificatio11.s, confor111a11t relationships are of secondary in1portar1ce for creating ti nity i11 tl'le repertor}' "vith \vl1ich borl1 R eti and I are concerned- that is, the re·pertory o f Y\' estern tonal n1t1sic fro111 1700 to 191 8. To begin i11 a son1e\vhtlt rot1ndabour \V~Y= tl1e great a.d vat1tage of coni.plex tecto1tlc structrtres, such as those de\1elo1)ed in V\ est:ern n1usic duri11g the period being considered, is tl1at patent cot1n·asrs and clifferences. on 011e level c

CONFORMA11.TT R.E LATIONSHlPS

. . . one of the salient ideals of ' I\!ester.11 (..'ltlture and a hall1nark of ''grean1ess;' in Western art, at least since the Renaissance, has bee11 that

of 11io1z1nnn1tality. To capture ru.1d corr1m,tulicate a ser1se of tl1e scope a.rid niagnitude of creaciot1-rhe variet)r and multiplicity· of tlill1gs, comp-0sers as well as a.rti.s:ts and '"''rite.rs 11.ave fotu1d it appr opriate to bring togetne:r a \:\.1 ealth of diverse materials, often placing tl1ese in sudden and viole11t juxtaposition, (One 11eed only think o.f a Bach Passion, a Beetl1ove11 sympl1or1y, or a play of Shakespeare.} 011e way of combitlli1g and uniting contrastir1g ideas i11to· a coherent '''l'lole, reconciling seenlingl}r in:c<>n1patible events, is to srtbsun1e cl1em wider some higher order-c·o embody• tl1em \Vitltln a l1ierarct1ic strucuire.23

Frotn tlus I'oint of view, Reci chose the '\Vrong 1·epertory to illustrate his tl1eo1)1 : he \\rould. have 'bee11 better off a,11alyzing recet1t serial rnusic or perllnps the music of ] a\1a. But even wl1en conforn1arlt relationsluf>S do pla) a prune role in c.re-ating a se11se of col1esion, similarity per se does not unite. For as we l1ave see11 the more nvo eventS ~tre alike, the n1ore ·che)7 t:e11d to be pe.rceiv.ed as discrete, sep:ara.red entities. .A .collection ()·f ide-r1tical burto11s possesses a high degree O·f tuilf·or1nity, but only an additive son of unity. Co1rforma11t relationships create the strongest inlpression of unity w hen they are en1bodied in some sort of fttncri.ot'lal process. Fron1 a :more mt1ndar1e and personal point of vie\~', I :find it difficult to bclie·,re that rhe kinds of correspondences, \Vhether obvious or disguised, w hicl1 Reti poi11ts to are the basis for unity or a cat1se of excellence in n1usic - for this sort of motivic manipulatio11 is so easy to do. Every third-rate composer of yesterday and todayr is adept at it. Having been one of them, I frtlly agree with T ovey that, ''Nothlng is easier tl1an to derive any musical id.ea whatever from. atl} other n1usical idea; and a long chait1 of such derivations is often sr1pposed tcJ cmbod}' the logic of 1nusic. In itself it can give us no se-curity t11at it Li) more logical t .h a11 a series of puns.'' 24 Tove)r s scat hing refere11ce to pttm is 11ot as negative as he perhaps took it to be. For it calls attention to an aspect of confon11ant relationships .n ot previously noticed : we take pleasur·e i:r1 tl1e act o.f perceiving trans• formation. Psyc1uc parsimony-seei11g a si11gle entity or id.ea do dot1ble dt1ty 1

7

23

Pre..~

Leonard B. l\lleyer, A·i 14Sic the Arts and Ideas (Chica,go: University of Chicago 19,67 )~ pp. 311- c3. A true believer in the exegetic method can still h.a ve the

final word-, arguing tllilt the contrasts an.d diversity are 11ot real. 24 Mr1sical Textures p. 50.


68

EXPLAL'OfING MUSIC

- i it1vol,red in the perceptio11 of coniorniant relariot1 tlips. J u t as \Ve delight in seeit1g a familiar landscape in 11e"\' lights-at different tin1e of day or ir1 different seasons, seeing ir as t he same and )' et different -and are e11chanred as a n1agician cl1a11ges a ha11dkerchief into a rabbit, so "\>\·e enjoy discovering that something fir r Ltndersrood as 011e lzind of e\?er1t can change its f u11crio11. F or i11sta.11ce, that son1ethir1g at first unlierstood as being a closi11g kintJ of cver1t ca11 be rransfor1.i1ed inro a begi1111ing kit1d. of event-as is dc>tle fro111 1n.east1re 19 co 20 in Hayd11's ''Sur11rise" Syrnpt1ony (Example 24) ~ind from r11easurcs 4l-43 to n1eastuc 44 i11 Brahn1s' Second S)rmpho11y ( E xa1111)le 3 i ) . For the same reasons, \\' e are c •hjlarated to f111d tl1at seen1ingly t1ncon11ccted a11d disparate then1es really '(fit togcther''-as in the Prel11de

to\ Vagncr's Die )l.J.eistersi11ger (Exan1plc 3 ) .

,_ /

I\

-·

•

~

t=l

~-

~

~-

-

-

~

,m. .. ·-

I

'.

•

:!J

J

~

j.) .

.

~

.......-

-

.

=:i

I

+

._

...

~

=t: '

1-

~

:i

~

:j'

'

@g)

I

-

I

I

• .

J

Example 38

The basis for rl1e u11iry of n-1ulcir11ovemer1t \\ 0rks is problematic. Thougl1 com})Osers are not al\vays a11 infallible gt1icie in critical a.11d acstl1ecic matters, tl1cjr \ i e\VS sl1ot1ld at least be considered seriously the Inore o "vhen there appears to have been a cultural cor1 ensus al>ot1r arrisric 111atcers . •~s F. E. Iurby l1as poi11ted out, composers a11d tl1eorisc of the eighreenrh a11d early 1

1

•

11ineteenth cc11t\1ries ofren r~1o ught of t1nit)1 i11 tern1s of a characteristic n1usical '' tyle. '' represe ntative of tl1is 'ie'''· l(irl)y quoces LJerer LichrenthaJ, ,,,ho in 1 26 '\'rote of the characreri tic S)'m~1l1o n)· as ''011e ''' hich proposes a 111usical picn1re; or a moral character, as it Distratto by Haydn, or an e e11c1 like Ln Cad1ltfl di F et01zte; or son1e phenon1e11011 i11 cl1e physical ,,,orld, as la T e111pesta, l'l11ce1idio, la Caccia, etc. To ,vf1i cl1 11elong tl1e y 111pl10·11ies pastorali, 111ilitari, etc.' ' 2G Not only do ·ma11y of tl1e v,1orlcs of I"iaydn and his co11ccm1)oraries belong to this class of co n1posirior1 bt1r so, Kirby argt1es, do q.t1ite a nu111bcr of Beethoven's \vorl•s: for instar1ce, rhe "Pachetiqt1e" ~n

Unpubli. l'l:cd r11anuscript: ''Beerl1ove11' Use of Cl1a racrcriscjc St)rles: A Con-

tribution to the P roblcrll of Unit}' i11 ri1e Large J:
Mate· al cor" dire tos autora s

CONFOR:l\i1AJ.'1T RELATIONSHIPS

Sonata_, tlte "Eroica1' Symphon;r, the ''Pastoral'> Syinpho11y, the ''.Les Adienx11 Sonat~ ai1d a nu.mber of the late '"'orks. Though a bias favoring the obvious obje.ccivity o·f a musical score has led mosr contem1Jorary critics and tl1eo:rist5 ro account for n1usic \Vholly in terms of pitches and durations, timbres and dyn.arnics, an explanation of u11ity in terms ·Of cl1aracteriscic ''sryles~ see1ns as conviricing as Reti,s ct1eory. The idea of unity fostered tluough cl'1aracterization is m1pported by historical evidence. A.n·d, equall)' important, sucl1 a ,riew is by no means as ''Sl1b.1ect1ve . ,, as 1n1g . h· tat first appear. E,ridence from n1any different culrures indicates that musi-c comprehension depends to a co11siderable exte11t t1po11 tl1e liste11er's k110~1ing a, traditional tonal synta.~ and a set of c·on\J'entional signs an.d schemata. (even tl1ough these n1ay be groun(ied in and lin1ited by the natt1re and capacit)r of the h.uman ear and mind). Ot1ce tl'iis is g.ranted and the actempt to explain music s·olely on tl1e basis of iruiate, universal responses is gi,ren t1p, the11 commo11 traditions and conventions ca11 be analyzed as objecti\•e aspects of a musil."al culture, just as irnages, figttres of speech, p.oetic genres, and dramatic conventions are in literature.26 As Kirby observes, \vhat is tneanr by a cl1aracteristic style ''is not necessarily a personal or subjective qua!jry, but rather somech:i11g objective ai1d co.a crete: for the particular expi-essi,re character is explicit and consists of a r1umber of distinctive elements that define it, a.m ong them particular n1t1sical f onn~ ke)rs, tempi, th.e use of certain liistrt1ments or combinations of instrUments, special melodic types, rhyth.mic patterns, d}rnar11ic qualities, and so on.' 2'1 The exist·ence of such conventions m·eans that there has been a continuing ttadirio11 of intisical representation in "''hich la:ter maru1ers of deli11eating a particular moral c:haracter, ~1. kind of e\1e11t,. an affective state, or son1e phenon1enon in the pl1ysical or mytl1ical world are based trpon at1d influenced by earlier ones. Just as there are l1istories of elega.ic poetry and of carpe diet.11 lyrics, of paintings of the At1n.i.:1nciation and of pastoraJ scenes, so tl1ere would seem to be histories of battle music, pastoral n1usic, Io, e music, lamentation rnusic, and so 011. In shon there is, or should be, a. field v.rhich 1

1

26 Here,t I beliC\'e,. is cl1e crucial it1istake of st udies S\l.Ch as De:I"yck Cooke's Tl'Je Lm1guage of Music (London; Oxford Unive.rsit;1 Press, 1962 ) and D-0nal
recognizes that the e-.rpressive characterizing power tl1ey .find in \IVest:ern music cannot be traced m i1ecessarv ~~natural''' sources alone, bu:c are to a g11reat extent a matter of •' learned convention. 21 "Beethoven's Use of Characteristic Styles.~,


70

EXPLAINING l\1.USIC

1night be called 111usict1l ico1zolog;1; ai1d, considering that art l1istory 11as to a co1isiderable eA.1ter1t been ;;\ inodel for rnusicoiog}' , it is strange that (to t'he best of nl)' kno\~ledge) little \vork i1as been done in this area. It ' :vou1d, l thi11k, be fascinating to study the history of the cl1aracterLmcs of, say, Haties (fury) a11usic or pastoral rnusic fro n1 the Re11aissance to the t\J\re11tietl1 centur)r, tracing comn1on features, describil1g changes of man11er tmd 111eans of representation and relatil1g these both to rl1e ttistory of n1t1sica.l style a~nd to t11e l1istory of ct1ltt1re ger1erall)r.

3· l.f the most in1portant fu11ctio11s of confo.rr11a11t relationships are creilring coherence a11d articulating structure, and if such sin1ilarity is n.o r the basis for 111usical u11ity, then \Vhat is d1e significance of motivic correspor1dence hetwee?i movements? For there is no dot1llt tl1at such correspon,den:ces exist-a11d ru·e specially commo11 in tl1e music of the ni11etee:n th cenniry. One explanatio11 is that tl1e}r pr<>vide ti1e pleasure of p~ycl1ic eco110·111y. lr1 ad.dition, the return i11 later i11ove.n1er1ts. to ideas presented i11 p1~evious 0 11es may on occasion have prograr11matic significance-as, for instance, in Beethoven's Ninth Sy.mphony or Berlioz>s S)rmphorue Fantastique. Of the other reaso.ns for increased emphasis 011 co nforn1aJ1t relationsl1ips during rl-1e nineteenth cent11ry ortly or1e need b-e me11tioned here: narnely, ·as the size a.t1d sco11e of inulci~movement ' orl
,

;

GI

_

{

. _ ..

_

_

_

_ p -

·

b'

Exan1ple 39


71

CONFORMANT R.ELATIONSHIPS

(Example 39). When he st1bsequently atten1pts to sl10"' that the, second theme of the Finale is a '
,

5

75

i

I

,_

p

i

.I .,.---•- -·_•v-wrr_'..-I,

a- '

'tW

' Vhy stop ""rith correspondences between moveme11ts? An aficionado of the ·e:.'tegetic n1ethod might easily go further. H e mighrt for instance, note that the changing-note n1otive so promi11ent in this Syin.phony occurs in ()tl1er ~ror1rs by Bral11ns, notably as an impo1·tt4nr eleme11t it1 th.e Finale themes of the First Symphony (Example 41A) , ·t he Third Symphony (41B) and the Fourth Symphon.y (41C) .

'

..... Example 41

At first,. this seems an absurd thing to do. Yet n.othing is farced-no to11es l1ave been igr1or:ed ir1 order to 1nake tl1e tl1emes similar. The confornlance is there for all to hear. And other instances of n1orivic correspo11denc.e between different works l)y a. single composer are not uncon1111on. In the last string quartets <:> f Beethovc11 a motive 111oving from leading tone to tonic and f ron1 the sL'7th 'iegree of the sc~1le iri n1inor to the fifth is t1sed as the main. them.e of the first moveme11t o·f Opus 131 (Example 42A), the opening of Opus I Ji (42B) and, in a sligl1tly varied form, as the i11trod11£1:ory statement of the fugue su.bject i11 Opu:s 133 (42C), trJ cite only the most

obvious cases.


72

EXP'L AlNING MUSIC

•

ff

Such confor111a11t .relacio;nslups are significant. But their significance. lies not in t he' area of critical anal7rsis, but ir1 t hat of St)rle a11alysis.. Tl1at is, the}r are relevant for the anal)rsis of a particular composer's idio111.- his special stylistic predilectio11s. Fron1 this point of \rl~t the theme of the Finale of Bral11ns' Second Syn1phor1y is r.ela.ced at least as much to the composer's get1eral idiorn llis preference for conjur1cr, lyrical tl1emes for finales - {tS it is to other 1notives i11 tl1e syrn1)hoi1y. In a si1nilar ¥Va)r, 011e could, I think, sl10'\V .the 111ait1 ideas of Bral1n1's fi1·sc 1110\reinents tend to be disjunet and ofte11 triadic. . One ca1i, of course, go ftirtl1er. For moti,ric similarities exist a·mong works by (lifferent composers. For exami1le, both the opening theme of the Finale of ~'Qt1i

Tcl1aiko~1sl{y•s

};"'iftl1 Sy111phoi1y (Exa111ple 43A) at1d tl1e aria,

sedes ad dextram patris, from Bach~s B-1\llinor J\.1Iass (43B) use motives lil'e the one "vhich begins Brah1ns' Second Sytnphon.)•· He.re, too, co11for111ant 11

Example 43 .relationslu ps are irr1portant for sryle anal)"S~ but on a higher level-that listeners to co1nprel1end and 1·enlember tb.e parcicttlar patterns in \vhicl1 rl1ey are actl1filized. These co11Siderations suggest: t11at a particular COffif)Oser's idiom is distinguisha.b}e from rhe style .of \:o.,h1ch it for111s a pa.r t because he te11ds to er11pJoy some possibilities available in a style with greater frequency tllan otf1ers.. 'Tl1t1s tl1otigh tl1e rnocivic patterns pr:esented in Example 42 '):li' ere ''stylistically a.v ailable'' to Beethove11 througl1ouc l'tis career (as tlte}' \'\'ete t o


C01\TFORMANT RELATIONSHIPS

73

other composers ·Of t he time), he. favored them specially in his late style

urluct1 they help to defi.ne. 111.e analysis of confo1·rn,a·nt relationships of ten leads to questions about the co111poser s in·tentio11s. Is the reserrtbla11ce bet-\veei1 the opening themes of t he first and last movements of Bral1ms' Second Symphony the result of tl1e .con1poser's explicit., coosciotts inte11tion, or st1ottld ir be ascribed to an unconscious use of commo11 St}'"liscic feattrres-or is it a chance similiuiry? Tl1e relevance ·Of tl'le artist's intention for criticism has received co11~ side.rable attention in the literature about aesthetics. And it has been argued 1

that even where the intention can be doct1meoted from relial)le outside

sources, its rele,rance for criticism is qt1estionable. 2S I do i1ot \\ ish to consider the qt1cstior1 of the relevai1ce of t11e composer's ir1tentio·n, bt1t rathe.r to, ask on '\Vha.t grounds base oux feelir1gs abot1t intentionality in the absence of 7

"'re

extrrunusica1 information. The excerpts in Example 44 are taken from the first move1ne11t of Beethoven's Piano Sonata in Eb Major, Opus 81a. The bra,ckered p-atterns are conformant in that eacl1 in\tolves a. pitch e.x-.change bet\>veen Ol1ter voices · or the exte11sion of such a11 exchange, as in parts A and B. A1thougl1 the conformant r ,elationships ben,veen A and B. and B at1d D look ''intentional," doubts arise a:bol1t the relationship ·of C to rhe orhers: H asntt the correspondenc.e been i1np<>sed, perhaps forced, it1 this case? "'' as Beethove11 a'\va.re of

the sllnilarity? W1th regard to the first question, tl1e answer is an t1r1eqtuvocal ''no." No exegetic ingenuity has been brought into play in order to show a corre-

spondence. No pitcl1es have been disregarded; nothing has· been inferred or imagined. The resemblar1ce is t11ere for aJl to see. That it is neitl1er striking nor obvious, is a question of the stre11gtl1 of rlie relatiomhip, not of its •

existence.

Tl1e ans"ver to ti1e seoo11d question is, 1 thjnl<, that in a borderline case such as this~ it is impossible tc> decide ,,rhether a relational ordering arose from the Sj"lltax of the style being employed or '\Vas consciotisly contrived. by the c.."Omposer. Since there is t10 51Jecific poir1t at \Vhich st)rlistic ordering ends and con1positional ordering begir1s they forn1 a cot1tllll1.UJn-it is impossible irtl principle to distinguish tl'lose conforrnant relationships attrib·u table to style from those -,.ve presume tl1at the con1pooer explicity dev:ised. Inten2a See \ ¥. l{. Wimsatt and N:I. C. Be~rdsley, ''The lnt)ention.'li Fallacy 1.11 reprinted in Philosopl:ry of AN 111:d Aesthetics, F. A. Tillnlan a,n d S. M. Caho. eds. (New York:·

Harper and Row,

i¢9)~

PP· 657-699.


EXP'LAINING MUSIC

74

+

•

r

·? · :.. ".If

,

.

""'

a

- lf

Exan1ple 44

tiona.lity is a fi1ncrjo·n of stylistic ilnprobability: tl1e r11ore i1nprobable it is tl1at a confo.r111ant relarionship could hav·e arisen for St}'listic r·easons alone, the i110.r e \Ve impute conscious i11te11tion to t he composer. For tl1is reason, tl1e 1011.ger con1parable eve11rs are, rl'ie rnore certain '"'e feel clu.r the simil,u·it)1 was intended:. And "ve do not cotlsider \Vh.a t is tnost 11orr11aJ:i\re i11 a style to ha.ve beer1 <'put t here'' by th.e composer. For ir1sta11ce, ·t e do not as a rule con.sider tl1e reset11blar1ces between perfect cadences as exan1ples of intention~tl co11fortnant relations}1i})S. TI'le similarit)' is attributed ro tl1e S)'Iltax of the style. But '¥ere suc~1 cadences to occur fro111 tirne to tir11e in a11 essentially serial "''orlc, vve \>\ oulcl probably conclude rl1at ·rhe cornpose.r explicitly related the111 to ot1e anoth.e r. Intentional and 11011i11te11tional cor1forn1a11t relationships, thert, differ f ro111 0 11e a.11other u1 degree rather t1la11 i11 lcir1d. Tl1e question is 11ot esset1ciall)' oi:1e of intention, but of at1dible resembla.r1ce. This depends not only upon all t:l1e f:.ict
:!9

See P· 49·


CONIZ:OIUV~'l"l:

.RELA"fIONSf-IIPS

75

tl1e e.'ttent to \Vlnch the event being com.pared is marl{ed-off as a se1larable entity. T l1e sinlilarit)r benveen part C of Exa1n~1le 44 and the other n1otives is difficult to percei\re because C is embeddeti in a larger and quite uniforrn •

patter:mng. Tl'1e question of the artist's intencio11 intrigues us and 'tvill, no doubt, co.n ti.t1ue to do so. For ot1r culn1re l11ls, at least u11til recei1tly, assigned tremendo·us importance to tl1e creative act- co perso11al expression ai.1d i11dividt1al discovery. Fortu:c1ately, ho"'eve:r, k11owledge of the c.'0111poser's intention is not t1ecessary f O·t critical analysis. lt is fortt1nate becat1se such • • • - 1) y in1poss1 • 'bll • • h f ·1.. • • lf 111tet1t:Jons are \1lrturu le to ascertam, e.it er ro m t11.e mUSJ.c 1tse or fron1 ex.tran1usical docu1ner1ra:tion. Even \''here docur11encacion exists, its n1terpretacio11 and reliability is p·.roble111acic. l{n0\1i7ledge of t11e composer's inter1tion is unnecessary l/ecause a relatiortsl'lip is a relatio1'1sl;·ip V\thether it: was expressly de\1ised by the composer, resulted f r·on1 the orderliness of sty1listic syntax, or in rare instar1ces \¥as the rest1Ir of ch~111ce. And a relationslu.p is a relationslrip 011ly in tl1e ligil:t of so1ne cc>gr1itive act, whether conscious or intuitive, learned or innate. T 'hougl1 every i·elationship is significant, it by no n1enns fallows that all are eqztally so, or that tliey ar·e so in the sar11e resiJ.ect. So111e c.011formant relationsllips a.re reL1tively l1nin1porta11t in creating col1erence or artic..11lacing structure, \vr1ile otl1ers are crucial. Some are con1positionall)r significanti111 po1~t:ant for critical a11alysis-\vhile others are most sig1llficant as .aspects of the com.poser,s idion1 and of style ar1alysis. Scil.l others ~tre rele\rat1c not .t o the analysis of particular co111positior1s"' bt1t co an u11derstanding of the psychology of creariviqr, t'J1e 11~'tory of ti1e composer, or the 11istory of style. The co11for1na11c relatlons1ups thtis far co11sidereti \Vere asSt1med ro be audible; often such. similarities are strilcing. Lil

EXPLAINING MUSIC

1

i\s a kind of ' ooda,," I should lil to '\\rhich only initiates or members of the it1-grot1p are priv)T. Somerin1es, as it1 jazz. su.ch. secret r·elacionships are audible in performance once or1e l1as lear11ed cl1e repertory.a-0 Tl1en tl1ey pla,y a sigr1i6cant role in aesthetic ap,preciarion. Relating the audil)le rune "vithin a florid jazz improvisatio:n to its hidden model. i11vt.)lves the pleastrre o.f psychic econo1Tr)' · In other cases, ho·wever, C<>' ert .relationships are inaudible and irrelevant for mt1sical experience. One tl'links, for inscance, of cr}rptograpluc de\rices in which the nan1es of friet1ds or o·f admire'i oon1posers are e1lcoded through the t1se of the letter names of pitches, Nlorse Cocle, or so111e othe.r C}rpher. Similarly the manipulation a11d orderi11g of pirch ai1d. tim.e rows in serial music are often inat1dible, i11 the sec1se that th.ey .are unrecognizable as suc!J. They ate relevant for i111d.erstar1di11g th.e precom110sicional activities of tl1e conlposer, not tl1e structttre and process of the con1posirio11 itself. Such inaudible and cryptic orderi11gs llSlJally do more l10\vever, tl1an satisfy the coterie instinct. First, by linriti11g tl1e nurnber of options open to . .111 a con1posit1on, .. · hi·. s . ~ J:>Omt t he composer a.t a parucuiar t hey f ac1·1itate 1

choosi11g. Tli.tlt is, if rl1e co1isrr:a:ir1ts present in cl1e style a.re i1o·c sufficie11t, tlle .composer will de\rise special ones of l1is OV\rt1, i\s I have st1ggested else-

where,81 this need for adcied constraints \\111.S perhaps invol\red in Schonberg's invention of the twelve-tone 111etl1od. Ir1 this case they were initially, at least, quite audible. But this t1eed n
a.o See Frat1k Tirro, ·~The Silent Theme Tr.tdition. in Llll, 3 (July, 1968) , 313- 334. st AdttSic, the Arts mid ld·eas, p. 24If.

jl:lZz~1~

TfJe Musi cal Quarter,y 1


CONFORMANT REl..ATIONSHIPS

77

i\ilore genera]ly: there is a clisti11ction betwe·en explai11it1g the ge11esis of a con1posirion and lin.derstanding the corn·pleted \vork.32 T o illustrate, let me take one final example frorn R eti's book. R eti wanrs to show that the theme o·f tl1e Trio of tl1e Scherzo from Beethoven's. ''Eroica', Sy1nphony is a transforn1atio11 of th.e first the11.1e of rl1e first 111ove1nent. He "vrires; If \Ve c.onsider the -rrio theme as it appears u1 the syn1phony itself [Exan1ple 45A] and the version ir1 \vhich t llis theme appears h1 the fourth, the last sketch [B], it is harcll}r possible to pnint out any actu_al afii1uty betwee11 these shapes at1d d1e \.Veil k.nown tlierne of the .first n1ovement C:

E~-ample 45

An attempt 011 this basis to prove that Beetl1ove11 formed the themes of the various movements from one common thought \Vould be refuted as artificial. Nevertheless~ the sl{etcnboolc lucidly den1onstrates the true process by Wtlich the T r io theme Clme into existence. Looking at tl.1e version in wttlcl1 the then1e appears i11 the fi1'st sketch,

"''e cai1 hardly belive our eyes. For th.is first original version of the Trio theme mirrors the main tl1eme of the sy1npho11y so distinctly i11 the his dl.stlnction, as well as that 1nade it 1 d 1e first chapter between b't)rle analysis and cricical anaJysis, is made by V\!illian1 Tl101t1son in t1usic Sy't11:posiwn, Vl (Fall, I966), 9'-93· 3:!· T


:EXPLAlNlNG 2\>1USIC

rhythn1 a.11ci spirir of its "vl1ole shaping cl1at t11e nature of the tl1e1ne'S as

two conceptions of one identical thought ca1111oc be dot1bced.33 Bt1t Reti's obsession with th.e u11if)ring function of cor1forn1at1t relationships makes 11in1 tniss tl1e crt1cial poinr. Trt1e1 Beethove11 begins vv.ith son1ething sir11il~1r to tl1e first 1noveme1l't: tl1e111e. f.'ar f rorr1 bei11g surp:rising or u11usual, it seems natural to start f.ro1n the ltnO\J\tll t1sir1g it ~lS a j\1mpi111g-off point, a \\'ay of gectit1g the compositio11al act going. But Beet.hoven vvorlcs fro1n this sin1ilar.it)r toward sor11ethi1ig r11-trnifestly ciitfere1:tt. H e co\1ld ha,,e ii1v·e11ted sornetlu11g r,elated by conforrnance, yet su.btly disgt1iseti. Bt1t that \11 as fXOt \1l Ork may T/Je Thematic Pro cess of 1Wi1.sic, p. 358. 3 ·~ Sketches can tlu:O\V light u.p on the con1positional process 01ily if they are interpreted in the light of theor y about the 11an1.re of mt1sica) relario1lships. For example, to expLUn. why Beethoven re1noved sforzm1di fron1 the fir~'t version of the OJ)e11i11g rr1ovem et1t of r:he F-l\tlajo,r Strir1g Quartety OJ>llS tS No. 1,. 011e must l·urve son1e hypothesis abo·ut the f1.1ncrion of tfor2;(11zdi in getlera:L O cl1erwise all one ca11 say is tl1at Beethoven clidn't like their efiect- ,vluch et1..'.Plains r1otl1ir1g. Ia1 tl1e 1lbset1ce of a background of t heoryr ( ho.\\rever inforn1a1), only description, 11ot analytic expla11:atio~ 33

is po sible. 011e 1nigl1t turn tl1is whole n1atter around and suggest rhat con1parir1g s1{etclles \Vitli tl1e finisl1ed composition \1/oUld give us fairly h~rd data against wllich to test our theories. Fo.r if theo.ries can explai11 v.rhy a composer t11ade 't l\e changes lie did- or in an ideal case, evet1 predict from a sl.cetcl1 or autograph \Vhat changes ~een' likely. and cl1eck these a.gainst the printed score----t.he:i1 our tl1e,ories '\\10\lld }1ave received a. kind of objective confirrnation.


CONFORl\IIANT RELA'fIONSHIPS

be illuminatit1g ps}"chologically and biograpliicall)t, but it is not the same as, and cannot be substituted for, serious analytic critic.ism.a5 35

This does nor contradict the observatio11 i11 iWusic, tbe Arts and Ideas tlrat we "11nderstand an e\rent o.r a,n objecrt partly at least, b)' u.n derstanding how it c:ame ro be wl1at it is . . .:r, (p~ 63; also see p. 89). Aesthetic underst:ancfu1g of the genesis a11d gro,vth of a musical partern is contextual; it depends upon a con1prehet1s.ion of the precedi11g patterning as it functions \vitlun th,e \VO.rk aod ,..,ichln its style an(i tradition. The ps.}rcho1ogical understanding of the act of cor11posicion depends upon. (>Uf corn~ prehension of rhe life of the composer1 a1id ,,,ill eertau1ly be enhanced by the sn1dy of sketcl1es a1id the lilce.


CHAPTER IV -----·------.........,.....--...,-· ·, --·- ·--

Hierarchic Structures Hiertfr,c,bic St1·itctures This cl1apter ''rill be. co11cert1ed \;i;1it'h hierarchic sr_rucn1res-the ha.sis :for their existence at1d tl1e l(u1d of orga.nizatio11 J>Ossible in ton,al m11sic. 1~hese consideratio·ns will, ir1 tur11,, lead to a disti11ctio11 het\\reen form a,nd process in n1usic, and to a discttssion of rl1eir interaction. Hierarchic structures are of signal irnpo·r tance because t l1ey er1able tl1e composer to u1 vent and cl1e listener to c<>111preher1d complex interreaccive mi1sical relatio11ships. If 1ntlSicaI stim.uli (pitches, durations, timbres, etc.) did 11ot fon11 brief, but partially con1pleted e\rents (1noti\1es., pt1rases, etc.), and. if these did 11ot in turr1 combine \vlth one a11orh.e r to forn1 more exte11ded, l1igl1er-order patterns, all relationsl1ips would he local arid transient~in tile note-ro-nore foreg.round. orihierarchic 111usic-tluit of John Cage, for insta:11ce-111-0ves, lil
0

T he Architccrtrre of Con1plexicy.'1 Proceedfngs of the A111ericmz P!Jilosopt1icaJ Society, CVI, 6 ( 1¢1 ).14'77. 1


HIERARCHIC STRUCTURES

1i

T o illusttat,e how hierarchic structuring worlcs, I have chosen a theme which is botJ1 exceptionally clear and very comp·act. It is tlte first section of the secon.d movement of Beethoven's String Quarter in Bb J\ilajor, Opus 130. TI1e music is gi1len in Exanlple 47. Hierarchic Structures suc.h. as this c.1n arise only if the series of stimuli are articulated itito 1nore or less discrete eve11cs 011 tl1e var:ious le:vels of the

hierarchy. In this case, the Uf)per ( i - 4) anal}'tic braces show rhythmic gr:onpings, and the lo\ver ones (d~a) , indicate forn1al relationships. These groupings are rl1e .result not only of durational relationslups bu.t of 111elodic, harmonic, tonal, and dynamic ones. as "''ell. The first measu1·e, £01· insrance, is an event -a more or less discrete pattern on "'rhat .is tnarked as le\~el a ( t ) . Tl1e measure is a single gesture. Tlus fi1·st event combines witl1 those u1measures 2, 3 &11d 4 to forrn a lugher-le\rel entity: cl1e a11tecedent p.h rase of le\rel b(z). The conseque11t phrase is similar to tl1e a:t1tecedent and botl1 cotnbine to form the mo.re complete phrase of le,rel c ( 3). Though it is constructed in a somewhat different way, the second half of the theme (measures 9-16) is also hiera_rchically org-m:ized. And the two t1alves co1nbine on a ·still lugher level (.level d, 4), creating a closed, stable shape---a rounded binar}r form. Finally the theme is itself a distinct event \Vithin the structure o.f tl1e whole moven1ent, which is a three-part, da capo form: a kind of scherzo-trio-scherzo. The section being analyzed is t:h.e ''tl1en1e'> -0r ''Scl1e.rzo." For a series o.f stlm·uli to form separable events \vhich can act as elements within a hierarchy, there m t1st be some degree of closure.2 Closure the arrival at rela.cive stability-is a result of t'he a<..~on ai1d interaction among the several p·anuneters of .n)usic, Because rnclody, rhyth.~ l1arm.ony, texture, timbre~ and: dynamics are relacivel)r independent variables some may act to create clostire at a particular point in a \~1ork., while others are mobile a11d on-going. To the eA'te!lt tl1at the parar11eters act togetl1er in the arriculati-011 of closure or, alternatively, in creari11g instability and mobility, they may be said to mo,~e congn.tnaly. ,Conversely, \vhen some para.meters foster closure while others :remain 01Jen, the parru:n eters are said to be 1201JCongr1iem. A decepdve cadence is a simple ixista11ce .of n-0n-congrue11ce: r l1ythn1 and. melod;1 act to arcict1late closure, but ltar111t>ny t..e1nains open and mobile. 2 T l1e ' 'ital i1nportance of closur e in the aniculation of forms and processes was m11de clear ro me by Bazba.ra H. Snlid1's boolc" Poetic Closure: Or Why Poe111s End (Chicago: University of Cl1ic'1go Press, 1969).


@·------------'~----...

•

Pw.lt.

..

'

,,

•

. •

• 1

.~ .~ , I

'

;

~

.

'

~

•

' j

~:

',

~

____...... ·

.-JI+'+r ,,,,,,,,..

If'

If~ !

•

~

~=-~_i:i~e , ~:a::~:::::=t= ' ~ ~ -~ ' l

,,

1-t ..

• • ~ '

'

'

:ii!

'

'

'.

'r .• • '

'

--'

..t~ -~----..::::::;.......~,=-=~=-=~~ -=-~===·=t~.............:,:,:::_...........~·

a

JS

'

I

I

' · '-----..:::::::=---=::::::_____...,:it.-_ _..,:===~----'

-·

... I .

. 1=.

• I

,, - '

•

'

-- <:::. .:;::.'

<:::;.

'~

::->

..::::::::-

' 7- ': >

'

' •:::

:6

. •

'

"" ::::..

.

'

...;:;;:::

~>

I

.

.

l

'~

~......:::::' ;:.;::;.. .

.

~· • -=::::

....:::: ::=·-

.

.

' . ';iJ' • . .. ... - .. .. ... .. ·J

'-~ ·

-

•

::::>'

•

,

..

'

.I

.. <:::::. ::;;::.

~ -.I ..

::;:...

i

:t:

-= :::::>

•

il

1. '" .

••

.

..

,,,

--· "'

J

8

'

•

•

.

-•

• ':II-· f=1i

.. ... - •

.

'

.

'

~

C:TC"- ·

"'

,-,,,;;;;;;_ ;;,,....1¥-.._,, , ' -

.:t:i .

•

'

~

. '

.

·~ ,. '

·f

'

"

_.....• -- .

,<'j ~

ti -·

'

'

•::...:i

:l !J • •

•

•

•

.

1

t

__ ,

- ... -:f

... .,. .

... ...

~· · ....

'1

. •.

'

.. i

I

•

LJ•

~ '™' ~--........::::::::;..........:i')Mf ...~...~J \~. .....Jl't•-•rt''-..............:i)..t''-------~-~2~-----===:::::..............,, -~ ..

%

F -_ -

•

:d

-

•

'

-• - .::."'' _, .. •.....::=-><··,.-,_ - - .. - .. ...

""Sp rd

'

•

...

•

'

'

'

" '

'

•

-~ [

•:

.. .:s: ::-..•

~-:f.

~- - k

.-

•

~

':w

•

'

I

.

.. ::::-

.,

.

~- . ,....c.u.

c;:;;:; :::it

W '

~

..

'

==-

":ti.

:;::::....

~

- . ..- . ·-

<=::::: ::::;:..-

.. t .

•

ll

.

.'

!:

. •

•

•

•

!

,- m'

'

_ ;@ -

- 2

1§

I

'

Example 47


BlEllAltCHIC STRUCIUR.ES

Closure, then, is an aspect of patterning. I ha\~e considered the nature and basis for patter1ting in t11usic at son1e length in E1t1otio1i and A1ef11lnng i-n Music. 3 Here, the barest summax:y 1nuSt suffice. The delineation of musical patterns is the r~'Ult of the relationships \.\i~ithin a.nd among a. r1u.n1ber ·of factors~namely.!

I) the presence of similarity and differe11ce bet\:vee11 successive eve11ts within. a particular parruneter. Both con1plete unifonnity a11d total heterogeneity preclude syntactic organizatior.1, and hence establish no srability-itlSta~ bility r·elationships; 2) ti1e separation of one event from ano'l:her in cin1e, pitcl1, or botl1; or througl1 clear diff.eren·ces in dynan1ics, timbre, or te"-1:ure; 3) immediate re1)etirior1, '\Vh·e ther varied or exact, of part or all of a pattern; 4) the Cl()n11Jlccio11 of pre\Tiously ge11erated implications; 5) harmonic ca.dence ai1d tonal scal>ility-r~ I-low these factors function to create patterns and arrict1late closure \Vill, I hope, become clear· as \
tl1e te111po "\\rhich is very fast. i\1.elodicaUy, tl1e first t\VO beats are understood as an F with a neighbor-n.ote., Eb, followed by G b \vhich functions as an ecl:>apee ·Or.nan1e11tal tone. In 0th.er \VOrds,. as indicated in graph b of Example 47~ the main melodic line moves from the Fin me,asure it to the Eb in measure 2,. with Gb acting as .an ornamental tone. B·ut perhaps t'his analysis is too hasty. f ,o r on the last beat of the measure Beethoven presencs the 'h armony-an 'Eb minor tria.d- whi.ch would have made tl1e G b a substantive, chord tone. It comes too late hovvever. The harmonic change does not really alter our t1nderstandu1g of the mocivic structure. ~(Chicago :

Universit)' of Chicago Press, 1956), Chapters ffi.- v.


EXP LAIN ING MUSIC

Th.e relative closure of the frr~'t rneasure is i11 ,i)art che result: of rl1)
first two megsL1r.es beer1 co1111ected ten1porally a11d melo,lic~tlly, as sl10\\ril it1 Exa111ple 49A,. the first measure '\VOuld have been perceived .as a d1screte e\re11t beca11se t he second meaSl1re is a. varied re11etition. Or, had there been no r epetition, but a clear separation ir1 cime and pitch, as in Exan1ple 49B, two

A.

s. "------.--- c:. •

'

&

~

!

I

_L

r""'

I

Example 49 diffe1·e11t, yet sep:a.r<-1ble events, "''oulcl have bee11 de611ecl. But \Vere there neicl1er repetitio11 nor se1Jaration, tl1en, as Exan1ple 49C indicates, the rvvo nleast1res would be understood as a single event- though w·ith some internal articulation. An.cl the col1esivei1ess of tl1e patterrung \.VOt1ld h.ave been reinforced by the repetition presenled in the ex.ample. The first n1easure (Exan1ple 47) .is stro11gly connected witl1 the second on the next .level of the hierarchy. Melodically•, the morio11 from F to Eb

c1·e(ttes a lnghe·r-level process vv.hich implies conrint1ation down to the tonic, Bb.4 The first two measures are also connected harn101rically. The Bb-Eb-Ab progression ll.1 tl1e bass l:>egiri:s a11 l1arn101iic process chr-0ugl1 tl1e C)rcle of filths. Note, ho,~ever, that melodic a11d har111onic processes are coor din~1te only on prin1ary accents, not 011 seconda11' or1es. Part ly for this reaso11 the Gb in measure 1 and the F in measure 1 ~ire understood as being ornamental. And this is "''hy, despite the cha11ge of harmony on the fottrtl1 beat; the main motion is 'b}' meas11res. Though t he m<>tio11 across the barli11e is supported by the syncopation in tl1e \riola and tl1e eighrh-11otes in th,e second violi1~ the cello does not stro11gly en1pl1asize tl1e h~1rmonic progression. That is., tl1e root: motion from Eb to Ab wot1ld ha\re bee11 111ore en1f)h,atic had tl1e lower Ab con1e on clte first beat of r11easure 2. The reSt1lt is that tl1e loiv Eb is almost heard as acco1n·pa1iin1e11t:al- like a kind of '(t1mpa.11'' bass. 4

'

Tl1e .na.tu.re of and basis for i111plicacive relationships a.re discussed .i11 the second

part of this book.


HIER.AR.CHIC S'IllUCTURES

The closure at the end of the second 1neasure, like th,a t of the :tirst, is melodically and rhythmicall)r unambiguous for essentially the ~ame reasons. Because it is sequential \Vith the first, both 111elodically and harmonically, the seco11d 111ea.ru.re strongly implies co11cinuation to the tonic. Indeed, beca.use the bass 1notion does move a fo1rrtl1 across the bar, t11e connection bet\veen me~
and 2. Measures 3 a.11d 4 are a sing1e ever1t. M.elodic:ally, there: is one 1nocionfrom Db ro F. Har1nonically, rhere is a progression fron1 VT to IV (with a ftlnt of ll-tl1e C 0 11 the fo·urth beat) to V. Rl1ycl111-Ucall) the repeated Eb· acts as a pivot linldng the weak i1art of n1easure 3- \vith 1neasure 4. The closure at the end of t his phrase is the resi.1lt, first and foremost, of rl1ythmic stru,cture. O·n the fusr rh)rtlu1uc lei;rel ( :1 ), e.'lch of the first two eve11t.s ex1ds on a n1obile \Veak beat e\re11 tl1ot1gl1 tl1e trocl1ee itSelf is closed. And this mobilit)' is emphasized by the final eighth-notes in the second violin part. But the final event of the phrase-n1easures 3 ~i11d 4-is decisively en.d-acceoted., and hence closed. Tl1is clost1re is empllasized by the rests in 1neasure 4 which separate the end of tl1e first phrase fro111 the beginning of the second. On the seconcl rhythmic level ( 2), n1easw·e 2 is an exact repetition of measure 1-in. durational relatio11slups. Because exact .repetition d.oes not create col1esi,1e patterning, these meast1res do n.o t create .a clear rhyth111ic group. They are connected melodicall77' and harmo·nically, but not rhythmically. \¥hat is needed is ten1poral differentiation or d.ynamic change. The former is provided by the t\VOpmeasure group v.rhich follows. Now the first two measures cohere because the;.. beco111e _part of tl1e whole pl1rase \vhicl1 1

1,

is an anapest group; or, from a fo.rmal point of vie\v {analysis), a barform: 'v-w'-x. As a rest1lt, the antecetient phrase is closed, end-accente,d on the first rnro rhytlunic le,~els. .Because rhythmic closure is forcef11I, the l1arn1onic progressio11 is tmderstood as cad.ential. T he progression from the subdomina11t ro the dominant strongly :Unplies contin.uation to the tonic and hence is on-going, particularly so beC:a1t1se tl1e bass moves linearly, vvirhot1t decis1ve root m.otion or disjunction. But rl1ytl1mic arcicuJati<>n makes t1s interpret measure 4 as a poi11t of relati,,.e scabiljcyr. To generalize: a setrricade11ce n1ight be defined. as one in which a n1obile, goal-directed, harn1onic proc.ess is temporarily stabilized by de,cisi\re rhythrnic closure. In other words, a sen1icadence is a case of par-ametric noncongn1e11ce w hich has become archetypal in the stylistic ·s-ync:ax of tonal music.

'


86

EXPLAINING MUSIC

The melodic closure of the a11tecede11t phrase, '\vhich is by no rneans forcefttl, is also prin1arily tl1e result of rh]rthmic articttlario.n. The F is an acceptable point of cl.osure parrl;r becai1se it is the relati\1ely stable fifth of tl1e scitle and partly because it was tl1e first 11ote o.f the rune. That is, if t.he in1plied melodic goal- the tonic, B·b is not reached, it is pr(!ferable to return t o .the 11oint at ,,,hjcl1 n1otion hegai1 than to ca(le.nc.e 0 11 some other IlOte.

Finally, the closure of tlte fu-st p.l uase is en1phasized b)' tl1e fact that t he sec,ond begins like a repetition. Indeed, one of tl1e he11efirs of antecedentconseqt1ent patterns is tlia.t the elen1ent of repetition malS caru1ot sl1ape l1ig h -level rl1j t1111lic grot•ping. hrfelod.ic-harmonic sy•ntax luis to shape the high-level rll)rthm. Norn1all)' the r11elody of the antecedent phrnse cadences 0 11 eitl1er tl1e second degree of the scale or the leading tone. For instance1 in rhe Theme from l\tfozart's Piano So11~ta ir1 A Ai1ajor, discussed in the second chapter, tl1e 111elody moves ·f rom th·e third of the scale to tl'1e superton.ic in t he antecedent pl1rase; in the 1

co11sequent pllr·ase che Sl1pertor1ic is resolved, motring to the tonic. Thtts the O\rer-~ll motion of both. pl1rases js basically frocn the tl1ird to tl1e second., to the to1lic. And, as in M·o zart's Theme, i1St1a.ll)r the. semicadence which closes the antecedent phrase is c] ~lrly art:icultrted by a reversal of motion and a skip in the bass. Bt1t ncitl1er of these tal{e place in Beethove11's tl1e1ne. The melodic n1or:io11 at the ei-1d of the antecedent J)hrase does not n1ove tOv\rard tl1e to1uc, but a\~ay fro.n1 it~ retur:r1i11g to l4"'. The n~1rn1011ic n1otion at ri1e beginning of each phrase is at least as scro·ng ~ls that at the end, and the F in the bass at measure 4 .is not preceded by either a reversal or ;a. slcip, 'b ut foilo'ivS from the downbeat rnotion beg11n it1 rneastire r : t}1a.t is., it is a concinuatio11 o.f 1

the scale, Bb-Ab-Gb-to F. As a result, :n ot only is the rhytl1m of le\1 el 3


HIERARCHIC STRl 1C.J.' UR.ES

.rather wealdy defined, but the l1igh-level melodic structure, inste·a.d of being unequivocally linear, is 111orc or less tri.1.dic--as grap·h. b in E."{ainple 47 shows. These considerations of closure suggest an e."\.-planation of the melodic and 1·hythmic strtlCture of the secon{l half of tl1is Scherzo section. Melodically, the first half of t'he section emphasizes th.e fifth degree of the scale (F) 1 with the Db as a Point of secondary importance. The Bb ar rhe end of the consequent phrase is et11phasjzed, but it .is not, I think, as promit1ent and stro11g as tl1e F. The triadic f.rameworlc (F-Db-[F-D~]-Bb-) '\\ luch forn1s the underlying Stt'UCtu.re of the antecedent-consequer1t organization is, in a sense1 a microcosm of the ~1 hole theme. For as grapl1 a of E."(an:1ple 47 sl1,ows, the first part of the rhen1e is a prolong:atio11 of F, the seco11d ii; a prolongation of Db, while the third is a prolongation of Bb. Rl1ytht1ucally, the highest level of tl1e first p-arr \Vas nor decisnrely defined. vVh:at is i1eeded i.s a strongl}r end~accented grouping to create unequivocal closure. A11d tl1is is what \Ve get ll:1 tl1e second section. The first of these phrases (n1eaSt1res sr1 i) is a. t'vo-measttre pattern exactly repeated. Both tl1e first and. second levels of the t';vo-measure groups are trocl1aic. Because they end on the u1zaccented part .o f the rl1)"thmic group, they are mobile and on-g·oing. Becatise th.ey are alil{e, they are not cohesive. Together they function as \Vcak, ai1acrustic .groups in relation to the four final measu.res. That is, rh.e \vho.le secor1d part i<; a.n end-acce·nre,d a,napest pattern (level 3); or, from a formal point: of vie\\' (level b}, a f)ar-form-m-m'-n. t\tleasures 13- 16, '\vhich are tl1e acce11ted goal to\vard \vl1ich measures 5f-Il lead, are also anapest or bar form ( """'-'\v'1 -x'1 ) on level r. That is, measures 1 3 a,nd r4, \vnich are essentially alike, ft1nction as weal\'. groups leading to the last n.vo measures which act as an accent ii1 relation to them~ Tl-US final goaldirecced patterning is intensified by the crescendo to fortissimo, bringing the "vhole section ro an emphatic close. (111 a sense., the unequi\ro~'ll anapest rhythm of tl1e secot1d A!-part of the scherzo is the realizatior1 of a potential present in the .first part. Tll:at is, while tl1e n1elodic-harmonic stru.cntre it1 the first part caused measures a arid 6 to function as pivots 011 rhytl1n1ic le\rel z, thus \Veakening the. S'COSe of end1

accentuation in both the antecedent and the conseqt1ent ph:rase, measure 14 is not processive in relation co 1 3, but only i11 relation to \\tllat f ollo\vs. To

put r·he matter 1nore gen.erally: in the first part o.f the scherz

88

E1'.11>L.A!NING tv1USIC

Finally, the rettrrn i11 meast1res 13 to 16 to a varied version of the rhythmic-rnclodic pattern of the first .part enlUinces the sense .of closure. A.r:. .noted i11 Chapter III, return is 11ot ·che same as repeation. Repeating tl1e first eight n1easures, for instance, does not enha11ce their closure. \!\That is needed is the te11sion of '(going a"\.\'il-)' ,1, of instabilit)' · Such instability, slig·ht tl1ough it is, is provided in. this case b)t n1easures 9-1 z. Tl1ol1gl1 the higl1~level clostue of the second l1alf of the tl1e1ne is strong, oddly enough the lo,v-levcl rhythm ends 0 11 a ,v,eak hear- is somewhat mobile. It is af'propriate tl1at closure be atter1uared, becat1se this is not th.e end of the n1oven1e;nt:. T l1e. Trio is abollt to fol\O"tV. v\'he11 the mc>vement does er1d, not only do all .rhythmic levels combine to ere-ate closure, but stability is emphasize.d by coming after t:l1e tension of an extensio11. 2.

I 11ave considered tl1is 1nt1sic i i1 detail in order to sho·\v that closure and .mobility are functjons of the action and h1teracrion a·mong all tl1c para·n1eters of music. At any point i11 a phrase or sectior1~ sor11e }>~rr~11neters '1vill tend to create closure, vvhile ·o thers "''ill pron1ote conci11uity. For instance, in the partial closure at tl1e end of tl1e first measure, pitcl1 and time disjunctions tend to sepa1·ate events, and harn1onic 111otio:n at1d seconcl-level mel<>dic organizatio11 make for concint1iry. ln other \Vords, and this is the important po.int, the paran1eters of music d.o t'\Ot as a tt'UC n1ove co11grue11tly. If tile)" did
.


HIERARCHIC STRUCftrRES

quite stro11g; but there is virtually no rhythmic closure on tf1,e same level at the end of measure 12 . Or, if the end. of the ai1tecedent phrase is contrasted with tha·c of the co.nsequenr, the latter is inore n1arkedly closed because. although the degree of rl1yrhn1ic closure is essentially tl1e same, melodic and harmonic closure are mucl1 more forceful. E\rery composition, tl1e11, exhibits a luerarchy of closures. The more

decisi\>·e the closure at a patticular point, th.e more important the structural articulation. Or, the smtctti,re o.f a con1position is so?-rz:etl:.ii1zg w/Jicl1 w e i?zfer fro111. the hierarchy of closures whicl' it presents. A composition continues..·is mobile and 011-going-partl)r because of the tender1cy of parameters to act independently of one anotl1er, to be noncongrt1ent. The end of a mo·vement is not n1erely a cessation of sound.. It is the point at
particularly those that are most important in the formation of patter11s ~ue themselves structured hier·ar·chically. Considering the melodic structure of the BeethO-\ten theme1 we per,c ei\re not only the note-to_;note 1notior1 witl1u1 meast1re r, but the motion of t he 111oriv-es themselves. On tl1e next level, w:e perc,eive the relationship benveen anteceder1t an,d conseqt1ent melodic structures. Arld so on. Tl1e same is true of rhythm and of harmony"-a11d to a smaller exten.t , of te.icture. 'T he \vay in whicl1 a, particular paraineter acts it1 artict1lating structure may be different on different l1ierarchic levels. F o,r example, on lower levels dynamics anc.f orchestration tend ro contribt1te to the artict1lation of rhyth... mic patterr1s, but on big.her levels they generally serve in the structuring of 1

large-scale fon11al relationships. Similarly durational relationships are crucial in the shaping of l(>w-level evenrs such as ffi(>tives and pl1rases; w hile tonality and texture are especially important for· tl1e org«mization of ltigh-level strnc~ tures. i\1oreover, the role played by a particular parameter depends not o:nly 1

tipon hierarchic level, but also upon sryle. Hannonic relationships play a central role in the strucruring of tonal music, but none in the ordering of most serial compositions. Timbre plays a very significant role in defirling relationships in \Vebern's music, but only a minor :role in the 111usic of Bach. Also the syntax of particular parameters tends to change as one mo,res from one level of the hierarchy to another. For instance, tl1e synr.ax of chord-to~cho.td progression in the foreground h.armon}r is different from that which go,rerns long-range harmonic structures-tonal relarionsllips. Thus the probability of t he tonic chord (I) being follo,ved by the mediant ~

.


90

EXP'LAl.NING MUSIC

fo.regrot1nd. hannony, but it is quite high i11 the succession of ronal areas-particularly i11 the n1ino,r 111ode \-vit11ess :n1easure.s 9-I 2 in the music '-Ve have been d.isct1ssi11g. I11 like rnai111er, it is r11ore like!)'• in the music of the second l1a1f of the eighteenth century, that a lo\,v-level melodic e;;,rent, a motive or phrase, \vill l}eg.in V\' itl1 a tr:iadic pattern tl1an thtlt the higher-

(Ill) is lo\v

i11

level connection between pl1rases \Vill be triadic thot1gh, as this Scherzo or tl1e 1Vli11uetto of ~'1ozart's Sympl1on)' No. 40 shows~ this is n.o t in1possible. I11 otl1er words, l1ienu·cllic sr1·uctt1res are nor1unifor1n a.tld discontinuous. .Just as the \.\1ays in \vhich chemicals uruce to forn1 molecules are different fro·m those n1volved in the o.r ganization of molecules into cells, so the ways in whicl1 to11es co111bine to for111 motives are differer1t fro111 the ' ''ays in. wluch motives cohere to create larger, 1nore con1plex mt1sical ever1ts.5 A motive, a IJhrase, or a period is defin·ed by sor11e degree of closure. On the level of its closure-the level on v\fhich it is understood as a separable e'\re11t-it is a relatively stable, formal encicyr. Tl1ougl1 it contai1is a11d is definecl by internal processes, once closed, it is not a process b t1t a palpable ''tlm1g.'' ' ;\Then ii1 rum ir combines "'"ith otl1er events on the san1e level and thereby becomes part of a 11igher-let-rel ever1t, it agau1 f.u11ctions a processive \Vay. l\11easure 1 of Beetl10\1en)s tl1en1e, for insta11ce, is a forma] entitydefined, as \Ve have see11, by i11tcrnal rhyth111ic, 1neiodic, and harmonic relationships. It combines in a processive, sy11Utccic \:\ray ' vith subseque11t eve11ts to form a higher-level enciry ·w hich is completed~parcially closed-at measure 4. O n the level of its closure; the antecedent plu-ase is also a for nlal entity; and it, in tum, combines ,,~ith t he co·nsequent phrase forming the :first part of a rounded binary structure. From rhis it a:ppears that the same event may be characterized as eitl1er fortn or process depe11dit1g upon the liierarchic context being considered. A ge11eral principle of hierarchically structured 111u!11c is that, as or1e moves from one level to an-0tl1er, there is alwa)rs an alternation of fun·c tio11al significance. \ IV11ac is processive ot1 one level (for instance, the 11oteto-note relatio nships \vitl1in the first 111easure of the the111e) becotnes f01·111al (a n1oti\1e) or1 rhe i1ex"t; wl10t is formal 011 or1e level ter1ds tc) becorne procesSi'ire 011 tl1e next. And sl1cf1 alternatio11 co11tintles until the highest level- th.at of tl1.e co111pos:ition () f moven1ent- is reached. 0

u1

This lnacter is cor1sidered ir1 rnore deta il ix\, 1\f.lusic1 tl>e Arts ttnd ld.e.1.S ( Chicago: University of Chicago Press. i.967 ). pp. 96-{}7, 157-159, 3o6-308, arid passin1. G£,rcncs n1ay follo'v on·e another 9;rithoi1t creating hierarchic s011ctr1.te, as, for insm11ce, in tf.ie successive st~ter11er1t of an osrinaro figt:ire . .An ostitiato patter n \Vill 15


filERARCH1C ST:RUCTURE-S

In complex composjcions the l1 ighe~t s:trt1ctt1ral le:vel that \.vl1icl1 is the basis for naming the form-is usually n1ixed. It is formal in that relatively stable the1nes are presented; ar1d it is processive il1 tlu1t sucl1 stt\ble events are fur1ctionally related to less stable parts. "fhe clearest example of this sort of organization is the .rounded bil1ary forn1. And the most complex rounded binary s:rructur:es are sonata-for·m n1-0ve1t1en·ts. Tl1ough one n1igl1t argue that the highest level of a sonam-fom1 moveme11t is essencially pro:eessive· wa single e11tity it1 V\'hich nvo sin1ilar sube,re11ts. tlte exposition and recapitulatio·n, are functionally bound togetl1er by the dei.1elop1nent-it is, I tt1inl£1 closer to aero.a l experienc
°"'

event, a nudd1e or pr,ocess Y\'hich moves tO" 'ard some goal, and an end- the arrival at: son1e sort .o f clost1re. The subject of the F Minor Fl1gue from Book I of the, Well-Tempered Clavier, as Example 50 showst is essentially a single jlrocess a chromatic decent from tl1e fifth of the scale to the tonic. T '.l1e changing-note ~gure, C-Db-C, is the beginning or generative eve11t; the conitsclf be hier'!lrchic, bu.t the series a·f identical patterns can cr.eate no higher-level org-amzaciou. The series of state11i~lots is add,i·ri,re, not processive. Th·us osti.o atos giv·e rise to 'illl'l1at Herbert Sil11on Iias called "flat hie1·a.rchie.s.. ;, See "The Architecture of

Complexity," p. 469.


EXPLAlNT,.; G MUSIC

jt1nct n1otion ro\,ratd F is rhe rr1edia1 process; and the arrival at the tonic in measure 4 is the conclt1sion, and iI1 this c'lse the beginning of a ne\-v event. rfhotl.gh the conjttllCt prc)Cess is i11terrt1pted by tile n1otio11 fro1n E to F, this is tinderstood as bcing separate fro1n the ''real'' 1nelodic motion. (The inter.. jection creates \Vhat \vill 'b e called a potential sttllCtural tone. Becat1se its

melodic promil1ence is 11or n1atched b;r its functional in1portance, stn1cnrral empll3Sls is called. f.Qf~ rfhis t:n:lpha~iS takes place-t}1e pOteIItiai is a<.."tllaltz.ed -~,l1e11 the ans\ver begins 011 cl1e F.) 7 T11e bfisic n1elodic process cannot be di,rided or labeled ''A') and ''B," etc. Compare tl1is V\rith the co11sequent pl1rase of the Beethoven example, \Vhich also desce11ds fro111 the fi.frh to the tonic in. the nlinor mode. The Bach is pl1re pr·ocess. In the 'Beer11o·ven, the same pr.ocess is embodied in a set of formal relationships which ca11 be labeled w-'\\r'-x-and classified as a bar-for1n. 8

r There tlre e\' eI1 corn.1-1lete pieces wl1ich are basically syntactic processes.

In the first nvo })relude.s of Book I of Bach>s ' lVell-Tem1:>ered Oavier, for instance, the.re are 110 closed stlble ever1tS ·m-0tives or tl1emes ·\Vhich are repeated on the highest lev.el. Neither 11relude is a form '\\rr1icl1 can be labeled as ,.i\-BA .o r the like. Bllt both /Jave form. Sy11tactic coher:e nce is a result of the fu11ctior1al .relati-011sllips between the begimill1g, wl1icl1 has a relatively ·clo~'ed harmonic sh.~rpe; the middle, in \vhich a. less st·able, sequential melodic.11armonic process moves tOV\ a,rd the goal of stabilizing rensiort on rl'1e donlinant; and the co11cJusio11, \Vhich consists of a. prolonged cadential progression. The only repetirio11- tl1at of tl1e 1tC)"board figure '\vitl1in. the t11eas11re, "vhich serves ro sustain the ham1ony-is not syntactically srrucrured. Irs constant 7

Porential tones are disctissed in Pan II. pp. l 96- :zo.1. Patterns v.1 hich are 11ot str(rngly st:n.1cture d frorn a for1nal point of view- \vl1ict1 a.re ptim-aril.)' processive-are specially u.ir:ab1e fo.r contraput1ra.l treatt11ent. T11ey become for1nal er1titi.es on higher levels, usually ,t firough i1nita.tion.. 7


l-fiER..\RCHIC STRUC1~R~S

93

"~'

~~ .1:..~m '

"

' -

I

-

'

~-

A.

,.

'

'

'

di

~

1'

:JI

7±4

' :!J!. .,,.,.,,.._.• . " ·.

'

~

.

'J

.. ,'

~

'

~ : .

-~ -=. ' '

'

:·:":: ::

ti

.

1· ...

.

-

'

-

W w~-

I

'

-·

-

'

-

-·

•I J .

1-

.

.,.. e,,.,

-

•

I

-

'

'

~

Jli!!

•

,. ~r,._

-

'

'

f

I

'

.. .-.fiL :...

'°'' • !' Oif"U'·

L

'

'

,~

'

I

'

~

I

ii

'

1.1.

~'lt

'

::!!!:

'

!!

J

'

-,,_,

'

'

-

•

"' ~

J

'

' Im

'

'

'

f f'H'

11'1

~,

•

...,,.•

!

'·

,~

•

•

..

'

.n '

f'

~

..-r..·..·.. .

~

~- · '

ii

'

'

~ f.,~ ,

'

',

'

'

'

•~ • ....

.• '

"

'

'

-

..;;

'

V""' -

-

'

'

'

~

'

'

-

.

;,

'

-

,, ~

-

'

~-

'

'

c.S,

.-

•,,, • • '

-

-

-~ -·

-

'

'

'

'

-

x ' '

·.

-

,, rr

~' '

-

.[

'

"

,,,.. -

-

'

,.,."*'~ -

.

'

'

-

,,

'

---- - - !!' '

-

-

,~~

"

~

'

-

'

'

~ '

-

. '

•

"

repecit.ion precludes process. Cotisequent:lj' it is understood as an active ground, mtl1er tl1an as an implicative p·atterning. Purely formal stru,ctures are additive. Though successi\re events may


EXPL.l\INING M'tYSlC

94

be related by co1uorrnance (as i11 a stro·phlc ·form: - · '-A'' . . . An) 1 110 syn.tactic processes create lugl1er levels of .organization. .l\11d because the)r do not readil;r give rise to higher levels of orgruliza:tion, such forn1al scrucn1res seldon1 occur on lo\ver levels. The single exceptio11 is the ostinato; and s\1ch unvaried }Jatt.erns generally fu11ction as grounds O\rer "'' hict1 processive patrerns appear. For insta11ce, the m.orive F-E-D-C# in Ravel>s RlJt'f'psodie Esptttpzole (Example 5 r ) is repeated over ~incl over, and, after one or two repetitions, is tlJ:1derstood as a patter11ed ground a.gainst which tl1e <'real tones are l1eard. Tl1e pecu li~:ir cirCl.1ri1sta11ce i11 rl1is case is tl1at the ground is aln1ost as well shaped as so111e of t'he figures. The result is a. subcle a1nbiguity \vl1icl1 llavel enhances by moclifying the orchescratior1 and dy11anncs o.f the ostinato so that at tin1es it beco.r11es t he focus of attet1tion and seems to ch~1nge 1

,

•

...

4

its fur1ctton.

At tirnes a:i.,-ial melodies ha'\re sornerl1i11g of the cl1ar~1cter of additive form. 0 The fust measltres of the Seconcl Moveo1ent of Brah111S' Fourth Sy111pho11y corisist, as Example 52 sho,:vs, of a11 axial E en1bellishe(i \virh upper and lower neighbor-notes. Though th.e e11lbellishn1e11ts suggest direcrior1s of rnotion, and tl1ough ti1e addition of octave dollbli11gs in the oboes,. bassoons and flu.tes pro,1ide variety and a sense O·f motion. a tr1Je sy11tactic process is not generated until rr1easure 5, wl1ere the repetition of the axial pattern a third higher r11a kes it probable tl1at tl1e cl1e,me \:Vill be triadic o·n th·e pltrase level. And rlus ii1deed proves co be tl1e case: the axial I'attern

n1oves to B i11111east1re 15.

Bttt the clearest instance of pt1rely forn1a1 stru.cttire occurs on l1igl1er :1rchitecto11ic levels-.na111elyr, the therne and variacio:ns. TI1e :first n.1oveme11t of Nlozartls i\ -Nlajor Pia110 Sonata, of whicl1 the Therne '~ras disct:issed in the second chapter~ m::t}' senre as an ex.a mple. The Theme, and conseq\1ently ~

A.xial tnelodies are discussed in Part II, pp. 183- 191.


HIERARCHIC STRUCTURES

95

each of the variations base.cl upon it, is both s}rntactically and formally structured (Example 17). The first part, an antecedent-consequen~t phrase, though inter11all}1 b)7lltactic, is a formal entity on the level of its closure. The next four measures ( 9- 11) are prin1arily syntactic in .fu11ccion, cre~1ting Sttfficient tet1sion and change to call for a return to the consequent p'hrase, 'vhose

clos·ure is enhanced by a brief extension. Though tl1e the111e itself is a mL'\:ttire of form an.cl process, the movement as a whole is essentially formal. Parts-tl1e the111e and its variations are related co one another 11ot it1 terms of any 0\ er-all process, but in terms of co1iformance, it1 cerms of harmonic stn1crural parallelis1n and motivic sirnilariry. There ~may, h.ov..rever, be in1plic,1tive connections benveen parts. \Ve saw, for instance, that the potential fourth, E to· B across cl1e first barline implies and is related to its actualization. in the first vai~iario11. But such fore1

g round co11necrio11s, ho\vever i11teresting and importan.t they may be1 do 11ot create ·o r depend upon the existence 0£ high-level processive relationslups. Observe that hec::1use tlle l1igl1est level is adclitive, ratl1er tha11 processive, the series lacks an internally strt1ctured point of probable termination. The number o.f parts is \7 ari.1.ble- n1a11y or fe\v-depe11ding t1pot1 the ingent1ity of the composer, the taste of the ti111e, and the patience of tl1e audience. Of cot1rse adc.i.itive strut'tures ,car1 be ordered in sonie no11-processive \\ray. For instance, variation n1ovem.ents often hav,e a te11de1lC)t , to 1110\re it1 the direction of greater complexity, faster te11l.pi, and louder dyna:mics. But these are ''rrended,' changes, 11ot processi\1·e ones. That is, a num.b er of \'ariations of increasing con1plexiry and so on ma.y be f ollo~ved by a variatio11 th.a t is simple, slo\v, and soft. Formal, additive sn11ctures n1ay be ordered in still other \.\rays: for instancet in terms of son1e set of ke;r relationships as in the Well-Tempered Clavier, the suite, or tl1e S)1 111pho11y; or in ter1ns of a cextas in a stropl1ic song. 111 ell.is last case, the text inay provide syntactic co1J-

nectir>ns \Vhich the mt1sic itself lacks.. In all these lcinds, ho\ve·ve1·~ the structt1re will be for-1nal~ 11ot processive, unless tl1ere is son1e sort of functional dilfere11ciation among parts. A t heme and 1rariationst a ~'trophic song, a set of prelt1des or erodes is similar to '\Vbat He1·bert ~/\. Simon J1as c'lllled a flat hierarchy. He O·bserves 1 thac 'a diaruor1d is l1ierarchic, for it is a ~'tal structure of car'bon atonis that C~l'n be further deco.mposed into protons, neUtrfJl1S and electrons. rlO\.Vever, it is a 'flat' hierarch.y, in whicl1 the nun1ber of first order st1bsysrems can be indefinitel}~ large. " 1~ From this poi11t o·f ''ie,v, th·e 'history of the rhen1e 1

10 Simon,

uTne Architecture of Comple~;cy,n p. 469,.


EXPLAINING M USIC

and variations in the eigl1teentl1 a11d 11i11etee11tl1 centuries migl1t be under-

stood as the search. for a '1vay of transforming a "·naturally>' flatt additive l1ierarchy (as in 1nost B(troqt1e a11d early classical variatio11s) into a·n arcl1ed, processi\!•e o ne- o ne 'tV>ith. functionally differe11tiated parrs (as in variations of Schu1nann, Fra.11ck, and so·me of Beetho,ren). Though the forn1 of a con1:posi.tion is general!}' classified ii1 terrns ·o f the organization of the highest level, for111al strt1ctures are found on all levels of any worl{ of i-easo11<1ble cotnplexity. And Vl itl1 sorne qL1alificacio1i, the same formal types 111a11 arise on differet1t hierarchic le,rels. The ·Beeth.oven theme given in Example 47 is a clear case in point. On the first level (~),both the antecedent and cons~qt1ent phr~tses are bar-forms. Unless, like son1e

cheorists, 011e considers a so11a:ta-form mo,.remei-1r tliat repeats tl1e exposition to be a har-fo·n11: Exposition - Exposition - Development a11d Re·c-apitulation A

A' B or one follows Alfred Lore1iz in discovering large-scale bar-forms in Wagner•s operas, sucl1 for111al ~1:ruct,ures are usually low-level; corllined to melodies or pa.rrs of melodies, On tl1e second level (b), the firsr pare of Beetl10'\1e11's the1ne is an ai1tecedent-conseqt1ent phrase. SL1ch f orn1al t)'P~ do at ti1nes for m the highest level of a compositio11- tlSually ratl1er short: ·o ne'5. Cl1opin's Prelude in E Nlinor Opus 18 No. 4, \vould be an exa111ple. l n this c~ ic1cidentally, the n.vo phrases are themselves essentially processi,re- withot1t significant in~ ternal formal structure as i11 Beetl.10,1 e11's tl1e111e. On tl1e higl1est: level (c) , tl1e scherzantto section of Beetl1ove11's rno,rerner1t .is a rounded bi11a:ry fon1l : •• A-BA. This is perhaps the most comrr1on form in to11al rnus.ic. It is the basis of cottn.tless nlelodies and themes \\rithin so11ata-form i11ovem.ents, rondos, minuettos, t he1nes a11d variations, and so 011. On lugher levels, elaborated by subsidiary formal events, rou.11ded-bir1ary orga1rizacion is t11e has.is for most dance forn.lS o.f the eighteer1th ce11tury and of sonta-form

itself. T he Beethove11 then1e is, of co11rse., pan of a 1noven1erlt \.Vhict1 is tert1ary, A-B,-A, in forn1. T.l1ough first arid last parts are related to one another by re·petition (that is, by confonnance) , da capo srructures such as this, lilce strophic structures suci1 as tl1e the1ne and variations, are essentially forrr1al, not processive,, on. th.e l1igl1est: le,rel. T·his is the case, n.o t 011ly beca1.1se clear closure and satisfactory stabili1:) are established at t l1e end of each part~ but 1


HIERARCHIC STRUCfUMS

97

because the parts themselves are not processively related: neither implies tl1e Otllet ill a, fl.1nctional way, as is ttle case, for itlStanc,e, '\Vith the developn1e11t section of a sonaca-f.orn1 mo'\1ement. For these reasons, even though. co1ifor111a11t relacio11ships ·n1ay furnish rhe ·comfort of familiarity, may act as an aid to n1emory, or may provide the pleasure of recognition and psychic economy, the parts of m-ost 111t1ltin10\1 ement \\rorks-syxn11l1or1ies and sonatas, dance suites and so11g cycles, or collections ·of pieces in tl1e same genre like preludes, ettides ai1d so 011-are related to 011e an:ocl1er i.t1 a, f<)rtna! \V-ay.•

i1any forms fall bet\'\'een the purel}r processive and the exclusively f.orn1al. Ritornello m,ovemen~ rondos, fugues, and characteristic pieces such as rl1apsodies, noctur11es1 and so on are nor, as a rule, ltlghly arched. For unlike sonata-form movenlents, they co11rai11 no centrally processive pai't. Nor, on the other hancl, are they strictly ac.iditive arid formal. Fo.r though :basically on the same hierarc.hic le·v:el, the several stable pans are processively connected to each orher by less sta.ble modt11atory episodes. Such sr.rucnires might be called co11junccive flat 11ierruvchies. Because th.ey contain processive parts, such conjunctive hierarchies caJ1 easil)' beco111e mo.re l1ighly arcl1ed- as in 01any of Bacl1's fugues or in the so-called sonata-rondo. Finally, falling ben.veen the formal and the processive> are contrapuntal forms whicl1 are mixed. For instance, in a gr·ound bass aria, such as the final lament in Purcell1s Dido a·1ui Aeneas, tl1e bass is strophic a11d additi,re, while the \TOice and upper instrun1e11tal parts are significantly processive. 111 canons, the iniitacive relaM tionships between the voices tend to preclude rnarked fu11ctional di1ferenriatio11 .bet\1reen parts. Yet cano11s a:re co11rinuot1s, i1ot ad.dirive. An interlacing of complementary, but like. strands of music creates wbat I ha,~e cilled a braided hierarchy, v.rhich is processive and. continuo1IS, yet at the same time quite £lat--'in the foregrot1nd.11 3.

The distinction ber,veen form and process is important for the analysis and classification of hierarchic structures and because they ar rimes function independently of one another. As a rule, formal organization-the part-wi1ole relationships of a con1posicion-and syntactic r>rocesses supiJort and co111ple11 \~ith in1purt:ant exceptions, most forn:1al tyI'>eS of organimti-011 can the111selves becon1e puts of larger w hnles. For instance, a canon n1ay be part of an essentially additive whole, such as B,ach,s An of 1J1e Fugue; or it may be part of a. prooessive scructuret like tl1e opening then1e of the lase 111ove1rlc.t1t of Franck's Soruita for Violin and Piano.


EXPLAINING 1\lfUSl C

me11t 011e another in tl1e articulacio11 of n1usical structt1re. For u1sta11ce, at the end, of the consequent phrase of the Beethoven the1n.e in Example 47, rhytl1n1ic, rnelodict and har111onic processes re<1cl1 clost1re ~111d at the san1e ru11e a clearl}r shaped formal e'\1e·nr comes to an end. In the second !1alf of the theme, .ho\vever>tneasu.res 9-12 are processi·\ e ax1d unsta.ble botl1 rt1ythrnicall)r a.nd 11a,rmo11ically. Their process carries o,rer into the next fot1r measures. But tl1ere is a clear formal tiivision between tl1ese pares. Tl1at is, the last four measures constitut·e a f or.mal entity, but are at the same tin1e the contin.uation of a process generated earlier. Because this ex.ample n1ight b.e ~1nalyzed in terms of the alternation of formal and proc~~i\re functions \Vithin a hierarchy, rarher than in tetms of tl1eir biftirc~1tio.t1 or division, it mayr see1n some\vhat equivocal. Let us, therefo1·e, consider other exa1nples. Example 5 3 is tl1e beginning, tli.roi1gl1 tl1e first beat of meJStu-e I 2 of Bach's Fugue in. F iVlinor fro111 tl1e seco11d Book of the '~' e11-1~en1pered Cla\1ier. Fro.n1 a formal poinr of ,,ie\v the subject is a bar-form__;not a onepart form like the subject of tl1e F-Mi11or Fugue of Book l {E.'X.arriple 50) . As the analysis indicates, the firsr event, ./' is an amphihrach group 1

[J2,}

\

""

.

.

-

ii: e:

"") Iii .

71!

lasting for four eighth-11otes. And so is the secon.d e·vent. These together imply that rl1e .next group "vill be t\:\~ce as long-a gest·u,re of eigl1t eigl1thnotes. Ancl this is b·asically "v l13t foJlo"''s, except chat the g:ro:i.1p httS no clearly defined closure. The n101nentt1m. of tl1e 5'ixteentl1-notes cor1tinues on into n1eas1ires 5 a.n d 6, despite the entra11ce of the answer on the la·st eigl1th

of 1ne.asure 4. ~1.elodically,

tl1e subject con.tai11s tl.rrec distinguisha:b]e stra11ds, as tl1e anaJ}rsis over the example sho,,vs. The first of tl1ese, graph ~ in·volves a morion fro111 F in rr1eaSltre 1 to the E in n1easure 2 \vl1icl1 is left uriresolved tu1til the ai1swer e11ters in 111easirre 4. {Notice that in lil{e ·manner~ the C to B of the ar1swer i111plies the second entrance of tl1e St1bjecr in 111easure 1i .) Srrancl b consiSts of tl.1e ur>l1eats to measures t and 2. Becat1se they are fu11ccionally parallel and conju1lct i11 pitcl1 . they are per·c·eived as a latentt subsidiary .Patter11 \:Vliose Iir1ear co nci11uation, pro·bal)ly to Eb, is i.n·iplied . .Like the F in Example 5 0~ t lie C and D b are potential strt1ct\tral to.nes. T }1ar is, their 111elodic pron1ine11ce st1ggescs tl1at they sh-0L1ld. he important strtlcturally; btit beca.use they aI·e rhythmically \veak, that ft111ction is denied them. Co.nsequently, they are iinplicar.ive of a stt~l1ctllftll tone vvithin the


HIERARCHIC STRUCl'URES

99

."-'---a u ·_....V,., .J:,,f.,. W----~''• IL.#

..

us;z

'><

""'

I

': -

-

_ V4 ' a ,i=, ""!

I •·

, - lp

•

. •

'

7

Example 53

linear pattern alre'1dy ge11er~1red. Thar implicaci<>n is r:ealized their potential is actualized-by the Eb in m~sure 5. Bach reii1forces the Db in measure 3 and moves tl1rough Dq in n1easure 4 to the Eb. The m.ain melod.ic line, strar1d c, consists o·f tl1e first beats .of ti1e measures, each of w b.ich is prolonged~encially for th.e whole measure. Thls line moves fron1 F to E, then, tl1rough a skip of a din1inished fifth, to Bb and Ab. And here ''theoret.icall)r•' the fugt1e subject ends. Except that it does11 t. Tl1e n1elodic process gene.rated in the first four 111east1res continues on as the Ab is f ollo·\ ,red by a skip of a fiftl1 to Eb \-vhicl1 111oves by conj'u.t1ct mor:io11 to 1


IOO

EXPLAINING &1lJ'SIC

the D. The E

is in1portant not r11erel)r becat1se it is the highest not,e and

accompanies the first acce11ted beat of the ans\:ver, but because cl1e main line 'O·f st1 bstar1rive cones ar1d the li11e of pote11tial tones co1iverge on that pitch. Though the melodic process ,of the subject co·ncii1t1es throt1gh t11easure 6, and possibly beyon~ the entrance of the ar1swer creates a. clear and unan1bigt1ous fo.rnr.il arricruation. This is the beginning of a ne\.V shape, a11d of a ne"v process. For n,vo n1easure.5 measures s ai1d 6-the co,ntint1arion of subQ )ect and tl1e a.ns\ver rnove ir1 parallel rnorion. The11 tl1e processes generated ir1 tl1e anS\\ter becon1e rt1e gt'liding force as both voices move seque11tially t0\:\ a.rd the cadence in 111e.asw·e t 2. 12 "fhis is a clea,r instance of tl1e bifur.catio11 of form ~nd process: a. second f orr11al-processive e enr begins before a previousl}' generated one 1'1as been coinpleted. What are called. elisio1is i11volve th.is l
1

1

n1e(iiant (C rruno.r) leads to the subdomi11t1n.t. 'T he r11ain inotio11-from

Ev

to A j,-leaves the harn1(11uc s:iruacior\ oper1 a.t1d n1obile. The liarrrionic in1plication9 of the iVla.esroso are ambig. t1ous. Rh.Ythmicallv, the first level-that "' of tv;10-n1easure groups-is croch.aic, as the sforza1idi of n1easures 1, 3 and. 5 ~·11ggest. 13 Or1 cl1e ne,~t le\ el, tl1e pairs of 111easures for.n1 a dactyl groliping, bt1t one which, becat1se of rl1e trill and the harmot1}' i11 measure 6, is open a11d n1obile. As a r esult, the sixt·h n1eastu-e also functions as an t,\Oacrusis.

.

.

vVhatever the an1biguiries of rh)rthn1 a11c1 harmony m-ay be, the 1nelodic structure is clear a,nd i1al1)able. As the a.n.aiysis over Ex.an1ple 54 i11dicat·es, it co11sisrs essentially of a tria.dic n1ocion., Eb-G~Bb, \vhicl1 i1101ires to C in the sixrl1 tneasure. Thot1gh the thoo.retic:il-psychological basis for the c,011te·ntiot1 12

Here, too, there is a dJ,rision between form and process--.tlbeit a slight one. ThougI1 the harn1ony and. tl1e process of tl1e lo\\ er n1elodic lir1e e11d. 'vith the firSt beat of tl1e subject n1 1neasure z1, r:he fourtl1. species counterpoint implicit in rhe upper tine conti11ues to the second. cigl1th-11oce, th:e F, i11 n1ea.sure l l--'Jnd perhaps beyond. 18 I suspet."t ti-mt tl1e sforzandi 011 tile syncopateci quarter-notes in measures z and 4 are tl1-ere in order to prevent the final eighth-n<>te.~ in these 111easures from fn.nctionliig as upbeats. 1


~

•.-

I,._•= '

... I

- J

-

..

l •

LI

1·-

I

'

' ~·

,.

I

•

..

. I '

~

L

r-. ,

-

...

I

I ,

..

n I 'P

~

,

'

•

r

I

':>

>

l

I

I I

' '

'

'

'

I

f

. '

'

'

'

· ~·

I

)

~·

I I

I

-

H
.

I

' . '1r' ,..,.,_

...

~.,,., ,. ~

'

l

....

,

'

'

. )

.,. .

I

t

'


EXPLAINtNG l\i1USIC

101

c:a nnot be explained here,1 ·t triaclic patterns can ft1nccion in t\\ro ways. Tney can impl}r their Ov\rn contintiation to a point of relative stabilit}r, usually• the octave abo,.re the first n.ote of the pattern-.:in chis case, the high Eb. And triadic patterns can also funccion as disjunctions (gaps to be filled) implying co11j11nct 1110tiot1 hack co the 11ote fro·m which they bega.n-in this case, the lo'\\1er E . Boe}) irnplic1:1tio11s can be actualize~d, as they are in this quartet. Tl1e u11plicatio11S gen.crated b}' tl1e tti.-itiic process of tl1e ~'1.aestoso are realized i11 the Allegro '~'h.ich folJo,vs. The lovely lyrical melody \Vruch begins the Allegro, cfescends conjunctly to the A b in .n1ea.st1r.e 9. It is temporarily deflected u.p to the B it1 a kind of spurious semicader1ce (tl1e harn1011y, I<.:: , is the don1mant of Ab and tl1e r11elodic pitch, }jp, is the fifrh of the tonic,.

Eb),

chat tl1e first fo·ur measures of the Al.le,gro function as a kind of antecedettt plirase. ~fhe conse(1uent follov.ts reaching a tonic chord \i(1ith G in tl:ie soprano. T his "anteced·e nt-co11Sequent' structure is repeated, chis ritne reaching t:he Eb in rl1e upper octa\re in n1easure 1 2 and. realizing tl1e irnplicacion of triadic conrin11ation. But tl1e decisive clost:ire-the end of the fust·cl1en1e group-eon1es 011ly in measure 32, ~rith tl'1e lo'\\rer Eb as the cadential pitch. A11d there, the second violit1 leads 1ine-arly th.rougl1 F ro the to11ic. In short~ the melodic, har.n1or1ic, arid rhythmic processes generated in tl1e iVIaestoso are conti11l1ed anti actualized in tl1e Allegro. No alternative eve11ts a!'e genented. But tl1e begi11rill1g of th·e Allegro, clear!)"' articulated by S<)

tl1e cl1aI1ges in m·e ter, tempo, dy narnics, texrure, expres.5.\011, arid cl1aracrer1 is nevertheless established as a ne\v formal entity. There is a biftll'Cacion of for1n ar1 d p.roeess.

T .hat Beethoven corisiders the .melody ,v}1icf'.t l1egi11s tl1e Allergo as an entity in its O\\.rn right is sl1own in the fa.er t'h at its returns in the recapitulation are not pr eceded by the J\1laestos-0. This inde11ende11ce is n1~1de possible b)r rl1e special nat t1re of the co11jt1nct part of the gap-fill structtu·e. Let me explain~

In 1.11ost gap-fill r l1e111es and they are legjo·n in tonal i11usic-the conjur1cc ''fill'' is 11or divided into paralell ~'trt1ctures such as. anteceder1t-coi1seque11t phrases o.r v\rbat Josepl1 l(erma11 l1a.s called. ''.doublets. ' J. 5 The subject of the D#-A 1inor Ft1gue fron1 Book I of the Well-Ten1pered Clavier is, as Example 55 shows, rl gap-fill n1elody. However, th:ougl1 it cont ains intert1.al 1

14 See t11e d.iSC'l1s.sio11 of "l)isja.r1ct pn·t terns,1 in Part II. and the nt111lys.i.s of t he melody of rhe fourtJ1 1novc:mcnt of Beethoveo*s Fourth Symphony, p. :18. l'J Tlie Beethoven Otlartets ( ..re~v York: Alfred A. l{nopf. 1967 ) , pp. :z.01 - 204.


HIER1\RCHIC STRUCTUll:ES

Example 55 artict1lation-the gap is rene\ved b)' tb.e skip from D # to G#-the pares have 110 rhythmic-structural indepe11dence. The subject is esse11tially a single gesture. On ·che otl1er hand, .if a t1Jelody is divided into parallel structural parts, then tf1e generating disjunction is usually repeatecJ at the beginning of tl1e second part. For instance, in the lVfinuetto from 1\iiozart's Flute Quartet (K.2,98) bo:t h the antecedent and the conseqt1e11t pl1.rases are pre.ceded by a triadic gap (.E."Carnple 56).16

~

I

If

Example 56 lr1 i1either ·of ·these cases does r:t1e conjunct: 'tfill,' itself form a quasi antecedent-consequent doublet structure, as in the descending m.o tion in the Allegro meJody of OpttS 117 . lt is because the ''fill 1' is itself clearly structured that the Allegro thern.e of this movement can become independent of the ~1laestoso· which first generated it. The reason wh.y it does in fact become separated from tl1e ~1laesroso-\vl1}r tl1e Maestoso d<)es 11ot retur11 ii1 tl1e recapit1tlacion- is conv.incingly stttted by Joseph Kerman: ' 1• • •. the iWaestoso r.1ever appears agaiti, neither to it1t.roduce its consequent in. the recapitulation, nor any,,ihere in the coda. It falls victim to the single-n1ir1ded lyric a1nbition The skip from the .first ( lo\'\~e.r) A to the D above it do.es not function as a gap to be filled but as an harmonic interval and a basis for ·ocm\re defi11.icio11. :J.6


E."t'PL.i'\.INING MUS[C

of the inovement as iI \.vl1ole . . . The rnover11e11t lives nor on contrast but 011 the inherent bea11r;r of the conseqt1ent doublet plrrase. t 7 I:n these exa.1n.ples of bifurcation, a pre,riously gen.erared process co11cint1es beyo.nd and transcends for111al arriculatio11. The re\rerse is 3.cl5o possible. That is, process tnay reach cl0St1re before the fo·nnal strt1cn1re has bee11 comp.Ieted. Echoes and extensions, for i11stance, often serve to i1ormalize the morphological length of a pl1rase or period after caden.tial clostrre.18 Since thi frorn D voHk s Piano Qi.rintet, Opus 81, begi11s v1rith the four-btir n1elody gl''en it1 Exar11i:>le 57 . T f1e harmo11;r r11oves .fron1 tl1e tonic to an altered subdon1i11at1t \\rl1icl1 is followed by the don1inant .ill <)Ver a tonic ped,al-'and then baclr to the to11ic, and is completecl in measure 3. So are • )!

•

,,

, v -

' Ex11mple 57

both melodic and rh}rthmic processes. Because the melodic accet1t is suppressed at the beginriing of n1eastire 2, as \~rell as for harmonic reasons, r·he second n1easure functions as a "''eak grottp linlcii1g me.as11res 1 and 3. !\tlelod= ically, tl1e seconcl mea.stU'e lea.ds back to, and etnpha..~izes, the Ctt a11d Fit previously prese11ted. \¥ha-c process tl1ere is, is clear! y closed b)r the erid of measlire 3. But the 111orphologic·al ler1gtl1 estal>lish·ed by the opening two-

measure group-at1d '\-Vhich in addition is nortnaci,re it1 tllis style-c.ills for ft11other nleasure.. To 11or1naliz.e the length, a fourth n1eaS1.ll'e-an echo- ic; added. The ecl10 is r1or part of the process; it is an aspect 011l}r of forr11. 111 oth,e r \\ ords, it1 this case form trat1scends p1·ocess, while i11 the Beetho en an,d Bach examplest process transcends form. Bt1t botl1 types i11,rol,1e bifurc~tion. 7

Beetho11e1i Quurtets, pp. 204- 105. B)' "anrecedene,1 Ke..riruin 111&lns- the Maestoso ·" gap structure; by ''conseqltent,u he means the desce11di1i-g conjunct fill." ts Prolongations are discm.-sed in P-art ll, pp. 126-241, 11 The

11

41


mERARCH1C S'rRUCTURES

105

I should like to close on a more ge11eral 11ote. The basis and nature of hier:archic structures is of more t han purely ,musical interest. In a climate of discipli.11ary diversiry· and specialization s11ch as the one ·we li,re in, the need tor COlllIDOU ruufy.ing COilCeptS· CrOSS-disc..~pli11ar}r fiel{iS-lS pressing. The young in particular, I think, feel that kno,vledge is f rag1nented and disjointed. One featuie which almost all (iisci1,lines l1ave in con1mon is that rhey are conce.r ned \\~th the analysis and u11derstanding of hierarchic structur·es. In the ph)rsical \VorlcL scientists srn(l}; relationslups ral'1gi11g from. the level of microparticles to tlia.t: of the cosmos.; biologists are conc:erned with levels of organization running fron1 microbiology to the ecology of the planer; in rl1e social sciences, the hierarcl1y includes a span ranging froin the ps}"Chology of the indi,ridt1a.l ro rl1e beh.avior of nations and culrures. Witllin the hunlanities, naming o:uly the rr1ost obv"'ious fields, literature and t~iStor}r as well as 1nasi:c are riecessarily concerned \vith hierarclU.c stri1ctures. l\411ch V\rorlc needs to be done. For even in the physical and biological scietlces, ,tl1e precise -u ays in which luerarchies arise and levels interact are not fully lmown. In the social 1

sciences a11d the humanities onl)1 the barest beginning has been made. But there is no real alternative. For most of us at least, the pate11t diversit)r of the world \\111 11ot be made co·mprehensil)le by the transcendental ·visions ·Of mysticism. Nor will it be united by trying to make the humanities more scientific., in the sense of striving for exh.ausrive systematiza.tion or exact quan.t ificatiou. Particularly, if one is. concerned \vith the explanation of particular instances, tl1en, as I argued in my first chapter, the exhausti\:re is impossible and the de.fitutive unattainable. Different disciplines and diverse conceptual fra1ne11;vorks will be b.rot1gl1t togetl1er throu,g h caref ul inquiry into problems and modes of organization '\Vhich are really comn1on and shared. 1ne nature of hierarchic structures. is certainly ttn area of such commonality~


·· wo ...

41

~

m . _· icat1on in ...

ona .


CHAPTER V

---·-----'-.-.. . . . . .,.-.-----y-·--lntroductio,n 1I

_

I.

Understanding music is

simply

a. matter of attending to and compre-

hending tonal-temporal relationsl1ips. ho\\r.ever sub·t le and. con1plex tl1ey may prove to be. Two kinds ·o f relationshlps-conformant and hierarchic onesha,re already been considered. Tllis p·a rt of the book \vill he primarily co11ce:rned with motlier kind of .relarionsl1ip which l sl1all call i'f11plicative. A·nd, as the title of Part Two indicates, the n1ain focus of attention will b.e npo11 ' licanon . as .1t o"Ccurs 1n. . ton al melo dy. unp There

is at present virtually

i10

viable co11ceptual f r.a1nework. for the

anal)~tlc

criticism of melody. .l\11oSt analyses. consist of an unilluminating amalgam of bla.tant descripcio11 (the melody rises to a climactic Fj and d.escends to a cadence 011 B), of rou.tine f orrnal classifica.cion ( tl1e first phrase is an antecedent, rhe second a consequent), and of a naive account of motivic similarity (~be motive of the first 1ne~sure is repeated in the third and is inverted in the seventh me,a sure) . O ne reason for this s-0rry state of affairs is that, at least of late, too mai1y °"'titers have attempted co discuss melody in general. But it is difficult; if not in1possible, to construct a theor}r \vhic'h will encompass tl1e melodic styles of Nlachaut; l\itozart, and Webern- not to cnention Ja''anese and Ja.paJ1ese music, an.d so on. \Ve liave a pretty good theory of tonal har1 .mony precisely bec~t1se we have not a.sl

EXPLAINING MUSIC

Il0

scale structure. Niy concern \vill 'be to explai11 tl1e foreground and its ad-

jacent levels~ This beit:ig the case, it is '''Orth recalling thar, as I argued in Cha1)ter I\T, the concer>tS and principles perrit1enr and illu1ninaring for the anaI)rsiS of one hierarchic level may not be t1seful for tl1c a11alysis of other [e,rels. Thot1gl1 I believe, perl1a:ps fonclly, that the concepts at1d principles de\relo1)ed in this part of che book are genuir1ely illuminating, there is no pre~ tense of completeness. I do not claim tl1a't the cor1cepts developed in this study will be t1seful in the anal)rsis of ail rnelodies,_~·ven all tonaJ ones. Fortunately, one does not l1a\re tO· be able to explain everyr'h ing in order to account for son1e things. Though some aspects of the physical world and the realm of biology ren1air1 to be explai11ed, \Ve 11e\1enheless are cortfident tl1a.t we do. have valid ex11lar1ations o.f ma11y aspects of physical a11d biological events.

The melodies analyzed her·e, are tl1ose v;rl1ich I thinl< I can explain. Beyond this, the}'"' ha\re bee11 chosen because they ill1istrate the conce1>ts ar1d ideas being develope.d as simply and u11an1biguously as possible. In other words, r10 atten1pt h.a s been made to "cover'> cl1e repertory of to11al mt1sic. If there are r11ore exa1nples fro1n th.e music of Be.e tho\ren than from that of Wagner, this does 11ot mea11 tl1ar vVagner's melodies cannot be analyzed l1sing tl1e same co11cepts as for those of Beethoven. Nor is any valt1e judgn1en,t implie·d. It is simply,. that, because their propartiorls ar.e more r11a11.a geable, the rnl1sic of Bach and Beethoven illustrates tl1e poir1cs to b.e n1ade more concisely and, at times, more clero.·Iy. 2.

i:tnplicarive reJacionship is one in '\Vhicl1 an event-be ir a motive, a phrase, and so on- is patterned in st1ch a wa)r that reasonable inferences can be made both about its, co1mections \.vith pr·eceding e\1ents and ahot1t ho"\v the eve11t itself i1ught be conrinue{i and perhaps reacl1 clostire and st~1hiliry. By '~.reasonable i11ferences>' I 1near1 those \Vhicl1 a. co111petent, experienced listener ~o ne fa1nilmr "\lies suc[1 a11d st1cl1, ' mlcing rl\e listener's con1pete11c}' for granted. A11

1

Malcing su.cl1 i11ferences-understanding implicative relationships-is

something '\\rl1ich all of us do mi1ch of the time, in reading novels or histories, in obser\ring changes in nature or in ht1n1an behavior. 1ne ri1mble of distant


INTRODUCTION

Ill

thunder and the piling-up of dark clouds suggests tl1at it will rain. De;nied a sweet; a child pouts and his eyes water-and we SlLrmise that tears will soon f ollo\v. The end of a consequent phrase moves to"rar.d t he cadence and a competent listener feels that the tonic '\ivill prob,abl)r follow. None of these implications may be 1-ealrled. The clot1ds may blow away; tl1e child hold

back his tea.rs; and the cadence ITu'l}' prove to be decepci1-1e. .But this does not mean tl1at the presun1ed consequents were not implied. Only that rhe implicatio11 was not in fact reaJized.- did not happen. As these exa1nples sug~ gest, patternings are implicati\7C sigrzs \vhich experienced observers know how • to interpret. These examples also call attenti.on to t!1e fact that our understanding of ternporal e'rents----0ur co11ception and cl1a.racterizacion of tl1en1 is both prospective and retrospective. It includes both an. aware11ess of what might have happened and our knovvledge, after cl1e fact, of \vl1at actually did occur. T he fact of implication, in 0th.er words, affects our understanding of both the ?.n tecedent and the consequent eve11t, whether the consequent was the one cho·ugl1t to be implied o.r i1ot. If tl1e sto·rmy conditions do not a.ctually lead to rain, then that fact is included in our retrospective understanding of those conditions: they implied, hut '\Vere not followed by, rain. And our understai1ding of the consequent is similarly 111odified: the pleasant day is one which had been threate·n ed by rai11 and is in tha,t respect different from fair days 11ot so threatened. And the same kind of change occurs if the hnplied event does ta.k e place. In retrospect, the stor·my conditions are unde.rStood not only to have implied, but actually to have led to rain; and th.e consequent rain is 11ot something "vhlch came out of the blue, an unexpected squall, but a

possibility implied by antecedent conditions. \¥ e tend to be aware <1f the co11tingency of te1nporal C\~ents. This is spec.ial.ly true \Vheo the antec,edent situation is ambigti.ous, 1-Vhen a nl11nber of alternative consequents seem more or less equally probable, B·ur even vvhen the implications of some sin1acion seem unequivocal, the possibilit)r of alternatives is so1nel1ow present___..;if only t1nconsciou.sly so. i\iloreover, even if the probable conseqt1ent actually occurs, \Ve are freq t1e11tly a·'-''ra.rc that thi11gs might have been othenlr:ise--tl1at it might not ha\re ra_h1ed. that the child might not have cried, that the ca.de11ce tnight have been deceptive. For our proclivity to c,011s1der alter.natives, though sigtlificantly related to the p,atterning .o f the particular event bei11g considered., also stems from a f·u11damental fact of l1t1mrui existence: na.rnely~ tli.e nece.c;sity of choosing. Because hu1nan behavior is not for the most part genetically determined, man must choose.


EXPl.-AL~O

II?

r-.1us1c

To do so successfully, he must envisage rl1e co11sequer1ces-understar1d the implicacions of alcer11aci' e cot1rses of action. Perl1aps for tllis reason con-

sidering alternatives is a deep human habit. The importance t hat we atta,clt to altern,a.tives is sh.o wn it1 tl1e fact that our understanding of a past event ofter1 includes not only our knowledge of \vhat actually occurred bur also ot1r a'''areness of \Vhar might have happe11ed. Particularly whe11 te1nporal e'lents are being ar1al/L.ed, we are prone to make co11t:rary-to-fact statements~to s;iy, for it1stance, that if it had not rait1ed, \ve would have gone ro the zoo; or, if the phrase had not been exce11ded, it \Vould h11ve re:acl1ed a cadence in tl1e eight~1 ineasure. F ro n1 tliis pou1t ·of vie·w, what is of interest about eo·unterfacruals is not their logical, btlt their ps)rchological significance. T'11ey are rhetorical devices \:1;1hicl1, by .callir1g attention to the co11tingenc;1 of alternatives, inodify ow~ understanding of events.

Cle;arly, however, \Ve do 11or alv\'a}'S co11Sid.er all alter·nacives. If \ Ve did, st1rprise \\ 0 11ld be impossible. Unanticipated events certainly occt1r. For instance, a. child ;vho seer11s quietly conter1red all at 011ce begins to \Veep. Searching for a11 explar1ation of this behavior, we recall llis earlier disappoi11tme11t 1

at not receiving a S\\ie,et. At the time perhaps t here was no implicative behavior , or it was not notic·e d ,o r h ilS beeri forgotten it1 tl1e meantime. But no"'' ' in retrospect, \ Ve recognize che relationship bet\ veen antecedent ~ind consequent ev·e nts.. Many detecci ve stories depencl upon our t1ot i1oticing or nor grasping tl1e implications of eve11ts e-arlier recou11ted. Only in retr(>spect do we understa11d tl1eir sig1Ufica11ce. Tlus sort of retrospective u.nderstandi11g is the basis for '\-vha.t has been called ''The Aha! t> or recognitio11 experie11ce. R.e trospective t1nderstandir1g is an aspect of musical experience as \\rell. Because patterns are seldo1T1 sin1ple and ''sjt1gle-rnindecl," alternative C0·11seque11ts or co11rinuarions are as a rule implied by musical events~ Someri[nes, howe,,.er, one alternative may be so strongly itnplied that ot hers are obscured and tend to g'O t1nnociced. At other cin1es, '
tenrening events. In. cases lil
be understood largely, perhaps ei1tlrely, in retrospect. Such retrospective u11derstanding ,of implicatio11s at first n1issed or but dimly divined is in part .responsible fo.r our feelin,gs about propriet)', tlnity, a11d even nece~-siry in a composition. 111ay


113

Nlost of the time a pattern can be fully comprehended and its internal relacionsl1ips analyzed only by seei11g "''hat follows from it. Criticism based upon this sort of m.eth{}d may see111 a kind. of ''post hoc, propter hoc'' ~ray of r easoning. D o·u ,btless there is so111e danger on this score; and we rr1ust, tl1erefore, ~1ork to r efine rheor·etical f ormulacions, st ylistic concepts, and critical methods. But I see no alrerna..tive to retrospective analytic criticism. We understand temporal eve11ts '"'rhether in the arts or the sciences, 11ot only in ter ms .o f where they have co1l1e from and wl1at rhey ar·e, but also in tern1S of their consequences-botl1 proximate: a11d remote. i\s a nun1ber of philosophers of history have pointed out, this is a common mode of understanding and explanatio11 in t:hls discipline too. Though v;re do it i.-v~th an1a7..ing facility and considerable acetiracyt understanding implicative relationships is a con1plex and subtle cognitive activ1

ity. Arid it is an activity of our whole being, not just that artificial abstraccion, the mind. The many facets of the human nervous system, physiological ch,anges and adjustments, motor behavior and rhe like, are all involved. For tlris reason, in1plicativ-e relationsl1ips may be experienced as kinetic tension and resolt1tion- that is, as fee.ling and affect. I~Iowever, tl1ough they can be experienced withou.t explicit conceptualization, implicative relationships cannot be explai11ed \Vithout concepn1alizarion- \vitl1out theories and hypothe-

ses, classes and norms. In1plicative inferences, then,, are like hypotheses which experienced lis.. teners entertain (perh,aps unconsciouslv) about the cotmections beN1een musical eve11ts--past ai1d present, present at1d fucur,e ones-on the several levels of hierarchic: organizati<>11 in a particular movement or \vork. Like hypotheses in other realms, reasonable inferences can be made about relationships only if individual sounds and groups of sounds combine in relatively orderly and regular ways: tl1at is, if they forrn patterns. '

.,1


CHAPTER VI 2 • --.·--,---r--.-............--w----... --I

De· .nitions and Methodology Melod.ies are impliL-ative because they are orderly patternings. Paracloxically, ho'\'\,e.ver, the 111ore regular and orderly a pattern is1 tl1e less cor1sciot1s we are that it is i111plicarive. The implicative relationships a1·e grasped with a kind of intuitive inrmediacy. 011,ly \Vhen a patte.r n pro,res to be prol1lernacic do " "e tend to become co11s·cio11sly a\\rat e tl1at it is in1plicative. 1 For instance, just before the end of the J\!1u1t1etto of I-Iayd11 s ''Lo11don'' Symphony (No. 104), a compelljngly goal-directed process is abr11ptly brol
"''",,. @>~ '

Tl1ere is, I

ri

~

~

~:;r---,---~,

!

think~

no (.i ot1bt wh
fact of implicatio.n. 1 Partl)T for chis reason I have chosen to describe tftis sort of r,elationsl1ip in rern1s of "i1111,Jication" rat11er tb11n i11 tem1S of 0 expectatio11"- the word. rn;ed in E1notion J~teamng

in 1W:usic to der1ote essentially the same son of cognitivt:: bel'lavior. A11otl1er difficuJt}r '\Vitll the latter ter1n is that expectlltions tend to be thought of as n1e.t1tal acts in which a single, exclnsive cor1sequeut everlt is envisaged. Because a par~ ticu41r a11teceden,t e·ve.nc 11iay be related to a n1.11t'\her of altert1:Gtive consequents, 1lnd and

l'>ecaose the paran1eters of n'lusic

111~iy

not act con.grue:nt.iy in the articulation of pro-

cesses and !>trUcrures, the term expect11rio11 is often a\vkJ.vard and at tim.es te11ds ro misrepresent the net of m1.1sicaJ unden.randin,g. For i:astance., to suggest chat: a p

•

DEFINT.I10NS AND M ETffOOOL()GY

One is not, however, rea1ly aware that the events \V.ithin t}1e simple folk tune \vhich occurs toward tl1e end of the Fjnale of Bart6k's Fifth String Quartet are implicative (Ex,ample 59) . Tl1e rnelody is s.moothly linear in the .

1.

~

.

.

~

.l

~&

'~

.z4 .

'

.&=

-,.

--Ji

.

~

JI ,-

S-

;;

4-

•

.'

p

. !!iii

•

.

•

.

"""

.

'

. '1. :

.

.

.

.

. . . .

.

.'

,#'"'-

,

•

'

.

W k.d
'JI

2L

.!

'

•

· ~

.,

7

't

~

JI

'

'

I ,

.

I

.,.

' '

,.,

. .

'·

7/1.

..

••

~

.

..

~

-,j_'

.

'

.

.

.

- •"

':f. '

.,; L

•

·•

• .

..J

'

.

'

,,..

.

~

.

••

...,.

'

.

-·

'

.!

I

•

'

.

.

.

Ls.

~

J

'

.

i

! '

.

'

0.

E."tarnple 59

foreground (graph r) , and regularly triadic on tl1e next }e·vel o.f organization {graph 3) . 'Nevertl1eless tl1e tu11e co11tains s-pecific and strong i1nplicative relationships, as can be easil)r tested by stop1)ll1g the melodic motion after, s-ay, the C# in measure 6. The tonic, A, is clearly implied {grapl1s za and 4). We <'hear'' that the n1elody s11ould desce11d through B ro rhe tonic. J\1.oreover, wlllle tl1e lower A is implied both b.y rl1e rende11cy of runes to ren1rn to starting points, parcici.1iarl)' if they are the tonic, a:nd by the cha.t1g·e .o f melodic direction which. begii1s in measure 5, tl1e high A reacl1ed in measure 1 3 is implied not only by• linear and triadic motions of rhe first two hierarchic levels (t:,1faphs 1 ~ind 3) , but by the n:iorio·n from tonic to fifth (E) o.n the third le\rel (graph 5). gnres rise to simultane-0us expec."tations . see1r1s strange, bu.t to say that a pattern implie.<1 alternat ive modes of continuation is 'Iuice natural. Sin1ilar ly. it is anomalous to assert that harmony leads t.he listener to expect closure \vhile rh)rdhn1 does not, b t1t it seems reo.s:onabl

I

16

EA'l>l..AINING

~I.USIC

This exarr1ple calls atte11tior1 to a Illlmber of irnportant matters of rrleth-

odology: r) Earlier it was st1ggested t l1at ir11plicarive relacionslups are lilce l1yporheses "vhich compete11t listeners en·certain about tl1e connections among J11usical events. To explai11 a n1elody "''lrich such listeners corr1prehend \Vitl1out co11sciol1S e.ffort, t}1e critic n1ust mal{e th.ese in1plicat.ive }1ypotheses expli.cit. H e must discover tl1e patterirings prese11t iti the n1elody, and l1e n1ust speculate-formulate exfJlicit hypotheses-about l1ow eact1 of the patteriungs nlight be continued to rea.c h the stability of relative closure. 1 or perhaps silence: the end of the patterning. To do so, the critic v·v1U often perforn1 a lrind of mental ''experi1nent. ') H e \\rill 'stop' ' the 111elodic .fio\v at particular points and cry to imagi11e wl1at cor1'ci11uacions seem probable. 1"'11is was do11e in ·Connection with the Ban6lc example when "ve aslced w hat ''-'as implied by the patterning up to the C# in n1easure 6, and b)' the morion of the opening 1

measures. c1~icic)s

poir1t of' '' ie\v, tl1ere is a problem to be solved: given some theoreticaJ premises, \\'l1at do ci1ese ii1complete patterns, tl1ese partial eve11ts, imply? Wh.at- gi\ren the style and mt1sjcal cc>ntext, as \1.rell as the struct·ure of the e\,,t11t irself-will probably follo\iv? T he critic ,~,,ill srud;r cb.e composer's score to see whether any c)f tl1e envist1ged ( alter11acive) concinr1ations acti:ially occur. If they do, his t111derstancling () .f the iinplicative relatin1isl1ips discovered i11 the patterning is probably correct. Very ofte11 t11is .is the case, even though the realizatio.11 of an implicacio.n 1nay at times be mucl1 delayed. If tl1e envisaged co11ti11uatio11S do not occur , tl1en tl1e critic's undersc1ndirtg ·Of tl1e 11relodic p·attern n1ay have been incomplete or si1nply mistaken. In such a case, he will restud)' tl'1e melody a11cl attemi:>t, w itl:>out 11todi~ fying his theoretical pre-mises>to formt1Jate an alter1tacive hy potl1esis. Often n ew insights can be gained by analyzing even.ts in .retrospect -in ter111S of Fro1n the

their known conseque11ces. For later e\rents ca11 c all atte11tio11 to aspe·Cts of earlier ones \vhich ma}r l1ave gone t111noticed or not been fUll y appreciated.

Here metl1odological considerations begir1 to arise. Fc)r if tl1e realization of a11 implicatio11 can be sig11ificantly delayed, then in a -vvork of even inoderate dttracion and complexity envis-aged co11seqt1ent events \Vot1ld be ''irn1ally bo·u n.d to occur- as a 111atter of statistical f)robability. St1ppose, for i11sta11ce, tll3t in a movernent in Bb, the opening melod)' is h;rpothetizecf. as implying motion from the fifth (F ) ro the upper to11ic (Bb). B t1t this 1nocio11 does nor occur \vitlrin the melody itself. As the rnovemenr t1nfolds, the it:nplie.d Bb is al111ost certain to occur in son1e 0011text- as part of a scale or arpeggio figure,


DEFINlTIONS AND METHODOLOGY

1 17

a harmony or another 1n.elocly: sim.ply because it will be an in1portant tone in the syr1tax of all closelj., related keys. To a consid.erahle extent, however, this danger can be avoided by stiptilacing as precisely as possible wt1ar the antecedent e\>·e11ts are understoc>d co imply". For the inore precisely the org.anization of each paran1eter in the consequent event is sp1ecified, tl1e less likel)' that an event rnatchirJg those specifications is the result of statistical distrit)ution. Tl1erefore, the implications of an ante cedent pattern should be mt1.;de as exact as possible wicl1 respect to rl1ytl1nuc position, ha.r.monic context, registral placen1ent, and timbre-as well as pitcl1 relationsl1ips. Tl1c possibility of t~1eorerical co11cepts detm11inmg analytic choices will be further diminished if the analysis ~xhibi ts a col1erent ar1d consistent order in which 1 ,0 11e part fits '1\irith and rl1ereby ''c·or1firnis} otl1ers. l\1ore difficul t 1neth(}
will ha\re ro be dealt \\"ith on an infor1nal, individual basis. At this point all that can be expecte·d is plal1sibil:ity. Implicaci
u1 a tonal-harrnonic context \\ l1ich is 1

i1ot tile 0 11e

called for by

the generating ,pattern. 2) More co111n1011 rhar1 the nonrealizatlon of pre·viot1sly ge11era.ted implication is its converse: the occurrence of an event not implied b)r or connected "«ritl1 precedi11g patrernings. "'"1\s a rule, su.ch l.Inanticipated events, taken togetl1er 't\Titl1 prece.ding ones aJ1d witl1 those wl1ich. follow, are understand-


I

18

EXPLAINING M.USIC

abl·e in retrospect as part of a l1igl1er-1evel ordering of e\rents. rlowe,rer J this is not al\lrays t he case. So111ecin1es, as \Vas arg~i·ed u1 Chapter I, an e\re11t can11ot be accot1nted for in ren11S of relationships \\rith precedi11g or follo"\ring events. It see111S convincing and effective, y et ren1ai11S in a sense anorna1ot1s. This is. I rhi11k, the case \Vitll Exarnple 59. The ttse of tl1is sir11ple f <:>lk rune '\c'Vitl1 its regLllar a11d unassurning acco1nparu111e11t just before tl1e close of a 1nove1nenr cl1aracterizeci by consid.erable i11tensit}; and co111plexity i11ust be regar de·d as an arbitrary decision b)' tl1e con1poser. One can suggeSt psycl1ological reasons ad l)oc or1es~ in rtti.s case-;,vlly tl1e passage 111alces 111usi-

ca.l-aestl1etic sense: a release fron1 rl1e tension of syntactic complexity, it (at tl1e san1e time) checks the for\\rard motion of tl1e mt1sic for a brief moment, rnal{ing t11e subseqt1e11t rest1mption of ter1sion a11d the dri,re to\vard final closiue particularly forceful ~u1d decisive. And l)eCal1se this 'tcon1posed fu-1r1ata~' .is not unlike places in other n1over1rients, it is stylistically l111derstandable as vvell. Bt1t to attetnpt to explait1 rhis pass~ige by relating it to earlier evenrs by· conformance, .i1npiicatio11 or formal ordering is, in ·n1y ju.dgnient, to 111~-:represenc its sig11iilca11ce arid to ni.iss the poil1r; 11a1neiy, th.at it is gen1

uinely aberrant and al1ormlot1s. 3) ln1plic..--ations arise becatise patterns are inco1nplete or t1nstable i11 sotne

respect. 111at aspect .of a patr:ern \vluch is tl1e basis for in1plic-ati,re inferences \Jlill be called a generative event. As e,re11ts f ollo~' one a11otl1er ir1 ti111e~ son1e of the implietttions of a pattern may be .rea.lized imn1ediately; otl1ers may be

realized 0 11l)r after other e\re11tS, \\rl1icl1 013)' be in1pliciti\re of alte111ari . e goals, l1ave intervened. For i11sta11ce, the li11ear n1otion of rhe firsr t\WO ineast1res of Exan1ple 59 .is a ger1erarive eve11t, i11'1plyi:.rig co11tint1acion to the E (grapl1 2), ~·luch follows \vithot1t dela}'· 011ce E is reached, a liigl1er-le,rel genera·t ive e\rent-a triadic pattern vvh.ich implies the lligh A- is f{)rn1ed (graph 3). But

the realization of tl1is in1plic11cion is ternporn.rily clelay"ed 11ot <>r1ly b)r the f>rolor1gatiot1 of r:l1e E, bttt by tl1e 111otion .frorr1 F# to E to D t wltlcl1 c hanges rl1e direction of tl1e melody and in so doing i1nplies descending 1norio11 to A. E:ve11ts v.vhich ge11erate sucl1 alternative goals \Vil! be called defiecti'.O'Jls . Beci1u.se rett1.r11 to the pitch of art initial acce.nted to11e~1Jartictllarly '1Vf1en it is the to11.ic-is quite probable, rl1is defiecrio11 i11 a sense onlj: actualizes \,Vhat \\1as already potential in tl1e n1elodic-ronal sttucrore. Noticej too, that the irnplication of motion d0\\111 to A is itself b1~iefiy delayed b}' a SL1hsidiary d.efieccio11-v\' he11 rl1e C~ mov,es back to E irlStead of desce11di11g directly to .A. The opening pattern is repeated, begil1ning ii1 meastrre 9. Thus reinforced by repetition, the second-le\rel generative event is realized as tl1e scalar -


DEFINITIONS .AND

~m·r110.00LOGY

I I9

triadic mocio11 reaches rern1)()rar}r clostl!e o·n th.e high A (graphs 1, ra, 3 and 3a) . Defleccio11s seldom cha11ge the itnplicacions generated by the initial, prin1ary patrernings. 1"hey crea·ce alternative goals wl1ich are as a role subsidiary. "/\ special case of deflectio11 is \:vhar I called reversttl.2 Partict1la_rly when they tend toward unifo.rn1ity, so tl1at llO de·cisive poil1ts of sttuctt1ral stabiliry· are established, patterns develop a strong internal momentum. In such cases. a n1arlred, u11equiv·ocal bre-ilc in process is needed if closure is ro he ·effective an.cl convi11cing. Since as a .rule sucl1 u11ifor1n patter11s take the. for1n of linear seque11ces, re.versa.ls generally involve a skip follo,ved by a c·hange in the direction of the motion- from des.cending to ascending, or vice \rersa. 4) Because the patterns \Vhicl1 generate implications are usually corn.plex, a n11mbcr of alterna'tive consequent events \\rill be il11plied. Indeed, even a single lnoti,,e n1ay in1ply alternative con,tin.uations. For i11Stai1ce, the skip of a third rna}r ft111ction. botl1 as a gap, in1plying motion to the pitch skipped0\1er, and as part of a triad; implyi11g concinuf1'. cion to the fifth or tl1e tot1ic depending upon the harn1011ic context. Because ine1odic events are i1ecessn.rily st1ccessive, implications 'trill be re~ilized one after a,n other. This being so, sorne realizario,ns will aln1osr al'\vays be delayed. Those realizations \\'hich occur before th.e main melodic cad·e11ce or before the end of th~ section con-· taining tl1e rnelocly, will be called proxi'l11ate re,1lizatio11J·; an.d those which h:ippen only after the e\rents generaci11g the1n h.1ve reaci1ed significant closure, \-vill be called rn11ote realizations. At times, such delayed realizations may b:e 'rery r·emote ind.eed; for instance,. i11 ca.ses \Vl1ere t:he ixnplications generated b)' rl1e ope1iing tl1e111e of a sonara-forni n1ovement are not realized unril the £rial coda. 5) Not all 1)atterns1 ho\\rever, are implicative. For instance, realizations wltlch co·n1plete and close a preceding pr'oc·ess ma.y not be so. And this is also crue of many prolongatio11s and extet1sions. Even begi1u1ing eve11ts such as opening themes ma:y be complete, Stable shapes '\\rhich, t11ough internall)r implicative. do not imply pai·ticular specific contir1uations. Because o.f their position at the beginning of a rr10,rerne11t, sucl1 decl1trative prolorigariorlS v.rill, of cotirse, in1ply continuation in the ge11eral sense that more mu.sic is expecredv Similarly, some medial even.ts-such as pnre12theses and internal extensions._ ge11erate no nev\r·goals. Witl1 some reser\vations, it might be said th.a t the folk 1nelody i11 Exa111ple 59 ( t:ogetl1er "''it'l1 its vTaried repetition in tl1e following 2 See

1956 ),

E?11otio12 and 1'rlea1Ji1ig in JW.i1sic ( Chicago: U niversit)r of Chicago Press,

p. 93 and passim.


EXPLAIN1NG l\>tUSlC

I?O

measures) is ~ln interpola.tion vvluch, "\v-hile i11te1tSifying the sense of goaldirected motion by inrerrupring it1 generates 110 alternative, no·vel irnplications. There are also ter111i11al even.ts \vhich_a.re i1ot in1plicative-for insta11ce, the echo which closes rl1e Dvorak rnelod;r (Example 57) disct1ssed at tl1e encl of Chapter IV. Ho\vever i111plicacive its internal processes n1a)r be; to the e..nent that a patterr1 is t1nderstood to be COffif)lete and stable on son1e luerarchic level, it is not irnplic:a.cive on tl1at level. In terms of the vie\.\'point develo.ped in Cl1apter IV, on t.l1at le\1el it is a formal entir,r, not a processi,re one. 6} Jt1st as p~itterns are structured hierarchically,. so are the implications the.y generate. 1.11 Example 59, for ir1sta11ce, rl1ytl1ntic-metric acce11ts on the relativel}r stable to11es of t~1e tonic triad gi,,e rise, as vve have seen, to a seco11d level of orgarlization. Because the E f ollo\VS directly from the preceding linear-triaclic patterning, continuation r:o the ,high A is in1plied. J\ 1ore concretely:: had tl1e relati\re tltliformity· of the prececlir1g patterning been broken, as sho\vn it1 E.xan1t)le ;91\ ., the implications of the second strt1cnual level '\voulcl have bee11 d.iff eret1t. The I·eversal of n1otio11 crec1ted by tl1e ~

..,..,f/:~4.l '

1

l

1

_'

·B•J' jdt Example 591\

skip o.:f a third (D to Fl) which the follo\ving E fills tlot 011ly n1akes the E a more Stable goa1, hut preve:11ts tl1e F# fron1 actin~g as part of the rising line as it \Vas able to do in Example 59. As a. resttlt, the triadic suucture of the first three 111easures 3 no longer in1plies contin\1ation to the higl1 -·., but only descent to the tonic. An on-going follrlilce ·melody 11.as been transformed into a. cadential figure. In otl1er \:vords, beca,use it is not a clearly articl1l~1ted, stable goal, t l1e E it1 Exarnple 59 implies co11t:inua.tio11 to t:l1e upper A. Tl1e prolongation o:f the E does .n ot diminish its. n1obilit)' ; i·ather the emphasis it pro\rides suggests a th.ird level of organization a penrachorda.l-tetrachordal di\'ision of the octave (t\'-E'-)-,..\rhich also implies A''. (graphs 5 and 5a). · orice chat the foreground scalar motior1, \vhicl1 implies E, is supported a11d reinforced by the secon.d~level triadic pattern., and t'hat tl1e triadic pattern, 3 The

E is still a structural

t o11e.

See l>elo\v.


DEFINITIONS AND METHODOLOGY

Ill

implying the upper tonic (A''), js in rur11 supported by ti1e even more fundarne11tal division of the oct"ave.

This discussio11 calls attention to a ;point ·Of importance: a single melody may exlubit different kinds of patter11ings oc1 eacll of its hierarchic levels. Differences in the organization of the several levels of a hierarchy are the rule ratl1er th.an tl1e exceptio11; the reinforcen1ent of lo,ver-level implications . times different levels by higher-lev·el patterns is by no means necessary. . ~t will imply alterna.tive tllodes of co11tir1u.ation. Stich i11elodies "'rill, almost by definition, be relatively complex. (Jr., put th.e other way around: tl1e simplicity of Bartok's folk tt1ne is in pan th·e result of the coordination of implications among its several levels of orga.nization. The analysis of luerarcluc parter.ns and, consequently, of tl1e implicati.,,.e relationships they generate, invol,res a methodological problem of co·nsiderable diffi.ct1lty: 110\>v to establish reaso11ably objecrive grotinds f·o r di&'tinguishi11g sttucru.ral fron1 orna111ental tones on a particular .hierarchic lev.el. \'.;\fithout so111e relatively rig<>rot1s ax1d explicit criteriat tl1ere is a .real danger that theoretica] preconceptions ,,,rill influence, if not determine, ti1e analysis of srru.ctl1ral versus ornamental. tones. And \vhen this occt1.rs, analysis becomes circt11ar and self~confirming. The problem is difficult because the structural importance of a ton.e on

a particular hierarcl1ic Ie,rel depends not only upon its piace and function within the specific seqt1ence of melodic events, b11t also upo11 the particular disposition and interaction of tl1e other paraineters which may be involved · ·rhytl11n and l1arrnor1y, d)rn.arnics and ·t imbre. Because these may not rnove congru.ently and because their relative importance may vary even \Vithin a single composition, the matter is a deti~te one about which competent criri~ may differ. More fu11dament-all)r, to revert to a point made in Chapter I: sir1ce each pa·t ten1 is a particular instance, it is ·doubtful \Vhether l1ard and fast criteria can be devised for distinguishing ornamental fro1n structural tones. The best one ca;n do at present is to suggest reasonable rttles of thumb. The most important ones used i11 this study· a1·e as follo·ws: a. Meter is .regarded as ·che prirn.e, thot1gh by no me.ans the only, guide to structur al importance. Tl1ic; is n.o t t1nre:asonable considering the basic regularity of m.etric organizarion and hArmonic rl1ytlun it1 the style of to11al n1usic. Tones \vl1icl1 oc.<.."l.1r 011 a n1ai11 metric acce11t are ai1alyzed as being on a higher strucntral level than tltose "vh.ich occur on secon:dary


EXPLAINING J\{USlC

Ill

metric acc,ents, with exceptio11s that. are .explaineti on ar1 i11dividual basis. Thus a tone co11tin'.g .on the fu·st beat ir1 % meter is in general considered to be. 0 11 a higl1er level tl1a11 011e '\Vhich. con1es 011 the tl1ird bear. T 011es occtirring on \Veak bears are scrt1ctur~1lly less importanr than either f)f these.. As r11ea.rures grot1p togetl1er, larger 111etric er1tities are forme,d, and rl1ese serve as gtiides for discingl:rishit1g stn1ctural frorn on1arn.e11ta1 tones 0 11 still lugher levels. b. -~, ge11er-al exceptio11 to this rule is t llilt goal tor1es- tones of resolution, lilre the Gin me-asure 4 of Exan1ple 62- are considered to be strt1ctu.ral on the luerarchic le\rel on which rl1ey ~ire goals, regardless of tl1eir metric posicio11, It follows from this tl1at appoggi.aturas, though 1netrically emph.asized, are 'fJ;Ot COilStrued a:S StrUC.."tllral t Ot1CS. 111 at1 a.p11oggiatura nguret cl1e structtiral to·n e is the note of resoll1tion- for Jnstance, irl Example 59A it is the E, not tlie F~ tl'1at co·mes 011 the acce11t, \Vl1ich is structttral. c. Relati,rel)'' regular p·at:te1·r1ir1g is the basis for i111plicacj,re inferences. Co11sequently, somerin1es a note that comes on a seconda.ryr metric accent vvill be assigned the same srr·uctural importance as o.ne coming on a primary accent. Thus in Ex;c1mple 60, the opening n1easl1res of Hay,dn's Strit1g Quartet, O pt1s 50 .o. 3, the G on t he fourth eighth-11ote of n1easure 2 is considered to be 0 11 che Saine structural level as the Ii ..vhich conies on the prirnary acce11t. Ancl tile sarr1e is tn1e <) f the Bb in 111easu.re 3. For tl1e motivic pa1'1lllelism n1akes it clear t l"ta.t the pattern is linear , 1110·\ril1g tltrough t:he scale fr on·1 Eb· to B·b. This analysis is suppo:rted by

" E.~a1nple

•

•

60

the h.arn10.nic changes at1d by tl1e dy1ianuc emphasis provided by tl1e other instrt1n1ents of t he quartet. Bt1 r even. w l1e11 rhe patterning of otller p~ran1eters provi

DEFINITIONS .A NO METBOI)()LOGY

113

even though tl1eir n1etric placen1e11t is different. Thus though the metric position of tl1e perfect fourth (C-F) \vhich hegirlS tl1e sub phrases of Till s tune (L~mple 85) ch.anges., tl1e fourths are analyzed as being 1

srructurally equivale11t. 7) Using the term in the special, analytic se11se suggested earlier, two b'.isic kinds of implicative {'problems'' cao be
functional importance-as was the case \\rttl1 rhe n~10 u.pbeats, C and Db, in the Bach Fugue subject anai}1zed i11 Cl1ap·ter IV (Example 53 ) . Or discrepancy rna}r arist~ becaus·e the event as a '"'hole implies a function not realized when it is first prese11ted; for itJStance, when a n1ovement begins with wb.a t is ur1qt1estioriably a closing, cadencial gesture (see Example r 1 5) . I11com.pleteness may be tl1e result either of the specific patteriung of a particular melod)r1 or o.f the syntax of to11ality characteristic of the style as a \vl1ole. 1"he for mer, "''hich "''ill be Ollt n1ain concern in wl1at f ollo~~t might be called processi\r:e incomp1ete11ess.; the latter, tonal incompleteness. Though ·rl1ese kinds of iinp]jcarive incompleteness often t,"On1plerrient one anotl1er1 this is n.o t al\\'ays the case. At least sorne of tJ1e implicatio11s of a melody 1\J..rill as a rule be realized before ro·naJ clost1re takes place. Conversely·, tor1al goals may be reached, thoug'h som.e of the irnplicatiorts generate
m.clodic process. 011 the lo\1lest ievel, the .conjunct mo·tion from F to D implies c-0ntinuacion to C, as graph 1 indicates. Bt1t this is a subsidiary pattern., as tl1e preced... ing quarter-note motion of the triangle, \.Vl1ich indicates the main metric 4

See, for instance., Examples r:29 an.d 130.


EXPLAINING l\11USIC

levels, h.as n1ade clear. TI1e .m.ore important melodic tllOci<>n is from F to D. Becat1se ir is t1arrr1onized by rhe ro11ic, this rl1ird srrongl)r implies rriadic continuacion to

Bb as an important srrt1ctural tone (graph

However, this possibility is not realized directly. Instead, tl1e D is prolonged~ a.11d tl1is ser,res to establish a higher-level metric structure in which the half-note •heco111es tl1e cl1ief measure of rrmtior1 and a two-m.ea.sure u1ut is t11e tnain morphological ? ) ..

__

I

.

,,

-

i

..

___,

I I

'I

•

J

-

·'

I I

I

E.\':arnple 61

lengtl1. As a result, the Bb in m.east:tre 2 d.oes not have the same strt1cu1ralrri.orpl1ologica1 in1portance-is not on tl1e sarr1e hierarchic level-as the D. For this reason (a11d because of tl1e lack of harn1011ic rnotion) ., tl1e .Bb is only a fl!olJisional, not a definitive, realizatio11 of the in1plicarions gen.erated by the preceding triadic patter11. T .he skill fro,111 D to Bb .has a number 0 £ consequences. It creates a gap (reinforced in .in·ve,rsion ar the end of 1neasure 3), ' vhich implies filling-in motion to the m1ssing tone, C-already i111pJied by tl1e first generative event. Secor1dt it deflects tl1e descending pa,ttern which might, had it been. regular, have continued sequentially as sho\Vtl in Part A of Example 61. And a.s the deflection irself becomes a. pa.ttern, the cornplementary motions sho~'Il i:r1 gniphs 5 and 6 are generated. To explain: jtist as the beginning linear motio~ F-Eb-D, in1pJied C, so tli.e reverse pattcrt1, Bb-C-D (n1.2) implies contin11ation ro Eb and perhaps beyoncl. This in1plicacion, togecl1er \\'ith rhe fact that the Ei, fills the gap from F to D (see grapl1 3), f1elps to n1ake the Eb a con1


DEFIN1TIONS A.N D l\iETHODO'LOOY

vincing ·beginning for the second phrase. In like manner, just as the failing in11llied Bb, so the rising third, Bb-D (graph 6) implies c~)ntinuation to F-a note ¥vluc·h as \\c1e shall see, plays an important role in the rev:ersal and clostue of tl1e tu11e.

third~

(gr~11)h 2),

F-D

The tonal rendenC}" of the D in this context, the opening conjunct motion, and the D-Bb gap in :mea:sure 2 all imply the C rea.cl1e<1 in mensure 4, makir1g it a strong point o.f arrival. Here, a lugi1er-level lii1ear pattern is generat·ed (see graph +), and it, too. implies the torlic. This implication is rein.force.cl 11ot only by the linear patterni11g (Eb-D-C ~ Bb) ~~Inch begins the

second phrase {graph 1) , but by the parallelism bet,veen the openi11g of the second phr'ase a11d that of the first. But .cl1e sequential conformance is broken in measure 6. Instead of skipping d.o\vn to A, paralleling the skip from D to Bb ii1 measure 2, tl1e C is repeated and then skips across tl1e barline to F, after whlch descending con.j1.111ct mo·tion leads towards the ·tonic. Notice that, as indicated in Part B o·f Exa1nple 61. h,a:d the second phrase paralleled the £.rsr exacd)1', cl1e tonic \-Vould havre been reached at the same poi11t in n1easure 8. But had this been the case, die n1omentun1 created by co11siderab1e melodic ttniformit)" a11d. 11ar·allelism wot11d l1ave re.nd.ed to carry the 1notion bey·ond tl1e Bb-perhaps down t:o tl1e lo·'\V F. Io otl1er 't\rords, the complementary motion which follo\VS the deflection in measure 2 establishes F as an alternative, subsidiary· goal ru1d thereby makes th·e reversal of the ongoi11g n1otion possible. Aithougl1 t11e se.co11cf phrase is not eKactly parallel to the first, a kind of higher similarit)" pre,rails: the breilc in phrase similarity (the re_peticion of the ·C ) '''hlch begins tl1e reversal occurs at the same place in th·e pru·ase as did rl1e defiectior1 wlucl1 presage·d tl1e re\ ersal. Fin-illy, it1 this context the gap from C to F i1ot only strongly implies linear return to Bb, but in son1e seri.~e t'surnnrarizes tl1e melodic .n1ocion of the nine as a \vhole. B. The first and last eight n1easui-es of the ('Soldier's l\·1arch" from Schumann s Allnl1:lt for the Y oung are given i:n Example 62. As was the case "With tl1e tune fro1n Petro11.,chka, tfl:e first accei1ted, strucrural tone is the third of the scale. Consequently, frorn a tonal po.int of view, descending n1oci.on to the tonic, G, is implied. This tonal tendency does not go unrealized, bur t he n1elodi.c pa,rte.rning implies altemati•.re goal'i as tl1e patterning in the rune from Petr.oucl:ika did 1.1ot. T l1e main generative event is the third°' B to D, in the. first measure. Two possible con.tinu.ations are it11plied: thirds, particularly ascending ·O tles su,ch as this, may function as gaps m~ru1g a conju11ct fill probable; or thirds may be undersr.ood as part of a triadic pattern, .and continuation to the third note 1

'

1


1z6

~lVSIC

EXPLAJ NING

of the triad

111ay

be :U11plied. \~ih,etl1er botl1 of tl1ese altcr11atives are probable

depend;;, as \'\! e shall s·ee, upon rhytI1n1ic relationships as well as 1nelodic ones. In Scl1un1a11n's melody t11e gap-fill patterning is .realized firs~ 111 retrospect the .D is u11derstood to have bee11 ''proloo.ged.'' .b), the neighbor-note, E, after \Vl1ich the melody descer1ds by co·11junct r11otion, filling tJ1e gap and nio,ring on to ri1e tonic (graph I }. At tl1e e11d of chis n1otion,, as B co.nti11ues down to G , tl1e tonal tencJenC)' of the tl1ird to reach rl1e ro·n ic is reali:led

ri

,...1%

t

.•

- ••

,,,, ' J.

.

I

.

"' •

l . '

'

.

•

~---

·~

"

..

'

•

. ~

ii I '

'

-· I,

'

'

...

,J,

.

•

• •

..

'

~·

••

:f··

.

... ..... •. _ ~- c"' Jtl rf¥2'i ':i ... ,,~..,.

~

.

I

•

.

••

•

.

.

J

'

~

•

'

.· '

.

'.f j;:; I '

~ ,._

•4

...

• .

;,: -•f

.

I

'

f

-:j. ..

13

~;.,,.,. ,-

.

..

'"

'

..

-.

'

d ._

"

I

. -F=

gl

'

.:;:3

'

l

'

•

.. --,

r

~ ~

, .

I

(ii) ~

.

•

'

'

_.,.

·~ '

. ..

j

'

,

.

' •

5 ,- ~ •

I

..

• ' '

·•CJ f

.

.

. ... ,.., •

-

-

...

·~

' '

.'

.

...

.IL .. .

'

'

'

'I'" !

,

.

-

!

~~ ~= *# ;

,

f

Em

•

I_

'

.

-

. I

Exi:1111ple 62


DEFINITIONS AND METHODOLOGY

(graph 2). TI1e seco11d alterna;tive t·h.at the triadic pattern \\rill lJe co1npleted b}r reachi11g the high G (graph 3) -is regenerated by the reperitio·n of the first t'\.'iro measures of the ru11e at rl1e beginning of tl1e seeond pt1rase (grapl1 3a). 'fhe l1igl1 G is presented. Bt1t beca.use it forms part of a Inobile, secondar}1 donlinant, ratl1er tlian a stable tonic liarmo11y,. th.e realrz;arion is only p·rovisional. The resolution 0 f this secOl'ldat)' dominant, in tum, generates a desce11ding conjtrnct motion 1~hich ftll.s tl1e gap that preceded the G (graph 1b). Like rl1e G in measure +, the D in measure 8 is a point of arrival and of relaci\re stability-a structural tone. Loo]rin.g at the nielo·d y in this "vay- in. ten11s ·Of begi11ning and ending str·ucn1ral tones- calls attention to the fact that the .n1elody is triadic on a still higher level. That is, tl1e fuse pluase ino\re.s fro111 B to G, and t!1is patterning ir11plies the low D (not shown in tl1e exa,mple; meast1res r 6- l4) around which the 111.iddle pm-t of the piece ce11ters (graph 4). Tl1e seco,nd plu-ase, moving from the Bin measure 5 to the Din measure 8 (graph 5), reinforces the rising tria.d1c motion generated in cl1e first and fil·th n1e~1stires (graph 3 and 3a). Thougl1 implied from the first meas11re, the reafu..arion ·o f a satisfactory high G-one which is part of conic harmony and which com.es 011 a11 accen.r- is ren1ote. It occurs only at the very end of tl1e piece (grapl1 3) 1 after being once again in1plied by the repetition of the opening measures (grapl1 3b), as the final cadential .n.ote. As Schurnann himself wrote: ''Tt1e be.ginr1ing is the rnain th,ing; 011ce one l1as begun, then the e11d comes of its own accord..,y ~ ]tlSt how important beginnings ma)' be, c.--an be seen by con1paring rhe it11elody of Sci1unla11n's ''Soldier's l\1Iarcl1'' \Vith that of the Scherzo from Beetho,1en.'s So11ata for Violin and Pia110, Opus 24 {Example 63) . The first eight 111easu:res of the. t'~'O melodies are al1nost icier1tical in pitcl1 contour. But the differences, thougl1 see1nir1glj" slight, ·ar:e by no 111eans inconsequential. Consider the first rhytJ111Jic grot1ps. I11 Scl1uma1u1's '~Nfarch," tl1e third of the scale, B, is an a.ccented. strltctural tone \Vhile the fifth, D, \vhicl1 follo\\-'S, is 'veak and inobile. Because the trochaic: gro11p is open and on-going, tl1e fifth of the tonic tri11d it11plies cot1tinuarion. to G. In Beet'hoven's Scher2Jo, on the otl1er l1ai1d, the first 11ote-the tl1ird, .A- is an. upbeat to tl1e filt~ C~ \'\rhich is the first structural to11e. Becar1se the C occurs as the relatively stable go:al of 1

p

'

,

•

Robert Schun1a~ Briefe, Ne11e folg.e ( LeiJ.lZig: Breitkopf und Ha.rte}, 1S86) t p. ).38. 4 'Der A.nfang· i~t die Hauptsaclie; 1'\at n1ar1 angefangent dann kon1mt Einem das En.de \Vie von selbst engegen. n I am grateful to rt1·y daughter, Gulm, foe C3.llil1g dlis qtl.otarion to my attention ~tnd for helpi11g r11e find the title for this 'bc.>ok~ ri


EXPLAI~'lNG

:N!USIC

an end-accented iamb, the tttlrd (A-C) scarcely implies triadic continua· rion. --r11e ope1Ung third funcrioris 01tly· as a, gapi and tl1e i1igh F it1 meastire 6 is understood as defining the area of melodic acti\rity an.d not as triadic continuation. In short, as is the case with couJ1tless n1elodies \Vl'lose first structural to11e is the :fifth of the s:c ale-parcicularly tl1ose '\'\ here tl1e filth is prec,eded by a gap--'1vhat is in1plied is co11junct descending motion to the tonic. 6 Because the strl1Ctl1ral to·ne ti;rhicl1 begi11s ea.ch. phrase in E." tl1e second phrase begins and ends on tl1e fifth (C). CQ11sequently, 110 triadic pattern is suggested; instead. the implicario11s g~11era·ted in tl1e first phrase are r·einforced. 111 co11tr·ast, then, vvith Scl'1umrum s ' 'i\tlarch,', it is not unexpected t hat, articulated by the clear closure of an e11d-ac.-cer1ted rh)rthmic group {measure 32- 35), the final cade11ce of Beetl10ve11's Scher~o is on tlie low F. One ta11gential poi11t l1a\tir1g to do '-''itl1 the relationshiI;s a.rr1ot1!$ S"yI1tax, tempo and dynamics,. ~1n d chara:cter. In a sense, Schumann has \:vritten che Ineter ''agau1st'' t he 11atural 1nelodic-har111onic patterning: ·t he first quanert1ote 'tshot11d'' 112:1,-ve been an upbeat, as in Beetl1oven's Scherzo. rlnough the organization of the Scherzo .is similar to rl1at of tl1e ''~Jfarch'' in the follo'\'\ ing 1

1

1

n1e-ast1res, it invol"·es no atnb,igttity; for, deSf)ite a ki11d of off-beatit1g effect.

tlie 'triple

lJel11lits

altei·11ative r11etric-rhychr11ic patternir-ig. But, except for the fi11al n1easures, rhe l>arlir1e i11 Sch11mai111's ''i\1arch'' riught ha,re bee11 placed one q \larter-note to· the left. I .suspec."t that Scht1ma1111 \:\1rires rhe ''J\lla.rch,~ in t1'1is '\va.y-nialcing the rhyth1nic groups trocl1aic-becattSe to play the piece properly (as notated), considera.ble stress mt1st be placed upon do\vnbeats. Othe1wise the ''natural'' mode o.f organizacio11-an ian1bic grouping-,'i.rill result. T his is specially the case becattse t h.e teinpo ( J=12.0 ) is 11ot reall}r fast. To pt1t the n1atter the ·Other ' "ray around: had the tempo 'been qt1iclc. tl1en n1etric accents \vould ha\.re been strong; and no particular stress \VOt1ld have been needed to make the groups trocru1ic.7 As it is., l1ovve,rer1 t he ' 'Soldier's tvl arch1' must be played \vitl1 sonle\vrui.t exaggerated en1pl1asis t1pon downbeats. ..i\.nd this h.'lS an it11portant effect ui1on cl1aracter. For the music becon1es almost tO
110

Se.e , f:or instance, Example 47. r See the (_llscus.~io11 of tCl:npo and grouping at tl1e e11d. of Chapter II. 6


...c!

-

•

IM ·-

~

""'

\0.

u __,

•

a.. E ~

~

•

-

I

, ,' ' ', I I• ,'

•

•

I

•

• I

'

'•

,

lI

IL.+-~ I t'

~ '

)

I

'

Material con1 direitos autora1s

I

30

EXPLAINING· i\>1US1C

(deciso)-like a small child's self-consciousl)' imperious and os-tentatious wa}' of exl1ibiti11g ''n1arcliir1g bet1avior. '' To st1mn1.arize. I111plicative i11fere11ces are possible because the regularity and orderliness of ~l patter11 sugg~i: }Jroba.ble conci11uatio11s \;vhich rhe con1petent listener ·uc1derstands and \¥l'licl1 tl1e C(Jnscier1rious critic atternpts t·O explain. This matter ctin also be stat·ed objectively- as though styliscicail y coheret1t ancl syntacticall)r structure·d n·1elodies the1nsel,res actually obey ed, or ''behaved'' according to, certain basic pri11ciples. 111 terms o·f proximate realizacior1s, the ger1era1 r ule wot1i(J the11 be: Once estahlislJeci, a patterni1zg te1ids to be co11tmtled until a poi'fzt of reJ,1tive t01t11l-r/rJ}1t/,111ic stability is reached. Prolongations and extension n1ay, as '\ve shall see delay closure; and deflections may give rise to subsidiary or alter11ative goals a.s points for closure. Iri t,erms of ren1ore realizations, tl1e ge11eral rule \~lould be~ C01zti1iuAtio1zs ?xot realized-o·r re11lized 01uy provisionttlly-before ti&rnificft,'JZt p~1ttern closure has take7i place 'Will probably be so sit.bsequ
partic ular ~1ttribt1tes of the pattern irself, but b ) r r1o rms of tl1e particular scyl.e

being employed-its synra;x, traditional f om1s, and cor1\renriona1 scl1emata~


CHAPTER VII -----. ._.--.rt-· --"V"'"-" -...___,.---

Melodic Structures There are but n.vo ltinds of pitch relationships: conjunct intervals ru:1,d disjunct ones. Though this distinction 111ay seem S
generating u11plicative r elationslups. Thongl1 some ${)ft of dassificacion is necessary as a \Vay of ordering the discussion, \vhat is important is the analysis of the individual melody. (~()N~J'UNCT

PA'ITERNS

Llne;ar patter11s are scales- diatonic, chromatic, or some mL"{tUre of ti1ese. The basic implicati,re principle, to repe-at the point n1ade at the end o·f the last chapter, is tha~ once be~ a linear, conjunct motion implies c.o ntinu.ation to a point ·Of .relati\re stabilir;r. To turn. to our fu·st ·e xample. 1.11e ope11ing melody of Nlozart's Diverti1nento in Bb tvlajor (K.z.87; New 2 7 r b) for strings a11d t\VO horr1s consists of rwo similar phrases (Ex~ an1ple 64). As graph 1 sho'\cvs, both phrases begin on ·che .fiftl1 (F), and de.scend by co11junct 11100011 tovvard the tonic (Bb). T houg h it co1nes on a weal< beat, the D in m.easure 3 ( a.nd the one in measur,e 7) is analyzed as


:EXPLAINING MUS1C

132

a strl1Cttrra1 tone both beca11se ir is the resolution of a. cadential harmonic pr·ogressio11 (II j \ T- \ T1 - J') and because, particularly in retrospect; tl1e simplest way of t111dersta11ding tl1e over-all n1oti-0r1 is as a conju11ct pa.ttern. The l l 6-l : -V1 progression in measure 4 is reso},red to tonic harmony in 111ea~\_rre 5. Btrt in.Stead of n1ovir1g to tl1e implied Bb, tl1e melody 111oves baclc to the fifth, as the second phrase begins. Tltls time the authentic cadence is complernented by· a clearl},. end-accented rl1ythm and b,y melodic c1os1.1re, 'vhen the B·b is reached in measure 9.

:-v

1

-

'· -

'"'' ¥'"·

..
'

. J ...-¥, P i\.i -

-

Jiil. '

s:•x

•· :a

"

l. '-"

f -c . ,_ f 5

,

,_. _ _ _ _ _ _ _ _ _ __

rfl1ot1gl1 tl1e in1plications of tile mai11 n1eloclic n:1otion are cle.arf tl1e phrase structure is complex and equivocal. When the n1ovement begins, \:\re assume that the fo·t te cho.rds a.re part of tl1e first pluase. Tl1e cadenria1 progression in i11easures ~ and 3, hovvever,. s.e etns to articulttte a half-phrase "'' hich 'vill be answered by a second t\vo-n1ea.sure group r't!aching closure in rneas\ire 5.

the tonic in measure 5~ the repetition <>f the opening chords rr1~1lces it clear that th.is is also the beginning of ·a new phrase: the motion from ronic co subdon1inant 1nakes the har1nony on-going, tl1e trochaic g.ro11ping creates rl1ytl1n1ic mobility, tl1e melotiic F is implic~tti\re and all this is einphasize.d by tl1e abru.pt ct1ange in dyna1njcs and te>.."tt:ire. Be~.use 1neast1r·e 5 ft111cti<,.ns both as the en.cl of the Ml: pluase a11d the begi11ni11g of tl1e second, it has the eff·ect of an interruption. The resulr is tlliat even though there is a cadence 011 the to-rue, the first plu-ase sot111ds tike an antecedent. When the second t)hrase reaches clost1.re in measu.re 9, Gttr feeling that che basic pl1rase structlire iiS i i is confirr11ed. Bue even in retrospect, the f·uncrion. of the chords in measrrres 1 and 5 is equivocal. Though measure 1 begins the main melodic pattern, it is at the same time se,parate from the basic mor1)l1ologicill strucn1re-botl1 because of the organization of the f ollo\ving phrase and because of the marked co11trast in dynan'lics and teA~e betweet1 Thot1gh harmo11y cade11ces

011

+


MELODIC STRUCTU!IBS

1 33

measures r and 1. Consequently, it is also understood as i11troductory. Tlle witty p lay on function is even n1ore co1nple..x it1 measu,r e 5: it is not only a beginning and ru1 introduction_, 'but also an er1d. Because tl1ese measures are set off i11 dy11a1nics arid tenure, separate·d from what follows synta.c.'tically, and eqt1i\roc~11 func·tionall)r, tl1ey are particularly promir1ent. \.Ve are a\.vare i1ot only of the descending conjunct 1notion from F to Bb w·hich begins here, b.ut of an alter11ari"V'e possibility. As sho~rn in graph t , t he low-level conj·u nct n1ocio11 from F' to G suggests linear continuation to the high Bb. Both. t he tnobility of the trochaic r hythm and the on-going i1armonic motion st1pport the implicario11 of a melodic A 1 bar· monized by a d-0minant, followed by a melodic

Bb complemented by

tonic

harmony. H owever, the high Bb never occurs as a stable goal harmonized by a tonic chord ir1 the exPosition section. Whe11 it occt1rs tO'\\<"ard. the end of the recapirnlacion after measure 238- it follows descending morior1 :f r·orn C. For: t~1e recapitulation begins not \Viti1 the melody we 11ave bee11 discussing, but with the moti,res \Vhic.h f ollotv th.at cadence in measure 9. The opening melody lioes not recur until the coda, where, as Example 65 s}10"\l\'S, it le.ads tlu-ough a melodic A. harmonized as part of a dominant s,evenrh chord, to an accented high Bb the note on \vhich tl1e moveme.n t en.ds. Not onl}r is the co11cinuation tl1e one implied by t he opening ti1eme, but t l1e relationslup be-

"

•

•

I li.'X'.ample 6 5 tween the chords of measures 1 and 5 and this remote realization is emphasized by the fact that in both pJaoes the violins play triple-stops. T hotigh l\llozart,s n1oveme11t ends ot1, t he higt1 B0, t:l1e main linear 1nociot1 - the one understood as being tlJe melody-is w1d.oubredly the descending one. .And if the examples tmalyLed in this book represent a, fair sample, it se,e111s that descending melodies are considerably 1nore cornmon in tl1e reper~

tory of to.nal mt1sic t[tan ones ,,,r}1ose mai11 111ocion rises. One important rea~'On for this is t hat ascendi11g melodic n1otio1is ll'lOlvc a sense of effort and con-


EXPLAINL.""TG MUSIC

tension (probably bec,it1se of 011r O\.\'Il kit1esthecic n1otor bel1a.vior) ; fallil1g melodic lines represe11t r elaxarion and are n1otions tO'\vard rl1e repose required for closure. Partly for tl1is reason r11elodies \Vhich begin on che tllird and the fifth of the scale tend to clesce11d; 1 and, further , gap- fill melodies t1sua.lly begin with a rising skip which is follo\l\ted by descendir1g conjunct • 1nor1on. Needless to ~'a)', there are exccpriot1S. The Scl1erzo from Schubert's String Qtlartet in Eb Nlajor, 01)uS r 25 (E.~ample 66) is a ve11r strilcing 011e. The n1oven1erl:t begins witl1 a cleai·ly defined t'flOtive ( ·tn)- an an1pl1ibrach rhythmic group in "vhich the first Eb, a·cti11g as an upbeat and creating ai1 octave gap, is follo"''ed b)r rising conjl1nc-r inorion, Eb-F, rhat St1ggests tl1at the gap is goi11g to be filled. The linear part of the. 111otive is co11ti11ued '~l1e11 the Fin 111easure 3 n1oves to G .i11 nJeasru·e 4. Tl1e co11junct rishlg mocior1, botl1 \\rithin and between motives1 generates the .irnplic'titions sl1ovvn in grapl1 3. The Eb tttJbeats are imJ>licacive not only because tl1e gaps they create st1ggest linear fill., b11t also because they are poterztial st1'ttct:ztral tones (see graph i ). Though a\irally co·nspicuot1s, tl1e Eb ,.s are \Ve.a.k be~.ts l~1cl{ing structural it11portat1ce. Tl1e incongruiry betiveen funccio11 and prominence is er11phasiz-ed by tl1e reite,ration of tl1e Eb's, \\rl1ich see111 uriaffecred b)' ;:iti.d. co,nsequently, separate from, the risir1g motion of the 111elodic li11e.. \t\Tl1at is implied -111ore strongly vvirl1 each repetition- is that Eb livill becon1e an actualized srructi.u·aJ to11e, as it does '.vl1cr1 it cor1·1es as the dovv11beat in 1neasi1rc 8. Tl1e arri\1al of the Eb in me;ast1re 8 actualizes the precedin.g potential rones a11d is a realization of rhe i1nplicacions of the preceding scale (see gra1)h 2). Bur .it does not constitute a sarisfactor3r realizacio11 ·Of tl1e in1plic~1tions ge11erated by the litiear n1ocion of the first foltr n1easures. For the satisfactory realization of atl imptie:atio11 is gove1~11ed by \"irh at 111.ight be c-alled tf1e law of l1ierarcl-Jic equivalerice: an eve11t is an adequate realiza.tion of an in1plicatio11 01tl)' if it is on a 11iera.rcl·1ic le\rel which is tl1e sa.r11e as,, or more exte11ded tf1an, the level of tl1e patrer11 ''' hich generated the in1plicacions in question. \Vhat is 1·equiredt if rhe irr1plicacior1s generated by the openir1g 'n1&1st1res of Schubert's Scherzo are to be sacisfat~c>rily reafrz,ed, is the arri\ral at a goaJprobably one of the t1otes of the to11ic triad- ,;vhich f ollo,<\rs fron1 and is on tl1e same luerarchic level as the lir1ear patter11 of n1easures 1- 4 (graph 3). ·rhough it contir1ues the morion in1plied l}y the o:pening n1easures, tl1e scale

cor11ita11t

1

1

This tendency is also related to tl1eir proxirr1ity to tl1e lo\ver roruc. T his is obviou . in c:he case of tl1e third. In. d1e case of the fiftl1 1 the thir{i, actir1g as a poinc of il1termediate stability; is proximate a.11-d inHtiences r:l1e probable direction of rnotiot1.


..,.

~·~

• . r

•

•

•

•"'"'

~

:I •

I

.

®

•

H!.!:: L1-

.

• • •

•

.

'

'

-d·

I~ '

'

f

L •

-~

w

~

,

-

.

-

fp.:;:::...

•

•

.

•

-

~

..,

'

•

i

•

•

•

~I

".

-

.

.t~ ~

. ....

•

IJ

t

,_

._

~

~ t: ~

~

-

~

·i. - -

-- ...~ ""p ~

-

,.

~

'

• '

. -•• .. ~; ~I

-

-

.

.. -

,.

•

"

•

I w

•

----

.

-

.

'.

f

• •

i'

1~

~·· .... f

•

j

.

. •

=---

,,

'

~'

Oi' ...

-~

•

~

-

' if::::.

~

-i

• ~

t ~

::.'!

:fR'.

•

'

J t' i

'

f •

~

•

:J

e

:J -

•

•

•

•

...... ·c • •

.. .

'

I

•

-:.

•

• ••

•.

I

• •

-

i

i 1. .

I

I

• •

-

, •

b.t:: __: '--

•

j

1

'

;;' t:i : t< ' - f:l •

.,

.. .5.=B· . .

I

I--!

-

.

-

\ir 4,~

jp '

•

f ,,

t

"T

-

••

•

j

-

•

f ·.

-

•

.

.i .

•

I!

j'

'1 •

•

,

i

,

'

••

-.,

.[

•

-

~

l "

:J

:;... .

.:

-

'1-

..

'

• . ,...

'

•

:i

f

.

~~

~

•

.

•

~

M: i lii.

~ & I

~

'

- ..

•

-

"""'

..... ._._,

.....

'

w ,

,... - - - - .. .

..

-

._

-.

l

•

•

'

'

i;

•

•

- ·~ ' .•

'

. ..

-

-

•

. ,.,B

L.

I-

•• •

~ •

'

-I

(

~

.,

ii .

•

a.

•

_,:

.1 ~

~

.,j

'

~

it'

.

'

::

- '.i s

.

..

.'

.

1

.! ....___. .,

-

-

.

-

-•

.!

.

'

n

..

w

....

-

.

~~~~~~....r:.==========:::::====~ > ~

•

L

•

:U

r

..-

•..,

l

'

......

H

I

~-

•

f1' >-

w,

-

•

ff

-

..

I

"

.

-~

..

'

l

-

'

,,

..,. .. ,.....

...... .I

-- ~

" ~

•

•

•

,:

•

... ...

-,. "

:i

.

••

•

t:l

••

:i.::J

' ~ L: -

'

'

f Exan1ple 66

Mater al corl" d1~c tos autora s

EXPLAINING MUSIC

i11 1neasures 5-8 is not on tl1e sar11.e i1ierarcl1ic tevel as the earlier pattern.!! For in the first fou.r measures, the fundan1ental le·~lel of n1otion is a t'\.VO-measure l1nit, \-Vhile in n1eas11res 5-8 the motion is at l~lSt t\vice as fast. F'o.r this reason, che scale is ortly a provisional realization of t he implications generated b}' tl1e earlier lu1ear pa.tter11. Qost1.re at c:he end of the ti,rSt part of the Scherz.o is de.finire and unmistakal)1e: the "''l1ole movement con.clt1des \Vith almost exactly the same cight -meastire phrase. .Harm<)I1icaUy, there is a full cade11ce. Rhythm is endaccented on the lo\-vesr level and on the highest one, v. hich is a clear a11apest group of z i + 4 rr1eaSl1res. Melodicall)r, tl1e scsJe creates foregrot.111d closure. However (and tlus point \.Vill concern us later), thot1gh it e11ds on the tonic, the linear rriotion is qt1ite t1niforrn- li\ritl1out reversal or significa.nt arricula:tion-and, for this reason., furt her mocior1 seems probable. The.re must be mo~re music. This for t\.\lC) reasons. First, because ·n o digression from ti1e stability of to1lic harmo11y arid col1esive ptitter11ing has taken place. ' hat is needed is the tension of tonal contrast ai1d morivic change; so that the return to stability cai1 create a sense of arri\ra1 and conclusive fi11ality. Second, a11d ou.r primary co11cer11 l1ere: the implicatio11s of tl1e n1ain n1elodic n1otion gene.rated. in the fust rr1easures ha,re been realized only provisionally. .!\dequate realizatior1 is still to co111e. Boc'h these require1

+

ments are· mer in tl1e secon.d part of the n10\1e1.nent. Tl1e li11ear processe..c; of the :first part of the Scherzo a.re contint1ed in the second. As g.raph 3 ii1dicates, the rnelod}r n1oves seqt1encially fro1n the Gth.e poi11t reacl1ed in t11casure 4-t1p to the D in 111easure r 6. In bolh the first

four measures and in these eigh.t, the linearic}r of the \'ioli11 1nelody is supported by parallel motion in the cello. Both lines-and the l1art11onic motion as well-could have go:ne clire·c tl)' fr.o m measure 4 to measure 9. Observe, too, th.at tl1e 1nobility of tl1e pattern is partly a result of the inscability of the se-

quence of first-in,rersion harmo·nies. Tl1ere is a. cad.ence in meast1re 16. But it is by no means .c.i.ecisi''·C. T J1ot1gh the rhythm is end-accented on the high~t level-meas11res 9- 1·6 are .a pyra1r1ided anapest: 2 + 1 + (1 + z + 2 )-tl1e lo\\r-level group is a n1obile a111phibracl1. Harrnonyt too, is on-going: both the laclc of root motion an.d t.he 1r:~-V progression i11 G strongly imply contini1acion. T he lo.w-level, fore2 Ic ni.ight be argt1ed that tl1e 'Vvhole scale-plll"aSe is a prolon_gacion of Et> and tl1erefo.re 011 the same level as the previous patterning. Bt1t in, th::it case~ th.e lc>w Eb i11 tneasure 55 is t:J1e tocie proloc~ged. The rest1lt would be a high-level neighbor-note n1otion, Eb(in.r ) -F(n1.3 )-Elt(m.5 )~ \Vhic l1 is s.c.-u·ccly a satic;factot }' conili1wition for the strongly linear pattern of the first four n1ea.sures.


1 37

1\-f'E LODIC STRUCTURES

ground melody creates a degree of closure., for tl1e skip of a diminished fifth from G to C# resolving to D is a cadential gesture. (As sl1own in graph i, this cadential pattern makes the G, a potential structural tone like the earlier Eb, part of the melodic-rl1ytlmlic strt1cn1re.) 011 the higher level of melodic motion, however, the unifor11lir}r of the c'hromatic line fr·o m B oo implies continuation. Thus, while the cadence it1 rneasure 16 has son1ewhat checked

the goal-directed mon1entu.n11 arriving at a kind of tension within equilibrium, the implication of continued morion is still strong. But the implied continuation of melodic 1nocion from t·he leading tone, D (measure r 6) to the tonic, Eb, does not follo"\v d.ire,ctly. The D is prolonge:d (and thus en1prl!:lsized) from rn.easure 17 to 111e<.:tsure 30 (see graph 3), ib ut in a context tl1at is u:nstable an.d hence itnplicatlve. Har1nonically, tl1e D functions as part of an alien tonality ( G minor rather than Eb). The harmonic .rhythm is. open and mobile because the temporary tonic (G) comes as the weak part of the rh)rthrnic patren1, rather tl1an as a g-oal. In addition tensl<)fl an.d n1obility are a result of d1e "'realcening of the 111elodic-rhythn1ic sha,pe. Tl1a.t is, the reiterated D 's a11d the repeated low-level iambs of the melody do 11ot give rise to l1igher level patterns~" And this lack of palpable patterning also implies change. An octave transfer at ·the end of 111easure 22 brings the melody b:ack to the original register. Because t11e srur1e 111oci,re is repeated an octave lov,rer, it is, I tl1itll{, a legitimate instance of registral transfer. That is, it is legitimate to analy-ze tl1e melodic line as contir1uing i11 the lower octave. Often, however, tl1e situation is 1nore problematic. In generalt analysis 11as posi~ed sucl1 transfers \vith unwarranted casu,alness. This is, in. n1y judgn1ent, a n1istake. For tl1e implicatio11s of melodic patterns a1~e in principle specific not only with 1 respect to pitch-elass (Cts, F~'s, Bb s, etc.), but "'rith respec..-,: to register as well. .At 1ne}tslue 25, D is h.armonized as ti1e tlilld of the do111inant-seve·nth ·of Eb. The fo.reground pattern rises through that harmony to the Bb \Vhicl1 is, so to speak, left hanging in mid-air-without explicit connection w ith t:l1e follo\i\r:ing pattern. Th.e progression through the dominant to the tonic constitutes a reversal o·f the hannonic i'>ro-ces-s and ere-ates high-level closure. 1\11elorucally the arrival of the Eb in ,measure 3 1 represents nor only the arrival ·of tl1e stability of tl1e tonic, but the realization of cl1e i111plication generated in th.e opening n1easw:es of the Scherzo. T he processes begun i

s See Meyer, E'tl'UJtion and Pressi 1956), pp. 160-196.

A~leaning

in Miu-ic (·Chicago: University of Chicago


138

E..n>LAINING l\
(with a11 octave tra11sfer) in li11ear f ash.ion from t he Eb in meast1re z to the one in meast1re 3 1. Since the first part of rJ1e move1ne11t endecl \vith a full cadence\ the realization in measure 3 1 is .remote. l\il elodic and t1arn1011ic closure, as ,;veil a.c; the returc1 to stro11gl)' structr1red rl1ytl1mic organization, are comp]eme11ted by for1n.al clost1re. For the repetition of the there f1ave

1110\ 1 ed

.6.rst part of the Scherzor follovving the harnflonic, mel()dic, and rhythnlic tension of tl1e iniddle pru::tt enhances the feeling of clost1re by pro,riding tl1e satisfaction of return. Bltt che first part is not repe:ited exaccl)' · T t1e Eb 1najor sC'ale (m. 3536), wl1ose linear cor1cint1atio11 ( n1. 7) ir1 tl1e :first part "''eakel'1ed tl1e se11se of closure, is bro1<:en off and 1·ei1la·ced i11 1neast1res 38 and ~9 by a disjunct cade11rial gesture. Th.is cl1a11ge accomplishes tl1ree t11ings. First, it est~iblishes a, co·nformant .relationship ·bet\¥ee11 the cadence of the middle p~rrt (m. 141 6) a11d the final closure. Second, it provides a continu11tiotl: for t:l1e Ab and Bb "''hi ch \-Vere left ''ha11gi11g in 111id air)' at the end of tl1e second p·art and leads them to tl1e ca,de.nce• .t\11d t11ost important of all, by re'\rersing the linear n1otion of rl1e precedi11g n1easL1r·e it creates deci,,'ii. e clos11re. Rotlnded bin.aryr forms, su.ch as this~ raise an in1po:rrant analy·cic-cricical question. If a particular· patterning of events- for instance? tl1e lit1e~1r 1110tion of n:1eastues 1-4 ir1 ·the first part of Sclluberc's Scl1erzo ii1·1plies co11riuuation. \Vhen first presented, \7i"hY isn't the restate111ent O·f the patterning in meastues 29-33 implic~ici\re even afce.r tf1e second ha.If }i.as been 1·ep eated? H ov.v can the

piece e11d satisfacro.rily? Fro111 a11 a·estl1etic-theoretical point of vie\V, 011ce an, implicario11 l1as bee11 adeqt1ately realized, the resolution ·of tl1e tensions of t l1e middle p ar t, togerl1er vvith the sAtisfaction provided l))' return., seen1 to crea.te a psycl1ological sin1atio11 in whicl1 closure cakes precedence over irr1plicatio:r1. Ot1 the practical side1 to er1111l1asize the cloSl.Ir·e of the secor1d part of tl1e forrr1, the composer may" extend t11e final cadence, as l\1ozart does in rhe Then1e of J1is A-1\lf.ajor Pia110 Sonata {see Ex.arnple r.7) . At runes a final coda, follov\ling tl1e repetition of tl1e \,rh.ol-e first part of a (!a c.1po form, serves to create a, sense of finalit)r. Thjs is the case i11 t'he Menuetto 010,rement of Beethoven,s Stri11g Quartet in C J\4{ajor, Opus .59 No. 3. The performer, too, ofce11 helps to make closure clear by slowir1g cl1e cempo or perl1aps cl1angi11g d;rnar1ucs.. A11d, finallyt the co1npete11t lister1er k110\vs, as a result of his stylistic experience, l10\:\1 such fornlS gener·ally '(behave - t hat tl1e secot1cl part of a ro·u11ded, bit'lllry strt1crure is repea.red, bt1r onl}' 011ce. Bec.'lt1se it is tl1e stable goal tO \~ ard wl1icl1 all other tones tend to move, tl1e totlic (in co11trast to the third ar1d. fifth) i111plies tlO part ict1lar dire·c tion


MELODIC STRUCTURES

1

39

of rnocio-t1. Conseque11tly. if the n1au1 melodic patterni11g begins on the to,nic, ascending and descending nlotions are eqt1all}' pro bable-at least in terms of

tonal tl1eory~ But a sense of ,direction t11ay l>e pro'\rided b)r other means. For instai1c.e, a prefatory gap n1ay i11dicate cl1e probal)le direction of mocio11, as the t1pbeac u1 Scllube.r.t's Schen:A> does. Or, the probable directio11 of n1otion may be suggested by otl1er means. Ox1e o·f these is register. T 'he n1elody of the ''Nottur110,, fro111 Borocli11 s String Q11artet in D Nlajor) for instance, begins with the to11ic, A,, played softly l:'Jy the cello (Ricample 6·7 ). The to11ic hannony provides 110 clt1e .as to tl1e p1~obahle direction .o f 1nelod.ic motio11. Because the cello enters on the seco11d beat of th.e i11eastue, ov,e r a g·ently sy11copated acco1n.pani111ent, there is a suggesti.011 of instability. But ttle parricttlar feeling of poigtlant tension, whic.11 tbe slo'>\' ten1po enables us to savor from tl'ie ·very first note., arises ir1 large part because, thoug·h it is rhe tonic, t.l1e A is 11011etheless implicative. 1

,. ~ ~ """"- ~-:

""'~

...

...

_,·· ~ft'·~ '" __..,,,. _..

Example 67

It is so, because timbre a11d register serve as substitutes for tonal tension. Relative to the. O\rer-all range c)f the cello, tl1e A is qt1ite high. Cons·equently, simply from a statistical poi11t of vi~r, descending n1ocion is probable. Had the (fsame'' A been played on a violin-sa.y , or1 the D-string-no clear direction of rrlotio11 would ha\re been li11plie(f. Bur ti1e sense of implie.d direction has a11 C\t e11 1nore im1)ort;1nt kinestl1etic basis. '~'e understand a11d respond to the effort jn, 1 olved in pla}rii1g tl1e {'high'' A with ot1r 'vhole being- l\.rith our bodies as. \Veil as our n1inds. Th.is feeling of tension is heigl1cened by the fact that the 11ote is played softl}' · Let n1e explain. '~High>' and '~loud'' are asso·Cl

EXPLAINING ~tUSlC

140

ister, dynamics, and timbre are, so to speak, protogene.rative. There is a. sense of diret.'tionality, but 110 specific g-oals are defined. V\Then the p·arternirig does generate l)articuJar implications, these prove to be both ·Complex and ambiguotis. The motion froin .l\ , t!irot1gh G# and E, to D implies ()Ontinuatio11 of tl1e descending line; the ascen.d ing pattern in measure 3 st1ggests a retur11 to the to1uc. Arid r1eitl1er of these is clearly dominant. Alloreo\rer,. bec."ause IJlelodic rehir:io11Sl1ips st1ggest ir1odalicy as mt1ch as t:onalit)7, neither of the implications is particularly forceft11. Though the stress created by tl1e gr;1ce-notes to tf1e first be~t of rr1easure :z. helps to clarify· the meter, tl1e.y also h.a.ve r·he effect of malcing tl1e G# seetn orna1ne11tal rather tl1a11 sttiicttual as tll(}t1gh it fu11ccioned fir5't as aJJ. appoggiattua to the F# and then as 3.11 ecbappee. This VleakenJng Of the leading-tone, the dt1ratio11fil e1npl1asis of the E and F~, and the absence .of a strong C:~ defining a tria(lic tonal mocion~all cot11plement one a11other in giv.ing tl1e 111elod)" its modal, quasi-pent1tonic flavor. And the com.bination of alrernative goals \Vitl1 tl1e some\vhat atter1t1ated and a111biguous i1np1ications of modality is to a cor1sidernble exte11t respox1sible for the st1bily static and contern1)larive lyricism of the melody. In .B orodin's melody, as weIJ as in Scl111111a.nn's 'attcrnit-ig irnplied alternative goals. B11t this need i1or })e the case. Different r-r1elodic processes may in1ply a single cor11n1,o n goal. 1"'lu.s is trt1e of tl1e subject of tl1e fugue tl1at is tl1e seco·nd mcJ,rement o.f Handel's Concerto Grosso ir1G 1\1finor, Opus 6 No. 6 (E.IDt111ple 68). ,

'-·..:·

.

•

c...,.,...,,..~ t'Ulc'"'° ,,...... ~"-

""-"°

.

•.

:=iii~ ,.,..4, ·1·e,..:c

'

.

•

...

-

-

-

:

.

Because ic is in a n101·e or less nor111al register for violins, the con.ic on which tl1e s~1bject begins is implicati.\1ely net1tral. The uniformit}' of tl1e


MELODIC STRUCTURES

ir1itial chro1t1atic pattcr11 is strongly iinplicarive of .conti11uationi as graph, 1 indjc-ates.4 After the first measure, tl1is motion is abr11prly brol<.en off, and a econd pattern begins on Bb. T lie forcgrou11d 1nocion of this pattern is patently linear, rno,ring up to Eb. T-Jo"ve,rer though the second level of motion (grapl1 2) -that of half-nc>te groups begins as conjunct motion, implying continuation to D , it skjp from C to E b. The gap thus created not only closes, but revc.rses the preceding rising pattern. Thus both lines converge to a comrnon. goal, D. St1c.l1 co1w.ergence, articulating structural poi11ts of importance, is common in bili11ear 111clodies.1:1 1

As is ofren the case vvith linear, descending fugue subjects, the last note of the subj.ect becomes t he first note of the at1swer. 0 The ans"ver, pla}red by the second violins beginning on D ( n1CaSllfe 3). not only concir1ues bot rene,vs the implications generated by the patterning of rhe subject. The alternative lir1es conv·erge on A, whicl1, l1arrr1011ically" as well as rn.elodically, implies rnotio11 to G-at \vl1ich poi11t the vio.las and a solo cello conti11ue th.e li11ear r11ocion as they restate the subject. Not onl)r can bilinear melodies that con\rergc be disti11guished fron1 those that do not, but bili11ear patterning should be differenrjated from \Vhar might be called ''bjlevel'' ones. The subject of Bach's Fug ue in C# Major from Book I of the ell-1 'empered Cla\1icr is a clear example of a bilevel melody (Example 69) .

'¥

•

'----------......1~...i..........l..._~

As the analysis sho,vs, the t\! o 111clodic strands are parallel, ~titl1 the upper one related to the lo¥ver one by latent fourth-specie counterpoint. Tl1e ''st1spensiotis" are resol\>'ed '\.Vhen both voices move to a c~ . The tonal tendency .of rhe fifth to descend to the tonic is the framework "\'ithin w hich patterning cakes place. The first generati\re event, the slcip of a six.cl1 to E#, both esta'blishcs a secon«.i strand of rnelodic activity and, at tl1e same time,. In th.ese respeccs alld ochers, H ar1del 1s fugu e subject is in rncirked contrast to Borodin's melody. 6 See Exan1ples 70 92, and 10 5. 1 a See Bac.h' Prelude and Fugue in G Minor ( BW \ T 541) for Organ. 1'

•

Mater al cor'l" dire tos autora s

EXPLAl ING l\l!USIC

implies the desce11ding rnotio11 of a fu.I. '"f h,e main ger1erarive eve11r. ' luch tno,res th,e l1armony away fro1n rl1e tonic a11d sets the coL1nrrapuI1tal patter11, is the n1otio11 fro1n G~ to F# iJ.1 th,c lo"''er line, Because the t1pper strand (grapl1 2) is cor1trapt111tally depet1clent upor1 tl1e lo\\rer 011e (gra,pl1 1), ther,e is 110 co11trar)r or obliqtte rnotio,n. Therefore, tl1ol1gl1 botl1 Jin.es r11ove to tl1e sa111c pitch-class (C~), there is 110 11r-ielodlc convergence. Incidentall)r lil{e the Ha11del fugue, this one is li11early continuous: the ans·vver begins o.n the lo\ver C# ('iVl1e:r e the St1bject ended), an.d, mo es dow11 the scale to G#, at '"'hicJ1 piccl1 t11e subject is repeated. Convergence 1na)r not rake place as directly a11d im111ediately as it does il1 Hat1del's ft1gt1e stibjecr (ExaoJple 68). The therne \Vhlch begir1s the exposition section of Beethoverl,s Se,1er1th Symphony illusr.ra.tes tl1is point. 1'"'11e 111elody, given in ExatTif>le 70, consists of at1 antecedent ar1d a conseque11t phrase. Tl10.ugh it is adequate, closure at measure 74 is scarcely decisive. Mostl;r it depencls llpon rhytl1r11ic rela.tio11shi,ps. As the rhytl111uc analysis u,nder the ~xa 111ple in.dicLltcs} e~lch part is e1Jd-a,cce11ted on the pl11~ase lev,el and tl1e cor1sequent phrase is so o.r1 the Jo'\.lest le,1e.l as "''ell. TJ:1e antecede11t phr·ase e11ds on a '\veak heat (B), but. because tl1e weal{ l:Jeat is lo·nger tha11 the acce11t, closure is eni1anced. Tl1ougl1 tl1e chen1e is a.lso closed ~1ar111or1ica11y, it is not forcefully so. For tl1e cacler1tiai progressio11 lacks empl1asis on the subdom~ ina11t. Generally speakittg, ther1, rt1ythrruc ar1d han11onic cl.ost1re are low1

1

level, a11d tltls is the case

~,.,G.£

&r .

~vitl1

1nelody as "''' ell,

.. .... .._ .........

"ii!

~-..

-- -- -

...

I

Example 70

lv.1elodicall)r, t11e foregrot1nd sm1cn1re is col1erenr and con1plete. But tl1e high level .is not. Like t he I-landel thcn1e, this .is a bilinear i11elody. As the graph a,bove ~'!ar11ple 70 sl10\~s, both the desee11ding line, E-D , an cl the rising line, G#-..A\.-B, imply COlltinuation to a comn1on c~. Becat1se it is an appoggiatu.ra, r1ot a st.ructural to11e, tl1e Cl u1 n1easure 70 cai:111ot co1111t as a rea1izi11,tio r1 of the io1plications gerler;:1ted b)r these linear e\rents. T l1e oot1sequent


MELODIC STRUCTURES

ph.r tse r.eaffi:r-ms the .i:rnplication, but the ci in measure 73 is even weaker. T11e in1plication generated by ·che first cheme proper is povverfully reinforced in the f ollo"ring n1e;lsur:es. After a six-n1easure prolo11gatio11 of rhe ronic (measures 75- 81 ), e1nphasizing A and E but not C~, there is a passage 011 tl1e donrinant (Example 7 1) . This n'loves r'h rough the triad to the seventh, D, and in. the same register as tJ1e D of the theme. Tl1is dorninanr-seventh chord, empl1asized by fortissi:tno d )r11amics and a fern1ara, uneqttlvoc-ally implies a resolution to ro11ic harmo11y \i\r:ith C# in rhe lipper liI1e.

.

..

..'

' \

But Clt does not f ollo"'· In fact, no satisfactory, structural Cl occurs until tl1e recapitulation. One can loolc through the transition passage, tl1e second key area, the closing section, and. the "''l1ole of the devclopme11t and not find a single C# \.\ hich earl fun.ction as a stable, goal tone in the appropriate 11armonic context a11d in the right register. 1~11ere are, of <.-ourse, so1ne important 'C#'s. On«~ o-ccurs it1 th.e tra11sicion passage of the exposition sectio11 at n1easure r 1 .1 (Example 71 ) . Howe\rer, thot1gb it comes at the en·d of a sequence, this is not a satisfactor}l" realizacion of the implications generated in tl1c first then1e gro11p. As ·tl1e l1arn1on)r sketched under E.ircatnple 72 sho,vs, the tonal context is nor the proper one. Instead of funetioning as part of tonic harmony. this Ci is the fifth of an F~-t11ino.r triad whicl1 itself ft1ncrions as the subdorninant of C# minor. i\4oreo\1er, this Cl is in the wrong register - an octave lower than that specified by the ge11erating eve11t. 7

Onl)-r

it1

the recapulacion, two i1t1ndred and thirty measures after it

'\-Vas

first implied, is C# realized as a satisfactory sm1ctural to11e. And perhaps beca11se it l1as been so lorig delayed, it is en1phasized again and again. The


EXPLAINlNG MUSIC

im1-1lication is regenerated by the return of the first theme group at the begiruung of the. recapitt1latio11. This time, however, the domitlant~sevet1th ferm,at:at D, is resolved to C~ (Ex.a.rn,ple 73A). And th.e resolution is specially strilcing-mark·ed by the preceding rest, ti1e sL"rteenth-note upbeat, and tl1e change in instrt1met1tation a.nd d)rnamics. Though the melodic C# is clearly str·ucrural- in t he right 1·.egister and part of tonic harmony-tl1e cl1ord itself is a rnobile secondary do·mina11t, \ ! of 1\T.

E~ample 73

When it occurs at the beg.u1rut1g of the second. ntne in the recapitt1latio11, 110,vever, the C# is StrucroraJ a11d the harn1ony stable (Example ·73B) . The plrrase also e11ds on the third. of tl1e scale, but as p:art of a si."X-four cltord. In a. se11se~ tlus is a function it sl101tld 11.ave liad in the opening theme-as is the case, for instance, at the end of t11e anrecede11t phrase of the Then1e of i\if.oza.rt's A-Major Piano Sonata :('Cl1apter 2, &uinple 17). This sL"{-four chord receives its nlain resol\1tion when the Cl moves t11ro·ugl1 B t<> A in me~'lire

346. Subsequently, a strongly in1plicati\1e, se,quential pa.<>Sag,e, r·eaffu·ming C# a.s a goal (Exa:111ple 74) is followed by a cadential progression, N -O- \ f- I (n1ea..sures J.68-370), 'vhich also arri es at a forceful srruct'ural C~. And tl1e movement ends '1vith th.e third of the scale as the .melodic note.

I • '

,,

(:1 t tc.

t°"'

Q..

!°"° - ·- - - - - - - - - ... ff

Exan1ple 74 DIS.JUNC1~

P'AITERNS

Disjunct patterns nia)' be implicati' e in tvvo \Vays~ A disjunct it1terval may be undersrood as a ki11d ·Of i11co·m.pleteness a gap-wluch implies that the note or 11oces skippe,d over ""rill be prese11ted il1 wl1at follo'\vs. 0 .r, when u11derstood as .Pru-t of a triad, a disjunct h1terval-sucl1 as a third, a fourtl1,


145

l\UIJODlC STRUCTURES

or a fif'th may imply continuation .o f the triadic pattern wicil a point of relative sta.bility is reached. A single disjunct interval n1a}' im.ply l>oth these possibiiities,.--as was the case in Schumannts '~Soldier's l\Jlarch1 ' (Example 62 ).

Gap-fill 111elodier Gap-fill n1elodies consist of two ele111ents: a disjunct-in.te.n l'al- the gapand 0011jt1nct i11tervals which fill tl1e gap. As a n 11e, gaps are not larger than an octave. Given this qualification, it is generally the ~se ·tl1at the larger rl1e skip, the 1oo:re strongly co11ju11ct fill .is ir11plied. A skip of a, sixth~ for instance, is more forcef11lly i111plicacive tl1an a skip ·of a third. For the larger the disjunct interva~ the n1ore noticeable the incompleteness it creates, and a triadic continuation of a large disjunct interval is not probable because the. melody \vould be carried beyond the octave,1 \Vhich would be unusual~ particularly in the early period of chis style. Rising gaps are 111uch n1ore co111111on than falling ones 8- probably because it is naniral for the teosjon of effort, associateli '\;\fitl1 both rising and disjllllCt interv'als, to .precede tile relaxatio11 associated '\vith d,escending conju11ct n1orion and with the approach to\v:ard closure. I11 Sl1cl1 rnelodies, it is not c>nly tl1e disjunct gap which ger1er,ates implications. As with other linear patterns, the conjunct motion functions as a gene.ratitre event in its o·urn right a:n.d, once begun, tends to be continued l1ncil relaci,1e stability is reached.. Fin.allyt an important

excep·tion 1nust be mentioned: an upheat inrer,1al of a perf,ect

fou:rth~

tnoving

to the tonic, does not necessarily functio·n as 11 gap, but niay be und.erstood as a rhythmic-harmonic event emphasizing th·e tonic on which the melody proper begins. Let us nO\V consider so111e ~xai11ples. Th·e subject of the fugue fro1n Gemini.ani's Concerto Grosso in E Minor, Opus 3 No. 3, is an almobt archetypal instance of a gap-fill xnelody. As Example 75A sho~rs., the main gap consists of an ocra.ve skip. Because it is an U()beat to the lower E,, the B emphasizes this disjunction. a.nd ar the san1e time acts as a subsidiary gap to the upper E. These gaps .are fallowed by iminterru pted descending conjunct 1notior1 to the tonic. 9 1

Fo.r in:stan.ce, the triadic continuation of a sixth1 E to C, would. cake ·d1e pattern oo the G outside tbe oc.ta\re above the initial E..

s Though there are notable

ex.ceptions~

for instance, the melod;r of Schubert's

Scheno ( E."

EXPLAlNING ~1JJSIC

A.

B.

Exac11ple 7s The only anal ytic problen1 tl1at arises it1 connection \vith cl1is exarnple is that of luera,tchic eqt1ivalence: is tli.e n1otio11 f ollo\11it1g t11e B ir1 ir1.easttre 3 linear or tria.dic? In 01)' judgment, it sl1ould be anal yz.ed as both. Tl1e tones on the primary and secondary ac.c.-ents- the B,. G, and E-are strucn1rally n1ore in1portant than the A and FI v;1 l1ich con1e ·on \Veale b,eats.. On the orher l1a11d, the i1iitial octave gap is so scrongly i111plicative of corr1pletioc1 that the fill n(.>tes acqt1ire a11 importa11ce '\Vhich they \\tould nor otl1en:\'ise f1a,re. fviore i1nporra11t still, the preceding linear motion (D#-C~) tends t<) s'hape our understandll1g of later events; it leads us to hear the conj u11ct rnotion as a conti11Lting patter.11. Consequer1cly, the conjt1ncr pa~t of the pattern seems a sacisfactOr}" realization of the in1plieittions generated by the ante-

cedet1t gap. Because ic is a single e,1en t "vitl1out n1arlced inter11,a l aroculation, this gap (E-E) :is ancillar}f: thot1gh it generates and gives direc.'tion to tl1e follo,vu1g C011jt1nct .m otion, the ''real>' n1elody is Utld:erstood to be cl1e co11seql1e11t

d.esceI1ding patcer11. G:aps can be even 111ore clearly ancillar}r than thisempha.sizing the probable di:rectiot1 of n1otio~ but not really bec<>n1ing part of the esser1tial patterning. 1"'his is tl1e case, for instance, tvit11 tl1e octave llJ)bettt to rhe melody of C11opi11's Prelude i11 E Mi11or;, Opus 28 No. 4, the first n1easUies of wl1icl1 are gi,1en i11 £.,~ample 75B. As the abst1·action give11 abo,re the example shO\l'S~ tl1e essenci~1l melodic motion co11sisrs of a. conjunc.-r descent, del~iJred by repetition a11d di,rersior11 fro111 B ro E . In addicio11 to empli.asizi11g the direction of rnotio11, the gap sen 1 es co define the basic area of melodic activit)r. Partl)r for this reaso·n and partly because of the restrai11ed .n1otio11 of the melody itself, the stretto marl is a11cillary is shown it1 tl1e fact tl1at tl1e accornt)anime11t do,es not begin until the do.wnbeat. ~Thether a ga1) is i1nciUary or is part of tl1e 111aiii ·n1clodic m<:ltion de-


MELODIC STRUCTURES

pends, thet-i, upon \vh.e ther itS r:ones are structural. This distu1ction can be made clear· by comparing rvvo menuetto melodies by wlozart both of \¥hlch begin with. triadic gaps. 1"he first~ gi\1en in Chapter 4, Exa1nple 56t is from

the Flt1te Quartet in A Major. In tins case, the gap is a.n upbe"3.t 'vl1ich precedes the first strucrural tone of the 111elody- the high A. The gap is ancillary, fo.r our t1nderstar1ding of the structure of the co11junct fill, which co·nsiscs of an. antecedent a11d a co11seq ue11t phrase, is not depe11dent upo11 tl1e presence of the gap. T l1ough the melod)' "''ould m1d()Ubtedly have been poorer '"'ithout the ga.p-for instance, l1ad. it begun witl1 a cha11ging-note ll p beat, A-B~ A- the ance,cedent-consequent structure '~1 ould nevertheless ha:\re 1nad.e musical ('sense.'' But in the melody of the !\1.enuetto f ro_m cl1e String Quartet in D Nlinor (K. 42 1), .as Example 76 shov;rs, the ga.p -a D-minor triad-is made up of structtual tones. Because rl1e descending fill is on th.e same hierarchic level as the generatic1g gap, and because tl1e r·e peated A 's in me~tre 3 act bottl as the end ·o f tl1e gap patter11 and as the begin11filg ·o f the conjunct motion,, the g-ap is .~rt of the fnndru:ne11ta1 melodic strucn1re. Tl1e conjunct fill does not form .a n independent pattern, understandable in its O'\vn right-as \\ras the case \vith the ~r1enu:etto fro·m the Flt1te Quartet (also see Example 54). Once be.g un, tl1e triadic motion is implic~rive both of its O\Vn continuation to the high A and of descending conjunct fill. The lower A i11 ineas·ure 1 performs tt'\' O important ftmc'tions. It esrablisl1es tl1e lower .limit .o f melodic a.ctiviry, and thereby makes it pro bable th.a t its octa\re in measure 3 is the up·per li111it a11d., therefore, a r·elatively stable goal for che triaclic mo·rio11. Also, it empl1asizes tl1e disjunct char.acter of the higf1er-level triadic piitte.rn (gra,ph i). Indeed, gaps i1nplying ftlrtl1er mot.io11 occur throughout the inelody: the A to F in n1easure 1, ·t he l(>w-Jevel triad, D-F-A, fron1 n1e-.dsure 2 ro 3, and the falling fourths and risi1ig thirds in n1easures 4 and 6.

Because the descending conjunct: m·otion (gr-aph 1a.) :is botl1 srrongly goal-

E~'PLA11'1'1NG

,. ... . t:""', •• •

. :iii

....

•

I

'

'

•

-·

....

"- ~ ·-

'

' '

;;i

~~ •

. '

f

~USIC

''

'

'

~ • .

' '

: ••

.

'" .

-

~

.. - ~ -

'I

•

l1 :9· '

It....

a.•

I

I

'

I -

I

•

•

:h'

'

1 ,......

'

I

.

I;~

'

f but here the eighth-note morio11 ar the bcgi11ning of the 111easure creates a closed ciactylic gr·o u,p ( J l J J J ) ~ a11d, as a res l1lt, rhythm is consider' -y ../! I ably less 011-going ~ind impJicaci\1 e. The clos11re of the \vhole part is assured by the arri al of the implied tonic, by the gap-fill strt1cture (E-[ A]~Ct co D ) i11 ci1e 1niddle-grou11d r11elody, b. cl-1e 111ocior1 of tl1e tertiary voice (graph 3) to rl1e tonic, and, fina lly, by the disjt1oct cade11cial morion of tl1e

bass. f 11 the preceding exainples, di junctiot1 \\·a explicit and ob\rious, and the

in1plicatio11 generated by eacl1 parcerniog '''ere basically sin1ple. Ho"vever, a gap consisr.ing of strucruraJ to11es 111a)' lJe embedded i11 conjt1nct foreground i11orion, ~11:i.d disjunct parrerns may Le con1plex-impl)ring a number of co11ti(1uatio11s. 1~11e St1bject of tt1e D- li11or l?ugt1c from Book II -o f tl1e ' i\f ell-Tc.111pcrcd Cla,1ier is a case in i1oint (~~xa mple 77). The sixreenth-l1ote triplets at ri1e bcgi11r1i11g create foregroL1nd conti1111ity. B11t they are essenriall)' orname11t:al ratl1er thar1 structural . s sho\.vn ir1 grapl1 r, there is a highcr-lc\1 cl pattern: tl1e eighrl1-11ores on beats 1, 2, a·11d 3 are strucrural to11es ot1tli11i11g tl1c ro1lic triad~ D-F-A, a11d tl1e. c are connected b)r foreground })assit1g-toncs. Tl1e in1portance of tl1ese strttctural co nes is due not only ro

Mater al cor"1 dire tos autora s

lV!ELODlC STRUCTURES

their nJetric position, but to their functio11 in the patter1Un.g: the D is empkasize.d because it is the first 11ote of the :melody; th.e Ft because it begins to repeat tl1e motive just presented; and the A"' becattSC it ends the triplet motive and is follo\\ ed by the first disjunct mocio11in the foreground. Tl1e triadic pattern rrtight seem to lead to rl1e octave, D, on the second half of f)ett 3. Ho't\rever, thoi1gh the D ni.1.lks the exte11t of the over-all motion-the probable area of 111elod.ic activity-it does not follow from or form part of tl.1e preceding tti.1dic patterr1. Becat1se the 0 , F, and A occur 011 rnaU.1 beats, not on \Veak eighth-notes., tlte D enters at the "'rrong point (too early) to .fit tl1e triadic patterning on the hierarcluc level on \vhicl1 it was generated. The A represents tl1e end of tl1e triadic pattern. I ts n1otion is continued. in the nein: measure "vhen A moves to, G and the11 to F.10 The co111plecio11 of this ·C011jw1ct til l does not occur until the penuJrimate cadence of the fuglte - fou.r measures before tl1e end. Th.e high Dis implied .nor by th·e tri'ldic n1otion, bt1t by the less patent gap-fill pattern showx1 in graph 1 . This pattern consists of che series of eightl1-notes "'rhich skip up a fotirth, creating a gap, a:r1d tl1en d.escend a seco11d- irn.plying funher linear motion. There are three statements of th.e pattern; eac1'1 begirming o·n. the last i1ote of the previous 011e~ tl1at is., D-GF, :Unplying continuation .d o'vn to E and D; F- Bb- A, implyi11g descending morion to G, etc.; an~ fina.Jly, A-D-C~ in1plying a still m,o r·e extensive linear fill .. A1:; the conjunct :motion descends from D, and the gaps are filledt the several lines converge and n1erge into a single r11otion. 1

,...,

.o:, -

Example 77

If th.e richness and. complexity of tl1ese i11tertwirring irnplicative suu.crures .is to be perccived1 the fugue should be performed as ''net1trally'' as possible. For instance, it \Vould be a mistake to rl1irik of tl1e second triplet 10 The last .eighcl1-note, A1 in 111easure :a is tiot part of this line, but a11 o:pbeat leading

to

the statement of tlie answer.


L"\PLAIN!NG ~!USIC

as a11 trpbeat to the F, bt1t it \\~ould also be ~rr(Jng to tl1inl{ of it as grouped ~rith tl1e preceding tri,plet thereb)r einph~ing tl1c triadic structure. Because none ,of the subpa,tterns should be tl1oughr of as t)eing do1ninant, 110 special artict1lation or pl1rasing is callee! for. Either of tf1e nvo basic lcinds of patterning embedded i11 this subject .n1ight fonn tl'1e basis of a separate melody, ~md it is instrl1ctive to see "' J1at happens v.•hen this is the case. 1"be first part of the aria ''Se ,,uol ballarei' frorn Nlozarr,s Le Nozze di Figaro, like the patterning of Bach)s fugt1e subject sl10'\vn in gra.pl1 1 of Example 77, is a triadic gap-fill 1nelody. Tl1e first eight 1neasures :are 1nade up of rvvo f our"measure phrases related to each other b)r c onform.an.c e (£,~ample 7 8) . T l1ot1gh each contai11s linear rnotio11, tl1c. lnair1 patterni11g is t1neqtu.·vocall}r triadic (see grapl1 1) bec:ause of tl1e pltrase relarionshi~ps:· the first begins on F and moves to A, tl1e secorid begins on. A a·n d rnoves to C, and tl1e tl1irdthat of the conjunct-fill-begii1s o n C. This !ugh-level triadic Strt1cture is emphasized byt the foregrt1t1nd parter1u11g; lJec'1use the first phrase returns to 1

F an(l the second to A, rhe major-third relationships ben\1een phr~ises is aurally explicir. And because tl1e criadic parrernit1g is so strong, the G in nleasure 3 and tl1e Bb i11 measure 7 are t~ nderstood as high-level pass.ing to11es I";;\tl1er than as srruett1ral ton.es. The triadic, g~tp structure of tl1ese first pluases

in1plies linear fill. The second part ,o f tl1e 111elody realizes tl1is it111)lication.. It begins b)' des,ce.nding conjunctlyt fillu1g the precedit1g gap as far as the G (graph 1 ). 11 Thot1gi1 the descent to G takes only four 01easures-tl1e gap-creat·i ng pa,ttern "''as twice as long- tl1e filling-in is a satisfactory completion fron1 a 1r1elodic point of \ri.e\\1 • For nor 011ly is e-ach of t he co 11ju11<..'t steps explicitly emphasized by rhytl:1m and tl1e articulation of the sequential pattern, bt1t rl1e first

ru10.-beat group in each of these rneasures ( tl1e falli11g third) is related by conformance to rl1e important pointS of structural arri,ral in measures 4 and 8. From. a ternporal point of view, 110\.vev,er, che ciescent is too sl1ort. To bala.i1ce the morphological Iengtl1 of the fu·st half, four m.ore measures are needed, B11c ' ve are gi,ren nvice tl1~1t nt1mber. This is r1o t beca.use the text needs more music-the tvords of meastires 15 and ,16, ''le Sl1.onero, sl>'> are sin1ply repeated in t be last four me,asures. Rather tt1ese ''extra', 1neasures are l l T he third phrase begins \.Vith the sa111e rwo not es, C

A, which ended the second. Ho\vever, though tl:ie inter\ra.I is che sruTie1 th.e rhythmic placement a.nd functior:1 of the 11otes is differe11t. An end has becon1e a begi11rting. When, as .is the case here, a single '{idea" is made to do d o11ble dut)'• d1ere is the plecas\1re of psychic eco.uomy rner1tioned in 0 1apter 3 (pp. 67f.). :111d


,.,.

'

~I~

""

' I

.... '

-·

I I ..

-

f

,_

\

'I

• •

I ·~ rtl

...

I ~~ ~'

Jr«

I I

1

,

1.l,

~

Illa.,.

I

.....

.

... H+tt..I

. •

J

I

-

· ~-

0

~

I

I

,

•

.-;

I

I

,

I •

' ' 'I "'-

• I'll

'·

t ~·

1 ~I .....

-

, I

~

f,;

11

~' 0

I

..

Material corn d1ro1tos autora1s

EXPLAINING IVtUSIC

appropriate because tl1e descending 111otion of measures 9-12 is so }Jate11t and

direct thar a cade11ce afcer 011lv four i11east1r.t~s-such. as is ....l?i\•en in Part B of E~'{arr1ple 7·8-\JlOuld have seemed flat an.d unir1teresti1ig, a11d closure 'vould l1ave bee11 " 'eak and ineffective. Tl1e last eigl1t 1neasures, \Vhlch fro1J1 011e point of \rie\1,r are little more than a prolor1gation of the G i11 n1easlU"e I 2, subtl;r combine an intensificacion of implication "'ritt1 a11 equally effective feelitig of closure. The sense of i11lplication arises nor C)nly because tl1e preceding descent stopped short of its goal, F, b11t l)f(.'"aUSe t.h.e lo"'er-level ech·appee figt1re. sho\Vll in g:rap·h 3, vvottld, if continued., ha,,e ino,red fron1 G-A t (» F. At1d this possible contit1uation of the set1t1ence i11 r11easures 9- 12 is repeatecily suggested i11 t he last eigh.·t measures. Seco11dly, tl1e sl
set of pitch-rime relationships: the tonic, D, ir1 n1easure 44 on a ~reak beat, 111ovi11g to the second degree of tl1e scalel E, in measure ;,'f. ], on an accentan
1-11e first 111elodicall)' generative event is the skip of a. f ot1rt~ from F to Bb. When .14. ,. follo"1-s contint1ed conju11ct desce11t, fillic1g i11 tl1e 111.issing 1


,,

I

I

~

@ L

•l

IC! a

I

.,

l±:W ) '

'

-

•

•

... •

..,._

'

'

I ·J iii

'

...... °" ...... IL)

-x

....0... d

~ I

-....•

'

)1 ...,. •

.++61 ' • ••

I

•J

.r

1

.'''- -

'~

!

Ir ,J

Mater al corl" d1~c tos autora s

EXPLAfNrNG

~1USIC

tone G a11d returning to F, is in1plied (g1·aph 3). This patterning is analogous to tl1e first eigl1tl1-note rnocio11, D~(;-F, i11 Bach's fugue subject. Al> i11 tl1e Bach, this itnplicacion is not re--alized directly. Instead, rJ1e gap pattern is repe~ted: tli.e A skips ·ro Eb vvl1icl1 r11oves dov.rn to D . Conrint1ation th1~011gh C to Bb is i111plied (grapl1 1 ) . 'fl1e second gap-fill patrer·n is rnore f<)rcef11lly

i111plicative than rl1e first, borl1 bccat1se tl1e dir1Unished fifti1 (A-Eb) is l.1arm.onicall)' goal-directed and because the over-all n1otio11 fron1 F to Eb rna]{es a progressior1 to the tonic, Bb, ''ery 1)robable. This ·motion does fo1Jow, as graph 1 indicates. But it: does so only after a dela}r of three eighth11ot·es: tf1e Bb comes O'll tl1e seco11d beat of meast1re 3, ratl1er thar1 ot1 tl1e second beat of me-asure 2 . Tl1is delay creates a [)re~k in proc~~s \.vhich l1as importat1t consequences later in the .n1elody, but it need 11<>t hai'e occurred. TJ1e implied Bb r'l1ight h ~ive follo,ved directl)r, as sl10\.\'n in E.xample 80. ~Had this direct descent to

,,.,, -

11

-I

Bb taken place, contu1uarion of the lin.e:ar patter11ing would ha,re been probabJe. And .suci1 continuation would 11ave led to the realizatic)n of the fi1·st irnplication ger1en1ted: the Bb A of measttre 1 ·would have..n10\red tl1rougi1

GtoF. ' N"l1at is crucial here is the relationshi11 between rh yth111ic structure an.cl n1elodic p:1tter11ing. As tl1e an}1lyses t111der botl1 these exarn.ples show, the groupi11g of the sec·o11d rh)rtl1mic level is an anapest, r + 1 + 2, vvl1ich is c.ompieted 011 the cl1ird eigh.t'h-note of meast1re 2. In tl1e spurio11s v·ersion (Ex"arr1ple So)~ the rhyth111ic closure created by the end-accented anapest is complemented by partial melodic closure 011 tlie Bb-the i111plied goal of tl1e secor,1d geT1erative e\rer1t. Bt1t i11 Scht1bert's rnelody rlus is 1xot the case. The rune is deflected a\:vay f rorn clo~ure by t'he prolo11gatio11 of tl1e D, "'' lucl1 rnoves tlrr<;ugh the Bb niajor triad to the lo'''er F. In other words, ch;ough tl1e rhytl1mic grot1p it1 Scht1bert'..s sor1g is closed ir1 111eastire 2.1 tl't e 1r1elodic pattern ren1aJns to be completed. 1l1e .res't1lt is ·rl1ar tl1e 1l1otive (D-C-Bb). which realizes the im:plica.tior1s noted in graph r, sc>unds both lil

MELODIC STRUCTURES

sounds like an e:nding because from a melodic point of view it is related to and closes the previous phrase. It see1ns like a beginning botl1 because the rh}rthmic closure in mea~ure 2 is quite strong and beciuse the two-rneasure mo.rpl1ological lengtl1s. established by~ rl1e piano introduction a11d the first phrase suggest th,at a. ne\v unir should begin here.12 And for re~sons of morphology, this ~ing meas•ure. This noncongrt1et1ce o.f .rhjrthm and melody is..itself implicari,re, as graph 2 indicates. What is implied is a patterning in '"vhicl1 the n1ocion fro111 Eb do\'\~11 to Bb occllrs 'vithout a defieccior1 or break.. In tl1e last t\\10 events o.f the melody, this implicacio11 is realized: melodic 1notion and rhythmic gro·uping form a single cohe,rent pattern. 1"'he rhythmic grouping potential in measures 3 and 4-an amphibrach on the first level and. a trochee on the second-is actualize.d in rneasures I 3 and r4 (and mea.sures 1 5 and 16), and the echo potenriaJ ir1 th.e earlier measures is made manifesr in rt1e higher-level organization. As Schubert,s marking of pia-1iissitno sh.o,vs, measures 1 5 and 1:6 are an echo of measures 1 3 and 14. On the highest level, th·e second part of Schubert's melody {measures 9-16) 11.1oves in t\<\ro-rneasure u11its. As grapl1 4 of Exan1ple 79 sliovvs, botl1 tl1e prinIDY line, D--C, and the sec·o·ndary line, 'Bb-A, imply Bb as a point of closure. On a lo\ver }e,rel, the G i11 meaStu-e r o and the F ir1 meastlte 1i are reiliz.arioos of the implications. generated by the first melod.ic event of the song-tl\e gapfill pattern, F-Bb· A (graph 3). Because they are g,oal-notes h11rn1onized as ten1porary tonics ar1d because they occur on a luer:a,rcluc level eqtnvalent to that of tl1e generatitig eve11t, ti1e G and F are satisfactory reali• zanons. But this is not true of the n1ain linear n1otion of rhe second l'an of tl1e song. Th.e Bb's in measures 14 and 16 are riot on the sa1ne hierarchic level as the conjurJct n1orio11 fro1n D to C. Nevertheless rl1e melody does achieve satisfactory closure. This is so for five r11ru11 reasons. First, the ecl10 repetition ar the end of the song acts as a sign of relaxation and hence of clostire. Second, t f1e .restatement of tl1e Bb ernphasizes its functio11 as a goil-tl1ot1gl1 it

does not change its hierarchic level. Ftrrtl1ern1ore, the slcip of a seventh, F to Eb, represents ~1 <.."-Ondensation of tl1e n1elodic n1otio·n of the first measure, and measures 13 an.d 14 (and their repetition,) are, in like man11er, condeJlS'acions of 111eas ures 2 ru1d 3. The sense of return thus created e·n hances tl1e 12 And suc!1 lengtl'is are tlortnative

u.1this st. le.


E'XPLAJNlNG MUSIC

feeling of closure. 13 Fourth,. tl1e harri1oruc 1)rocess-the cycle of fiftl1s progressio11 begun in n1easure 9-react1es a f till cadence i11 measures 14 and r6. A11d, fi1ttilly, the high-level rhyth1nic strt1crtue, i11clt1ding the repetition of tlte first four measures, is closed, as Exam,ple 81 sho'''S;· .

I

'

i

, ~

'

!

l.,y~ ,. ,~

-

4 .,. 4 u

,,

'

4YJ• \Y"C

''""}

'

o

Yt ' ,, t' ~ 1 '= -~ r,c(W)1-\~· '9e' ,, ,_...._ '

I

'I .

-f 1

1

,,

'

t

'

;

F

z brs

I

J

l z

"

0 0

4

I

\

1

,,

14

f

'

--h' 'I f..--: nr-

-~ '~' \\--==;:__-·---·- -·-··-·-·'t·r.:..!. ··<~·- - -..-'' "'-"_____,;~ r/ \,

tL

s

,I

'+'. -·

i

14 ¥

1

tt

5

tJ t•~,~ '1t: _t

.•

==----~"-'=:::#~' s af

•

Example 81

T11e a11alyses !'>resented in Examples 79 an.d 81 indicate that rhythmic and n1elodic closure con1plen1e.t1t one a.r1other. The lo\v-le,rel r11elociic sn'Ucrure is closed, but tl1e high.est level-the D tc> C rnorion generate..i:, w h" . . . . l sic .Jc h·' emphasizes t he co.nt1nu.anon o f ·'' wa11der1ng,. t.1e strophjc n.attJ.re of t he song, \\1l1ich must rct11aii1 some'\'hat o·pen to allo"'' for repetition,, a11d the fa.ct that tlus is tl1e first so11g i11 a cycle- then tl1e la.cl< of 11

absolute and unequi · ocal closure see111s entirely appropriate. Other points rnerit consideration in co1111eccio11 "'rirh this song. At tlle outset, two different gai')·fill itr1plications ~1re generated. The secorld of these -tb.e co11ju11ct t11oci-011 fro1n Eb to Bb- is realized, after a brief clelay, before the first cadence. It is proxin1ate. T l1e first generaci\1 e evet1t is realized ren1ocely-in the seco·n d half o.f the melody. 'T htts the realization of gapge11erated implicarioris 11111}'• as was tl1e case v\1 ith li11ear pa:tter1ungs, be significantly delayed. Also, thougli large sl{ips, like octaves ar1d sixtl1s, te:ncl to be inore forcefully in1plicative· than sn1aller ones, even modest disjunctions

may ft111ction as gaps in1plying fill. This is true not o.nly of the fourths '\Vhich begin tl1e n1elodic pa.tter11ing of ''Das \ ¥anderr1Js; it .is also the case '\.vith some of the thirds ir1 the 111.elod1r (Exar11ple 82). For instai1ce, t l1e closure 13 Despit"e cl1e gap of ~ se\'enth, motion b,eyond the B~ is not i111plied. Partly this is becan.~e th.e F can be 1.mderstood as mo,ring " harrr1onically .. ro the Bo-as do.n1i.Ila11t ·t o tonic--mld more importantly because the irnplications generated by a con1patable earlier even't (in measure I) have alread}' beei1 reaiiied s-atisfacrorily, and, as noted. abo'ltC (p. c38) . once a set of it11plicarions ltave heer1 adequately realjz.e d, closure takes precedence over .implication.


MELODIC STRUCTURES

4.

s.

the Bb in measure 3 is enl.1anced by rt1e preceding skip of a third from C to ../\. As g:rnph i indicates, this is a gap- albeit: an unobtrusive one-wl1ich the Bb fills. H~1d the C merely been repeated as a sixteenth-note, closure would have been slightly \\reaker. l\Jfore important: ii1 the seco11d half of the melody, tl'1e ·descen.ding 1notion ·Of ttleasure 9 is given i1nperl1s by tl1e skip from the preceding Bb (grapl-1 3a) .. Similarly", rhe follo"'ing third fro.m A to C implies motion do\vn to F (graph 3b). ·rrurdly, the in1plicacions gen:erated by the opening fourtt1 and those ge11erared in tl1e second half of tl1e melodj" converge on the G and the F (graph 2 and 3a and b) . T lus convergence facilitates the penultimate closm·e. Fi11ally, despite the dominant sense ·o f linearity (st1own in graphs 2 and 3), tt1e 111iddleg:round structure is triadic (graph 4) . 011

Triadic 11wlodies Som.e disju.n ct interv"ttls-sucl1 as tl1irds> fourths, and :fiftii~ are implicative f->eeause~ "vithin the style of co11al n1usic, the}7 are understood as poSr-

sible parts of larger, sy11ractically normative 1>-atter,nings-namely, triads. 111ot1gh tl1ey are unco1nn1011,, otl1er regttlar, disjur1ct IJatterns do oc,cur. For instarice, in rr1easures 5- 8 of Scl1ubert's Scherzo (Exaxnple 66)~ fc>.reground linearity includes a higher-level disjunct motion: the two adj·acenr tetracl1ords, Eb-Ab and B~-Eb {see graph 2) . 1~l1e symmetry of the relationship, reinforcing tl1e tendency of the Eb scale to continue itS linear m·otion, evidently implies the high Eb. However, beca,use triadic patternings are by far the most important in generating i111plicacions i11 the repertory of tonal music, no11triadic disjunctions 'vill not be considered l1ere. The extent i1nplied triadic morion-\\rl1etl1er a .P attern will 111ove to·


LXPl.A,INlNG ~tUSIC

the octave abo\re and, if so; vvhether the octave "''ill be a strucroJ·al

to·1 1e--

depeI1ds upon tl1e parcict1lar str·ucture of the generating ever1r. A pattern tl1ac begins on the root of a rria.d may continue only to the filth, toucb1ng the l.lpper· oc·t~1 e as a ' ¥ay of defir1i11g the area of melodic a,c tivity, but not as a structu1"'al to11e. This; as \Ve l1ave seen,. is the case in 1\11ozart s aria, ''Se \ri.101 ballare'' (Exan1ple 78) . l"'I. oweve.r, .some triadic rnelodies begin on tl1e root a11d imply <.-oncinuation beyond tl1e fifth to th.e tipper octave as a structural ton.e -the 111elody o.f tI1e second n10\1e1nenr of T ele1nannts Suite for Flute ar1d Strings in A l\tli11or is an example. T o· con.sider wh)r the tipper oct ave is irnplied. as a strtlCtl.ltal tone in Telem.ann 's melody~ but not in l\1!ozan's is i1ot onl)7 ii1srrucrive, but a clear instance of the need for a,t /Joe explanatio1i. i\l!ozart>s n1.elody begins with two sirrular phrases: the first on t 11e root .o f the triad, F ; t tl:e seconci on tl1.e tlilrd, 1\.. As a rest1lt1 ·t he tfrird of the tria,d receives the sa111e emphasis, is on the same strucn1ral level, a.s tl1e root. 1"'i1e fifth of t he triad, C, is structurally important 11o t ·only for t11e melodic .r easons co11sidered earlier, but for rl1}rthrnic .ones as ' vell. Because of their filn1iliarit}J', the first t\vo phrases are un.derstood as elen1ents il1 a high-level anapest rhythrn: 4 + 4 + 8 -plus a. four n1easure intern.al eA,'tension. C-0nse'

V

7

7

"'-' -

? __

fl

quently rhe C, \\rl1ich begi11s the ''accented>' phrase, ft1nctio11s as the goal of the preceding 111otio11. Because each ele111et1r in rl1e triadic r11o·tio11 is stable, with C (in n1easure 8) as a point .o f closure, cootit1l1aciot1 to the octav·e as a strU(.'tl1ral to ne is r1o t strongly irntllied. In Tele111a1m's n1eloclyt on t he other ha11d, the tor1es of the triad. are not structurally ·equal (Exami)lc 83) . Ti1e .root (i\ ) ~1nd tl1e fi:fth (E) ~ \ Jt1tuch come on .Prir11ary accents~ a.re on a higl1er leve] than the third (C) . As the proximate goal of both lit1ea.r and triadic p·atterns, and ~:is the acce.nted fifth of tl1e scale, cl1e E is pote11t'ially stable. Bt1t tl1is potential is ciimirristled l)y three circun1Stances. First, because it is rhyt'h111ically '\;\r~k iln(i is x1ot articula.ted as a separable patternii'lg, tl1e C is mobile ·m uch more so th.a 11 the third in ·Iozarr's rnelod)r. Si11ce soine of tfus 111obility is, so to speal<, carried over to the E., chere is a tendetl.C)T for tl\e triad to be cor1tint1ed. Seco n.d, i11stead of closi11g an. end-accented rhyth1n_ic group or beginning a\ ne\v e\.rent, the E 1noves 011 to a weal< be~1t 0 11 the lo\\' A. The stabili ty of tl1e E is so redt1ced t hat there is a strong possibiliry of cortti11uatior1. ·11d third, the Strt1cn1ral in1porta.nce of the. root and fif tl1 emphasize the pentachordal relationships in,1plicit in the root-position tria·d, tnaking ir proba.ble t11at the tipper tetrci-


t-H.I .

.....

'

H.1.1•

....

•!..LJ

u_ _

.~

J.

••

'

'

·tl·

M

.~

•

•

-·

.. -

® ~-

'

'

'

.. ..-.

'

.,

~

•

· ·- ~ ,•.....

@ 'i ~ LI

..

· ~

-·.

...

- -. ..

I

~

~

I

•

I

•'

w.. ,..

•''

,,

(9 ..

"t-

4

t

·~

• "

.. l ...9'\

•

>~~ )

•

I

. J,

. .I

.I I

•

~

•

I

e

I I

,_

·I I . I ... ...._, ' l

I .;. to(

••

J '

I

Material corn d1ro1tos autora1s

i6o

EA.'PL1\INING 1\tiUSIC

chord will lJe c<>n1pleted. For ail tl1ese re'lls-Ons, continuation to tl1e high A is Strongly in1plied. The high A is noc) holrvever. realiz·eti in the first part of tl1e move1nent. l t1stead, the moric>n to the lo'\ver ro1uc., iruciaci11g a series o·f gaps of "vhicli the SL"!tl1, A~F, is t:l1e most i1n,porta11t~deflects tl1e r11otio11 do,Jimv\.ra.rd (graph 2) . This, together ':vith t he te11sio11 created by rl1e internal extensiont leads co co11j t111cr r11otio11 a11d closure on tl1e lovv ../\.. The high A is realizecl 01lly ren1otely- " 1l1e11 tl1e n1elody is restated at ct1e er1d of t11e movement. As in i\i1ozart 's r11elody, the arrival rlf the octave contribta.res to tt1e sense of closure . The realization of tl1e preitio11Sly impliecl A is marked' melotiically b}' the disjunct n1oriot1 ,;vhicl1 surrottnds it. Rl1ychmicall)r it is emphasized because it is both tl1e l)egir1ni11g and e11d of ai1 u11equivocal e11d-accented g.roup.11 Both Mozart s melod)r at1d Telen1a11n's reacl1 the octave abo\re the first suu.ctw'al tone. Bue c:llis is n.ot: alwa;1s the case \Vith triadic patten1s,.._even those that begli1 ';\rith the root, The n1ai11 melod)r of Sn1etana's J7Jtava (The ~1Ioldat1) (Exarnple 84), for insta,nce, is similar in a n11mt)er of ways to the one from Telemann's Suit:e. T11e first structural note, tl1e t:otuc, is foll·O\\red by cor1jt111ct foregr
,..,... c1111•"0>.o ,.....

•

I'•••_,,..... •

..6) . <.

•

•

•

•

. '

··-·

r-t Telenlalln's melody is si11illar, at least in general inoriont to the folk mn~ fror11 Bart:ok>s Fifth St'rin,g Quartet (Example 59). In t11e Qua:rtet., the 111,e lody is repeated beginning o n the higl1 A of meas11re 1; - and agilin beginning OL1 Bb. In this final statement, B:artok taltes care co arti<..·ulate t l1e leading-tone to tonic prog:£le.~or1 so tl1at

clear closure is assurecl. 111 lt1

all these cases the n1atte.r is o ne o f relative probability. It is more probable tha,t Tele1na,n n 1s n1elod)1 will rnove to tl1e upper conic tl1atl. that S111eta.n-a's \WL But Sn1emn3's r11elocJy might ha\-'e done oo anyway. I-lad this been the <..-ase, our undetsmnding .o f the earlier events w.o·u.ld be cl1ar1ged i.n retrosr>ect. Our understanding ,o f t!1e begh1ning of the Telen1ann tune is also changed in retrospect w.hen the implied high A is ren.liz:ed. For the "h)1 poth~is" g·enerated by the op~11it1g p«ltlern has now

been ueoufinned."~


M&LODlC STRUCTURF.S

161

Bec1use of the phrasing and the structure of the acco111paniment figure, the third of the tria.d (G) receives so1ne articulation as a separate event. TI1ougl1 the articulation is not 11early as forceful as .in .f\1fozarrts 1nelod)r, it is so.me~rhat stro11ger tl1a11 in Teler11ann>s. Tl1e fifth (.B) is a. clear stable goal 1 \vhich, instead of being weal{ened by a femj1line e11djng as in Telemann s n1elody, is reinfo.rced t)y prolongation. More-0ver, the neighbor-note pa.ttern -fro111 fi.fth to sixtJ1 to fiftl1 ir1 tl1e rni11or n1ode~1nakes desce11dirlg cot1jl1nct motion probable. Finally, and per1'1aps 111ost i1nporrant, rhe fifth of th.is triadic morion is ar1 octave abo\1e tl1e initial \1pbeat. Conseqrienclyr, "'r.h.en the upper. B is reached, the nlain area of melodic activity is understoocl as: defined, and. contirrt1ation to tl1e higt1 Eis 11ot f eJt to be prob:a:ble. 16

The first .melody fron1 Richard Strauss, Till E1de1ispiegel is also triadic (Exa:mple 85). Like Smetana's 111elody, it begins \Vith a skip of a fourth and n10·1;res through the third. of th.e scale to cl1e fifth; \Vhich is the goal of the preceding t11ocion. However, despite these surface sin1ilarities, the tunes are 1 ob\ri
''main'' pa.rt of tbe melody-the ,descending conjunct mocio11. I t1 the Strauss inelody, on the other ha11d, a triad is tl1e l>asis for tlie patterning of th.e whole event. It is a. note of the triad- the fifth, C-\vhich .is d.elayed. And wl1at the .. disjunct pattern in1plies is not complementary co11junct fill, bt1t conr.inuarion withi11 the triad. Till's tune begins \.vith a skip· of a fourth, C to ·F. 17 The probability of triadic conti11uatio11 is si1ppo:rted by tl1e a.ccon1pa11ying toiu.c harmony, v.rhicl1 is sustained throughout the melody a11d which helps to emphasize the structural importance of all the cones of the triad,, The implied continuation ro A is at once d.elayed and srrengthenecl by tlte ''interposition'' of the G. It is delayed beca\1se the A might ha,re followed th.e F directl)r; it is strengthened because the linear inorion, F-G, is itself .Unplicacive of A. In a similar way, the following GI perfo1--ms a dual role. But he.re, because ·of its greater duration and its function as an appoggiarora1 the delay is so emphasiz,ed that we become explicitly a\var,e of tl:1e fact of in1plica:cion. 1

ie This is m exrunp1e sho·w ing tll:at a upbeat fourtl1 need 11ot function as a gup. 17 As i11 Smetana•s melody, the fonrtt1 is a.dju11crive: it is understood. as part of th~ triad, but not as a gap. GotnpGre these adjunctive fourths \Vith rhe one wl1ich begins the first nielo

E.tU'LAINING MUSIC

L ·

.i""""'~ ii '

I 5 -

'

.

ii

....

-

Once the A is heard, ft1rther contint1atioc1 is forceft1ll)1 in1plied for three mai11 reasons: the urtiformit)' of the chromatic motion creates no t.1ea.r point for closure; the A co1nes 011 a '\ eak heat ~lnd is r~1ythmicitll y i11obile; and ascending triads '\;\rhich begi11 on the :fifth te11d to reach closure on the upper octa.ve. 'NevertJ1eless there is partial closure-becatlSe the A is part of the harn1011)' , which G and G~ are 11ot; a11d becatise the A js follo,;1.red by a marked pit·Ch disjt111ctio11. v\ heL1 the pattern is repe1ted, the triadic in1pressioo is stro11ger. For the F no'" comes on a second,ary accent. Not 011ly are melodic irnplications reinf<>·rced by tl1e r·epetitio1~ btlt a clear sense of higherle'lrel .rl1ytl1tiuc srructltre e111erges, and this, toot is implicative. 1"'11e sirililarity of the t\.VO phrases, a and a', r11alt1tes to tli:e rel~1ci.\re stal)ility of tl1e C, ·w l1ose structural force is also rhe rest1lt of rhe fa.c t tha1: it Ls the first goal tone \~hich cornes on a pri111ary tlcce-t1t. Tl1,e acct11nt1l~1t:ed tension of the asce11ding triatiic motion with its conjt1n,cr delay is, so to speak, :released '1-Vhe:11 r:he C is reached- is realized as cl1e goal .o f the precedu1g p1·ocesS=--and its Ino11lentu1n car1·ies tl1e 'pattern throt1gh the descending triad and beyond the confines of 'vt1at \VOt1ld l1ave been irs iwrrnal n1orphologic~l length. In the ar1alysis over: , xa111ple 85, nvo levels of tr.iaclic r11ocion are distingi1ish.able, thol1gh the}r are not really as distinct ~1s tlle graphs n1ake the111 appear. The first level (grapl1 1ti, b and c} consisrs of tl1e patterni11gs \

MELODIC STRuc·rmms

itnply the C in measlue 1o. Th.e second (graph 2) is a hlgl1er-level, yet some~ ho\\' less forcef1tl, motion. fr,orn A ro C ro F. Tl1e foreground pa·cterning has already been. discussed. The h:igher-level n1otio,11 arises because the A and C a.re co11Siderably more importa11t structl1ra11y than tl1e F's in meas:a.res 6, 7, and S.. T he importance of the C has already been considered. vVhile th.e eighrh..note F's are parts <>f the 1()\V-level triadic motion, they a.re clearly subsidiary pou1ts 011 ti1e \vay to tl1e A.'s. T11e latter are more irnporta11t not only because they a.re goals at1d ends of phrases but because they are em .. phasized by appoggian1ras. Tl1e F's in measur,es 11 at1d 12 are Strt1cturally important for botl1 tonal and rhyth1nic reasons. 18 Despite cor1siderable foreground complexit)'"-rhythmically, becat1se each phrase begins on, a diff.ere11t part of the beat; rr1elodically, because of the ch~1~01naticisrr1 .and tlle long ap1)oggiamras-the Strll'cture of Strauss' melody is essentially sirnple. T 11ere is a singlet basic patterning, a single primary goal. Alt11ough evet1 more patently triadic, tl1e first part of ·the th:eme o.f tl1e slo"v rnovemenr of Haydn's Syn1p.ho11y No. 97 is much more complex. For u11der dec·epri,rely sin1ple foregrot1nd relationships, analysis reveals an intricate J1envorlc of alternative patterrk') ai1d goals (E.x:_ample 86). The first motive (m) is (.'<>njunct on the lo,iiest level, and perhaps im~ plies linear descent. On the next level, the A-F is part of a triadic patter11 implying· continuation to the lovv C. T l1ese in1plicatio11s are not realized until measure 3, \Vl1ere a return to F is follo'\!ved by conjunct foreground motion and second-Le,.rel triadic motion to the C in i11easure 4 (grap'h 1a). Th~ motive is repeated a third hlgl1er (in') , and again., after a modest dela)r; the implied patter.n.iugs are continl1ed-the co11j unct n10tio11 0 11 the G at tl1e e11d of i11easure 2, and the triadic n1otion on the F of measure 3 (grapl1 1b). A third versio11 of the moti\7e., beginning ·on the lugh F, moves triadically frorn cl1e F ru1d li11early fro·n1 tl1e D dow11 to the C at the end of the 11Iirase (graph 1c). All rhree linear and rriadic p~tter1rings con\rerge as they ap-

proach the C. The 1irst nvo statements c>f the rnotive ( m ai1d m') create an alternative, higher-le,rel patterning which is alc;o triadic (grapl1 za and ib). The stru.ccural tones of this patterning rnove f:rom F-A to C. Because it is the chief accent of an anapest rhythmic grot1p, the ,C func..'tions as the immediate goal of tl1e l1igl1-level triadic patter11 (sl10"\Vt1 as un.filled notes i11 the a11al)"'Sis).; Both. th.is lugl1-le\rel triad .and ir:s atixiliary pattern (graph 2.a) i111t1l)r co11tin:uarion to, and completion at, the i1pper octa,re. But arrival at th.is goal is is Because cl1ey follo\V from t11e prolonged to 1tic triad of mea.sare 10, the F's

s:e>Cm to be octave o:ansfers, as sho\Vn in the analysis.


EA'"PLAINING

~iUSIC

dela;Ied u11cil tr1e low·er-le .el implications i.lre realized (graphs 1a, 1b, ai1d c) . The co11ve.rgence of the descer1ditig co11junct a11d triadic i1:atternings in n1easures 3 and 4 is accompanied by the first harmonic changes tl1t1s far. Tl1e subs.eque11t inelodic mocior1 to tl1e goal, C, is co111ple111ented by a strong cad·e ntial progression ('\ / V- V ) . As a result, the C becon1.es a str·ucrural tone o.f such n1ajor in1porta:nce thai: it not 011ly f L1nctions as the e11d of the first pl1rase,. but, as grapt1 3 shOV\'S, it also participates in t he triadic patter11ing of the second phrase. And, once tl1at triadic motion has tal{et1 shape, a structural C an octave lligher is impliecl. But jt1st as the realizatio11 of the lugl1 F was delayeci in o.rder tl.1at the altern.aci e descending patternir1gs might be continl1ed, so 11ow cbe realizatio11 of the triadic n1otio11 f ron1 tl1e lo' :\! C to the highe1· 011e is pt1t off s·o that a11otl1er; previou.sly~ ge11erate~ i11111lication can be realized- na1nel)r, that generatec1 hy the bigl1-level triadic motion of the first plu-ase (graph 2a and b) .

-- ·---....-·)'--- - -

-

la. .- ........_ _ _ _ _ _...

-

.

'

'

'

-

Y-• ,.... ,

' '

• '

•,,.c--"'" -'

QJ

-

7

-

Example 86

\i\Tl1en tl1e l1igh A a.nd F are reach.ed i11 tl1e second plirase, tl1ere is i10 doubt that the opening motive, here displaced an occa-;re (n1'''), implies and is followed by conjt1nct descer1cli11g rnocio.n to C. 'T he irnplication of conjunct desceodi11g motion is reinforced by tl1e l1igh-le,rel gap from rhe A in measure 5 to the F in 111e•t~'Ure 6. Tl1e desce11ding conjunct 111otio11 not otil)' tl'lalres patent and direct "\i\1 l1at "''as latent ar1d delayed i11 the first pa.r t of the melody, bt1t it also co1lStitutes the realization of a f<>regrr>11nd irnplication generated earlier. For , give11 tl1e conj1111ct motio11 witl1i11 the firSt t\:\'O n10tives, the skip .fron1 F to D at the end ·of measure c is un.derstood as a foreground. gtip. The l1ighly cor1junct moci<>n from F to D it1 ·measl1res 6 a11d 7 fills t hjs ga11 and thereby co11tributes ro tl-1e sense of clos11re created by th.e cade11ce in measure 8. The closure of tl1is first part of Haydn s the1.ne is assuretl not onl,y


MELODIC 51.'RUCTURES

by tl1e strong harmonic progressio11 to rl1e dominant, but by the convergence on C of the ltlgl1-leve.l fXittemings-that is, by the fact of .realizatio11 itself. Of .co·urse this is nor rf1e end of the ti1en1e. R etu.rn to the tonic is in1pliedboth melod1caJly and harmonically. And. iI1 contrast wirh the intricate and predomina11tly triadic n·1ocion. of the first pan ·Of 'the 111elody, the second co11siStS of r.elacively si111ple conjunct n1otion1' so1ne\.v h.at delayed, from tlle C reached in measure 8 down to the tonic, Fi in measure 14In tl1e abo,re examples of triadic 111otion, disjunct morion took place Vlitltln a single triad. But this need not be the case. Because triads a third apart share a cotnmo.n inter,ra.l of a tlilld~for il1sta11ce, th.e third F-A belor1gs to both tl1e F major and the D 1.11.inor triads-different triads 111a)7 follow one another yet create a single disjunct patte.rning. T \VO kinds of triadic

overlapping may be disringuished: linked triads and conti11uous triads. 1.) Linked triads occt1r u1here the patrer:nuig co11taii1s both thirds and fomths in alt·er11ation. The '''ell-kno\.v·n melody ''and He sl1:all reign for ever and ever'' from the Hallelujah Chorus of Handel's itlestiah illustrates this kind. of overlapping triadic patterning (Example 87). The fore.ground motion consists of three rising fourths separated from one anothe.r b·y falling

sixths. The pattern's convi11ci11g coherence co.mes fro1n this patent intervallic regt1larir;~ (-whicl1 is significantly dt1e to rh)rthmic strUc'.t11re) and from the orderliness underlyi11g tl1e succes.~on of disjunct intervals. For when the pitcl1es of the melody are vvritten as g single "line''- \\Tith the si.xths inverted to thirds, ·as in tli.e first staff of t11e ana1)rsis it is e\ren more e'rident tl1at tl1e .Pattern is one .o f o,rerlapping triads: tl1e D and F # of measure I are not only part of D· major, but of B minor; ar1d the B and D f measure 2

•

,_

I

'


r66

EXP'LAL.'ITNG

~1USIC

are 11ot only part of B n1inor, bur of G n1ajor. The forceful ,p atterning of the n1elody comes not onl}r fron1 tl1is triadic linking. but from tl1e higher-level r11orion whict1, as indicated in g.r aph .1, is also, triadic. On this level con~ t.U1t1atio11 to cl1e lo\ver D is ir111)lie'Cl. Though 11ot directly relevant to t he problern at ha11d, rwo aspects o f this melody merit consideration. Though rising f 011rths1 whicl1 begin on weak beats need not function as gaps, impl}ring conjtmct fill, they d,o so here. For, u1sread of beir1g ~ldjuncrive, like the t1pbeat ir1 Sn1eta11:1,s rr1elody (.Exan1ple 84), the fourths .in this case are part of the n1ain tnelodic patteri-1. T l1e nn,,'t t\\ro are ~ 1 filled,i ~ as it '\\'·ere, by proA.7 \vhe11 t l1e fourtt1 from D to G is f ollo'\\red by descendi11.g conju11cr motio11. Perl1a:ps to cor11pensate for tl1e lack of fill for the .precedi11g fot1rths 1 ar1 extension e1nphasiz.es the con;unct motio11. For the t\;vo .similar iambic groups of t he beginning i1nply rl1a.t the lugher-level rl1yrhm will be an a.napest group of four-beats duration. Had

this been the patterning, the D \Voold ha,re come on tl1e first beat of nleastire 3. -r11e nvo-beat CA'te11sion 11ot only ser,res to en1pl1asize the conju.n ct fill, but to syn1bolize the r11ear1ing of the texr. It is clearly appropriate tha.t the wo1'd '(ever'' is coupled with rbe srrecching of the phrase create
f'

.,,,,,.. '

I' ..

)

fl'

-

-

.

.

'

'

·• u~

Exarnple 88


MELODIC STRUCTUllES

As the abstractioi1 O\rer the ex.ample sl1ows (graph 1) , the b11sic tria.dic sttt1cture is exactly that of Handel's melod)l ! I11deed, eve'n the rising~fourth falling-sixth pattern is the same. But this simple orga:nizatio11 is veiled in foreground embellishme11t. The:re is bu.t one explicit disjunct motion-the falling-sixth n1 i11easure 1. All cl1e other inter\7a1s are connected by conjunct motion, emphasized by a. pluasi11g wluch calls for smooth continuousness. As a. res11lt, tl1e fundamentally end-accented organization, so empl1acic in Handel's mel<>dy, is muc h attentuated. i\tloreover, the rhythmic-formal organization is complicated by the fact that the G in measure 3 is an accent, not an upbeat conforming with tl1e prececli11g patternu1g- as \Vas the case \vitl1 the equivalent D in Handel's tnelody. As a result, the C is so much less emphatic than is t he G in Handel's version of this schema t hat tl1c analysis n1ay seem

''forced'>in order to en1pl1asize the similarity bet\veen the nvo themes. Bttt tliis is r1ot, I thlllk, really• the case. Mal1ler mak;es it clear tl1at he thinks the C is importa11t b)r em-phasizing it \Vith an accent and calling attention to it ""vith t he grace-11ores. ~loreove.r, tho11gh it c-0mes ot1 the downl)eat, tl1e G is unstable because it .is harmonized by a six-four· chor d wllich n1oves to and is ''resolved' , on cl1e dominant ch.a t l1arn1onizes tl1e C. Finally the C is brought out because it is the turning p<>in:t in a ttaditional and familiar cadential gesture . as a comparison ben,1een tile last measure of l\t1ahler~s melody and rhe closing moti\ e of the slo\v movement of l\ f,ozarc's ''I-:laffner'' Symphony makes clear (Example 89) . 1

1

Moklv

:t ...

.i } Contim4,fFZIS Triads cor1sist otuy of tl1irds (or, t>y inversi(>11, of sixths)t r:at her rl'lall of an alternation of thircJs with fo11rths or· fifths. The melody

which begins the second key area of the first: 1novement: of 1\1.ozart,s. Sonata for Violin and Piano in A Major (K.305) is a <..."'Onspicuous1 but by no means uncomplicate,d instance of this sort of orga.nization (Example 90). The Cl in 111easure 14, \\7hich begins the patterning, is followed 'by a s11ccession of 1


r68

EXPLAINING MUSIC

thirds: C~-A-F#-D#-Il-and so 011. As staff A of the example shows, tl1is creates a series .o f overlaf)ping triads: f~-dl 0 B-g~-etc. (The s11p·erscript '' mean.s diminished; lo'\.ver-case letrers desigr1ate minor triads; uppercase letters, 111ajor triads). This disju11ct, f oregrol111d patterning is the basis. for a higher-le·\1el CQnM j11nct morion. The groups of f ow: eighth-notes create an implied bilinear organization, shovvn in staff B of the example, As the ('tie'' in the analysis indicates, die fallit1g se,•e11tl'l (Ct do\vn to D~) cre~1tes atl in1pliect dissona11ce sucl1 that the following B acts as the resolution of an imaginecl St1spcnsion. In otl1er \>vords, the patter11 is one of late11t fotrrth~species co1111terpoint-as, for instance it1 :Example 69. Both the f oregro·Uild ar1d higher-le\1el patter11h1gs are co11ti11ued to tl1e Ft in measure 27, after which tl1e rno1nentun1 created by uniformity is slo\.ved. do·wn, an.d brougl1t to stability by a reversal. The 11oncong ruence of nieLodic motion :mcl metric structur.e makes the passage con1plex and. e·qtlivocal, creating subtle, yet effe,cti,tet te11sion. As the braclcets aho\~e and belo\>v tf1e mt1sic ma1
+

s.

, - -~·1 \;-:vv l'i'~s~f

'-- - -1• - -

· .. ~

__,/

, .. ,., . r;

V'

I,, .. "''"'

1

·. ~

5_ZiLJ , , ,,

-

,

"'""'


MEL-OD.IC Sl'RlJCTtJRES

in measure 26, there is little s·ense of stability and re;tur·n because the Gt is not a beginning, preceded by ru1 upbeat figu;re (as the Cj in measure 14 was), but is part of an on-going process. Tllis noncongruence of melody and 1neter s.ho·uld, I think, be brought out b7r the performer: the low-level triple meter must be made evident by slightl}r stressing the prin1ary arid secondary metric .a ccents--beats i a11d 4-witl1in each 1neas·ure. That the passage is without accolnpa11iment u.ntil the second half of 111eas:t1re 27 is irnportant, for this a]lo\?1.TS the metric-melodic conflict to be f.ully effeccive. Nloreover, l'>ecause no accompartln1ent emphasizes one of the triads latent m th.e u11iform succession rather rhan anotl'1er thus creating a sense of progression, the harn101lic stru.cture of the passage is ambiguous. Because rnelody st1gges.ts one patter11ing and meter a11other, wt1ere harmo,nic events begin and ,e nd is nor certain. Fron1 a melodic point of view for in!>i:at1ce, t:he D# at rl1e beginning of n1east1re 25 is part of the precedit1g patterning, Iargel}r beca.use it is separated from vvhat follotvs by a larger disjtmction-the skip of a. sixth. f\,1etticall}r, ho\vever, the D# is the begintnng o"f a triad and should b·e grouped with the weak beat.~ that follo1>v it.. Sim-

ilarly, the follo\ving B a11
1

ous. Not onJy is the connection between substantive tone, dissonance, and resolucion veiled by the intervening thir(ls, 19 but the metric placeme11t of 1& Compare: with Exan1ple 69 w.he.re this is 11ot the case.


EXPLAl~ .

tNG J\.1USIC

dissor1ance and resolutio11 l"eeps sl1ifti11g. Mo1~eover, here1 too1 co11tint1atio11 is implied- at: least for anotl1er 011e a11d a l1alf n1easures at '\\ihicl1 poi11t a cade11ce to the totlic \vottld he quite strongly im11lied..20 Becat1se botl1 th.e foregroltncl and higl1er-level nlelodic patter1li11gs, as well as the rhythn1ic-111etric organization, itI1pl}r co11tinl1ation, a decisi\re and u11eq·u i\ ocai re,rer'!>'a! of -cl1e 011-goir1g n1otion is reqltired if tl1e pl1rase is to :r each stability ai1d closure. The reversal .is .t11ade clear b)r a nurnber of cl1a11ges in t he musical organiz:ation. T 'he 1110St importai1t of these is t'he et'1d. of tl1e tri.'ldic 111o·cio'!'1. \tVhe11 the foregt·ot1nd becon1es conju11ct in 1neas.ure 2 7, tl1e t'\:\' O strands of the melody are connected ·b }r a scale a11d converge to the E in meastu·e i S. Second, tl'1e ar11higuity of tl1e rh}1 tl1n1ic structure is resolved i11 i11east1res 28 m1d l9i \Vhere the fo.t"eground dac.'tyls beco1ne part of a higher-level arlapest group. Fir1ally, tl1e .e ntrar1ce of the violin and piano accotnpani111e11t in 1neasttre 27 not only defi11es the harn10·11}r, but clarifies the rhyth·mic r·elatior1sl1ip bet\:\,ee11 tl1e })recedi11g rneasl1res a:11d the cader1tial ones. T t1e domitla11t-seve11th cl1ord ft111ct:ions as an i1pbeat to the 11eA~ tViro-n1easure group. arid partly through this the anacr,u stic ftr11ction of the earlier r11easures is n1:-l.(ie explicit. The reversal is en1pl1asized or signaled-by the trill. o:n the D jt as welt as by the cha1lge ir1 tenure and the begin11ing of rhe accom.panin1enr in tl1e violu1 and pia:r10. The t1lcimate cl0S1.1re of the plmse is a..c;sured botl1 by tl1e etid-acce11ted rhyth:n uc group a11d by tf1e srro11g ~111d explicit 11.armonic progression to the dorninant. l\lfelodic~i.lly, the F~: is a sacisfac,cory goal because it h as bee11 in1plied by cl1e precedil1g gap o f a tl1ird, E- G#l ·a nd is delayed and e1npl1asized by a11 ~1ppoggn1tura. A contint1otts triadic scruc..'tt1re t1eed 1101:, however, be patently tmiform and a1nbiguous in process. vVhen l1armonic fur1·crions are cl.earl)' defi11ed and rhycl1nuc-melodic orgar1izari-0n is decisi ely patterned, a st1ccession of foreground thirds may lJe rl1e l1asis for a scroiig and unequi,7 ocal n1eiodic structure. This is true of tl-te r11elody \\'l1icl1 begit1s the first rnoven1er1t of Bral1ms1 F ourtl1 Syrnpl1ony (Example 91 ) . Not only does the t'h)rthmic structu,r e distit1gt1isl1 ber"ve.en srruct11ral a.nd orna111enral rones. on se,reral levels, bt1t tfiis di:ffere11tiarion is supported ~JY a clear and normative harrnor11c progression (see cllor d an~tlysis t1nder part B of che example). As i11 f\1101,art's melody, t l1e foregrot111d successio11 of thirds is tl1e basis for higher-levels of melodic parterr1ing. The highest level (tl1e '
fiad the pattern been continued, the tonic, E, \'Y•ould ha.ve occurr.e

MELODIC ST'RUCTU:RES •

concint1arion to

E. The middle-level 111ocion begin.s u a. series of falling

fifths (or rising fourths), G C-Fl-B;

th.e process is altered in 1neasure 4 after \vhich the patter11 consists of rising fiftl1s an.d falling ~i"rths, E- B-D-A- C. This alteration is the result of a change u1 both the direction a.n d th.e timing ·o f cl1e underlying triadic organization. . As line A, of Example 91 s11ows., tl1e first four measures are b·a:sed upon a descendi11g series of thirds (x), and the second JX?tt of the n1elody (beginning in measure 5) is based upon a11 ascendi:rzg series of tllird.s (y). vVhe11 tl1e second part of a n1elod}' has the san1e basic patter.tung as tl1e first bt1t cl1anges tl1.e directio11 of motion., the first part "ivil1 be called the n·1odel, tlte second the complement. The whole p.a ttern \\rill be referred to as a complementary melody. Such melodies \Vill be co11sidered in the next section. The m.anifest m.elodic organization1 ho"''ever, is :n ot essentially complen1enrar)r. For, despite the change in rl1e directio.n .of triadic s11ccession from falli11g to rising (see li11e A), tl1e n1ain aural patternitlg- the lugh-level conj1111ct n1otion -continues in the second part of tl1e melody. F:or ·tl1e coi1junc.'t pattern to continue, the rate at \\rhich triads succeed one another must be n1o·dified, a11d this e11tails a co11comitant change in tl1e mjddle-level pattern. Instead of a rate of four di:fferent pitcl1es in eight quarter-notes as in tl1e fust part, the octave repecitio11 of tl1c ~ D, and C across the barlines slo\VS do¥vn t l1e rate of cluinge so tliat t'here are 0111)1 tluee di.fferent I>itch-elasses in eight quarter-notes. 011 the high level ffi,e consequence is greater emphasis bt1t

~·- ·--

x - • -· s - z

Example 91


.E.XPLAtNU G l'vlUSIC

on conjun~ct n1otion. 011 the rr1iddle level, though the alternation of rising and falling ian1bs continues1 tl1e i11tervals are changed fron1 fourths and filths (as in che first p·art) to fiftl1S and sixths .in tf1e second. Tt1ot1gl1 tl1e cl1a11ge f ron1 falling to rising triadic n1otion is sig.nilica11tly rnaslced by the dor1tinant conj t111ct rr1otion, explicit c.01111}len1entary relationship is not entirely abse11t. 'fl1e ·b ~tss. line, v;1J,icl1 was static {a sustait1ed E, over '"'l1icl1 tl1e l1arn1011y chru.1ged) it1 tl1e firsr pare o·f the 1nelody, begins to mO\'e in n1east1re 5. ft ascer1ds tl1rough the cycl.e of fifths (C-G-Di.-A), corr1ple·menting the rr1iddle.-level cycle of falli11g :fiftl:1s of tl1e first fo11r rneasures in the melo·dy. Had the series of fifths i11 tl1e bass con.tinued, the next bass 11ote '\vould ha,re been an E, bt1t tlus possibilit}1 does not f:ollo\v directly. Instead. tl1e seqt1ex1cial clis}t1nct fifths are f ollo\ved first by a. chror11aric risirig line i-11oving from 0$ to A (measures 9-14 ivl1icl1 are onl)r sketcl1ed i11 tl1e aria!)tsis) and tl\en by an a11t:l1e11tic cadence. IV-0-V1 / \ ' -l!-\T7-I,, at the enci of which the implied Eis realized. A11 an.alogot1s delay occ11rs in the ttiadic 'n1otion \Vl1icl1 underlies tl1e n1e!ody. Just as tl1e first four measures of the n1elo--ures 1 3- 19 are on the sanie hiera,rcilic level as the preceding pattern, \Vl1ich rnoved in two-111easu:re unics, se·veral aspects of tl1e :passage st1ggest tl1at these measur-es are a strocturall).r eqt1i,rale11t co11cinuacion of t'he first part of the n1elody (Examt)le 9,1 ).

a) The prolo11gacio11 of the C delay1'ing tl1e previotISl)' generated process, lends IlsychologicaJ i111porrance to tl1e linear n1ocion i11 tl1ese measures. b) The cadenrial n1otion in rneasures 17- 19 beg.iilS with a marked ga,p


,

MELODIC STRUCTURES

°"'hich strongly implies and thereby calls attention to th.e descending • COllJUUCt pattern. c) The iambic rhythm and bilevel melodic organizacion of measures 1 3 and

14 rela.t e this part of rl1e melody to tl1e opening n1easures, as a kind of \faried dimin.utio11. The patter1ung is linear and, as indicated in line A of tl1e ana]ysis, botl1 the primary and secondary stra11ds Sl1ggesr conjt1nct morion con\1erging to t11e E in me~ure 1 9. d} Both strands-particularly tl1e subsidiary 011e-seem to be continued in measures 15 and i 6"-tl1ough 'to nonsubstantive E's. The succession of descending linear motions toward E , li as a cumulati\,.e, summationill effect-as thot1gh a single desc.-ending line \Vere spread ,over these five meast1res (arialysis line B) . e) This suggestion is not as fancifttl as it migl1t at first appear. For had th e conjun.c t motion in t\vo-measure units ( the rate of motion in the first ni11e measures) been continued in the second l1alf>then, as line D shows, the pattern \vould llave r,eached the E precisely as it d.o es ii1 the acrual then1e. Accompanying the t1nderlying triadic patterning is a fundamental and co nsistent rat e of pattern change. Thougl1 Brahms' 111elody cr eates a stro11g and persistent sense of linear-

ity, the two-measure units of the second part do not, in facr, m.ove con-

~··.

I

.

l

,.

-

.

r·

8.

.

~

:/,,.

.,

.

-

.

-]

I

'

.

-

ff

-

. . . -. ·~

. =.;I ..

11•

\Y

?:~

~

•

r

F

•

.

<> W;; ·;~

#. '

Ii)

~·~'· . .

.

'

•

~ -

!ii; ~

:l

~·!!:'

.

.!

:if

•

tl -

t

'

-

a"

-r

.......

r

:r •

'

~ ....~ !::;t''-' .... . . .

-:fl.

(r/J) "$.#?

•

..

'

--

•

p

.

--

I

-

.

'

•r.n <> ·-

'

.

p

~I

.

·t "

.

·-•

.,..,,,.,

~

.

•

1 1: ~ ,

(.

.

·-.

•

--

'

'

-- ~:• '

.

'

.

. '

i .•

•@ .

·~.

,

..

.

·~

'

ii

•

.

.._


EXPLAINING MUSIC

1 74

juncrly.. Rather, as graph C (Example 92) shows, tl1e momenn1111 of the conj·u11ct 1r1otion wlucl1 lecl to tl1e prolonged C is cl1ecke(l b)7 11 re\rers.il, On both the middle ar1d the high level, a ski1) fron1 C to A not only breaks the pr·eceding lit1ear pattern., b11c cre~1tes a gap ir11plying cor1junct fill. The fill follows \Vhen th:e A 01:0,res clu-oma.tically (tl1rough A# in r11easure 16) to B-tli.e first high-le,,eJ asce11di.l1g 111otion in the t11elod)7 • 011 tl1e high level, the B tJ1en inoves i 11 disjunct motion dO\l(rn to the tonic. Tl1is arri.culatian of tl1e melod ic process is accom1)anied by a reversal in the 111cJtio11 of ·the lJass. 1""11e chro1natic li11e l)egt1r1 in measure 9 e11ds ii1 rneasure 14, arid a clarifica21 tio11 of th.e l1a.rmo11ic syntlx is f ollo\.ved by art a1,ichentic cade11ce. That this melod)1 is })ased 11pon a pattern of con.t inuou.s triads is si1own u1 tt1e r,ecapirulation. In the restaten:aent of the firsr part of the n1elody, BraJ1'.l11s elimirl.ates the bilevel patter1ur1g, as \.ve!l as the ia111hic rhythm ~vith which it '\ Vas associated in the exposition. As a result, the underlying triadic stt1.Icture is t11ade i11anifest in tl1e audible foregro u11d (Exaxnple 93).

I'

,

~

' \.. -

.... -

.. ...

~ ....

-

..,. ._ ..... -

.. ...

~

... .. .,., ... ... J

SYMMETllICAL P ATI'ERNS

In n1ost of the melodies C'C>t1Sidered t hus far, successive e\.renrs \Vere related t<> one axlothe.r by: reperitior1--as in the lll1ear 111elody from l\11ozm's Divertilnenro (Exan1ple 64), or tl1e triadic rune from Strat1ss' Tilt E1:tlenspiegei (Exanriple 85); cor1t.inuati.011-\'\il1etl1er co11junct, as i11 tl1e Scherzo from Scl1\1bert~s String Quartet (Exa111ple 66), {)f disjt1ncr, as ii1 the r11elody frotn Handel's iVi essial:; (E.\:a111ple 87); and contrast of ft1nction, as in gaf1-fill melodies like the su'b ject of Bacl1's Fugue in D i\1u1or (Example 77) .. In S)r111metrical melodies, the relationship l)etwee11 successive e1r,en.ts is .st1ch th.at one e\1ent 11nrr·ors the patterning of another. 1 11 other '\vords, there is a balance of motion and, counr.er1not:ion. vVhen st1cl1 cou11t ertT'l.otions are primarily 21 Agaln the dual

charnc~r

of critical explno:atior1s is clear. For, though rule reasons were freque11tly en1plo)1ed iI1 the analysis of tl1is melody, con11Jr1o:n-sense1 ad hoc reasons "vere ind.ispen.'>able. 1


MELODIC STRUCTURES

1 75

.melodic, st1ccessi,re C\rents are similar in it1ter,rallic relationships, and often in rhythm as \veil, b11t opposite in direction of morion. This is true of co·mplementary and axial melodies. Or, the nurroring n1ay be primarily harmonic. Then the e\.re11ts invo.Ived may i11ove in the sa111e melodic directio1i, but they have an 11ar1nonic patte.r n of, as a rule, I-V, V-I. at1d a 11igh~level n1elodic stru.c ture ~~rl1!ch r11oves by conjw1ct i11otion :'l'\\'
Co1't1plernenta:ry 111elodies Tl1e first ther11e of Braltrns' F'ourth Sympl1ony is based1 as we have seen, upon a symmetrical pattern of two continuous triads n1oving in opposite directions ove1· a rwo-octa'\•e rarJge (Exa1nple 9 I). However clearly setised this ,complementary relatio:nship n1ay be, i:t is not explicitly presented. The 111ani.fest lnotion is p:redominanrly linear. Bu.t con1plen1entary patterns can be explicit and manifest. T he cle-arest iristances are those li1 vvhich one melodic event is an exa.ct inversion of anotl1er, and in '\vhicl-1 durational relationships are the san1e for both. Anorl1er n1elody· by Brah1ns tl1at '\vhic.h begins tl1e third movement of hls First Symphony- is such an unequivocal instance (Exan1pie 94).

The analysis shows tl1e main {Ycltterning of the first phrase as being linear -a coi1jt1nct morion from Eb down to Bb (grapl1 I). Because the C is the most important tonJ~ in measuxes 2 and 3-as the i\ b :pedal in the second clarinet and the bassoon, and the cotmterpoint in tl1e horn ~tnd celli, mal{e clearthere is also a hll1t of triadic motion. But rl1e metric position and repetition of cl1e Db, as well as the en1pl1atically oonjt1nct r11otio11 of tl1e foregrou11d and of t11e bass tine, suggest that the pa~ter.n is essentially. linear. Both this linearit}t and the prototriadic strtlC't ure (Eb-C, \Vi.th tonic harmony) srro11gly imply continuation do"rn to Ab. The second ph:rase~ an exact inversion of t.l1e first employing the same d11rational t'elacions[1ips, rises co.11junctly frorn C to F (grapl1 2). This patterning, too, i111pJies conrint1acion-ro tl1e upper tot1ic. Hovvever, because the interval fto111 tl1e first note of the plu·~tse, C, ro the potential goal, Ab, is a sixth (rather than. a fifth as in the first phrase), in1mediate conti11uaci-011 is i1ecessary if the ends of tl1e t"vo phrases are to correspond; th,a.t is, if the second is to reach the leadir1g tone l>elow Ab, as the first reached. the supertonic a.bov"'e it. Tllis continuation, '-''l1ich takes place '''hen tl1e next pl1rase begins on G, reiitlorces the implication of motion to t:h.e high Ab.


LXPLAINING 1\'flJSl C

. (MO

-L Exa1nple 94

T'hough tl1e stabilit)r o·f tonic harmony is reached ar the end of the \\rh.ole tf1erne ( meast1re 19) a11cl a,t the beginning of its re1Jecition in rnea.w:re r09, the melodic Ab's., implied by the opening phrases and by their St1bsequent repetitions, are delayecl t111til the beginr1ir1g of the coda (measure 1 54) . Th.ere, as tl1e example sl10\~lS, boch lines 1no,1e li11e.-irl)' tl1rough octave transfers to their respective goals. Yet eve11 11ere, tlle rhythmic place111ent ·of tl1e t1pper Ab ke,eps ir; fro111 flmctio11ing as a srrt1ctural to·ne. Only in. the final cad.ence, as a st1stained G moves up to the tonic, is there satisfactot~y closure and congrt1ent arri\val at i1nplied goals. Tt1e n1elody wl1ich begins the exposition se.cti()n of the first n1oveme11t .o f i\do.zart's t'Linz'' Syt11phonyr (K.4i 5) consists of cor11plemencary phrases which a1·e linear on tf1e lo,:\1est level and triadic on the next (Ex.a111ple 9S). Despite its see1ni11g simplicitjr it is so corr1plex tl1at to do it justice \\rould require a disci1ssion of the \Vl'lole rnoven1ent of v\1hich it forms a })art. ' ;v'hat follo'\vs must, there:fore, be regarded as illustrati11g a type of organization, rather tluu1 as an aclequate analysis. The melody begins vvith a foregrou11d pattern of conju11ct "'/hole-11oces {moti"Te x) . Because tonic ham1ony persists rhrot1ghot1t the first four measures, 11owever1 the E and G a.re u.nderstood as structural, 'vrule the F functions as a passitJ.g-ro.ne. The main patterniitg is t herefore triadic {graph r ) in1plying continua.tion to C~rather than lit1ear. But the C \Vhich fol}o\VS in meast1re 22 is not a s-Atisfactor}' realizatio11 of the irnplied triadic motion, for it is n-0t on a hierarchic le\rel equivalent to tllat of tl1e n::vo-n1~1sure lengt.hs wluch g~nerated the patterniI1g. Because the melod.ic prominence o.f the C is not matched by its sttucturaJ imporra.11ce, it is a potential str·tlcttiral to,oe (grapl1 3) \;vhicl1 stre1Jgthens t he in1plication of a structural Ca Tl1e C in measure 24. '\vhich occurs in the rigl1t register a11cl at the rigl1r point in time, is a satisfactory realizacion of the implied triadic patterning. t


MELODIC STRUCTURES

Ne,rertheless, it is only a temporary goal. For, thougl1 the triadic patterning is confirmed an.d thereby reinforced, the C is not srable, either harmonically or rhy·chimcally. Consequently, tria.dic continuation to1 and. co1npletion ont the E an octave abov:e the first note of the tune is a strong possibility. At the same time rhe C ft1nctions .as tl1e beginou1g of the second pl1rase. The descendit1g, co11junct, and \vhole-note n1otion (r) '''ith w~hich the t1ew p.hrase begins is a C"<>11formant in\rersio11 of the opening pattern (x), and. as a resu.lt the complementary rela.cionship bet'\veen the phrases is clearly audible. The bass, moving in parallel tenths ~rith the melod}r, strengthens the impression of foreground linearity. The harmonic progression 1

1.

IV6- V !0/ IV- IV 3

establishes the C a11d A as stable, structural tones and creates the complementary triadic pac:rerning (gra.ph 1a) 'vhicl1 implies cunr.inuacion to F, and perhaps beyond. This possible patter1ur1g is t1ot, llowever, realized immediately. LJ d. -or perhaps ((an:racted''. b· y alterr1at1ve . posSl'b·.ili··nes. t he mel D ei1ecte ody cha.oges direction, rising conj.unctly fro111 the A in measure 26. The E, implied 'by the triadic patternu1g of t11e t11odel, is reacl1ed in measiire 27. It f·unccions as a r elatively stable goal, 11ot only becat1se of the satisfaction of octave completion~ but because it is preceded by a gap-fill pattern (graph 4) which articulates closure.

''"

.

'

'

.

.. -

..' J

E."(ample 95


EXPLAINING MUSIC

A desce11dir1g tr:iadic pattert1 follo,:\'S t11e E. Its n"Iotion is both n1ore explicit (the linear connectio11s o,f the f oregrot111d are l:tnequivocally ornamental) and rnore rapid (tl1e structttral tones occ11r every~ l1alf-n1east1re, rather tl'1an ever)' other nieasure} than tl1at of the 111odel a11d tl1e begit111ir1g c>f rt1e co1n1,le111e11t. Tl1ese two rn.easures (27 and 28) join n1odel a11d complement in a single p~tterr1-mo.v ing f ro.m E wluch belo·ngs to rhe model, to C \vfucl1 is comn1on to 111odel and co1npler11el1t, a11d tl1e11 ro A \>vhicl1 belongs c>itly to tl1e comi'Jlemenr. The F \\1 hich fc>llo,vs is the continu~ition of the trialiic morio11 of the cor11plen1e11t begt1n in n1ea.sure 24 (graph ra) . The p1·obable co11tint1ation of rhe rriaclic p,attern begt1n in tl1easure 24 is to C ir1 the lo\li'er oc.'t~r,re. But this C is not reached in th.e a11tecedent period. Instead, .the motio11 continues t o D, where the appoggia.rura progression (!:~ ) to the senucadence creates tempot'ary cJost1re. 1"'he COTtsequent period (of '"'hich only the firsr me1\sure is give11 in Ex:imple 95) begins \\1ith an octave transfer to th·e ttpper £. This tra115fer, as 'veil as the octave tnotion of the n1odel phr:lse, is made ex·pl.icit a,ncl, t\S it were, con1prehe11ded by the ocra·ve leaps u1 the first me~lSttres of tl1e period. In the consequent periocl the sn·ucn1ral C wl1icl1 begir1s the cornplen1en.t is delaye.d b)r a varie,d repecitior1 of the pattern in 111east1rcs 22 at1d 23. This is unpo.rta11t because the reiteratio11 of tl1e pote11tiaLstructural tone and the extension of the pl1rase le11gth rr1ake tl1e C a m.ote emphatic a·n d stable goal. As a resLilt. continuatio11 to the higl-1 E is 11ot strongly im1>lied.-and cioes not in fact follo1,v. Insteaut because the inrervallic strticrore of tl1e several hie.rar.ctuc levels is basically pre.served in tl1e i11versio11 of tl1e pattern. But this 11eed 11ot al v ays be tl1e case. Fo.r the salce of cotnparison, let tis consider at1orher melody by Mozart- tl1e 011e ,;\frlucl1 begins i:he Me11uetto of the Strii1g Qua11'.e't ii1 A Nlajor (l{.464 ). It, too, is complen1er1tary (E.xan1ple 96) . On the l1igl1est level-that of t\:vo-1neasure units-the risirig conjunct


1\11EI.!ODIC STRUCTURES

179

morion (A--B) of the first part of the phrase is apparently mirrored by the descending oonjunct morion (E-D) of the second part. Even on a lo,ver le,1el, the beginnings ru.1d endings of the nvo-measure eve11ts (x and. r) .complement one another. Each of the first rwo e'\1 enrs ends on a '\Veale beat a perfe.c t f 011rth abo,re its 6.rst note, and in the seconci pa11'. each e,1ent closes on a. vveak beat (but tl1e. third beat rarl1er than the seco11d.), a perfect fourth below its first oote.. T l1ere is anocl1er sort of symn1erry as V\'ell. For, if tl1e pattern is cor1sidered to consist of strllcturall)7 equivalent tones, then the second part of the phrase .is a retrograde o.f the first, and the whole phrase is synID1ettical around the E: i\ - D- B- E / E . -B-D- ...i\.22 Th.ese relational similarities tend to be n1aslced, 'hovvever, by differences in fo.regr·ou11d patter11ir1g. Melodically, the motives (x.) of the n1odel create an ir11pression of triadic rnotion, 1-vl1ile those of tl1e c,0111pleme11t seem predon1inantly linear. Rl'iythnncally, tl1e differences are even more striking. On the lo\vest le,rel ( 1), t11e rl1}rthmic grot1ps of the model are unambiguous, beginning-accented trochees; 23 tl1e g.roups in th.e co,1nplen1ent are 1nore dor1btf11l. For the 11armony suggests that tl1e groups are begilu'ring-accented dactyls, but the re1)eated notes in the first measure o·f each pattern te'od to become anacrustict creating 1.atent ( ;w = end-accented groups. On the second level (z) the contrast ·bet"''ee:n model and compleme11r is clear. The rhytl1rri of the tnodel is beginning-accented; tl1at ol the comple111ent is endaccented. On the third level ( 3)-V{here pitch synunetries are most apparent - the rhythmic org-anization is ttocllaic in both parts} so that the impression w

"

-,

)

1

"'-

12

CT? ..

i

'=

t~ ~ ·•• ·"-

4

-.....:

,

)ti ' -:".

-......

*-' \._, .

z lrl' rod

\

I\ - ~;

J

\,;--z' -£ Q

1

' t

~ y»Q~,

t

"" f"

•

1 ,

~'

•

mJ 'a _;

\~

-2

J.---=--~

Example 96 22 This S}rmmetry is preserved eve11. if the ne.n l.ower level is in.cluded.- i.e., if the notes given \vithont ste111s in the analysis are counted. 23 For a discussion of the 'basis foi: this atulysis, s:ee l\11eyer, 8'fflOtiot? and lY!.eaning in Mtaic, p. ro7.


F..xl'LAINING l\i!USIC

180

of con1plementary structure is certaitll)r felt, if ·n ot consciousl.j:r recognized. ~fl1ere is anotl1er difference bet,veen this linear, con1ple111entary melody

and tl1:e one fro1n Brahn1s' First S)rn1phony. ln Brahms, melody r.he plirase.s ir11ply alternative goals: tf'le 111odel n1oves conjunctly tCJV\'ard the lo\i\fer ~i\.b, hile the co111ple111.ent n1oves t:O·~iard the Ab a11 octa,re higher. In such cases -a11d the n1e.lody f.ror.11 tlle ''Linz,. Sy1n1)hOO)' is similar in this respectthe compleme11tary motions may be divergl!11t i:r1 the sense tl1at they i111ply ' "1

no comrnon tnelodic tr1eeting point. I11 the inelody from Nlozart's String QuaJ·tet, on tt1e other hand1 model and. complement <.to imply n commo11 meeting tone, the Cl assuming tl1at for the mon1e11t "ve n1ay disregard t11c fact that the linear motio·11s occl1r in differe11t octaves. Such cotnple1ne11taryr n1elodies ought be Cliled CO'l:l'Verge1z&e ty·pes. The si.nillariry between tl1is converge11ce melody ru.1d one. anal}rL.ecl earlier-the theme Vl'l1.ich begins the exposition section .o f t.l1e .first move-

ment of Beethover1's. Se\1enrh S)r111phon)r (Example 70)-will not have es~ caped. the rea.cler. Not: only are cl1e .higl1-level patterns very similar, l).u t even the keys are the same. The obvious difference is that in M<)Za.rt's inelody

the asce11ding conjunct morion (A- 13) precedes the descending motion (E-D), while il1 Beethoven's rhen1e tl1e opposite order (E~D, A-B) pre-vails. In i\il ozan's Menuetto, as in Beethoven's movement the arri\1 al of the implied C~f is considerably delayeci. No satisfactory, stable C# o ccrr.rs either i11 the re1nainder of the first part of che row1ded-binary £orn1 or .in the sec;0nd part- the recapitt1la.tion. Onl)r at rhe begiru1ing of tl1e codett~t, after tl1e return of tlle n1elody l1as rege11erated the implicative process, does C~ OCCllf as a structuraJ tone (Exa111ple 97). •

Exarnple 97

Because the n1odel an.d cornpleme11t are pre..ll)ented in different octaves, thcr·e is a q11esrion \Vl1ether the C# in n1easure ·65 is a satisfactory realization


MELOPCC STRUCTURES

·o f the implications generated b.y rL1e immediatel)r preceding patterning and by the original prese·ntarion.. Three cc)nsid. erarions su,ggest that it is sadsf actory. First, the relationship between tl1e end of the inodel and the b.eginning of the comple1nent irr\rolves a tninsposition to the tipper octave i..vhicl1 ch.e change of register at measure 65 appropriately mirrors~ Secon~ beca.t1se t he recapitulation of the melod)r, beginni11g in measure 55, is an OC:ta\ e liigher tl1an tl1e u1icial pre.~e11tat:ion, t l1e registeY ()f the C# is medial, linlcing earlier and later staten1ents of the melody, The C# is in the proper 1

register ir1 relation both to tl1e co1nplen1entary p·attern (E-D ) of t he initial statern.e11r ·Of the melody a11d to the model patter11 (A-B) of tl1e recapirulatio11 version. Finall)r, and m
c,

·Of its it11plicative processes \\rould ·entail a discussion of the '\Vl1ole n1o:venle,nt. Onl}r the imtnediate and essential oon1pleme11taI)' and implicative relations.hips will be consid.ered 11e:re (Rican1ple 98). Tl1e antecede11t plirase consists ·o f t\vo pans: the model and its ·COmple:mei:1c. As t he graphs 1 l) and 2 b sl1ow, botl1 are predominantly triadic, and in both the direccic>n of motiot1 changes covlard the end of the 11attern. At this point,. too, both model and complement become mo.re linear. T his pa.rallel Tl1e C in these measu.r es is basically part of the co1nplement pattern: coining f.rom D, it act5 as a lower ueigllbor to the followiag C~. i\t th e san1e tit11e:, there is. a slight sens~ that it acts as. a chrom.'ltic passing-torte bet\\ een B and C# in the 1nod.el. 24

1


EXPLAL"lilu~G

}.;IUSIC

f

change l1elps to make the structi1re at1dible and e\r:ide11t. 25 lr1 addition, as tl'fe anal)rsis tinder cl1e example sl10\;\ , the rt1y thmic Strti.ccure of the t Vi' O parrs of the plrrase is ·virtua.il)' i.de11tical. A<> a resl1lt, tl1e con1plerne11rary relationship bctweet1 the pirch patterns is }1aten.t a.nd palpable. Because tl1e root of rl1e triad does 11o t begin tf1e patterning of the n1ociel, concint1ation ro the stabilit)r of o·cta\re identity on the t1pper Bb is probable. Bue instead of mo\1ing directly on to rh.e upper B0, tl1e pattern is deflected down co E. 26 As t lie ar:u1lysis it1dicates, tl1e implied Bb is realizecl only at the end of tl1e cor1seqt1enr }Jlrrase-afcer tile generari11g triadic pattern has bee11 repeated (grapl1 ta'). Tfle triadic strt1ctu.re of the complen1er1t, st1ggesring dominant~se\rentl'l hannony, also in1f1lies Bb (graph zb) . T he tonic follo\VS 2-0 Tl1ottgh it is obvious!){ mo.re conj 11r1cc tl1an the model, cl1e c.onlJJlenlent is written so tl1at the disjunct aspects of the first two measru·e.s are Cl'l)l)tlasized. rrt1e Ei, at t11e beginning of the con1plel'nent is strongIy stressed b;1 the large skip, tl1e grace~ nores, and the fcr'te dynarnics. As a r esult., tl1e trochaic grollp in tnellst1re 5 is closed; th.e D is understood as coming frO'lll rl1e Eb ratl1er tl1an ilS inoving to the c~ a11d the irltpression of linear-it}' is tl1e.reby rninimized. l "l1e rh·y clln1ic group of tlie nex't rr1ea.'iure is similarly closed. Not onl;r is tl'le C stressed bjr the rrlll, bur che potential li11ea.r fl.o'l:v is brok:en by che echappee, D. 2 ~ Tl1e E is equ.i'i1 ocal. On the orie ha11d, the linear n1otion leading to it iin.plies cor1tinuscion, and, after an oc·cave transfer, £~ a11d conjunct foreground morion. foUo,1ir. T he chron1acicism Jntent in tllis relationship ( and eA'Plicit i .11 meas1tre 7) has COllrequer1ces later io the rnoven1ent. On tile otlier h

MELODIC ST:RUCI."URES 1 1 '',~ rong '

as the consequent phrase begins, but in tl1e register. Tl1e realizatior1 is only provisional. A satisfactory realizatio11 is delayed i1nril the e11d of the co11seque11t pfu~ase, in v;rluch the version ·of th.e con1.plen1ent also implies tl1e conic. For, tiw11gh their order is cl1a:t1ged :1rJd they con1e 011 ¥\reak he.:itSi the tones of the con1plen1e11t are nonetheless present and emphasized in measures 1 3- 15- t\VO .of the tones, C and Eb are resolutions of seconda.ry domio.antst and tlle A resolves the ,dissonance of a ~~ progr·ession (graph 2a.'), Thus the implications of botl1 111odel at1d complen1er1t converge i11, and. are realized by, the Bb whicl1 closes rl1e 111elody in measure 16.27 1

Axial 1ne/.odies Axial melodies, as the name i11dicates, consist .of a main or axis-cone embellished by neighbor-notes abov'C.~: and belo\v,.28 They do not, as a rule, occlir on high structural le\rels. Like compleinentary melodies,. they 11ave two parts: a, model a11d its inversion or ~mirror., .Both the model aJ.1d the mirror n1ove from the axis.-tone (A) to the neighbor-note (N) and back: 1

?'a:A

A

'Nv' n1odel 111.:irror 1.""hough the foregrot111d patterning of tl"1e axis-tone and irs ncigl1bors n1ay suggest probable concii1uatio1is, tl1e relationship between model ar1cl 1nirror is, as rl1e diagran1 sho\.VS, pri1narily fornial- that of r·eperition by in\ e.rsionrather than processive. 29 The artiet1lation of the parts of rui axial melody is partly· th:e result of rl1e. lovv-level closure vvhicl1 t'.lkes place {IS eacl1 ·neighbor1

11ote renuns to the stabilicy of tl1e axis. " On the highest level, ho,vever, closure is we.ak and implication absent.

In1plicatio11 is abse11c becal1se, since ~"'1al melodies are essentially prolongations of a single tone, no high-level processive r·elationships. are possible. Closure is vvealc bec-ause, wicl1out itll})licative processes, tl1ere are no goals to a.c t as point~ of stabilit)r and con1pletion. 111 sh<>rt, continuati.011 is likely, because of the lack of established Points of closure rather than because of the ge11eration of specific it11plicacive relario11ships. Axial melodies., the11, are ar Much ·more migl1t be said. about this rnelody. But th.e continuity of the general discussion must take precedence over mo.re detailed a:naly!>is. as J anl inde,lJc:ed to Professor Eugier1e Nar1tlOttr botl1 for calluig Ill)r attention to tl1is kind o.f or,g anizario 11 a11d for the t:e1i11 ''axial..,, 2 0 See Chap:cer IV, pp. 93-94·


EXPLAU..~G

l\if:USIC

r,elati\rel)' st:able· e·ve11 static- patter11ingsl yet, a:t the s:ame titne, n1ore or less open-e11ded.30 Tl1e firSt t.he1ne of t~1e last m<)ve1nent of D'rofak's ''Ne':v ·w o·rld>' Svm" phony is an axia.l melod} (Exan1ple 99). T he essential elemei1ts of the patter11 are established

i1,i

tl\e first fo .u_r rr1eas:t1r.es. As graph 1 r11alces clear, the

tonic, E, is e11lbellished by an 11ppe:r neig~1bor, F#, a11d a lo\ver i1eigl1bor, D. Each of these is in ttirn en1bellished b)" a neighbor-tor1e on a lov;rer hierarchic level: tl1e F j by G~ and tl1e D b)r B. The melody is repeated in 111easures r4-17, bt1t the penultir1mte measw·e is varied, e1nphasizing closure. The wl1ole eight-111easure I)eriod is ti1e·r1 rest.'lted. A single two-me.1sure 1110rive has thus becon1e the basis for a dot1ble period of sixtee11 rr1easures. And there is n1ore to come. But before guing o·n, let us consider tl1e in11)licative relatiorisLups '\vitl1in this period. rf}1e n1elody is tlOt rnarlcedly ~1rocessive. Ot only d(JCS each . plrrase begin a11d end on a tOI.L:"llly sta.ble t<>nic\ bu.t, as the harn1on}.r ir1 the other instrun1e11tS of cl1e orchestra (see score) rnake:S ·clear, tl1e Il() tes i11 betvlee11 are essentially ornan1ental. This is obviot1s in the firsr two r11easur·es for tl1e FJ and G are botl'1 noncl1ord to·nes, a11d it is also clear in mea..rrore 1 :z, Becat1se they are understood as mirroring, a11d functio11allj" analogical to, the so Unequivocal incStances of axial me1o
to

disco~ter,

O'Ccur in the .repe:rto11r ·o f tonal tTius.ic before the middle of the nineteenth century. vVitho\1t pretending to account for dlis in an adec1l1ntc w~y, the follo ..ving

observations seem perrh1e·nt. During th.e rm1eteen:th centuI)1; t lle size and 111a.ke~up of the concert au.dier1ce ch~nnged C«)nsi,ierably. In.<:li idual concens tended to have larger, bllt rather n1or.e hete.roger1eoust al1dienc<-.s tl1an: (fid those of the eighree11r.t1 ce:11tl1r'Y. Tl1is c }mnge, complemen.ted by the literary preclile.ctio1:i,s1 pl'lilosophical inclina.t ions, artd aesthetic taste of l)om co,nposers and their audiences. led to a growth in scale and length of «p\tbliett compositions. ( C-0mposers conci11ued to \Vl'itc sl1orcer, r110re ir1timace \vork:s .for ~'hon1e const1n1prion.'') The larger instrumental fom1s (su.ch as sonam-forn1 n1ovements) tei1decl to beco111c, if r1ot n1ore cottlplex, qltite a bit longer. Probal:>Ly·, as the size and hete.rogeneit)' of tl1e audie11ce increased, the general level of musical traini11g and sophlstica.tion declined. Greater length coupled \Vith less 1nusical experience created ,t he need, mentioned i11 CJ1apter III, for Striking a.r1d e'3.sily ren1embered th.en1·es. A..nal rnelodies have t.he advan~ge of co111bining lengt:h and memorabi:lir.y. Be· cause they are not decisive\y closed., tl1ey are easily repeatetl-eitl1er at the same or at a. new pitcJ1 level. In rhis \Vay four r11ensu~e.s of n1usic ca:n be exten_ded to eigllt, sixteen., or. e·ven 111ore. Yet the amount of n1aterial- t he nu.01ber of

w

.

•

'

@

~

'

'-('~

•

I

••

cOlf ·

I

I ·1

•..... I .

•

:(:'cm

I

I

.

-,.JJI

t

•

!

I

~

I

LI

:t '

. 18

rJ_ •

I

·•

I

I

..

...

,:~

'

, H '

i ~ "'1' ++41

I.

. .. '

I

'

.

J .

•I

'

I

li .,.l ~

I I

•·r

f

I~

'I

· ~

' '.

I

I

'

.

~)

~

,""' I

I '

I

.•, I

.. -.! .t

,.._

•

....

-

·I

...•i@

,~

,

·-

lI 'f

~

L~ r+-+tl.

i. ..

Mater al cor"' d1·c tos autora s

186 F~ aL1d G, '\.Ve take the D

EXPLAINING l\1'.USIC .

a11d B to be or11arner1tal. J\1loreo\rer, though the D arid B are chord tones, the harrn()ny \vh.icl1 accompanies tl1erl1 is itself an e111b·elLishiI1g cl1ord 'wl1icl1 immediately returns (at the begi1111ing of i11easttre 13) to tl1e tonic ha.rmon}' and tl1e sr)nori ry v.1 hich preceded it. Evert ii1 r11.eaStire 16~ ' vhere more patent i111:plications ar e ge11er,1ted 1 the se11Se of tnotion to differenc goals .ist so to spealc, o\1ersl'\ado,~ed by tl1e cadential progression. The predornir1ant etl1os is one of assertive a11d assured stability. A.nd the unpression that cha11ge 'tvill follo·\\ t he seco11d state1ner1t o,f tl1e period is as much a result of tl1e feeli11g tha.t ftirther repetitio11 '\i\'Ould be supererogatory as of rl1e strengtl1 of specific implicati,,e processes generated b)r tl1e fJilttei1ling itself. Yet specific implic~1cive relationships ar e J1ot e11tirel)r absent. As graph ~ i11dicates, the linear p;1 tte1~ning of tl1e f oregrot1nd it1 111easr1.res lo and r4 implies continua.rion:-though not \.'\1 it l1 sucl1 urgency t hat the varied re.petition of t he parten1 a11d the restatemet1t of tl1e period ere-ate a11y real tensio1t. On!y after bei11g restated arid rei11forc.ed in 1neast1res 26 at1d 34 is t11is implicative relationsllip .realized, ' t1e11 t11e meloi::l)r moves cc)njl111ctJy fro·rr\ E up ro B in 111east1res 3B and 39. The patter11ing of 111east1re r 6 is inore patent ly 11nplicative. En1phasized b)' the 11ovelt}' of disjt1nct lnotion artd the grea·ter spee.d of the triplet .figure, both the foregrot1nd triadic pat tern and t11e second-level pentachord~1l relatio11Sl1ip imply cor1cinu~1cion to che 11igl1 E (gra11h 3). T11ere is, in addiri.on, as grapl1 4 indicarest son1e sense t hat the pot e11cial of tl1.e B-which is pro111inenr i 11 itself and Str~-sed l1y the precedit1S' disjur1ct-triplet rnotion- .has 11or bee11 ft1ll}1 acttlitlli.ed. It is acrualizeeriod, wl\en it beco1nes t he axis-tone i11 a sligl1tly varied version of the n1elo

MELODIC STllUCTURES

while the lo't:ver strings enrich the texture with figuration. The role of the brass instruments is initially ha.rrnonic and rl1ytlm1ic.. But just whe11 th.e linear co11cin11ation in1plied in tl1e first period is realized- \vl1en the melody t110,res bejro·nd G to A a11d B (111eas\1re 39, see graph 2 )-the first hor11 ru1d trumpet, accompanied b)' the otl1er brass itlStifuments, joi11 in playing the melody, a.n d in so doing, they ernphasize the continuation implied at th.e beginrling of tl1e passage. The high B can act as a stable goal not (Jnly becat1se it is: the fifth of the scale and _p ar·r of a closed trocl1aic rl1)r-tl1111; but because, as the note an octa.ve above cl1e upbeat which begii1s this consurn~mator)1 stateme11tJ it defines tlle litruts of melodic 1norio11.3 1 To baJa11ce the greater motion of this version of the n1elody a11d t f> emphasize t.he closure of tl1e \·vhole section, tl1e cadenrnil pattern '"'hicl1 follows tl1e B is extended for three me.'J.SU!es. But the E to \.Vhich it rr)oves, tl1ough I1ar111on.ized correctlj', is in the '\vrong register. Not until the coda of tl1e 111o·veme11t is a C'flde11tially sacisfactor}r E presented in tl1e right register. One final poir1t. Despite the syr11111etry o.f tl1e patterni11g in '\1Vl1ich n10rion above the axis is ba!a11ced by motion beloV\r it, \vhatever implications there are seem to be tho.se gener11ted by the p;atterning abo\1e~ rather tl1m1 belotv, the axis tones. Three reasons help ro explain tllis. First, be.cause the axis-tone is the tonic the neighbor-note below is understood as part of a cadential patten1. E'ven thougl1 tr1e D in 1neasure 11 is a \vhole--step, rather than a half-step, .below the tonic; its leading-tone funccion (e·mphasized and strengthened by its O\Vl1 motlo11 to the do1ninant, B) is clear and unequivocal. Second., because the n1odel is neitl1er strongl}7 sllaped nor clearly closed1 the 111irror is under~tood as a dependen,t reflecrio·n -co1npleting a.nd halan.cing the pattern-rather than as a..tl e11city in its O\Yn right. And t:hlrd; the psychology of patter11 perception may lJe such tl1.at, generally speakit1g, motion abo\re a }Joit1t of ref·cre11ce--for i11st.111ce, atl gxis-tone-see1n,s more important, more Iltarked for consciousr1ess, than n1ocion belo"'r it. The significan.ce of these ad J1oe reaso11S will become apparent in the analysis whi.ch follows. The main melody of the seconc.i lde up of n~ro fot1r-measw·e pllrases. These are related by axial symmetry. 111 the m.odel, the fifth of tl1e scale, Eb is en1belli.sl1ed by .a succession of upper neighbors-a half-step, F b; a "l'l1ole-step, Fq; and a di1niniSl1ed 1

1

7

1

31

In clUs connection, see 'E xample 84.


188

EXPLAINING M USIC

fotrrtl1, Abb (see grapl1 i ). As the ru1alysis below tl1e exa111ple sho,vs, the r t1yrh111ic group is a.n anapest ( ,v-....,,_ - , ) . Fron1 a formal pou1t of vie'"', the pattern is a bar-for111 (n1-m.' -n). In the mirror phrase, the E is embellished b)r lower neigh·bors a.t tl1e sa1r1e i11ter,ra!lic dista11ces. It, too, is an aru1-i1est/ bar-forr11 stn1cture. The seco·nd i)e.riod {1neasure.s i 3 2-1 39) is like the first, except tl1at tl1e tliird i1eighbor-note (Cb) in tl1e model phrase is 11oticeably furtl1er a\.va;r fron1 rl1e axis-tone-a minor sixtl1, rat.her tl1an a diminished fourtl1---..tl1an \ Vas previously tl1e case (in m.easurc 1 z6) . Both beca11se it is marlredly separated frt)1T1 the a.xis-to1J.e, and because the li1rge skip stresses it (recedes ther11. J\Jlelocli.cally, the)r are pro111iner1t bec:at1se they e,-ome at the encl o·f a. r l1)7chntlc pattern. The melodic salience of tl1e tl1ird neigl1bor-notes (the Abb in tl1e xnodel, and the Cb i.11 the rrrirror ) is even. inore parent. For tl1ey ~ue rnore noticeably sep,irated. fron1 tl1e ~xis-tone a.nd, bec~l'use t11.ey a.re approached by disjunct motioni receive additional stress. But "vhat ma.lees the neig hbor~

.

nores the co11tinuing focus of .atu-al actention is tl1ar.~ since the structural, atlt;-tone does 110 .n1ore than l)ersisr, th·e y create tl1e ortl)' melodic n1ot:ion, and that rnoti<)n takes the for1n of a strongly shaped a11d r111rtl1mically regular

li11ear pattern. Yet, though the patterning is pronounced, \Ve are aurare that it is niade of orna.n1entat tones. And this discrep~111cy benvee11 audible promine11ce and syntactic f u.n ction i1nplies contit1uation to an actual Str11ctural ·cone. At this point, two differe11ces bet\.\1een this rr1e:lod)' a,n d cl1e one from D'1~oral{'s ''Ne"v '~lorld '> S)' mt1hon)' should be noted. In this melod.y th·e a."tis-tone is not tl1e tonic, but tlie less sta.ble fi.frh -of the sc..'ale. Conseqt1ently, motion to the tonic is implied b)r tl1e tonal orgartizarion of the pattern. Secondly, in Franck•s n1elody, both 1nodel and :rtu tror are rh)rtlunically close·d and clearly structured from a forrl1al poil1t of view. As a result, though the mirror is related to the 111odeI b)' m\1ersion, it is relaci,,ely autonomous. For both thes.e reasons, mocio11 belo\cv the axis, as w·ell as motio11 abo,re it, shot1ld be implica:tive.


.MELODIC STRUCTURES

......

.._

....

......

-

- - .. - - -:i.-.. ·' - .. --

~

Ex.'tmple

-

I 00

Ho\\,ever~

thougl1 rhe descending pattern of neigl1bor-notes 1nay be irnplicative. it is not forcefully so. Probabl)' this is because mocio11 belo'v .an axis to11c is aJ.\\1ays less pronot1nced tltai-1 £11o tio·11 above it. Possibly, t oo, beca11se descending patterns ter1d co 11ave less urgenc;r-to seem less goal-directedthan asce.n ding c)nes, Nevertheless, ihe C in n1~isur.e 139, and the Cb which follo~vs in 1ne1\Sure i40, can be considered actualizations of tl1e implications generated hy the preceding patter:n of or11amental ron.es (graph 2b) 32 for a nun1ber of .reasons. First, and n1ost iinportant, the linear patterning of potential to.nes is so patent cJ1a:t a1n1<>St an)-r strucrural tone -vvhi-ch firs into tl'te conjunct patterning '''ill be interi)reted as an acu1alization,. As the first sttt1ctural tone belov-i· the uis-line \Vtnct1 l1as n1ore tl1ru1 an auxiliary function, the C fulfills this conditio11. A.n d even thougl1 it enters bel°"'' the whole-note Eb~ '\vhich ends tl1e repetition of the rrurror, the c is pronlinetlt beC'3.U.S:C it begins both a new voice and a oe'v patter11mg---{l dialogue, throt1gh inversion, between th·e first and seco11d \riolin (su pported by t he cello and \>i.ola respectively). Second, tl1e foreground n1ocion ·Of tl1e ·viola in measttre 141, '"' l1ich co:r:1dens·es and sum..r11a.rizes the patt ern of ornamental tones, le{lds to a structural Cb. And finall)r, the higher-level motio11 from strucrural C to structural Cb can be understood as part of the descendirig lli1ear pattern. 32

If a succession of potential tooes creates a pattern implying a more or le$

speci:fic continuatio:n, then arri.val at: a stroctural tone congrue11t wirt1 tllilt pattern wiJJ constitute a sa.tisfacto.t)r acnroliz.atio11 of all the preceding tones. Se-.e p. 197.


EXPI~NL~G

ll'v1USIC

After a v•aried repeticio11 of the dic'llogue ben,ree11 first and second ''ioli·n, the linear patterning .is a.ppare11tly co11cinued as the pia110 n1oves cor1ju11ctly ·dow11 to the Ab (measuxes 14 3-1 45 ) . Bt1t the c-onnection is tenuot1s at best. For a decisive change U.1 so11oriry, texn1re, a11d even melodic strucntre aa 111ark tl1is as the begirulirtg of a ne"v section. In shorr, thortgh the descending on1an1ental co11es of tl1e mirror are a.cn1alized b'' a strucrural tone, the .im~ plicative reiationsllips are 11ot f orcefult ru1d the arrival at the goal is not striking. j't1st the re,rerse is the case \Vitl1 the ascer1ding p~tttern of potential strucru1·al tones iI1 the inodel (graph 1) . But acrua]w1rion. is dela)' ed. A11(l tl1ough lu1ear n1ocion to structural to11.es does occur in the clevelopme11t se{:cio11 at t11easurt-'-S 2 34 and z 38, the most sustained az1d i111pressive instanc·e is that which occurs at irvl1at I take to 'be cf1e begin11i11g of the coda. T l1ere, as ;

. ~

-

- -

-

-~~

.

~ - -

.

Q~

I \.

Example 101 sl10'\\rs, the final note of rt1e pattern of potential rones becomes the strucn1ral axis~tone for the next statemer1t of the model (g:raphs i and z). This relationship is ntade expiicit at tl1e end of eaci1 pattern, w here

the ()receding axis-tone ·moves by a sltt.rred skip to an anticipation of the next one-for instance~ in lneasure 387. As t11e seqt1e11ce contin1.1es, a higher-level pattern of disjunct 111otion b)r n1ajor thirds, C-E-G#, is.. establisl1ed, and c·ootinuation is in1plied (grapl1 .1 ) . Consequ·entl)r, eve11 tho·ugh the third staten1ent of the n1odel is brolvhe11 the beginning of tl1e pattern is presented 011 tl1e high C iI1 1neasure 400. The stat;ility of tl1e l1igh C is de}'endei1t t1por1 ,che inter·r'uption of the patterning 'vhlch u11plied it. Because it is completely trr1ifor·n1, tl1e 111otion by rnajor thirds establisl1es 110 po:int of sy11tactic closure and stability. It is 31J Tlwugh not in rhythn1, or even in register~ becai.1se the pia110 part is iu the san1e regisrer as d1e \riola at1d cello.


M_ELODIC STRUCTURES

circular and might continue il1defi11itely . Consequently1 differentiation is 11eeded if the l1igh C is to functio11 as a go11l and point of completion. The interr11prion creates. tension tluough delay a11d digression., so that when the C is presented, it is understood ro be a point of realization and completion~ A,n(l it is also for this reason that the re.-;tateme11t of the neighbor-note figure is understood as a .rerur·n and a closure, r11ther than as a continuation of the

uniforn1 patterning. Cl1anging-note 711,elodies ,Changing-nQte melodies are st1per.ficia11)r similar to axial ones. Like axial ones, rhey begin and e11d on tl1e sanle pirch\ w.hich is '
E.-xam ple r02 34 Althougl1 axial n1e-lodie.~ are rebtivel}' comrr1or1 in d1e music of the nineteenth

century, cha11ging-note ·melodics a.re not. J\1,aybe, because tlie)' are closed, they ,1l~re not co111patible with the flo\ving lyrictSI11 favored at the time. 011 the other ~nd, ch..'lnging-note melodies are freq.u.e11t in the music of tl1e cigl1teenih and early nine.teCllth. cer1ruries. This is not, l sust1cctt because of a desire for s;rn1rnetrical elega.nc,e, but because the dran1a of tension a.n d arprise~ of delay and v:ariarior1, ca11 best be realized '.>Vhen implications are clear artd strong.


EXPLAINh"llG }\.1USIC

111elodic p,attern and rhe latent ha:rn1onic progression in1ply recu,r11 to tl1e ton.ic. As indicated in the grapl1, the se11se of 'gt}al-directed 111otion, is strengthened by tl1e gap from Bjf to E, wluch impl.ies co11j u11ct fill moving to the tonic.35 Tl1e first nloverr1er1t of Mozart's Oboe Quartet in F J\tiajor {K .307) begins with a sin1ple, u11pretentio11s melody ' vlucl1 ca11 be co1nprehended and remembered v;rithot1t difficulty. Its to11a] n1ateria!s are con1monplace, its form is regrtlart and its processes archetypal arid fa1niliar. Yet behi11d this a.lmost folklike facade lie relacionsllips \vl1icl1 are botl1 intricate ai1d subtle. It is not the orga,niza,t ion of at'l)' particular level of patte:rnin.g thac is complex, f10'\i\rever, b11t the relationships among tl1e several levels (Example l 03) . The openi11g eigtrt: 111eastires-=-rl1e :rnelody prope.r-co·nsisr of t\VO phrases: a 1r1odel a11d its parallel (graph :c and 1 ~1). The .first phrase begin.s 011 F and closes on G; the second, \Vhich js sequentially related., begins on E ancl closes on F. On tl1is level the t u11e is a Sj1Tnn1etrical cha11ging-nore pattern. Indeed, it is an exact inversion (or a retrograde) of Bach's fugue Sltbje·c t (Exan1ple 102) . Irs h:armo1ric strticru.re, like that o-f Bach's fl:tgt'l e subject thou.gl1 different in detail, js. botl'1 t1na111big11ous a11d implicati,re: 1- 11- \ T-I .. TJ1e n1elody is Ilatentl)r a11 itlStan.ce of a c11anging-11ote organization,. Each phrase of th.e larger cl1angin.g-nore organizar.i.011 is in its turn n1ade tip ·o f nvo parts: a i11o(lel and its complen1et1t. In the first phrase, the model consists -o f a rising third from F to A, ru1d tl1e comple111e11t 1·everses tl1is motion; falling a tllird fro111 Eb to G (gr~iph 2). In th,e second phrise, the model mov.es from E up to G~ an d the corr1plem,e nt descends a third- from A t o F. Ho\vever, bec.a,use it c~1n be underst ood as a cor1tinuatio11 of the descending triadic ,p attern of che first phrase 1 t he E it1 n1easure 5 belongs to both phrases of rhe melody-functioning tis rhe end of the prececling con1ple111er1tary plrrase (Bb-G- E) a11d as tt1e begit11u11g of the n1odel part of t t1c second pllrase. Moreo,ver, tl1ot1gh ti1c}r are on a }o\\t.er l1ierarchic level, tl1e C in measure .2 and the Bb in n1e.asu.re 6 so easily fit into th.is schen1e that they too form part of the patterning- making tl1·e complernentary relationships fully triadic (graph 2a). If the E belo·n,gs to the patterr1ing of the first as well as the seco11d phrase. then, thot1gh the high-level changing-note relationship and the partial complen1enrary orgariizacion {graph 1) are completect in measure 8, the Iar.ger con11)lementary Strl1cture {graph .2(\ ) is nor. Just ~s the descending third, a:; Because the E is r1ot a harmor:tic to11e in rela:rior1 to tnocion is not ir1Tplied.

Bf., continued

disjunet:


MELODIC STRUC!URES

193

Bb-G, was 0011tir1t1ed to E, so the descending third, A-F, of the second pll!'aSe, implies continuation to D. The D is presented in meaSt1re 9 (graph 2b) and though ir is the resolucior1 of a long appoggia.tura, C#, it is structurally equi,rale11t and temporally parallel to the E in meas11re 5.36 This is anotl1er instance of th.e bifurcation of forn1 and. process. The 1nelody proper reaches closu.re in ·me:asure 8, but ti1e implicative processes generated by tl1e larger co·mplementary patterning of the seco11d pl1rase contio11e beyorid tl1is point of cadencia.l articulatio11. It is now appar·e:nt-tho11gh perhaps largely i11 .retrospect~tl1at a still l1igl1e:r-level melodic process is invol,red: a lir1ear patterning from F to E to D (graph 3). This linear m.otio·n also i1nplies co11cin11arion. Thol1gl1 the linear motion is more ra:.pid in the following ineasures tl1ere is no l1ierarchic dis~ crepa11cy because the seque11tial pattern tnakes clear that tl1e ·C in measure 1o is eq.mvalent to the preceding D. An 111.1equi~1ocal octave transfer througl1 j

0

. .

,.., -

-

-

.

.®

•

""

-,

· u. ·· ·-

.

.- •

Exsn1ple

1S

103

na The pa.rallelism is n1ade more palpable by the conformant relationship bel'\veen these two points in the melody- that is, because t:he motion from the F at the end of measure 8 to the C# and D is analogous to the liighe.r-le\~el pattcrtung fro:m G ( m4) to E(.n-..5) to F ( m.8).


EXPLAINING l\'lUSIC

the tonic ttiad is follovv·ed by continued conjunct motion thro·ugh Bb to A. 37 ~fl1e11, after a delay of fi,,e rr1easures (during \vr1ich. tl1e ap1)oggiatt1ra pattern of measures 9 a,nd 10 is repeated a11 oct;ive l1igher),~ tl1e rnelody moves, as thot1g+1 recapin1L1ti11g a11d st1mmarizing tl1e n1otion of n1eastrre 11; to a strong, err1bellished G \vhicl1 moves t<> a cadential F in measi1re ao. Here tl1e x11e.Iodic processt~ ge11erateti by cl1e higl1est le\1el of n1otio11 react1 clo5me. This melodyr is as clear an exan1ple as 011e could "'risl1 for ()f a srronglyshaped J1ierarchic strt1crtlre. Or1 tl1e highest level, t11e n1otion is that of a complete and relatively regular descending F-major scale (graph 3); on the ne..n level~ the patterning cor1.sists of a S}1111n1etrical cl1angi11g-r1ote relationship (graph r and ra); and or1 the level of the l1alf-plrrase, the orgilnization is patet1d1r complen1et1tary (graph 2a and 2b) . Each of tl1ese strt;1ctur es is

an instar1ce of a ·well-lcnown, traditional scJ1err1a. Yet their con1binariot1 is co.mplex and elegant. Tl1e v;1l1ole is said to be strongly-shaped because the several levels of structure s11pport and rei11force one anotl1er. What about t11e lo,1;rer levels of organization? The note-to-note foregrom1d of tl1e first phrase cor1sists of ascendi1Jg co11jt111ct motiori from F to D., follo,ved by descending conjt1nct motion back to G. But tl'1e predomin:mt patterning is clearly triadic. The G in meas11re 1t tl1e Bb ir1 rrie-asure 2, a11d so 011, are passing-tones. This parrerrii11g is begt1l1 a11d e111phasized by the upbeat slcip of ·a f oiirth. N om1ally there wou1d be little qr1estion abc>l.lt tl1e implicative relario11st1ips of tlus sort of patterning. P~1rtict1larly whe·n preceded byr a11 upbeat in t·he lower octaver tl1e fifth wo·uld te.n d to function as t he relatively stable upper limit of n1elodic 111ocion (se.e E.xam·ples 76, 84,, 85 and 98) . ln this case, 110·\vever , lineru.· arid t1·iadic co11ti11uarion to the octave a·b ove the tonic, F, is 11lore pro'b al1le. Ir is so largely because o·f rl1e n1elodic and rhythmic position of the D in 111easure 1. Because tlle pa.rat11eters do not n1ove cor1gr11er1cly, t l1e fu11ction of the .D is ·eql1i~.;.oci1l belo11gi11g partly to prece
37 That tile Bl> in measure 11 and the .A in n1easllJ;e 12 are Strucrurnl tones is clear fror11 t.l1e harm<>n)r, \Vl1iel1 is sketcJ1ed in below the example. S$ These rrieasures. are not given in Exa·mple 103.


195

l\if.ELO'DIC S'ffiUCTURES

tonic is almost equally probable.39 Tltough ti1e D can be understood as ,part of the follo\ving triadic patte:rn (D- B0.-G) ir is more p~tently part of the preceding conjunct motion-follo\\rir1g fro1n the C, r·ather tha11 1noving toward t h.e next structural tone, Bb. Ten1poral relationships, ho1'\rever, are the cn1cial pan1n1ete.r. As the analysis under £'(ample 104 sl1ows, th.e prevailing rl1ythn1 ·Of the fu·st t\lro rneam1res is that of ai1 eighth-note f ollo,ved, by a longer n:o te basicc.1lly a dotred quart er-note. H~1d this 1)atten1 been continued, so that the D came as tl1e last eigl1th-11ote of 1neasu.re i, tll.en t11e D would l1av·e. been a relarivel}"' inconseqt1e11tial or1utnental tone (an upper neigl1bor) \-vhich, for reaso11-s of pro·xitnity, would ha,ve been grou ped as an upbeat vvith the follo\:ving patter11. Because it enters to-0 soon, breaking the establishe,d iambic rhythrn, the D is not only emphasized, but connected \vith. the 1pteceding C. As a result an unpression of lit1ear tnotion is, so to 51)e.ak, left hanging in micl ait-\vitl1out satisfactory co11tinuacion. l\1Ioreover, becallSe tl1e D is a '\Veak a.ftetbeat conling frotn t11e C, the C itself is made mobile. l11sread of bejng tl1e goal of an end-accented ian1b, it is also tl1e beginning of a: trochaic group. Tl1e resulting mobility raises the possibil.. 1

ity of triadic c·ot1tinuation. It '"'onld be a ·mistaJ{e to exaggerate tl1e forceful11ess of this irnplicacion or to suggest that it is accompanied by a i1oticeable sense of incompleteness. Tl1e stren.gth of the. higher levels of patterning is st1ch that the melody is

felt to be satisfactorily closed in r11easur:e 20. Nev·erthelcss, the listener is; I believe, someho;w a·ware of tl1e possibility of lirJeart at1d perfiaps triadic, concint1ation to the higl1 F. And it seen1s clear that Mozart 'vas too. For thougll long delayed, the lugh F does occur j11st be.fore t:he final cadence of the movement- and it is reached by explicit and unequivocal conjntict motio11 from the D (ex.ample 104) .40 3DTltls differenc,e is 'p robably a result o f prox.i111ity .r elationships. In the tni.nor

mode. the sixth is onl)r a semitone a.Wa.-}t from cl1e fifcl1 and. is separated, fro1n the seventh by a \Vhole step in t he <·'melodi,c't 111i11or a.nd by :an angme11ted secon,d in t he ((hru.,11ooicn 111.inor. lo the n:1ajor r110de, ot1 the ot:h:e~r h.'l.n(t, t l1e sixth is e.quall}t disrnnt f r

L'tPLAINING MUS:lC

'

'

• ,, r· &.:&.., ' r·

~

Example

104

CONTEXTUAL DISCltEPANc:Y

Context "''as not an important consideration in the rnelodies analyzed in the preceding sections. Attention ">Vas directed to the i111plicacio11S generated by an order in the pattern irself. \IV hat '\'\1 as implied "vas. the continuation of a process until so·me relatively stable goal 'vas reached even t:b.ough ar1rival at st1ch. a goal w·as at times deflected. or delayed. In this section \Ve vvill be ooncerne(l, though all too· briefl)r, ¥v:ith a different ki11d of implicative relarionship---0ne in vvru.c h a. discrepancy bet\veen rhe stn1cn1re of an e\1 ent and its syntactical or f orn"lal context su.ggeStS that a change of functlo·n is i)robable. In1plicacions are ge11e.rated not becatis-e a. pattern suggests contint1ation to an internal goal, btit ratl1e.r becau.o;e; an e\ret1t is felt to ha\re a potential ft1nction '\vhiC:h has not 'bee11 satisfactorily actt1alized.

PoteTitial tones and r}J3'thms A potential structi:tral ·cone, to repeat the de.finicion given earlier, is on.e wl1ose melodic pro1ninence is not 111.atched by its structtrral importance. More specificall)1 , it is a "''eakbeat marked for conscious11ess in some 't\rayoften becat1se it is approached and left by disjunct 111otion, a11d ge11e.rally speald11g, tl1e larger the disjw1ction, the n1ore pron1inent and m·ore strongly irl1plicarive the note. P:otential structural to11es may occ·u_r singly or in a patterned series. In the former case, tlte n1elodically 1)r-0minent weal\:beat is actt1alized \\rhen tl1e same pitcl1 st1bse~t1enrly occurs as a.n accen.ted structural ror1e. In the latter ,case, the \\rhole series is ''resolved' 1 vv•l1en a s-ubseql1ent n.ote understood as part of the patre:m is an .acc.ented stn1ctl1ral tone. Because several potential structural tones have already been discussed.;l1 only two '\vill be analyzed here. ~ 1 See

Chapter IV, Exa111ples 50 and 53; a11d Cl1a pter V) &-:t111ples 66, 95 1 and

100-101.


.M ELODIC STRUCTURES

197

The Burlesca from Bach''s Partita No. 3 for Harpsichord contains a clear exan1ple of a series. of potential structural tones (Example 105),. The meJo~dy of the soprano voice, 'before the cadence that closes tl1e first part o.f the movement, is a bilev'el structure. The primary level consists of a conjunct pattern 111oving from C through D to E (graph z) . This pattern is embellished by a subsidiary line o.f potential tones a sixth above. T ·he A and B in. n1easures 1o and 1 2 are aurally promine11t but are strticturally unimportant we;akbears (graph 1). Continuation r·o C is implied not only because tl1e secondary level is itself a clearly patterned conjunct motio·n, l>t1t because it parallels the primary line. Consequentljr, '"'hen the E occurs in n1ea.sitre t J; irs sll.rth, C, is itnplied. Instead of co·ming on tl1e third beat of measure 14 (the previous merric position of the secondary li11e) > the C comes on the a.c ceot. It is a structural tone, a.ctualizing the i1nplications latent in tl1e preceding series of pote11tial tones. Botl1 fo·r this reason, and because it is an octave above the C in measure 9 'vhich began the melodic process,4~ the C in measure 14 is an important goal. Indeed, the rising line of eighth-notes in measure r 3 st1ggests the C as

the goal of ·the primary level of p.atter1ung also. Bt1t this .is n:0t tl1e case. The implicative relationships are more complex. In the first place, the patterning of the prin:mry level of "morio11 has bee11 .in nvo-measure units; a11d the eightl1-note ''continuation'' is obviously on a much lower level. If the hierarchic levels are. kept straight, th.e n1otion from E to C .must be analyzed

.as disjunc~ and tl1e C \Vould not be a goal buc part of an implicative gap of a inino.r sixth. Tl1ough tlris relationship ut1doubtedly l1elps to define the 1,

F

·-

... -

r

S,

-

E

E.xample 1o5 42 Tha:t b.oth Cs ar·e hart.nonized b}' sub-dominant chords in root po.si.t.io11 eotphasizes this relationship.


E.."\PLAINING MUSIC

direction of motion in measures r4 and r 5, the E~ reinforced hy its 'nic re11sons: it is accon1parued by· iUl u11sc1ble, first-inversion triad, and tl1e tr:tad itself is part of a uniform chromatic bass-lin.e "vJ1ich lacl\:s sy11tactic articulation. vVhlle the n1-0tion of the primary level is thus suspencled, a degree of

111

bilevel patterrili1g 11evertl1eless co11tinues. For the desce11diJ:1g, C0[1junct motion from C to E also contaj115 a subsidiary le,rel (graph 3), thol1gh one tl1at is n1ucl1 les-s obtrusive rl1an. that cre.ated by the line of potential strt1ctt1ral tones. And tJ1is subsidiary, per·haps rertiar}r level., seert1s ro derive in :part fro111 the risi11g eightl·1-notes in m~11re 13. 43 As the an.al)rsis inclica.tes, all tl1ese levels co11verge to the E "vhich closes thls part of the Bttrlesct. Though the actt1alization of potential srrucrt1ral tones is, as a rule, proxit11ate, it n1ay be remore--it nlay tal{e place after the closure of tl1e main n1elodic pattern. Sucl1 reinore actualization occurs in the Min11etro inovement of Haydn's ''London', Sy1nphony (No. 104) (:Exa111ple ·ro6). Consequently, co explain ho'\.v the li11plications generated by the potential sm.1cn1ral tones are actualized, lat«:!r parts of the movement, as \Veil as the opening melody, muSt be considered. Tiie main n1elody, whicl1 opens the moverr1er1t, is an ir1teriesting mixrure o,f different ki11ds of structu.r al relationslups. The first pllrase is essentially axial- ai1 A en.1bellish·ed by neighbor-notes \.Vhicb are stressed by sforzandi (graph .2) .'" The :first nvo neighbor-11otes, tl1ough stressed, ate not melodically pr.01nit1ent, 'bttt tlle tl1ird D, is-eve11 thot1gl1 it is cot1:nected throt1gl1 the B. to tl1e axial A. For not only is it separated from the axial tone by dis· junct motion, bt1t it is bo·rl1 the upper to1uc to which the fifth., A~ might 11a\r.e n1oved, and it: is an octav·e above the upbeat which begins the melody. Because ar the same time it is also rhytl1mically t1nstable a11d . Strltcturally weal;:, the D is
'18 It is related by confo.rma11ce t

MELODIC STR:UCTUR'&S

between these pitches in meaSt1re 3. But the functiot1 of the B as a potential structural tone is not t1neqruvocal and explicit ·until m,easure 6. There, though it rernains rl1ytl1mically unstable a11d structurally· " 'eak, it is n1elodicall)l" prominent (graph 1a.), for it is both stressed by a sforza':tzdo (relating it to the earlier D ) and above and separate
The B is also in1portant, and her1'ce en1phasized, rhythmically. For· though the last grot1p in tl1e fust section, lil{e the preceding ones, is an an1p'hib:racb, it is twice as long as the others. As a rem:tlt, the sec-011d plirase is an end-accented anapest (level 2 1 in tl1e analysis under tl1e exam.p le), and the \Vhole melody is an end-accented iamb (Je,rel 3). The melody is closed rhythmically as "'"ell as melodicall}'· ,~.t ~ f'll>-.(P$1J '

-1

. '

'

'

-------=

..Ju.Si-.1.ST

,.,... ,

p .; 17

Ct••-<••••>

. . .t<'r

3

2

Si

..

>a-·.... R-..i..:..· _ .._}

~·

'

a.f-l r,.ti 1T, .9 r, f ..

\

,-,


lOO

E;.."t,P LAINING MUSIC

Yet closrue is by no means definitive. Most important, there has been no tonal deparrure and onl)r minimal structural tension in relation to \vhici1 return a11d stability \:\1ouJd create strong clostire. But tl1e1·e are other reasons as \.vell, and tl1ese are effeccive precisely because no higl1-level processes have been generated. The first is a matter of l1ierarchic equivaler1ce. T .he established unjr c>f melodic-rltytlunic morion in the fi rst six measures· is three quarter-notes long. Bur the Fl i.n meastue 5 is only a qt1arternote; it is not a structural 't one on this le,1el of n1otion. Consequentl)r, the conjunct fill implied by tl1e initial g-ap is only proc1.risionally :realized. 111decd, tlus aspect of the patte.r ning of the seco11,d. phras.e it11plies, not the D \Vhicl1 follo\\1S, bt1t continuatio11 to A- \vhich proves to t)e the first s.trt1ctural note after t he double bar (gr·aph 4) . Not u.ntil after the 1nelody is repeated in rhe se'cond pare of the 1noverr1ent-as th·e reader vvill discover if he tur11s back to Example 58- does the Fl occur as a structural tone. There th.e pat~ ter11 is so strilcing and the octave displacement so clear that the change in register does not obscure tl1e realizatio11 of irnpl.ications: the conjunct motion fron1 G to F# in 111easi1r es 41 and 42, and from E to D from mea.o
n1easure 18, tl1e potential tone in1plied earlier is realri:ed in measure 20 (gn1ph 1a), where the B co·mes in the proper regisrer, as the goal .of a cad.encial progression (V/ VI- ' 1), a11d as tl1e final note ·of an end-accented rhythmic group. The D occurs as an accented note in t,he prnper register in meaSt1re 11 (graph r). Bur as the se,1entl1 of a secondary clonunant ( \ T7/V) ., rather than the ro-or of a, to11ic cllord, it lacks stabilit·v. It is only .. a ·provisional actualization. The stability o.f the D is also weak.ened by tl1e tem,poral 01~ganizatioo. The .D is not the goal of a closed, end-accenred group, hecat1se repetition preclt1des the possibility of unequivocal patterning. All that remains is n1eter, and even that is threatened. For " ' hi1e tl1e r epeated D~s are still pre~ swnably grouped in th;rees, tl1e accompaniment, indicated by the s-ketch t1uder t he staff, is in rnTos. ~


M .ELODIC STRUCfURES

2 ·0 .I

The kind of implicative relationship inv,olved here is 11ot, strictly speaking, an aspect of melody: it is wh.at I have called. tl1e "'reakening ·Ot shape."5 Not only is the metric orgmlizacion weakene'l by tlie cross meter; but n'ldodic patterning disappears it1 the succ,ession of repeated D's. The lack of distinct and regular I)attenling creates a sense of tensioi1 a.n d t1ncertainty, itself implicative of a, return to the ?>)"'Chic security of stable, unambiguous

shapes.46 Tl1e .in1plications of the D, provisionally acrualized in measure 11, are more than regen.erated \vhen the melody is re1Jeated in the second part of the Minuetto. For there the D's are, if ru-1y·rhir1g, both more pro1nineot and less Structurally important than in the original stateme.n t of the melody. But tl1e function of the D as a potential saucttiral rone is even n'lore striking in meas.ures 42 and 44 (Example 58). In both measures, the Dis marke-dly separ·ated from the main melodic n1otion, and in tlle second, it is followed by a. grruid pat1se left \vithout immediate continuation. 1ts contintiation and itS actualizarion-and that of all ·me preceding Potentially s.trucn1ral o·s-takes place at the very end of :the movernent, \vl1en the final D occurs as the goal ·Of an end~accented rhytl1m, with tonic har:m.ony, and, in the right register.47 Th:0ugh poten,tial rhytl11ns are consider1\bly less common t:han poten~ tial structural tones, the last m-0\~ement of this same Ha}rdn Syn1phooy con-· t-ains a clear exan1ple of this kii1d of implicative process. It oc·cu.rs in the passage wluch begins in measure 19, f (lllo,ving the second statement of the main melody of the movement (Exan1ple 107 ). The melodic-rhyrhrnic pat·cern of the passage is derived from the penultimate n1casure of the n1elody (measure 17) . The second. half of that 1neasure (the E) acts as a pivot linki11g · it to measttre r 8. It also groups \\ti.th the precedirig A for both harn1onic and rl1ythmic reasons: harr11onically, because rl1e E and A belong to and are harmonized by the same triad; rhyth111ically, becat1se the trochaic grouping established io the first t\vo ,measiires of tl'1e phrase tend to be co11tint1ed. (level I). The E and F# a.re also .grouped ' vith the following downbeat both 1

0

-

because of their melodic proximity to the D, and because the inore rapid morion of the final eighth-notes malce both tl1e subprimary (i) and primary level {1) trochees mobile and a11acrustic.

'*5· See E1notion a:rz4 Meani;ng in MtJSic, Chapter 5. Harrnony wh.ich is patently pattenl.ed i.o; specifiailly implicative~ nnd it gives direction to the implications generated hy the v.1e.~ke11h1g of die other parameters. 41 Perhaps because of its promiI1et1ce, Haydn actualizes tl1c B as a structural ro11e in n1eas:u1"e 44-even though its previ-0us acru:t.Lizacion in measure i8 should 1111.ve 40

sufficed.


EXPLAINING M .U SIC

202

In measures 19-2 I, ho,,rever, the second half of the m.easu1·e does I10t a.c t .as an upbeat to· t he follo~ring measure, for two main reasons. First, as noted in C!1apter IV, repetition does nor make for cohesion bt1r for separa·rion. The rhyrl1mic identit}r anti intenrallic similarity benveen these measures is such that each is perceived as ar1 ir1divid11a.l e11rity. Seco11dly, tl1e rnarked pitch disjunction bet\\1 ee11 tl1e e11d o·f 011e. pattern and tl1e beginning c):f the i1ext pre,1enrs the eig11th-11otes from f unct:io11ing· as upbeats. The nvo parts of each nlea.sure cohere 11ot only because cl1e)r belo11g t:o che same harmony, but becal1se tl1ey a.re pate1:1tly croclllic. The trochaic pattern is strongly reinforced by the o.rchestration. As tl1'e abstraction of the acco111panime.nt rl1ythn'1 (grap11 4) sl1ovvs, cl1e b,rass and tympani (stems do,vn) play a half-note on the first part of each me.a.sure. By thus stre.ssin_g the accent, they help to create heginning-accente·d trochees on level one ( r). The wood,;\>in.d s and lower strings (stems up) reinforce

this organ·ization and, in addirion-t create trocl1aic groups on the sub·prit11ary level (i) ~ Nevertheless, despite all t11e aspects of t11c pattenili1g "vh.ich inal{e us percei1le the groups as. rroc.hees, \:ve are aware, both because of their relative speed and beca11se of their previous ft1nction as pivots, tl1at tl1e final eighth-notes ir1 these me.1.sures sbo1JJd be t1pbeat:s (gi:aph 1). Tl1ese potential rhytluns imply a patte1~n n1 "" hicl1 tl1e final quarter-11ore of each. measure '\:vill be grouped with tl1e follov.ring downbeat in an end-acce11ted rhythm. .t •

ii

--t,,,,~~---.,~·

.:/::1-·

-

f

'2 · ,........---· --;· _.,, ,· ' ..iJ,

•

t=:=================:j

1--~~~~~~~~~~~--t-i-~

;

" I.

z '1

,.at

~

,ii9 '-'J

\~"1

__,!"_,,_ "1Ji -

\.•

.

\~==---~"':--1'

\.. V ,t

.

-

~

'\~ Vatc ·~' \a-:4'/J t-:V J\ ,,.0 V/'f:"" Jt:-t.1.l\::(;b3Jt:'ke:At'.: 1 'iflt..~ _ ( ...,

\;M: 1 . b:;1~ 1,,...

\

v <. -

··:-~

I

7

v" ,.•:=a
y, ,,_

v

I

'

-:--

.-., ~,

v I

--~-----~ .. -j..J ___ _ ,,1.,_._,~ . =:::::...._~--~ ·'-'~ .. ~----'I

'

Exa.mple

107


lVIELODIC STRUCTURES

The implications generate,d by the rh)rtl1mic organization are C<»mplem.ente.d and strengthe11ed by the n1elodic patterning. For though the rhythm of the paired m.easures ls identical, malcing for separacio11, there is melodic motion between th.em. Because register i"> c·onspicuous and melodic-rl1).rthn1ic activity is gre'Ster, the second 'half of each of rl1ese r11easures c-aptur·es attention. Fro1n a melo·dic poii1t of vie\v, they are n1uch lik.e potential stru.ctural tones. As a result:, we are speciall)' aware of the r·elationship betWeen t he G a,nd B of measl1re 19 a:nd tl1e Fj arid t\ ·O·f measure 10, 48 and the linear pattern thus established implies cor1tinuacion to a point of relacive stability. And the fi1llll eighth-noces of e~.cl1 n1east1re create a gap \v·hich implies that conjt111ct n1otion in die opposite directio.11, fillin_g th.e gap, will follo~r (grapl1 3) , All these implications are :realized in the f ollo,ving measures. The final eightl1-nt)tes of meas11re 21 become acroalizeti upbe>ats1 perceived as connected \.vitl1- 1no\ring to~the a,c ce11t i11 111eaStue 23. This patterning is supported h}' a change i11 orchesn·atio11. For che fuse time, the acco111panying instrtunents play a patrern \.Vhich n1oves from tl1e last beat of one measure to the first beat of the follo,ving one. The flt1res play" st1ch. a rhythm from the laSt beat of measure 22 to the first beats of 111easu.re l 3, an.d then all the instru.ments play the rhyrhn1 fron1 cneasu.re 23 ro n1easure 24. The foregroun·d gap is tilled as the A at tl1e e11d o·f n1e:asure 22 n1oves co G, and the 11igl1er-level linear patterni11g conru1ues to E-a.nd beyo11d, 49 Tl1e resoll1cio11 of the pe·n t-up tension, caused by the delay of goal-directed melodic and thythmic processes, requires time and the co111pensa.tory en1phasis of reirera.. cion.50 Not only are both the upbeat ,rhythn1 and. the gap-fill figure repeated as the higher-level pattern continues) btit meaStrres 2 3 and 24 are theinselves twice repeated before the impetus is sufficient!)' dissipated, allowing the p.assage to mo·ve to temporary closure on the Din measure 29:. Closely related to potential rl1ytruns :are those vorluch can be cl1a.racreri S The n1otion between first bears ( fron1 B to A ) seems so1nei10'\' subsidiaryalmost as tho11gh the t:e111poral order of ru1 ~·oornpah" bass p11ttero h_ad been .reversed, with the repeated n_o tes of the acco111pattl1ne·n t figt:t:re coining on the acc·en:t ID.stead of 011 the " 'eak beu. 49 Because. the E cannot act as a goal. th.e lirle.a r 111orion cot1tinues and., after two repetitions of measur·es :z 3 and 2.4'. r eaches te11.1por:a.cy' st3bility on D. This continuation is indicated. by the art-o\v a·r the end of gra·ph-line :z. lW Because this is the beginning of a ttansicion passn&,~. Iilatked clo~'\U-e of a i·ev.ers:al would be om of })lQce. The accumU'l1l'ted 1no1ne11tunl of preceding events tnust be allowed to slow down of itS owi1 accord.

me


EXPLAINING MUSlC

ri:ed as incom1)lete. Rh}rc:hrnic inco111pleteness is t1ot ortl}' in1plicarive in ics ow11 rigl1t, btit qualifies a.net co1nplerne11ts the implications generated by mo.re specificall;r melo·dic retarionships. H.o\vever, since this l{ind of patterning is disct1ssed else,v1'1ere,51 a brief account and a single example "vill suffice l1ere. Tl1e existe11ce of similarity a11d difference is, as 11ote·d earlier,62· a11 esse11tial condition for patterning. This is as true c>f temporal patterns-rhythmsas ir ifi of melod.ic .or harmot1ic e ents. If a series of e\1 ents is so utuforn1 t11at it provides no basis for clisti11guishi11g acce11ts fron1 weal~heats, then the r hythn1ic structure i;\rill be incornl)Iete.~3 And because it .is iJ1oon1plete, it is implicative, implying the probabiliry- even the necessity of a. differentiation r:o \vruch, the stlccession of lilce events can be related. Because the repeated note or patterr1 is i11 this sc11sc goal-direc.-ted, the serie,s is generally understood to 'b e anacrt1scic-a suc·cession of weakbeats movi11g to\var,d an. organizing, differencia.ting acce11t.l'latterr1i11g and at the sa:me rime makes it difficult for cl1e list ener to i111f1ose St1bjective diffcrer1riation 011 d1e l1igher-le\'el t1niforn1icy · as he can easily do \\1hen, say, a single note is repeated. Because they are unsta;ble and goal-direcced, inconiplete rhythrris are 111ost freque:n tly e:r:icounterecl in cor1jw1ctior1 v;rirh the 1nelodic rnobility ai1d ltarmonic tension characteristic of tral1sition passages and development

sectiotlS. But the}' occur else\vhere as v-rell: eve11 v\ritlm1 ·pritnarily tl1e111atic 51

Cooper and f\
p:issim. GI? P. %60 . Ga Temporal uniformity m11st be co1nple n1ented by unifor1nity in other paran1eters, for changc.s ir1 pitc~ dynamic , harmo11y, and ti111bre can bec.ome tl1e basis for ri1;rtl.unic differen,dacion. Indeed, changes in other parnn1eters "vill te.a.d to be specia1Jy n.oticeahie p.recisely· because temporal relationships are unctia11gccL 114 Anapest rf1ytl1ms, frequ.e ntly the basis for pl1rase st:ru.cture in the exan1ples in this boolc (see Exarr1ples 79, 86, and .1:00 ), are na. 'ecru: inco111rllcte rhythillS. For \.vhen a p.attc.rn is repeated,, ,even at a d:if'fe.r.ent pitcl1 level, the sin1ilarity bet\veen eve·nt:s te11ds ro "veaken temporal pa.tternin.g and is raken co imply morion it o a larger group chat serves a.s an organiz1n.g acceo.t a.11d goal.


MELODIC STR.UCTURES

20)

statements though not as part of stable, cadenrial ones, st1ch as closing themes. This. kind of implicative relationship is much more common it1 the dramatic style of classicisn1 an.cl romanticism than in the Baroque style,

for ex,act reperitiont \\rhich rl1ytl11nic i11con1pleteness requires, is not consonant with the fio,v of contrapu11t:al voice-leading favored in the earlier period Thus incomplete rhyrl1ms are characteriscic of Beethoven1s music-for instance in the firsr movement of rh.e ''F.roica" SymJlhony {measures 128131 and 272- 179)- but not of Bach's. The ScherLo of Brt1:ckner,s Se\re11ch Syrnphony begins \vit~1 an incomplete rbyth.m- four measures which are exactly alike (Example 108). The lo,vest level (i) is clearly and strongly srrt1cn1red. Accent and weak beats create a relatively closed dactylic grouping: the initial eighth-notes make the second beat see111 stressed and tl1ereby . dirninish. tl1e tendency of the last quarter-note to act as an anacrusis to the f ollo'\'Ving measure. On the next level ( 1), however, tl1ere is 110 different.iario·n an·d, consequently, no basis ;

for grouping. Because beginnings are generally' marked for consciotl.Snesst the first measure is assumed ro he acce11ted-but 0·11ly 1I1ome11tarily. Wl1en exact repetitio11 follovvs, our understanding ·o f the first measure is changed. I·n retrospect it is understood, together v.vith th.e next three measures., as anacrustic and goal-directed. An e\rent with which the un~tterned repetition can be patterned, and which will serve as an. accent for the series of firstlevel ( 1 ) weak beats is implied. The motto melody, played by the crumpet in measures 5-8, provides the nec,essary differentiation and is the goal of the preceding measures.

. &wi,.t • .. p .

•

'

1,

\

- _ ,_,-~· t.h - -_ • w

'

_--.aJ ,,

, f"fI~. ---a...J. .. -v~,. ~~w- ,. . .

1

'

• • , ,_ l

-

--

l

l

~

'?

...

-

-

--

..

'O ., ,. -

Exan1 p.le I 08

Tl1e first four measures are .implic-acive 1nclodically, as well as rhythmically, in two ways. Negatively, we are sure that rhls is not the t•real'> '


106

musical substance- a theme or 111elody-but on1y

LU>LAINlNG l\'1USlC

ostina.tolike introducti.on. V\Te feel tl'tis both bec1t1se; taken as a \:vl1ole, the relationships are addi~ rive and formal rather tha.r1 S)rntactic a11d processive, a11d because otrr experien.ce ~rirl1 tonal n1usic tells us tl1at this lcind of structu.re does oot normally conscit11re the mai11 ·basis for n1elodic orga1iizatio11. On the positive side, the rela.t ionshif)S "''ithin. the repeated moti\re are syntactic, and they generate specific implicacions. 1~11ot1gh these relatior1sl1ips may seen1 11Unin1al-almost lost in the i11siste11t reJ>ericio11 arid oversruldowe.d lly the e11trance of tl1e n1otto n1elody they are not iI1co1isequ.e ntia:l. Since tl1e smallest unit of .regular motion is a q11arter-JlOte, tl1e eigl1r~1-11ote Bis 01lly a passing-tone. Consequentl)', the first rrlelodic relacionslup is cl1e thir·d from A to <: The third implies disjurlct continuation-co the fifth, and perhaps the octave a.s \veil- and,, ft1nccio11ir1g as a gap, conjt1nct motion in the opposite direc.cion. The second of tl1ese possibilities is realized \!\tithin tl1e 111.oci\re : the B on rl1e third beat fills tbe gap. Tl1e linear pattern (from B to· C) \.Vhich results is itself iinp.licative suggesting continuation to A, ~nd perhaps her yond (gnipll 2 ). Both tl1e disjt1nct continuation through E to ..A;. and the conjunct descending motion t:.o A and beyond are provisionally realized in measures 6 and 7.63 E.ven this provisional realization ist hovvever~ overshadowed by tl1e entrance of the morto n1.elody in rl1e trun1pet. Satisfactory realizati.ons of both implicative relatio11ships occur later in the movement, but in a key other than the tonic. Bu.t in the coda. of tl1e Sche1·zo1 realizatior1 talces place in the ronic: first, the C<>ntinuation of tl1e cor1jun<.."t pattern leads to the lo\ver A (graphs 2 and za); then tl·1e oontit1uaticJ.11 of the disjunct 1nocion leads chro11gh E to a11 octave A and there the move111ent ends (graphs 1 and 1a).66 a11

JUelodic gesture and pctentiatity Earlier in this cl1apter, .melodies '\\;"ere grouped according to the kind ·Of patter11iI1g \Vl:1icl1 ge11erated inlplicativc relationships-that is, as bei1ig linear, triadic, gap-fill, con1pletnentar}", and so on. But 0th.er modes of classification are also possible. Not sw·prisingly, melodies which perform the sam·e general ft1nction in a particular lcind of composition o·ften h~i-\re 56

The E is onlj' a provisional realization because it corries on a \.Yea.k eigl1tl1-c1ote; the linear continuation is so, becau.'e it does not follow from t t1e genet•acit1g conjwict •

lllOt:lOn. 56 Beca11s:e

violas, celli and basses play these patterns one~ ~vo, and three octaves below the pitches played by the violins ( tvhich are the ones given in the E.~ample),

registtal hnplications a1·e satisfactorily realized.


l\{ELO.DIC STRUCTURES

207

corrunon characteristics, a kind of fa1nily resemblance. So1ne n1eloclies see1n typical o.f the beginning of sonata-£orm mov"e1nents; others are members of the class of .closing the111es. Some n1elodies are cl1aracreristic fugue subjects; others seem like themes upo·o which. a set ·of \rariations might be built . .A good nt1mber .of melodies ca.nnot be classified in tl1is \vay because they arc more or less neutral \.Vith respect to function. At prese11t there is, to the b~'t of n1y knoVtrledge, no adequate typology of the compositional gestures ·of tonal n1usic-though a nu1nber of scholars are, I believe, doing '\vork i11 this field~57 Such a tyl)()logy is sorely needed ir1 analytic criticism. Creating it \.vill require the close cooperation .o f the disciplli1es of style analysis and music theory. The account given here does not pretend to be. even a beginning. Rather, after a number of prelimin,ary remarks~based, unfortunately, upo11 too small a sample and too little sn1dyit seeks merely to illustrate ho\v this sort of implication 'v.orl{s. 'Typical composiriona.I ge.smres do not appear to be defined primarily by .inter11al melodic structure. 011e kin.d of melodic patterning-say, a ga:pfill structure--can evidently be tl1e basis for a nu1nber of different kintls of coinp-osicional gestures: a fugue s:ubjecr, a sonata-allegro theme, or the melody of a nocturne. Special rl1ytluns, usually in co:i11bination with a particular kind of interv·a.Ilic structure, often se.r\"'es as a sign of a type of gesture: for instance, a fanfarelike 1nelody or a march tu11e. Harmony, too, may play an important role in specifyi11g cl1e function of a n1elody. Forrnal orga1uz.ation,, 110\\rever, appears to be n1ost irnporta11t of all, defini11g a11d llirutir18' compositional possibility. For instance, ru1 antececient-consequent srrucrure can function a-S the opening melody of a sonata-form movement or as the first part of a theme t1pon whicl1 \1 ariacions "vill be based, but it is too patently dividect into phrases ai1d coo closed to ser\1e as the subject of a fugue. To a considerable extent, then, our untierstanding of the function of a compositional gestuie depends u:pon. its own characteristic :features at1d inten1aJ stru<..i:l.11"e. But it also depends upon conve11rion and p1·actic e and the impljcarions generated by gestures are i10 less influential because this is the case. A typical melodic gesture is as mt1cl1 a component of a musical style as are tonal sy'ntax and archetypal melodic or rhythmic patterns. A competent listener t1nderstands nor only the irr1plicatio11s ge11erated by to11alt melodic; and. rl1ythmic relationships withll-1 a pattern, bu.t th.e functional potential of tl'1e pa:tt·ern as a whole. This is partly the reSl1lt of his having learned, tltrough 1

Becau egestu.re and ethos ar~ incit1bttely connected~ Frank Kirby's stnd:y of the "Ch.1.racterisric S)"mphorqr11 may 9]so· be i1nporUcnt. See ChaJ>ter Ill, pp. 68f. 61


EXPLAINING MUSIC

208

exper:ie:nce as a listener and per.haps perforn1er, ho\v different kinds of gestures tend to beha.ve. Characteristic tnelodic gesrures usually occur in appropriate and familiar contexts. '¥hen they do, a more or less specific con1positional procedure is implied, and perhaps f orr11aI orga1i.iz.ation as well. Such il11plicarions may be realized directly, or they inay be clelayed for a cin1e. il1 so111e cases, ho'\Vever, there is a discrepa.n cy berween t:l1e normal function of a. gesru_re and its actual use in a COIDJlOsition. Wl1en this occurs,, \vhat is implied is tl1at the gesnrre \vill probabl }r be presented in a oontext consonant witl1 its customary and characteristic ft1nctio11. The Finale of Scl1urnann>s Piano Quartet in Eb Major, ·0}-JUS 471 begins with a melodic gesture ~vhich anyone fanuliar with the style of conal rnusic "vill immediately recognize as calling for conrra.punral developrr1enc-prob~ ably a.s a ft1gue or ft1gato (Example 109). Ho,vever, pol}rphonic textures and contrapur1tal procedures are not comn10·11 ii1 compositions of tl1e period. The question is: ·what is ti1e basis for tius recognition of rnelodic potential?

••

,, .............________ ~1(

•

Example 109

First of all, it is 11ot pri1llilf'ily a n1atter of te...xture. The first thre,e notes are accompanied by full, sor1orous chords. But even had. tlie movement begun with a tnonophonic o:r uniso11 textt1re, this \Vo11ld have been ot1ly a contributory clue, Llot a s1tfficient conditio11 for tl1e recognition of gestnral potential. l\1iany inelodies u1 both eightee11th- and 1tlneteenth-cenn1ryr com.. . . 111 . unoo11 . pos1oons ar:e f. or solo 11JStrumer1t or e1ise1n bl· e pIayu1g or octa.ves. Bt1t polyphonic creatn1ent is by no mea11s irnplied. Three exan1ples co1ne to mind; the opening melody of 11ozart s Sor1ata f(Jr Violin and Piano in E J\1inor (K.304); the beginning of the first moven1e11t of Bral1ms Quintet in F ~1inor for Pia.no and Strings; and the n1aia melody, played by a. solo flute, of Deb11ssy s: Prelztde al'Ap·rts-1W.iai d'u1z Fa11:12e. 1

1

1

The gesture suggests pol)rphonic rex:ture because this lcind of patterning-~rith its assertive gap a11d contrasting, runnir1g fill-is the basis for countless ft1gue subjectS and conrr·apuntal con1po.sitions of the Baroque period: for insaince, tl1e subject of the C-lvlajor Fug11e f.ron1 Book II of the


MELODIC STRUCTUitES

Well-Tempered Clavier (Example 11 oA). And in our own time.>Stravinsky employs this kind of struc.rure in the fuSt movement of the Octet for Winds, whe·u he wants to create a Baroqtleh""ke contrapu11tal texrure (Exan1ple I I 0B). But prevalent conventions and persuasive traditions are not as a rule arbitra.r y. Repeated use .indicar.es tl111t the gesn1re is es1Jecially suite,d to indeed, 1

calls for-co.nttapuntal treatment.

: ·· -

' ~.

.

ll ..

~

'

'

B.

.

A -

.

'

' 19· ' ' ·

F

llir

Exan1ple

T he m.elody is

=

,

o

11

~.,

gap-fill structure (Example 109). The parts, the gap and th.e fill, are distinct, so that tl1ey can be t1sed separately as well as together; and e-ach has a distincti,re profile, so that it w ill be recognizable even w hen the texture becomes complex. Moreover, the parts are different not only in ft1nction, bot in rate of foreground motio11:. tl1e gap consists primariiy of q.'uarter-notes, th,e fill of sixteenths. As a result,. gap and fill can be contr1tpunmlly con1bined, yet retain their id.entity (Exam.ple 11 3). The marke,d mocivic contrast has another virtu.e. It allo'\vs the parts to be clearly defi11ed without tl1e n.eed for inter11al closure. The pattern is a single process \Vhose n1omentum is not diminished by c~dencial articulation. The stru.cmre is potentially conr.inu.o us-a desideratum in polyphonic ml.1sic in other ways too. The fill part of the n1e1ody co11sists o·f a series of thirds, C-Ab-FD-Bb, \Vhich ends when a re\rers.al crea,tes--a.lmost forces-temporary closure 011 .E,b (E,xa,mple r09). Follo"ving this, the vioJa begins the fill part of the pa.t tern and extends it s~quenriall}r (Example [ 1 t ). Thotigh the pat, tern is modified, b-Oth by the skip of a sL~th and by a cl1ange in the rate at wl1icf1 thirds succeed one anotl1er; the t1nderlying organization remains a co11tinuous triadic morion (graph 1). The sequential patterning of the second-level stru.cture "''hicb is also triadic (graph 1) , eril1a11ces tl1e sense of·

on-going

fi
Ai1d ,this is true of the lughest level as well: there, 'the


EXPLAINING MUSIC

lIO

melody is lu1eat, implj. .ing C·Ontinuarion ro Eb wl1ich is realized a measure later-first in the ·viola and then in rhe piano (grap'h J) .

Example r 1 J

The gap part of the n1elody also iniriates a linear pattern.:-from Bb to C. \Vhc:t1 the F comes. on the accent at the beginning of the third n1easure, the possibility of sequential continuation, a falli11g fifth follo\-ved by a risi11g si.A'th, is evident (Example 1 r z). Tl1is possibility is actl1alized in the passage

%.·--- ---------Exainple r 1 2

given in E.-xample r 13, where the essential components of the rnclod)r are presented as part of a sequence, a,r1d in imitacio11 at the octave and fifth. 58

•

1 •

£~mple 11 3

ns Sequential imitation of a .som.e'\vl1at different sort occurs earlier in the movement - beginnii1g ac the end of nleasure 71,


MELODIC STRUCTURES

lII

But Schumann. reserves what he cletlr1y considers to be the culmination -the ultimate actualization .of the melody's potential-for the begin.ning of

.

'

,

•

.

, ,. .

._

··-

If -

'

f

•

•

E.\'.ample

114


EXPLAINING MUSIC

1 I 2.

tl)e coda. There the gap pan of the f>atter:n, eA'tended by seqt1ence and by .a rhythmic moci,:re fro·m a transition passage, becomes the st1bj.e ct in a canonic-fugal exposition \Vhich is acco111panied b,y the sixtee11th-noce fill pattern (Ex~UTiple 114). Th.e phrase given in Example 1 15 is 11nmistakably n cadential gesture. The hart11onic pattern, ending I! -V 1-I, is. a traditionally established closing progression. TI1e rise to the tonic in th.e seco11d x11easure is a common pe11l1lcimate motion in closing themes- for .instance, the one \vhich comes at the end of tl1e slow movemet1t of i\4&zart:'s St1,ing Quarret in D Major (Ex.ample 132). The rema~inder of the pru~ase a descending rnocion from fifth to tonic, is equally cl1aracter isric: a sin1ilar pattern bri11gs the slo\\<1 move111ent of ~1ozartts Haifner S}rmphony to its finitl .cadence (Exa1nple 89) . Yet this is ·the opening phrase ·of the slow introduction \.Vt1icl1 begins Hay·d n s Symphony No. 97 in C l\11a;or. A .··

~·

Exax-nple

1 1;

0ea.rly there is a discrepancy bet'\'CCil tli.e C.'UStOr11ary, cadentiai function of this cl1aracteristic gesrt:tte and. the t1.se, riaydn n1ak.es of it here. It should co1ne n.ot at the· beginning of an event, bt1t at the end. And it does: it closes tl1e slow intr·odtiction. But the discr epancy is so stril{il~ tl'1a:t it is not fully resol,red by this prelin1ll-1ary closure. As the eigl1teentl1-centtlf}7 theorist, Johann Ma·ttl1eso11_, \V:rOte : ''One ca11 make use of many ordinary and well-known devices. Cadencest for exampl~ are quite c:ommon . . . and. may be found i11 e · C!)' piece. vVh.en, ho\i\i·ever, tl1ey arc used at the beginning of a piec,e, they become somet hing specia.l, since they normally belong at the end.,. r19 4 That Haydn, toot consid.ers st1cl1 use ' sornethlng special'' is sh-O\vn u1 the fact that this melody· renirns as tl1e closing tllen1e of the sonata-fort11 Hans Lenneberg's translation of portions of Der v o1tkom1J1e11e Capell'11teister ( 1;39) in his article~ Hj ohann 1\1actl1eson on Affect and Rhetoric in Musict" fotLrnaJ of t\.1.t,sic T heory, II, 1 (April 1958), 70. ri& !l'rorn


MELODIC STRUCTURES

structure. Ar. .rl1e end of the expositio11 section it is ttansfonned into an even more obviously c:adential pattern-a lcind of daocelilce figure (E.xample 116). In the recapitulation (measures 240-262) irs importance is emphasized when it becomes the basis for a passage whici1 moves rhrougl1 a series of keys to the first dominant liarmony of tl1e coda.

•

'

.

p

Exanlple 1.1.6 ARCHETYPAL SCH&\•f,l\TA

From rime to time throughou.t this book reference has been made to archetypal patterns and to traditional schen1ata. Though not so labeled, all tl1e n1elodic types analyzed in this cl1apter might be thought of as archetypal patternst and tl1e com.positional gestur·es discussed just above 1nigl1t be regarded as traditional sche.111at~i.. Tl1e subject is so significant that it merits a section o.f its own. To recognize th.is significance, recall an observation. made in the first chapter: ,partiet1lar eventS are invariably un.derstood as rnetnbers of some class. Archetypal 1)atter11S and traditional scl1emata are the classes ''tl1e

rules of the ga.ine,,, in Koestler's phrase-in terms of "vhich particular musical event.IS are perceived and oompr'ehended. No n1elody, howe,rer origulaJ and in·ventive, is an exception. to tl1is principle. Indeed, the concepts of originality and invention-as distinguisl1ed fron1 tl1e .ecce11tric and the biz-arre-presume relarionshiJ) bet\.\reen a particular instance and th_e class or 11orm to which it belongs. Furthermore, as I have argued else\>vhere,00 the delight of intellige11t tnental play a:nd tl1c excit:en1e11t of its complementt affective experience,, are significantly dependent upon the deviation of a p-Articular mt1sical event from the archetype or schema of \~:hich it is an instance. Su.c h norms are abstractions. One cannot find an archetypal gap-fill melody or an ideal cadential scl1ema it1 the literature of t·onal music. Bot eo Em:otion and Mean,ing

flt J\f1isic,

Chapter r. and passin1.


E..~PU.fN!NG

214

MUSIC

it does not follow t.lat tl1ey are remote or detacl1ed from tlcn1al 111usical experience. Juse ~1s tl1e speaker ·Of a l1lrigt1age u11derstat1ds and responds to verbal t1tterances accordi.rig to tlJe t}rpes t,o '~1 hich they belo11g- prose or poetry, e111otive exclamation or reasoned argun1ent, declarati,re a.sserrio·n or inrerrogarive alternaci\reso tl1e compete11t listener lJndersta11ds and respor1ds to t:l1e melody from Strauss' Tilt Euletispiegel (£.,~ample 85) partly~ ii1 terrr1s of its triadic strucn1re, ro tt1e subject of &.1ch s D-f\4inor Fugue (Example 77) i11 te1111s of irs gap-fill a.nd triaclic orga11ization, an.d to the then1e of Scl1u111anr1's Piano Quartet (Examf11e 109) in tem1s of its pot·enri~il for ooncrapu11a! treatmer1t. To· talce th.e a11alogy to la11gu1'1ge srill ft1rther begin11ii-1g tl1e 111overr1ex1t of a sy1npl1ony \},.1 ith a cade11tial gesrure, as fla)rd.n does, is lilre beginning a story wit h the words: ''a11d they lived happily ever ~iftcr. ti I11 theory it is possible ro disringlilsh ben:vee11 archet}1 pal patter11S an<.1 sche111aca. The former v;rould be tf1ose patterns wlucl1 arise as the result of pl1ysiological a11d psycl10.J()gical consrants prest1111ed i1111ate i1J h111nan behavior. The latter '\¥ot1ld be those norn1s \\truch \\1ere tl1e rest1lt of learning. Bt1t t he distir1ctio11 breaks down ir1 practice. For 111ost cradiciotbtll)' est:ablisl1ed t1orms have son1e basis i11 il.1nate co11sta11ts, and, 011 th.e. other h~ind, patterns deri·1.1ed from innate consta11ts becon1e 11ar:ts of tradition. TI1is bei11g r~1e c"lSe, the terms will be usecl n1ore or less i11terchar1geably, and tl1e phrase ''arch.etypal scl1err1a>' "'ti.ll be u.se,d to refer ro the ge11eral class of stylistic norms. The diffict1lty of disringuishlng betwee.n teamed a11ci u.111ate p·a rterne irlgs is illustrated b·y th,e main, :,111d opening, mel0-s gestttte reaches its close vvithout signific-anr delay, in1plying change of ftlnccion but not further 111elociic nlotion; Brahn1st gesture does not. Conseq,uer1tly, both co11tinuatio11 to realization and cl1ange of function are itnpliecl. Pres11n1ably innat,e modes of i1erce1)tio11 arid patte.rnit:lg rend to cl1aracter~ ize the gest11re. Tl1e 1nelody begi11s on tJ1e tl1ird C)f the scale, G. Fron\ a tonal point of \rie"
-

1

1

1

61 Tria.dic co11tit1ttaciQtl to

tl1e higl1 Eb is not ii11plied because the Bb is relatively mblc-the goal of an end-accer1red rh)rtl1n1-rather tl1sn \,V eQ}( and mob:ile. ln this co11nectior11 see the d iscussions of E:xa1n1,les 61 and 63, above.


'

•

.

••

I

'

.

"

. I.

-

'

I

.......

....

-

'

--.... C)

c.. r

'

'

w

.....

~

•

'• ..

'

'

,,

-

.u

I~

'

Ill . .•

f')

x

..... r

~ .. '

':'ii

I

I

.

'

I

•

Material corn
EXPLAINING MUSIC

once rh.e Ab follo'\\IS ~virhli1 the firsr quat[er-11ote of tl1e n1e~tStue, '\:Ve suspeet (in retrospect) t11at the Bb should be u11derstood as an ''unconst1mn1ated appoggiarura-a sum1ise confirrned when the .morive is re,p.eated iI1 rneasures 4 and 18 (Exan1ple i 18). The t11ost regt1lttt and, ''nor 111al' 1 co·11tinuation ~·ould f1ave bte11 conj unc.t motion directly d<:l\;\rn to the tonic, as sho\-vn il1 pan B of E~"tample 11 7. Avoiding this ob\r1otis at1d du1ll possibiljty, 110~,rever, the moti,1e skips dovln to r:he F, creating a complerne1ru1r}' gap which is filled by the follo\\ting G. Fclr botl1 rl1)rtl1rnic and 11arn1or1ic re111ent, tl1e melody rect1ms (E.xarnple 1t8). T he implied Eb is agai11 a'\i'oided. ~f'hen a seqt1enoe 11

tr.ised upon rl1e first tlrree notes of tl1e motive n1oves th.e pattern up a third. "Th.is enlarged n1otion n:mkes the repetiriot1 of the F- G pati ern in measure 2 1 parcici1Iarl;r prominent. Once n1ore, ho\vever, G is su.bstiruted for the prob-

able Eb. Acr1.1.alizatior1 finally t:al

MELODIC S1--aUCfURES

1.1 7

__________

. ~~--.._ --------------------------------------•>-

•

•

+ :: -

~i

-..,_, _

.· :;>•

t -

11

-,

Exrunple

.1 18

However natural or i1mate the im11licarive r elationships generated bJ' the structure of the m.elody, their force and spe.cificity is also, and significantly, tl1e result o-f learning. The implied continuation a:nd the .mocivic ac~ tualizacio11 can b-e avoided again and. again because our repeated encounters \lrltl1 tl1is sort of patremn'lg rnal{es t1s con.fident that it is c"adencial tl1a.t it will sooner or later reach closure xnore or less as expected. Two ex:amplesi chosen because they a.re in the same lre.y and the same tempo, indicate that this is a fan1iliar and traditional cadential gesture (Exan1ple I 19A and B). ,·.

.

~,,,~ ~

,

J~

w"""'

:Example:

1 19

The first (A) is the closing figure froin the slow 111oveinet1t of' Moz.art~s String Qua.re.et in Bb .l\l!ajor (K.458). No explication is needed: it is selfevident that the bracketed figure (n1) belongs to the same family as Brahrns' gesture ind-eed, it lacks only tlte ecl'Jn.pee, G. The second example (B) is 1 taken from tl1e Cavatina of Beethoven s Stri11g Quartet in Bb l\1ajor, optlS 130 -to my mind, one of the 111ost exquisite and 1no,ring m.oven1ents in the reper-


EA-PLAINING MUSIC

tory of Westerr1 inusic rrl1e four n1easures (8~11) give11 at tt1e l)egi,nning of the exainple bring the first melodic pattern t o a. close. J\1east1res 8 arid 9 belong to \ovt1at might be called rlie fainily of cadencial-ecl:7apee gestures (as does Bralm1s' melod), )-except thar tl1e ecl1ap·ee functions as a!1 a11cicipa.rion until tl1e cader1ee at the end of the nrSt part of the m-0vement. Ti1e n1otive (n1') in ineasures 1 o ~ind 1 1 is a member of the san1e fifth-to-tonic cl~r.;s as Mozart s cadential figure ar1d ~1s Bral1rns, melody. Both ki11ds of closing gestures-the fifth-to-tonic and the cade11tial echl1,pee--are joined together in meaSll!es 63 a11d 64' wl1ere the}'' form. the n1elody of the last autl1entic cadence of Beethoven's Cavatina. Archetypal scl1e1nata r1e,ed not, of course, be cadenrial gestures. Tl1e last n1ove111ent of Beethoven's Fourth Syn1phon)rt wl1ich Sir George Gro\re characterized as a < 'pe-rpeti1.ur11 1rzobile>' M begins \-Vith a gesture \~those infectiot1s '' erve and piqu~nt '\\ it suggest cont:ir111arion, contrast, < lnd developn:1enr. ~fo a1J1)reci~1te tl1e complex and s11btle relariorlSl1ips an1011g e·ver1ts "\X.:rithi11 rl1e 111elod.y, 011e must co11sider the in1plications of ti1e initial s.LYree11th-11ote motive, Jn. (E.xa1nple 12ol\.) . The moti\re consisrs o.f the tonic, Bb, ornat11ented b;r a lo\ver neighbor-note ~t11d follo\l. ed by rhe slcip of a third to D. TI1us reduced to essentials, it is a 111ajor tll.ird ( 120B) '1vl1icl1 can be t111derstood either as part of a triadic pattern irnplyir1g co11ti11uarion. to F, tir1d perhaps t l1e tipper to11ic as \·veil ( I 20C); or as the disj rn1ct }Jan of a ga,p-fill pattern implying C, 1

1

1

a.n d p.erllaps seque11rial n1o tio11 co D a11d be)rond ( 120D ) . Such sequer1ci.al

motion is prob:able because, once the C is reache~ a h.igher-le\rel linear pat~ rero arises, and it in nirn in1plies co11ti11uation ro a point of rela.rive stability or closure. f\'1oreover1 to the eA.'te11t tl1at ic is n1ore than ornan1e11tal. tl1e neigl1bor-n.o te pattern from A to Bb seems to imply linea.r concil111aci<1n, for this sort of patt:er11 occurs in variot1s guises tl1foughouc the period of tonal 111usic ( 1z.oE). Filially, tl1e in1plicati\re openness of the motive cornes in 110 small meastue fro111 the D,s .rhyth1nic \veal\:11ess an.cl functiot1al mobility ( i 20B). Had tlte r1eigl1bor-1 1ote figt1re been at1 upbeat and tl1e D an accented goal ( r. 2.oF), triadic cot1tinuation, especially, "''ould. have been co11Siderably

less probable.

8.

c.

ll

s. I

Rlatnple rzo 64

p.

Beethoven. tttid H is N ine Sy·J-1tp/3onies ( Ne"'' Yorl<: Dover Publications1 n.d.),

112.


MELODIC STRUCTURES

Here tl1e in1plicative inferences of a con1petent listener may he based upon the patterning per se. But once again, it is in1possibte in pra:ctice r-0 separate 'C\rhatever innate processe.5 111ay be inv<>J,red, from our familiarity wicl1 the bel1avior of this traditional scl1en1a. For tl1e motive, at1d oth.ers belonging to the san1e family of gestures, is con1n1on. in the 1:epeno.r:·y of tonal music. Let us consider so·n1e exan1ples. First, triadic continttarions. 111 the first n1ovement of Hayd;n's String Quartet in Bb I\tlajor, Opus 76 o. 4 (Example 121A), the mori\.re (m) occurs ''"ithin a triadic pattern and is continued itnn1ed:iately to tl1e i1nplied :fifth, F, and th.en in r11easure 4 to the. octave, Bb. 1-:Iere, as in other Classical period examples, the moti\re is set off and called to the listener's attention. by the grace notes \Vllich en1bellish its begin11ing. The n1ori,re also occurs in the seco11d measure ·o f the fot1rtl1 n1ove1ne11t of Scl1ubert's Octet for Strings and \!\finds (Example ti 1B) . 0~ The continuation is triadic to G-but .i n tbe \vrong register. The re~iliza.tion is, tberefore, only pro,r:isional (Pr.) . The trliJdic it11plicatio11s of tl1e n1otive are Stttisfnctorily realized, 110\vever, \\Then the n1elod)r retur11s in the seco11d part of the rounded-binary form (measure i .r ). Bt1t histead of tl1e prob·able G1 ~~ follows as pan of tl1e do111mw.t of D minor (\ T/ ll) . Schubert ''·ackno"vledges'' the special, deviant character of ·t he A '\\ itl1 a f orte...piarzo n1arking. 1~he poignancy of the A is the result not only of its de\rian.r cl1aracter and its ham1011ic context, but of the fact that tl1e larger inte.r \ral-a sL\"tf.1, r:atl1er rl1a11 a fifth-acts both. as a triadic continnacion and as a gap implying descending n1otion to"vard closure. 1

. -·-.

-

____ ... -- .'1 ...

•

'.A • ·--

· ·- - · .--· - -

Example

-

.

. c •

-

··-- - ·-

·.·- ,;;ii!

---

-

121

I futve purposely chosen a. \VOrk \Vr.itteo after ne.e thoven's Fourtl1 S)'lllphony t'O emphasize that our under&'tG11ding of schen1ata is aten1poral S.chuberc,s use of this traditio11al patren1 it1-finences· oar undersm1'ltential of a fuguel&e gap~fill pattern (Exarnple t 10B ) affects our undcrsrand.ing of sinillar patterns in th"C music of preceeding comp-osers. ln rhis connectio·n. ~ee the dis.ctis~ion in Nl eye.r , ~1:usic, t/Je Ar-ts 1111Ji, ldens, p. 47f. e;s


.EXPLAINING MVSIC

220

Gap-fill, seque11tial concinu~1rions are ~ilso common li1 t l1e literature. 1"'he 1norive occ1.1.rs 11ear tl1e beginning of rh.e first n1ove111e11t of M.ozart's Piano Sonara, h1 C Major, K. 279 (E.xa1.11ple {22) . In this case, instead of moving triadically to tl1e G, tl1e third (G-E) functions as a gap and is follo·\Ved by D, \Vc1ich is the fill (grapt1 1). Sequent ial co11tlt1uacio11 is implied not only by lrigl1er-level linear n1otio11 fr om C to D, but by rl1e foregrotind rltlrd. D to F, whicl1 pairs 'vich tl1e precediI1g :C to E. T'l1is, hov.re·v·er, proves to be pa:rt of a caclential figi1re Iea.dit1g to a re11etitio11 of the first t\\tO measures. The implied line~1r c·ontinuacio11 does n<>t occtrr L1ntiJ twelve 1neast1r.es iarer, where the n1oti\re moves in co11jt1ncr n1ocio11 frorn. C ro G (graph :z.a). T l1is passage not only represe11ts rl1e rea.lizaci.011 of the implied co11junct co11tint1ation, b t1t, when G is re:a,ched, tl1e triadic potential latent in tl1e n1otive (and regenerated b)' its repetition at t~1e t)eginnlllg' of the seq.11ence) is aJso actllftlized (graph 3a). In additio11, as th.e motive 1noves on stro11g beat:s (3 arid 1) froJ11 C to E 1:0 G, a still hjgl1er-le, el triadic motior1 is generat·ed -vvhich carries. both patter11i11gs beyond G to tl1e octave-definin.g stability of the high C, 'Ine same motive, embodied it1 a somewl1at n1ore elaborate figurel is the basis of a 1

,__ __ ,

z. _____,.

______,,

·f Exainple

I al

sequential, gap-:fill pattern in the Gigt1e from Bacl1's Partita No. 5 for Harpsichord i11 G t\ 1ajor (Example r 2 3) . Partly because tl1e whole p.a ttern includes a con1plete triad and partly r>ecause the triple n1eter " 'ealcens the mobility of •

'm

.•

:

£"(.11111ple 1 z 3


MELODIC STRUCTURES

221

the second beat, ther·e is ·evidentl;1· little implicatiot1 of triadic continuatio11. Openir1g melodies generally establisl1 the cl1aracte.r or etl1os of a movement. This is certainly m1e of tl1e last mo,re1ne.c1t of Beetho,l'en's Fourth Sympl1ony. From the first n:1easure, the affective tone is one of pert, yet not hostile) venre and c-Jprice. For the nJost part this is the result .of syntactic relationships-the sudden cl1anges in register a11d dynamics1 and the sharp breaks in implicative process. Bt1t ti.le 11acure arid c11stomary use 0£ the motive itself also· contribute to character. F(1r, as the preceding exam·ples indicate, the moti,re usttally occt1rs "vitlU.11 a melody " rhich is already t1nder way, or is included as part of a. transition process. Here, ho-vv.ever, it is the first gesture. 'Nitl1out benefit of phlite preparatlo11 or customary, introduction" the beginning seems brusq.uely witty a11d genially vvillful (Exan1ple 1 :z.4). Tile ope11ing 1notive (tn) in1plies triadic motion to tl1e fiftl1, and perhaps to the upper octave of the tonic (graph 1) . The fifth,, Ft foll.o\VS,, but not in the implied register, so tl1e realization is 01lly provisional (indicated by pare11theses in cl1e anal)'Sis). Tl1e F is also iinplied by the linear pattern of potendal structural tones (graph. 2) : becal1Se the D in measure 2 and the

•

1

Eb in measure

are tnelodically pro·minent weakbeats, not effecti,rely connected with the f oUo·wing low-level accents, C<>ntinuation to an accent is cAlled for. As rhe analysis indicates, the F 's thus implied are not realized until meas:ure :zo. Bt1t the most important aspect of th.e triadic patterniI1g is harmonic. When the tonic (I) and supertonic ( II) triads follow one another in. root position (as they do u1 this case), th.e syntax of tonal harn1ony makes it h:igl1l)r p!iobable that, if the linear patterning of fust nor.es is continued, the next chord \vill be the ton_ic triad in th·e :first inversion. not the improbable triad on the m.edia11t (DI). 1'"11.e opening n1easures of Scl1ubert,s Sonata for Violin and Piano in D i\1ajor are. a clear illustration of ·r.his typical 1>rogression (part B of Example 1~ 14). 'F or these reaso1lS, the late11r triadic motion of the first two measures of Beethoven's theme itnply tl1e p~tter1ling in measures 1

1

•

xg-i r ·continoatio11 to Bb (graph 3). The second implicative relationship generated by the opening morion - the gap-fill pattern- is pro·visio:i:ially realized ,vhen the motive is repeated on C (graph 5). The realizatio11 is not entirely satisfactory be,ca11se the till should follow rhe gap directly. as in tv1ozart's Piano Sonata (Example r2z, nieasure 1.4f.). The gap and fill are brought together in the proper register in .meastire 19 (graph 6), and he.re, too, the D implied by the second foreground gap, C-Eb, is realized (graph 5a).


2 .2 2

EXPLAINING IVlUSIC

••

•

jp•

__,'t: ,_

•

5

•

2..~". . ' ' .....

.. 5.

<

--

-

....

_.. _

•

'

it= -.

!USU

-

";;:::::::/.

'

.,

.

t.

·_ b p·~· .

....

---

a

·-·

4'~ ~IW

..

.M

~

I_I I

•

•

.

.r

-

.

..

0

~

... . .

5

~

......I

.

• •

-· ....·- -- ·- -

"f

e.

I

w

'"' ·~

....-

.

5

,.• -

5.

'.' IL

5

;ii .

. • iai: )-Ii ·.

.

Lo

1'

•

•"

m

I

· -..

~

.

..... ~

-

r--

J . '% T

- -·-

-

•

-

5

%

•

.

5

.

,.

.•

'

I; ! J

:Ji·

5

. .

5

i

.

--··----- -

.

, ga. t -, • !.

----

-

n ••

5

-

-~ •

-

-

' st

The relatio11ship betwee11 tl1e rn.ori\re (111) and its repeticior1 (n1') creates a higher-le\rel ·Co11ju11ct p atter11 ¥ 11.icl'\ Ul1plies conlint1acion. to D . After two n1east1res, D arri,res. Bt1t :, ·011ce again, register is \~ro11g. A sacisfactory D is presented in measure 19 (grafll1 4) , wl1ere it also ft1r1ctions as th.e beginning of rl1e in1plied firsr-inversion triad (graph 3) . In the preceding discussion, the triadic in1plicarions of the second staterne11t of the mori e vl.rere r1ot f1lil)~ C(JI1sidered. Because tl1e G-, implied b)' tl1e foregroun d patterrw1g of tl1e ·n1otive (C-Eb), comes in rhe lo,:ver octave in measure l,. its r ealization .i, only provisional. T hough it tends to r)ass t11.1r1-0ticed becaltse th.e triad to \\'l1ich it belor1gs is pa.rt of a higher-level process lead,i11g to me~strres 19-2 I (gra tJh 3), i10 eXf)licir realization of the G occurs i11 the exposition sectio11 statement of tl1e rnai11 theme. In. tl1is respect tl1e G is unlike the correspo11tii11g F~ vvhich is satisfactorily realized in meast1re 10. 1~11ese roles are re\rers,ed in tl1e recapitulation, '\vl1ere a s.arisfact:ory G is realized. both explicitly and \\ritl1ot1t delay (see Example 12 5, m. i 86) . ·Ot o:nly is the Gin n1eas\Jre 2 in the ~1rong register, but the followi11g sixteenths inove in tl1e v\rro,n g directio11: tlle)r (}O not parallel the equivalet'lt patterning i11 tl1e first rr1e.asure. I~Iad they do11e s<), a ri5ll1g lin·e '"vo·u1,i l1a,1e


MELODIC STRUCTURES

led to D 011 the first beat of meastrre 3, and tl1e pre\riousl)r established li11ear and triadic p:atternings \\ ould then have been con.tinued. Instead., as if l1eedless of preceding events, the line seenlS to plt111ge downward. into 1

an unrelated and u11explored register. TI1ougl1 rhis bcadlor1g descer1t continu.es in the lo\ver stru1gs (as the reader ';\'ill fin.cl if he consults the score), the listener)s attentio11 is pexempto.rily e11gaged by tl1e entt'ance of a rrew gesture- played forte b)"' the 'vhole orchestra- \Vhose even-paced ernphasis

ai1d rising conju11ct n.1otion sets off and cou11terbalances the do\"vn.ward rusl1 of sixteenth-notes.

Before c.onside1·ing thjs new gesrt1re a11d its implication, ·a ivord al)out its character ai1d kind. T11e impression of n1eiody onl}' three notes in co11j.unct sttccession-is roo n1inin1al ro be characteristic. But the three clipped chor<.is pL1)red by the full orchesn·a (witl1 violins pla}ring double a11d triple stops) and t1'1e V -1-V ·hannony t11a · e a f an1iliar soun·d and so·n ority. \¥ e l1a.v'e heard this cor1figuratio11 counr1ess tin1es i11 this nlusic-son1e~ times as introductory gestures, but more oft:e11 as oodas to s;rmpho11ic .move1n.ents. Its use l1ere, u1 th.e 1niddle of a n1elodiousl)1 proc.essi\"'e .first tl1e1ne, is unust1al arid surprising. In conlparison with tl1e pre,rious breaks between the motive and its continuarions1 the e11mince ·O,f this assertive gesture seems abrupt-an al:n1ost arbitrary int.erpolation <)f a11 alier1 idea, contrasting "'rith the pr·ecediog pattern in almost e·very '\'1ray. This se·nse of anomaly is both conlirme
motion and tl1e CC)IJj u11ct r11otion o.f the gesture would be ob,rious, not only l}eca.use of the c·ontinuity of conjtmct pitches, bl1t because the metric position of rl1e A is the same as that: of the F a11d G (grapli 7), 00 E,~en discounting this connection "''ith earlier patterning, the conjunct .morion of 'Urha:t l shall call the coda figure-to discing·uisl1 it from tl1e motive (n1)-i111plies continu.a tion to D . But precisel;r because of its chordal, cadencial character, return to Bb is a possible alter11ati\re... The D which 6~ Tt1:at

this suggcsti.o·o .is 11.0t so farfetcl1ed as migl1r at :first appear is sho~"ll in the recapitulation. !~here, beca.\1se the G is not dh.--plnced dO\Vil, tbe connection seems clear and a1ore explicit'.


EXP LAINING Ai.USIC

follows in n1easure 5 is not, ho\Ve\rer, a sarisfactory continuation of the coda figure.: nor 011ly irs register, .but its dyn.at11ic level, o.rcl1esrra~rion, a.nd, above all, irs .mocivic content pre'\rent it f r
-•

f/tdd. .j

-·

"'"AfV..U

a . .

'

.

.Jilllf..,.

B

Exan1ple

115

61 ~1easu.res 7- 18 have been omicced fro:rll E.xamf)le 1 i4~ In those meas·ures, first . oontmues . tlle seque:r1na ' 1 patterrt. begun i11 . 111easure 5. 1"trlen, .m nieasu:re 1i, a t he monve l}~rical n1:elody, 'vhich contrasts with both the pert verve of the motive a11d the asseni"V·e force of tl1c coda fig;.1re, is presente{t over a dominant pedal. The repeated F's. ''rhich e11d this rune are picked up .i11 11leasure J.O, \vhere the)1 lead to a .resolorjon 011 the torli.c ..


1\lELODIC STRUCTURES

225

was in tl1e exposition. Instead of leading to a sequence based on ·the motive, as it did in measure 5 (Example 114), the coda figure continually returns to th·e beginning of the melody. Fo.r instance, in the first statement of the melody in the recapirulation (m. 185f.), the 1n:0rive part is played by the hassoo11.; the cod.a figure, performed pizzicato by the violins (m. x8788), leads to a forte repetition of the san1e music an octave higher (m.r89). This change has a number of impor tant consequences. 1) The motion of the coda figure to tl1e tonic (with A and C acting as neighbor-notes), which was inferred in the expositio.n, is here made explicit. 2) Because the passage which \\1ould have repeated measures 5-2 1 is missing, 11either the seque11tial co11tinuation in which the gap of a third ·was ·followed directly by fill nor the F implied by the opening motive . s and• potenna . 1·1t1es . are inst c._ occur \V.t.th.·10 t h· e . t h· eme. Th ese .unp1·1cat1011 1

reaJized at the beginning of the coda, where a chromatic sequ.ence, slowed by intern:al repetition, leads from Bb to F (Ex:an1ple 12 5, part B; measures i81-z94). Subsequently (measures 335-343), a compressed version of this seqt1ence leads to an even more em1,hatic and stable F. In th.e course of both tl1ese sequences, the D implied by the motion ·of the opening motives is rea1~. l

·-· ~ ..l

3) But the D implied by the coda figure is not. For to realize the in1plications generated by that pattern satisfactorily, the D must occut in the proper register (which it does not do in the sequences),. be the resolution of dominant harmony (witl1 root pr·ogression in the ba~), come on an accent, a11d, ideally, follo.w the generating figure directly. These conditions occur twice in the coda. Following t he stateinent: of the theme \vhicl1 'begins in measure 301, the coda figure moves conjunc..--tly from A ro high ·G , as

shown in the sketch. at the end of Example 125A. That this is a version ·of the coda figure is clear not only from its relationship to the opening motives and its own melodic p.attern, but from the harmony and bass motion. Nevertheless, the :realization leaves something to be desired, partly because the clipped ehords associated with the figirre nave been replaced by a s-moother,. more co11tinuous succession of iJitches. Mostly, however, because

the D occurs within a larger n1otion, the f acr of arrival and realization lacks appropr iate emphasis. The D is, however, appropriately conspicuous and emphatic wh.en it occurs as part of the final cad.ence of tl1e movement (Ex'" a:mple I 16). For it is the goal and end of the quarter-note rhythm of the coda :figur,e and of the sL\.'teenth-11ote morion of the bass. And in these cho1:ds,


226

too, the .potential inherent in tl1e cl1aractex of tl1e gestt1re is unequi,vocally

actualize·d.

.._

........

Exa.1nple

116

I c:onclt1de tl1is section witl1 a. final exan11>le, a question, a11d an obs.er•

\raaon.

Tl1e exa11't.Ple, from tl1e beginnil'1g of the tta1lSition passage fro.n1 the Haycln Qttartet move1nent quoted earlier (Exan1ple 12 1A), is:

Rumple

r2 7

Tl'Je qzJ.estio1t: If ·seetho,ren's music can be characterized as original,· \Vhat is the b·asis for this qt1a1ity?

T l'e (J.bservation: Given the pre,ra1ence and central importance of arcl1erypal sc.hemata in tonal music, originality in musical art,. at least until recently, corlSisted not i11 the i11\1e11tion of no·vel means and syntactic relationships, bl1t it1 t he inventive use of establish·ed relatio.nships and shared conventions. PROLONGATIO

~S

Not all 111elodic relationships a.re implicative. Most notlimplicative events ~lI'e vv.f1at are cor11monly called prolongations.. There are in1portant differe11ces atnong kinds of prolongatio~ though tl1e;r are of te11 treated as a single, mo1101ithic class. Some prolongations cor1tah1 clearl)' .defi11ed implicativ.e processes; others do not. Some are implicative rhythmically or har1110nically, but not melodically. Some serve to create balanced m'.o rphological le11gths; others stretch establisl1ed ler1gths and rl1creby heigl1ten the effective-


MELODIC STRUCTURES

ness of in1plicarions already generated. ~1.lost prolongations are related to preceding evenrs by confor,mance, but sorne-fo:r instance, wha:t \Vill be caUed parentheses are not. The follo~ring discussion is an attempt to distinguish a few broa,d subclasses.

Declarative prolongatibns 'T he basic n1orive or thema.tic id.ea ·o f a rnelody or composition may itself be a prolongation. The first four measures of the first prelude in Book I of the Well-Ten1pered Clavier, for instance, are a prolongation of a tonic----C Major-triad (Example 128A) . Be.ca.use th,e melody begins on the tllird, 111otion to tlle tonic is probable. But this implied tonal motion is c1ot rein£orce.d by the pat-terning of the parameters. T .he seco11d an.d third measure.~ imply closure and return, rather ' tllan <..~11tir1uatio11 a.n d mobility. Melodically, the F's in measures 2 and 3 function as the upper neighbor of E. Harmonically·, the :progression is e:adential, 1-II~-v:-I. For this reason, end~accented.

on the phrase-level. In short, these measures are understood to be a sr-able, closed shape a statement of motivic materials, texture, and tonal center.. The implicative ,processes which shape the morion of the Prelude are not generated u11til n1easure 5. rhythm appears to be

I

;

1 ,

:rzt

, ;

\

7 72

\ !!! 3

I

_ -12

,

I

. . . ,

Ji fJ J_IJ

- -

-

Example

_,.

'.

1 28

Tl1e first theme of Beethoven's ·rhird Sympho11}' is esser1cially a stable, anal melody (Example I i8B). Eb is I>rolonge.d, anci the

.EXPLAINING MUSIC

strong: not only is

Bb the relatively stable fifth of the tria.d and the goal ·Of

the patterning in measure 3, bur it is defined as the probable upper li1nit of melodic activity by the Bb in measu_re 4. RhythmJcall.)r, ho\vever, the theme .is open and mobile., as the analysis under the example shoV\rs. Ne\. erth.eless, like the Bacl1 exar11ple~ this tl1em.e is i1nderstood prin1arily as a statement o·f the main ''sc·uff'' of the con1position, rather tl1an as a patter11 generative of specific implications. It is a declarative prolongation. Although they are not specifically implica.rive, declarative prolongations usually nave an aura ·Of latency about them-if only because th.e}' are t)eginning e.vents, and \Ve believe there \.vill be n1ore m11sic. This feeling ·Of anticipatory tension is specially strong when tl1e declarati,re event does not itSelf c
Smecana,s Vltarua (The M old.au; Example 84) o.r the one fron1 lV!nltler's Fourth Symphony (Ex.an1ple 88) are the clearest cases of si1ch declarative prolo11gations. Finally, 1nany, but not al] melociies which are the basis for

theme-and-variations movernents are also examples of co1nplete, stable events.

Norrnalizi11g prolo1tgatio11s Prolongations frequently occur in conju.n.ction with implicative processes. \ ¥he11 this is. the case they rna.y perfor1n a n.u111ber ·of some\vhat different functions. One of these is to m.ake the phrase or period fit with a previously established or stylistically normal t11orphological lengtl1. Often the normalization. O·f lengtl1 is achieved thro·ugh simple repetition. The r11elody of the second movement of Schubert's Trio in Bb 1\llajor, Opus 99t is an antecedent-consequent structure, though the co.nseqr1ent phrase is r~ro measures longer than the autecedet1t (E~'lmple I 29) . The fust phrase begins on tl1e third, G (rrieasure 3), ai1d frorr1 a t o11al point of view, 1notio11 to the tonic is probable. Nlelodic patterning begins with conjunct descending motion which also implies continuation to Eb (gra·p h 2).


MELODIC STRUCTURES

Tlus implication is reinforce.d '\\'hen, after a. gap of a third" the melody again descends conjunctly, fron1 Ab to G i11 ineasure 4, and then con,tinnes to F in 1neasure 5 (graph 2a) . At the sa.me rime, the second level pattern from G to Ab implies) and moves to, Bb, the rela.tively stable fifrh of the scale (grap'h I ) . The foreground patterrring of the melody is partially closed at the en.d of measure 5. Harmonically, the progression from I to V ~ is a semicadence. Rhytlunically, tl1e change from sho·r ter to longer note ,,a.Jues--:from a sixteenth-note to· an eighth, .a.nd from eighth to quarter-note-creates a clearly closed, trochaic groi1ping on lov. er .levels. i\1elodically, the G .and F c.onstitute a return to the opening pitcl1es, and the se·nse of closure is enhanced by the gap from Eb to G, wlucl1 is filled by the F. 1

1· .

~§'

'A

""-- - - - - - ' _ ___, ,' ,_- r r -

.

I

, r••:•£.t,....-'""' ~ f \ !

Exan1ple

2I

,

, .,;

_____.

----~

129

But inore ·n1usic is needed. T \vo..-measure lengths have been established as the norm-in the melody itself. and in the introdu.c tory measures as well. T he varied repetition in measur,e 6 does not change or even noticeably .reinforce implicario·n.68 Rather, rhe sixth n1ea:sure serves to create a n-0rmal, regular phrase le:ngth. This is w instance of the bifurcation of form and pro·Cess disct1sse.d at the end of Chapter IV. Here, as in the Dumka from Dvofak's Piano Quint.et (Example 57), process reaches closure befor·e ·morphological lengr:h is corn1'>lece. The norm.alizing prolongation completes the morphological le11gth. The construction of the second phrase is similar to that of the first. Like measure 6., measure 1o is a normalizing prolortgacioo. But th.e:re are interesting differences. Tl1e implication of descending motion is somewhat stronger 08 ft does:, however, emphasize closure bjr so1newbat intensifying the rhythmic motion towatd G, and by suggesting m.o re con1plete linear p11tterning, Bb-A-G,

down ro F.


:EXPLAINING MUSIC

230

at the be·ginnir1g of the co:nsequet1t phrase because the gap fror11 ·F to

Bb is

larger t11an the cor:responding gap frorn F to Ab i.c1 tl1e first plirase, an
Bb functions like an ap:poggiatura.

the other hand, che low-level patterning at tile enci of the phrase (n1easures 9 an<.i 10) do·es not imply direcr conjunct motion to the tonic,. as the G-F inocion at the e11d of t11e first phrt1se did. Instead, tl1e m·e lody descends to Bb wl1ich defines the area of n1elodic aetivit y and rl1ereby Stabilizes the upper fiftl1 makin.g linear or triadic contin11atio11 a.ho\,'e it less lill1s, converge to the Eb. To balance ri1e le11,gth of tl1e sL~-measure conseqt1ent phrase, Schubert begins the m.ove1nent \.Vith a t\vo-n1ea.si1re introdl1ct:ion. T l1ese n1CaSl_1res might be regarded as a kind of 11or111alizing prolor1gatio11 /;efore the pl1irase. I-lowever, be~iuse they are obviousl)r a grou11d-a11 arct1etypal accompaniment figu.re:, rather than a well-patterned melod)' the rneasures are in1plica~ rive. ~re expect tha t a srrot1gly shaped melody "'rill be presented si1ortly. Though the regi:Uar.ity of the rt1ythm ancl the siznplicity of the harmo·ny, dynarr1ics, and tempo suggest that wl1at f(Jllows will be lyrical, '\\"'C l1a,re 110 way of knowing vvhat the rnelody will be. lnlplicarion is not specific. That this i11ttod·u ction is in1plicative it1dicates that the effect 0£ a 11onnalizi1~ prolongation depends in part upo11 \Vher·e it occurs in a n1elody. 011

When it occurs. at: tl1e end of ~1 pl1ras-e, as in n1eastire 6 and i o of Scht1berr,s melo·dy, a prolongation \iv:ill i1ot as a rule be impl:icati,,.e. But vvhen it occurs at t·he beginning or in ·rhe t11idtile of a phrase, it ' vill tend to be so-cl1011gh. perhaps for rhythn1ic rather than melodic reasons. The melody 'vhich begins the second r110,1er11ent of Nlo1iart,s String Quartet in. D· 1\1ajor (1(. s75) illiistra.tes this point (Exan1ple 1 30) . Like the melod}r of Schubert's Trio, it consists of an antecedent a11d a consequent phrase, each divided into t\VO parts. The implicative relationships \vithin the antecedent phrase merit attention- both bec·aL1se tl1e effect of the prolong'ation (.i.epends in part upot1 then1 and because t11ey are of interest in their o\:vtl r.igl1t. 09

Because it is conspicuous tl1rougl1 it.s abser1cet D is implied as a probable goat. The sltip from Bf> to Eb in measure l t in particular implies D as a fill. And ~·hen the n1clod}' is .repeated, the D (r11easure ll) follows the Eb and mo,res to\.vard a ca.der1ce 011 cl1e d.or11.inant.


.MELODIC STR.UcroRES

2 31

The antecedent phrase (Exan1ple 1. 30A) l)egins \Vith a four-meast1re prolongation of ·rhe third ·of the scale, Cl (graph 1) . For ronal reasons, morion through B to the tonic, A, is probable. This implic11tion .is reinforced by a suri)risin:gly st1bcle, yet basically simple1 melo.dic stn1crnre. In the foregrou11d, nivo descendit~ conjunct patterns, each preceded by a gap, imply conrin11aci-0n to the tonic; the first moves d.o'\vn from E not F#, an appoggiarura (graph 2); the second, f ror11 A (grapl1 3) . 011 the next leve11 the ~'t1ucrural tones are th<>Se ·Of the tonic triad, and the second part of the phrase (measures 5- 8) is incll1ded in this pattern (gra.ph 4). Again,, the lo'\\1er tonic is implied, n,o t onl)r becau.se by m.easure 4 (and more patently by measure s) the triad is con1plete except for tl1e lower to11ic, but because the gap-fill s.trucrures suggeSt descendi11g 111ocion. Consequently the probability .o f triadic contin11ation ro the tipper C# is lo"''· Once again seemin.gly slight differences have in1plicativ.e significance. In the co1ISequent phrase, t l1e first n1easure is varied so tha.t tl1e relationship between Cl and E impljes continuation to tl1e upper octave (Example 130B); in meast1re 9, a weak>mobile eighti1.- note, E, follows Cl directly, so that further triadic n1otion is probable. And,, after the A in measure r 1, the seco11d part of the phrase b-egins on tl1e l1igt1 C#. 70 (See grapl1 above and rhythmic analysis belo\v in Exan1ple r 30 B.) In the second part of th.e antecedent ~'ll1rase (Example r 30), C# is prolonged until measure .s, '\vhere it beco1nes an. appoggiarara in a semica.dence mo,ling to B. The motive n1easr1res 5 and 6 is a li11ear rise to E, hut f t1rth.er n1orion is not .implied. The E is the end of a relatively closed grottp: rhyth1 r11ically, it is preceded, by a sixteenth-note '7vhich acts as a l{micro-le\'el ' upbeat, and, consequently, the hig!1er-le,1el dactyl is closed; melodically the skip fro111 B ro E (in a Il - \r l1armoruc context) makes the E sound lik:e the co1npleri.on. of a group. The interesting confo r111ant relationship ben:veen this motive and the pattern.ing of meastires I and 1 is sl1own in grapl'l 5.

u1

Turning .now to our chief concern-prolongation. Wha.tev·er the 111elodic i1nplications of tl1e patter11i11g tluot1gh measure 5, tl1ey are in no \vay changed by th.e repecicion tl1at takes place in measure 6. From a melodic point of vie"''' mea.~ure 6. cor1tributes 11otl1it1g. Its contribution is fo.rmal .and calls attention to an important point in critical method. A specific cl1ange .cann.o t be adequatel)r explained o.n the grou.ncls that lt provides variety. Cert:tlnlyt ·th.e re is a need for \l'ariety, but ther·e are it1nunler11ble \:vays of achieving it. Wh111t: is required is an explanation which suggests wh)' this particular. way of aetii.eving variety is app.ropria.t'e or \Vhat its format or implicative consequen.ces are. 10 This


EXPLAI~-riNG

232

("" •\ ,~vM , ·-'!' ,P,., , -,~

MtJ"SlC

""-- ,,..,v,,v I

'

,-.. :· ··;a

d

I·

..J

I

Example t JO

rhythmic. From a. formal point: of \tiew, it norrnalizes the morphological length of the seoond part of th.e pf1rase'-S(J that like rhe first it will be four

measures long. Rhythmically, tl1e .repetition is processive ancl i1nplicatlve. Because they ar·e identical and c-ome at tne beginning of th.e second part of the antecedent, measure s and its norn'.lalizing prolongation, measure 6, make it pro'bable that the whole subphrase will be an an_apest, 3 + 3 + 6 beats, on t l1e seco11d level. And this implication is realized in measures 7 ai1d 8. which for1n a pivoted rhythmic group. 1"'he sense of goal-directed n1otion created by the rhytlimic structure depe11ds i'n part upon me context of strong n1elodic implication. At the same ti1ne) rhythmic implication compensates £.or th.e temporary suspension of melodic lnocio11 t O'\ Vatd the tonic.


MELODIC STRUCTURES

233

Extensi()ns

The conseq11e11t pltrase of tv1ozan>s melod}r is given in Example I 31. If the changes in melc)dy and register already considered and those in the closing measl1res ( i 6 :u1d c7) are disregarded, tl1e phrase is regular a11d t)rpical, rea.ching the expected cadence 011 the tonic in measure 16. Just '\Vhen stability and closure seem assured, ho\vever, the 'riol~1 and cello begin a repeticio11 ·o f part B of the phrase. The \'1.olins are, so to speak, obliged to follow along, elaborating a ne"' counte1--poinr. T11e resttlting lack of closure is e1np.hasized b),.;· the fact that the tonic t:riad is in the first inve1·si.on, rather thai1 root position.. This repetition, t:oot is a prolongation. But instead of being 11or1nalized, the morpholot,rical le:tigth is stretched. In ocher '\vords, tneasures 16--19 belong to the class of prolongation l{no,vn as extensions. Not only is the niorphological length stretched, but the e>.."'te11sion begins 'before the co11... sequent phrase h,as finisf1ed-reacl1ed irs normal length. l "'l1ere is an elision: measure 16 is both the end of the main n1elody and the beginning of the extension. The ''n1orphological dissonance', which results from this ov,erlapping creat:es a tension \v·hose .resolutio11 e11li.ances the sense of satisfaction and closure " 'hen an .u,ndisrurbed, cadence is reached at measure i 9. In.deed, the tension is such that some of tl1e accumulated energy spills over into the bridge passage which follo·'-''S. J\1tost extensions at the en.ci of a phrase do not, however, i11volve elisions. They occuI after tl1e morphological length is complete. The coda of tl1e l\4o.zart n1ovement we ha\re be.en analyzing provides a cle.'lr illttstrarion (Example I 32) . A.fteJ· the the1ne presented in tl1e t~wo pre,ceding examples, a moduladon to the don1inant le~1ds to the statetTienr of a. secondary melody. This leads back witn virtually no dela}r to a restaten1ent of the ru1tecedentconseq11e11t melody. 11lougl1 details are varied, the mt1sic is basically tl1e same. But now no bridge passag·e follows to absorb tl1e tension of the elision and to arrest the momentun1 built up by the delays in the melod)r· The coda of the movement accomplishes these ei1.ds. It is, so 'to speak, a c,omposed ri:tardando. As Ex.ample 1 32 sl1ows, the coda begins after a full ccadence ir1 rhe tonic ar m.easure 61 . It is an extension consisting of a four-nre.asure pattern (A), I Il6- l!- V7-I, which is repea.ted (A') , Tllis is followed by a further e>;tension . a two-measure authentic cadence (B), whicl1 is also repeated (B') closi11g the movement. Since the extension. is essentially a melodic prolongation


' I

J I

,

~·

I ... •

--

r -r,

11.)

-a,_c:: '

J

~

c

w

)

t

>I> I

. •

Mater al corr dire tos autora s

MELODI C STRUCl'URES

r•r

_ , ..,_ _

of the tonic:, A, n1oving th.rough three oct~rves, are generated.i1

i10

r-J''

- - -.J

implicative relarionsl1ips

Extensions also occur within phrases. Though they do not influence the ·d irecti0n of the implicative ·process, such u1ternal ~xtensio11s often 11eighte11 the listener's sense of implication by delaying arri\7 al at implied goals~ A strilrin.g instance of ~ln internal exte11Sio11 \vhich intensifies tl1e feeling of goal-directed motion occurs in tl1e firsc: section (measures 1- 17) of the

Prelude ro Tristan und Isolde. The larger strucnire and ha.n nony of these m.easures have been expertly ai1a.lyLed by William ~f· !V!itcl1ell.12 l'ily concer11 will be the ·co1nplex relationship:S witlun the prolongation,. wlrich occur betv;ree11 i11easures 12 a11d 16 (Exan1ple i 33..i\). The passage is forcefully in1plicarive l1arn1011ically and n1elodically. Harmonically, each element of the seqt1ence consists of the ''Tristan,, chord, '\vl1icl1 functio11s as a se:coadary don1i11a11t and resolves to an unambiguous domi11ant..se\renrh chord. As indicated un.der graph 4, the first element ends on the dorrunant seventh of A n1inor (V7 ) ; the second, on th.e do111intu1t of C major (V7/ lII) , the relative major of ~f\. minor; and the third, on the -dominant-of-th.e-domi11ant (V7 / V ). Each is therefor:e internally processive and i1nplics co11tir1uation. TI1e sense of goal-directed mocio11 ge11era,ted by this foreground patterning is enhanced by"' a higher-level triadic structure created by the roots :of tl1e chords of .resolution, E-G1f- B. This structure implies continuation to E (graph 3). 71 The

cl~sure

in different octaves at t11easures 71 and 73 is one of the results of the rnotior1 of tb.e consequent phrase co rhe upper Cl. Other resttlts :n1ay be fotmd in the middle s-ection of the ·moverne11t. 'fl! '•Th.e Tristan Prelu.d.e: Tecl111ique mid Structt_ue," in Willia:t11 J. Alt.itchell and Felix Salzer,. eds. T.IJe M .u.J'ic Fo1U1n, Vol. I (Ne\v York: Columbia University Press, 1967) I

pp.

J6l- l03•


The melodic patterning) \Vhich. col11plen1ents this harmonic sttucn1re1

is no less implicative. 011 tl1e lo\vest, note-to-note level, it consists of a chro1naric scale \-Vl1ich rises fr.o.n1 G~ to F# (111easures 2- 1 r). ·On t he next level eliminating those to11es. \.vhicl1 are ob,1iously appoggiaturas or l'a.'\sing ror1es, tl1e n1elody is a diatonic sc.ale in A minor. As sho\v11 ir1 gra pl1 r, botl1 tl1ese motions imply continuation co the Uf1per G~-ru1d 1 because the processes are quite uniforan, tler1mps beyot1d.73 011 a still J1igher level~ tl1e passag·e ·cor1~ sises of a. series o·f thirds G~-B-D-F#; continuation of this patterning \vould lead to th.e higl1 A. R11ythmic organizatio11 st1pports arid enha11ces the goa1-(lirected processes ge11erated by melody arid harmor1y. 011 the prin1ar)' level ( I ) , each t)lm1se ends on the iveak part of t11e, rl'lyth1nic group. Hence, even though th.e lo\vel)t le\rel is a partiall.Y closed trochee) the basic structure is inobile. As a result, the appoggiatnra pattern ·\v ruc }1 ends eac}1 phrase ren1ains inobile and goal--directed, even '\ivhen iris cletacl1ed an(l begins an nl1pulse, as it does 1

in r11easuxes r4 ar1d 1 5. On tf1e next level ( i a), that of phrases~ the varied repetition of a single pattern weakeris the impressio11 of structure and st1ggests that the ltltimate organizatio11 of t}1e \;Vhole section \vill he son1e sort of a.n apest grottp. As a res.tilt, the seco11d ai1{i third grou ps--a11d tl1e first ir1 retrospect-are perceived as 011-going anli in1plicacive. These strongly? goal-c.iirected processes ''get stt1cl< 1 after measure 11 . I11f:t' tead of contint1arion. there is obstinate r.epecitio11~ N ot o nl) r is the re.alizatio11 of tl1e processes ge11erated by tl1e patte,r1ling of t he first ele,ren measl1.res delayed, but morphological le11gths are stretched ~11Jd r11etric contint1it}r is disturbed.

'!'he analysis of these n1easw~es is some\vl1at proble111atic: are they all extensior1, or p~1rtly extension and partly r1orn1alizing p.rolongation? The)r might be considered a complex extension_, for had the melody conti11t1ed to the cadence in i11easure 17 'vithout aJ1)r repetition "'lh.a:ci;oeve.r. che result would ila,

MELODIC STRUCTURES

Nevertheless, for rl1yrthrnic reasons, a11d .harm.o nic ones as well, these measures also seem to involv.e norn1aiization. The organization st1ggested

as basic in part B of the exan1ple is far f rotn satisfactor)' ; it is too regttlar and reacl1es the cadence too quic'kl)r. Not only· does tl1e 1nelodic rnotion, \Vhich heretofore n1oved ":vith inte·nse deliberatjon, no-vv speed ,~ ith ,un\va.rranted ease, bt1t the arcict1lation of tf1e crucial cadencial harmonies is ct1rsory and. cru,i.1al. Above all, the equal, four-1r1easur·e phrase lengths and sequential . uniformity create no l1igher-level structure. F'.o r all these reasons, ·closure

at the end of such a t'1\relve-measure pattern \voulcl have bee11 ab.r t1pt and wealt. Though these measures a.re best an.al )l"Zed as part e>..'tension a.nd pa,n normalizing prolong'ation" it is riot ~lSY to deternline '\:Vllich. is '\Vhich. The analysis give11 in tl1e exarnple is based on a nt1mber of considerario'ns. Be-cause they function as upbeats to the closing ca.der1ce, measures 14 and 1 s create a srrongly end--accented anapest groupu-ig on. the pr.imary level ( i). Th.is patternuig supports tl1e meloclic closure ar rneasu.re 17 and empl1asizes tl1e bacrmonic articulation of the de.ceptive cadence. In addition, they re-establisl'\ rh.e four-measure phrase lengths \vith 'vhich the passage began. For tl1ese reasons.• n1easure 1 5 seems to be a norn1alizing prolongacio1t. Tl1ough measures 14 and 1 s are si111ilar in structure and function to measures 5 and 6 of l\4ozan)s String Quartet (Example 1 30A), they are much more pate11tly implicative, for four reaso.ns. Fir~'1:1 iD: Wagner's Preli1de, rl1e half-step motive, E#-F#, is part oft and by reitera·rion reinfor·ces, tl1e implications previously gene1~aced by the n1elodic patterning. ln the 1\11ozatt, 011 the other hand, the comparabJe measi1res are not derived from,, and do not support, the basic descendi11g melodic motion. Se-oond, harmonically these n1eaSt1res are in1plicative because of their previous association witl1 goal-dire'Cted processe:s-the donlln.a.nt-s·eventh of the do1ninant (V1jV) in ineast1re 11. In i\11.ozarr's n1ovement, l1owever, the :comparable n1easures are part of stable, tonic hanno11y. Third, when first presented in measures 11 and 13, this motive \\ras a '\\1eak, mobile part of the rh;rth111ic grou.p. Because it retains the function tl'.1.at it ~'as originalljr understood to have (ei,ren in the absence of the accent in relation to \.vluch it "''as a ~vealc ele1nent) 1 the repeated motive is rhythmically in1plicari'\re. In the M,ozart, measure 5 is ini\,ially understood as accentc;d and stable, and is rl1yrhmically implicative only in retrospect. Finally these rl1}rrhmic and m.elodic implica·tions are af:firn1ed and underlined 'by tl1e" crescendo ( \Vithin each n1easure) \Vhich ''points to', (.acts as a sign of) an organizing accent. 0

'


I

I

if' ,.. ti

. I

II

'

,

'

I'

'

11....

1

I'

~

I

I I

• ,

I I

l

I I

;

I

I '

'

' I

I

' •

I

I

,

I

. ... ', '

...

I I

'' I1

•

..

, ..

!

(

f.

I

'

JI,

•

1 ..

•

"14 '' '

•

' '~ .. · -~~'tl ,I ) 1~. ~

•

~

)

,

. .J

~

I

' ... '

•

~

Mater al corr dire tos autora s

MELODIC STRUCTURES

l\1easures I 2 and, I 3, then, are an exten~'ion. rfhis is evident if the third eleme11t in the sequence is i"normalized'' as in part C of Exa:mple r 33. N<:Y\V each phrase begins on the last eighth-note of the measure, and the final four 111easures ' 1·fit:,, perfectly into the regular pattern of fottr-times-four n1easures~ And this is precisely \\.~hat is wro11g " rith t l1e structtrre: oot only :is it tedious, but no higher-level r.l1ytl1mic structure arises. It is so unj_form that clost1re 'vould be '°'' eak and an:ticlim.atic. A longer, con1plemer1tary patterning is r1eeded, and such parrerning has already been potential in the basically equal morphological le1igths of the first t\Vo elei11enrs of the sequence. The elision welds the· third eletnent of the sequence ro the last four measures so that, in a broad sense, the structure of the \,\' hole section is a bM-form, with an extension, both on the n'liddle levcl (2b) and on the.highest le,,el (2a). 1

Parentl3eses Internal extensions, like the one in W agne.r s Prelude, take place witllin, and are continuations of, pr·ocesses gene1·ated b)r the precedi11g n1elodicrhythmic patterning. By stretching n1or,phological lengths and delaying n10~ tion to implied goals, tl1ey dist11rb continuit}r, bt1t tl1ey do not intern1pt it . Co1iseqne11cly, tl1ey are ·ur1derstood as intensifications of the existing patterning. .Bue there ru-e ir1ternal prolongatio11s wl1ich,. while not affecting e.~tablished implicatio11s, interrupt the n1usical structur~e., usually after arrival at some poi11t ·o f pro,lisi.011al stabilit}~- Because tl1ey do not really ''belong'' to tl1e pr·e,ceding and following fY.ttternings, s11cl1 internal interruptions l1ave bee11 called parentheres. 11, T his kind of discontinuity is discusse
is perhaps even more surprising thru1 are the ·e.ffects of links a11d. overlap11inb7S· For two tJtings l1apper1. First, suspense res\1lts, and suspense inte11sifies \.Vhatever co11cinlliry there mt:iy b,e. It llla.y be several seconds Earlier theorist.s such as A1attheson (sec foo-tr1ore 59, p. 211), "'·110 took langua.g,e as a modelt osed th.is term. But I a.111 n,ot sure wl1cre the modern U..'ie c-omes fronl. 1 '1'1

have., however, f r,equently discussed tl1ese 111atters with, and leart1ed much from,

Professor L\lwxence Bernstein.


EXPLAINING MU:SlC

after the lightriing fiasl1 tliat the sound of tl1u11der r·eaches our ears1 but

kno\.v the thunder will follo\,r. Second, there. is a certain unreality abo11t the interroptior1. Tr is nor part of tl1e •'real" piece, '\Vhich \¥ill resu111e as rhougl1 1\od1jrtg J1a.d. happer1ed whe11ever it .is allowed to. In or1e se11se, of coiirse"' the cac_lenza is part of t he piece because \v·e expect it to t>e r:here;. in another very real sense, .hO\\'~ ' eve.r , it is not. The.re are so.n1e'\\1hat ar1alogous cases i.t1 both art and literature. A painting is supposed to have a frnme; tli.e frame is r1ot part of the pli11ti11g-bur it is. A story 'vitl1 a fr-ai11e- for exan1ple, a play~ \vitllli1 a :play or a 110\.rel \Virh a. flashback- includes and does tlot include the frame.. .E xtraneous coolie i11teri't1des are ar1d are 11ot .p art of a serious play. A ballet it1 an ·opera does a11d. does not belong to that opera. And so on.76 \:\!hen

\Ve

see the lightning

\Ve

.

-

·r11e melody wllicf1 begi11S tl1e last 111oven1ent of I-Iaydn's String Quartet it1 Eb Nlajor, Opt1s 50 No. J, co11rains an t111equivocal example of a parentt1esis. T t1e theme is an anteceder1t-consequer1t structure. The antecedent pl1ra.se and. rhe last two n1east1res of r:he consequent are given u1 Example 134.

Ttle patte11ling of the high.est level is linear, moving up· from Eb t<> G a11d rher1 returning to F in the antecedent phrase and to the tonic, Eb, in the consequent (graph c). Because of the patterning of the lo\\1 er levels, the G \vhich continues the processes generated in the first fotir meastrres is clearly not tl1e one in rneasure 5. b·ut: the one in mean1re 9. The ·m iddle-level pa.ttem (gra:p h 1) consists of a series of thirds, EtT-G, F-Ab, w·hich should conci11ue ~Tith G-Bb. But Bb does r1ot follow in r11easure 6, as it should Instead, tl1e direction of 1nocion is reversed- Inoves do'vn to the D. The fail11re of the. i1nplied patterning tc) concint1e is ernpha.&ized by the repetition of tl1e parenthesis figttre. 1ne rr1elod}' continues ir1 n1east1res 9 and 1ot where rl1e tension btiilt up by the d.elay carries tl1e n1otion to the high E b for ru1 instar1t. 111 The man1 are~ of melodic activity is, l1owever~ d.ef:i11e.d by the octave frorn the opening u.pbeat Ilb to tl1e Bb ir1 measure 9. The low·est le el (grapl1 J ) is also li11ear. And again the pattertling is brol{en ii1 measures 5-8. The concinuacio11 in m.easure 9 is 11ot in dot1bt. And. the momenn1m. of the .li11ear n1otioti c-arries t l1e lllelod)' to t·l1e si:ttl1 degree of 16

P" I 49·

ra Altt1ough thee high Eb is har111onized by a subdotnina.nt triad, it is also part of tl1e oonic melodic triad ( graph 2 ) . For the triadic 1notio11 latent in each of the first tvto tllirds suggests tonic and su.pertoni.c patterns i11 roo.t position, and, acco.rding to the probabilities of tonal synta.x, die next men1ber of the series ~vould be t he :first inversior1 oonic triad. (See the discussion of Ex3'mple 114A, graph 31 a11d 114B.)


MELODIC S'fRUcrurutS

the scale (C) which1 mo\ring to the second degree in the antecedent phrase and to ti1e tonic (harmo1uzed b:y a ! chord) in t'he conseq\1ent, creates a tradition.al ·closing gesmre-mucl1 like cl1at of the c~dential extension in Example 132. From a rhythmic point of vie\v, the goal of che first four measures is clearly the last four, 11ot the mid.die ones. j\ileasures r- 4 a.re made u,p o·f two similar rhytl1111ic J>atterns. A four-measure unit in relation to \Vhich these two events can be grouped is implied. But n1easur'es 5-8 are even more :patently divided into two identical ev·enrs \vruch forrn 110 higher-level rh)rthmic sn·11cture. The implications of th.e first four measures are ~is tl1e m1alysis under the cxam1>le sl1ows, realized in rhe last four~ 1

-

oda.c ti<~~

r·-·---·---------_

-··-

s P,11fo !

\._"'-;y:..-....Ji....~~-

.

?---.-_-'_1). -·=-----=----...:-=.....·=-=--_-_-_-_-_-_-_.........

_ _.,.... ""'"

- - - - - - - - - -....... -- ~ ...... ~

a-i

Exan1ple 1 34

For all these reasons, \Ve recognize at once that n1easures 5-8 are not part of tl1e ''real'' melody. And. such ree-0gnirjon is facilitated by the char~ acter of the pattern and by its Jack of rnorivic (conform.ant) relationship ro the opening 1)attern. For the opening 1nelody is emphatically conjunct, but tl1e motive of the parenthesis is pri1narily disjunct-almost like a thumping bassr mar1ring cin1e. And. tl1e real 111elody is cl1aracterized by goaldirected motion; but the parenthesis is static. It .is as thotigh a person purposefully striding to\va.rd some objecti\'e should suddenly pause, perforn1 a dancelike caper, a11d then continue to his objective~ T he need for the pai·enrl1esis is nor hard to u11derstand. The 111otion ·o f the first four measures is so apparent and palpable that had the goal bee·n re-iched without delay-ha

CHAPTER VIII ---..__.,,,,_.-w·-......,...--·.........~:-·, _...-.-"'

A Summary Example By

of re\rie\·ving sorne of the theoretical ctS ar1d analytic methods cJeveloped in these 'Essays and Explorati<)llS, let us co11si<.ier tl1e first t\\'enty-one measures of the first movement of Beethove·n's Piano Sonata in Eb iVlajor, Opus 8 ra·- ·'
1•

A competent li.stet1er perceives at1d responds to music "\1rith his toml being. ,.l\ s tonal Stimttli, filtered ancess a11d structure of music~1l eve11ts. Throt1gl1 such empathetic idenci.fication, music is quite literally f ett, and' it can be felt \.Vithout tl1e rnediation of extramusical concepts or linages. Sucf1 kinestl1etic sen.si11g of the ethos or ch.aracter of a .n1t1sical event is \Vl1at the term ethetic refers to. Bec~t:tse llutnan experience is not cotUE)arttnentalized into n1usical an.ti n-0nmtJSical, aestJ1etic ac1d. r1onaesthetic, the etl1os of a 1nt1sical evenr will ofte11 seen1 sinUl~ir to an.d suggest so1ne aspect of the extra,mtisical \vorld .. The inusical event is felt to b.e sad or joyful, restrained or e,"{uberanrt cal111 or agita.ted, and rh.e like~ And SL1cl1 cl1~rracteristic stt1tes of lJei11g t11ay in turn be associated '\Vith more specific circun1Stances and ideas: a summer evening's calm, tl1e gaiety and bustle of a social gathering. lvlo:reover, wl1e11 it explicitly


A SUMMARY EXAMPLE

imitates e~-tramusieal sounds- -.1s in birdcalls, wind and thun.der, and the like or is established as part .of tl1e tradition of '~' estern tnusical ico11ograph,y, 1 a musical pattern may denote quite specific kinds of eve11ts, actions, and ideas in the extramusical world. Although they can perhaps be differentiated iii theory, in p·r actice ethetic relationships are i11separable fron1 inlplicative and l1ierarcluc ones. T he ethos of a n1usical event, based in part t1pon the 1.nore constant parameters of music sucl1 as ren1po, d}1!lan1ics, register, mode; and the like, influences and (1ualifies the listener's sense of' ho\\l tl1e event 't\7ill probably proceed syntactically and formall}'· Conversely, the syntactic processes and for1nal structt.l!'e of an event- \V.hecl1e.r regula:r or sporadic, balanced or asy1runetrical, i)redictable or capricious-play a crucial role in defining its ethos. And just as our preliminary opinion ·of an individual is revised and modified in tbe light of l1is subsequent behavior, so our impr~ion and underStanding of the character of a musical event is often modified by its use and ·variatio11 later in tt1e ·\vorlc. Tl1e begu11ung ·of the Sonar-a ''Les J\die\L"(,, is a ·case in point. Tl1e first event-the ' Lebewohl'' motto \ne (Example 1 36A). That ·w ords cannot adeq·uately express the sin1ple and unpretentious, yet co·uc.hlng, sense of ·"vistf·ul regret at1d resignation does r1ot gainsay the importance of the ethos of the rnotto. Its p·a.rcicular ethos is the result of a .con1binacion of facto.rs. Register and. so11ority, tempo and dyt1amics are obviotisly crt1cial. H·ad the same pitch-time relationships been presented in a higl1 register, at a fast tempo, a11d \Vitl1 forte dynamics, cl1aracter \\-'Oul.d ha.,re been very different. Tl1e action of these param:eters is complemented b)' rhe patent and regular n1elodic, .rhytht11ic, and har·monic structt1re of tl1e n1otto-despitc the deceptive cadence closing the event. This cadence further defines t:he ethos of the motto, bringi11g ''tl1e eternal r1ote of sadness in>1- ar1d perhaps suggesti1:1g that the parring is nor final. Feeling-tone is also tl1e result of the de\riant use of an archet ypal schema, horn .fifths: a conventio·nal patterning-\rirtually a formula-used in rhe eigl1tecnth century by iurural brass itlStrurnents (without valves) to p:lay authentic ca.deuces. Their occurrence just before tl1e end of the fourth move1nent of l\1ozart's Symphony No. 39 in Eb Major is typical. As Example 1 35 sho,ws, they follo\v a harmonic progression, II!-V', which 1 See pp. 6.;.f. 1


EX.P L1\ININ G MUSIC

·s trongly implies morion to che tonic. The horn filths are the resolution of this progressio11, a11d rhey prolong and emphasize t he cadential character of the pas.sage througl1 an alterna.tion of tonic ai1d don1inar1t-sevenrh chords.

Because th·ey a.re 1nelodically and hannonicaily c-adencial, horn-fifth patterns normally occt.1r i11 closir1g se·c tions or codas of fast movements and, as .in Nlo~artis sympho11y, they are frequently played forte. As a rule. ho\vever, tl1ey are not'. tl1e r11ai.t1 melodic substance, bt1t support a11d reinforce the cadenrial use of motives talS. 111 so cioing they act as signs of iinpending closrtre~ The use of horn fiftl1s in the first measures of Beetho\ren's <'Les Adieux'' Sonata is u11usual in al1nost every "''ay. Instead of coming at the end of a fast mo,ren1en4 they are the l)eginning ·of a slow i11troducrion; i11stead o.f being accompanime11tal, rl1ey are the 111ain subsca11,ce; an.d instead of reaching e1npl1aric closure 011 cl1e t:onic they end in a tisl1i·ps, st1ggests something of the probable co\1rse of st1bsequent C\Ientc; in r}1e 111ove1nen.t: it: estahl.isl1es at1 a.tn.bience inim:icaJ to tO\.verir1g de11relopments, l1eroic contrasts, and capricious Stlrprises. The aLnost apl10.risric specificity .a nd paJpal)ility of the Inotto (E.~mple 136A), taken together V\1 ith t.he fa(.'t tl1at it is deft:et.'ted fron1 its to11ic goal, makes its st1bsequent 1


A SU Ml\iCARY EXMtP LE

ubiquity both appropriate ar1d cor1,ri11cing. s it returns in 11e\v co11textS and different guises, the \ arious facets of its character arc revealed. It function in \ 1aricd form as t he forceful, dri,·ing first theme of the Alleg ro (136B), as the begiru1ing of the second key area ( 136C), as the basis for the .fiowirig themes of the closing grot1p ( t 36D ), 2 as the sot1rce for the ambiguous, almost hesita11t linear motio11 of the dcvelopn1ent seccio11 ( 136E), ar1d as t he in.a.in 111aterial of the coda1 wl1cre it occurs in irs p11re arcl1et)rpal for1n, first i11 .i:ts closing theme ( 136F) a11d t hcr1 in ca.r1.onic imitation ( x36G). 1

Etl1eric relationships are 11nqt1estio11ably i1nporta11t-p-articularly iI1

•

4

i

.

.

I

.J

'

•

l

.a

-

•

-

•

I'

-

I

F.

f

• .

--

•

I

,~

-'

'

'4

-

.~ '

1:11!1

'

J. J

~

;

'

T

·---.

~

"

-

'~

'

t

- '""'\

.

.

\., ...

i '

.

I

r

~

I

' '

. ...

...,...,..

I

-

J

~~ • ,

'

~·

.

Example 136

2 For the i.11ke of con1parison,. these ar·e given as they occur in tl1e recapirolarion1 where they are in the tonic.

Mater al cor'1 dire tos autora s

.&."'CPLAINING MUSIC

some styles and in some compositio11s. And the ''Les Adieux'' Sonata. is, I thlr1k, one of these. Howevert though they are notoriot1sl}' easy to discuss in casual, plausible fashion, ethos and affect .are hard to anal}rze with rigor a11d precisioxi. Partly, tlus is because la11guage ca1111ot adeqttatel}r ciistingtush ben veen and delineate subtle shades of character and nuances of feelingtone. f\ . more ba motto. Tl-1e vario~1s and va.ried returns of rhe motto 1narlr important poinrs of structural arciculacion. Ar the sai11e tin1e tl1ey create a col1erence which rends ro lessen the sense of explicit contrast in w hat is. a basically dran1aric form. But the main reasons for t:he pervasiveness. of rl1e n1otto are s;rntaccical and in1pLicative. ·"

2 .•

T'he ho r11 fifths which ope11 t.his mo\rement are as clear an example of a Strongl)7 establish·ed scherr1a as one can hope to fir1d. The r>atcerl'l is specified melociicall}r an·d i11tervalli<..~ly, harmo11ica.ll}" and, \Vith some,vh,lt tnore latitt1de, .rl1)rthrnically. We have no dot1l)t as to 11ow it should sound. Conse·quet1tly, \\re are acutely avv-are tha,e the lo\v C t1sed to ilarm·onize tl1e rni11or sLxth (G-Eb) at the beginnitlg of measure 1 is an aberra.t1t in1position, a11d ' ve presu1ne that the corre.ct, archetypal version of tl1e •formula \tVi11 occur la.ter i11 tl1e n1ovement. !-Iowever, tl1ot1gh it appears in a rnyriad of ,rariar1ts, the schen1a is i1ot preset1ted i11 its pt1re, horn-·fifths form until the middle of the coda. Approp,riacely rlus archetypal versior1 occurs im.mediatel)r follov1;r~ ing what can be consid. er·ed tl1e ''solt1t:iot1,' of the C-n1inor problet11, tl1e deceptive cader1ce, '~h.ich. gave rise to tl1e implicati\re r elario11sllip il1 the firsc place. 3 Io music, and parti CtlL1.rly in instrt1mentl'll conlpo itio11s, t l1e con11ecrion between successive ethetic staees is 1.,robabl)' largely con\rel1tior1al. Tltis does not tl'1ean, h-0'\'ever, that cha.ra.cterization and feelit1g-tone are less J)ers-unsive and .c aptivating. But it does ·inean that impomnt work 1nust be done h1 the area of St )rle a11al) sis lJe:fore sucl1 .relationships cati he ~cif.factoril)r e.Ypln:U1ed.


A SUMMARY EXAMPLE

kno·\\~

Bec·ause the scl1em_a is so \Vell

and specific in its pattern, ti1e

effect of the alien, C-mino.r harmon,y is particularly powerful. This is no mere deceptive cadence; it strikes us as expressly anon1alous.·1 For thi.s reason, " re se:nse, thoug.h perhaps only intuitively, tl1at it is significant. And so it is. The implications of the inrrusive C-minor harmony reverberate throughot1t the movement. Let us begin by considering some of th.ese.

I) C-minor har1n-0ny has an importfillt etfect: i1pon t'he n1elodic tendency of the Eb which it harmonizes. For the linear morion generated by the descent f ron1 G no'"' has less tendency to stop on tl1e Eb. \1Vhat the listener ''kno,vs'' ought ro have been a sta.ble tonic is experienced as a. mobile thirdc} possibly implyi11g motio11 '"'ithin sl1bn1ediant harmo·ny to C, or. implying motion co the leading-rone D. The first of these possibilities is r·ealized in the closing th.em.e version ·of the pattern, where the ''Lebewo.hl'' motto is followed by a skip to t.he note (C) a th.irtl .below (Exa1nple 137 ). 1

__...______,....__,~~~~~~ -

--,:----

... .

'

.

••

.. .

'

The tendency of the Eb ro move to D is the result of harmonic as well as melodic relationships. The syntax of tonal harniony makes it probable that a I- V- VI progression such as this will m_oi.re to tlie leading tone, harmonized by a dominant chord, and then baclc to t:he tonic. Such a progres~ sion is shown in Example 138. Here a.n intel"esting psy•chological paradox aris_es-one wl1ich l c a11 describe, but not really explain. While we JtnfY'UJ that th.e deceptive cadence is abetranr and in tbi:s sense l1nexpected, we ~rould be at teas.t -as surprised- e·.,.-ren taken aback-if the fu:st b-eat of m.e asure z were an nnembellished tonic triad. 111 other words, tho·u gh we a.re conscious of how the sche-i11a should go, we -are also aware that a tl obvious and predictable aat:l1emic cadence ic; improbable at the very l>eginning of a composition. In the same 'vay t we botl1 expect and donit e.~pect that the cadential partero w.hich begins th.e slcr~" movement of Brahms' Violin Sonata in G Major (Example r1 7) \Vill move to the ronic at d1e beginning of the second measure. 4


•

EXPLAINJNG MUSlC

•

!

TI1e in1plie.d r11otio11 fron1 Eb to D ~lnd back to Eb is realized provisionally "'' l1e11 tl1e prolonge,d Eb (see belo\-v) in oves to D and back to Eb in 111eaSlire r .2 (~'t:11n1ple 139A), and. n1ore forceft1lly in rhe tl1ird measure of the Allegro the111e (Ext1111ple I 39:B) \Vlucl1 is a varied versio,n of ·che earlier pattef'rling.

-

"""'·

.Exan1ple 139

Bur the in:11)lied cadentia1 progression does not occur in the p roper register ar1d \Vitl1 an u11.equivoctlly end-accented rl1yythn) t1ntll the pe11u.ltimate cadence of the coda (Exan11)le 140). For implications are spe·cific, 11ot only "vith respect to schema, bu.t also to regjsrer, l1ar n1011}''" a11d rl1ythm. 1

~

I

.

c

!

'

-

, . }fr.. .

,,,

I'

•


A SUMMilY EXAMPLE

1) The proximate conse·quet1ce of the imposition of submediate (Cminor) harmony at the ;end of the horn-filths formula is the prolongation which follows. TI1e i>rolongatic)11 is important both because in this case it is itself inlplic~rive and be·cause it pro,res to be a basis f.or the Allegro theme wt1ic'h begins th.e exposition section of the sonata. The prolongation begi11s witl1 the slrip from Eb to Ab. When ·this gap is followed by G, conjt1nct continuation to the ·conic is im.plied (Example 141, graph 1). The fill does not follo\v directly, ho"vever~ but comes in measures 7 and 8. TI1at this a,na1ysis is not fantas)'~ is sl1o~vn by the hartnoniution of tl1e G in measure 7 (grapl1 za) and by the fact tlar it, coo, is preceded by an Ab upbeat: Beethovei1 specifically connects the G at the

beginning .o f measure 3 to th.e one at the beginnit1g of measure 7 :by using essentially the same l1armony in both plac:es.~ Thougi1 the G in measure 7 does n1ove to Eb as implied, the cade.nce is again deceptive- and in the \Vrong m.o de as ·\vell. Consequentl)' the realization is only provisional.6 Instead of motion to Eb follo\1i~11g tl1e G in me-asure 3, the gap :figure (x) is repeated (x') beginning on G. Tliis repetition. generates furtl1er implications. Like the first statement of the figure, the second also implies descending fill. A11d this implicacio.n is realized \Vhen rhe C moves to B in measure 4, and through Bf;r and Ab to th.e G in measure 7 (graph 2) .7 Once the C in meaSttre 3 is reached, the melod.i c 111orion is clearly triadic on. a somewhat higher level (graph 3) . The high Eb implied by this patterning is realized aln1ost i1nmedia;tely. The Eb clos-es out the triadic mocio.n b-oth because of the satisfaction of occave completion and because, as the analysis: under Example 141 sho,vs, it is the accented goal of both lo\ver and higher rhythmic groupings. Be·cause it involves disjunct motio11, tl1e tria.dic pa.tterning is also implicari.ve. The first gaps, wl1ich are filled as st1own ll.1 graphs 1 and z:, have Because this ftatmony is tlllttsnal, the confo.n11ant relationship see111s to be intentional. fi

The fourth, Eb t-0 Ab, migh.c be thot_ight t-0 imply tri3dic continuation to ·rl1e C in rneasure 3. However, thoU;gh t11e C dGes follo,\•. the implication is \.veak. Because the; C-minor harmony is strong and. the Ab is onl)r a sixtteenth-note, the .Ab tends to be understood as ·au ornamental tone. lending co the G and irnplyin.g co11junct fill 6

1

rather d:ian triadic continuatiot1. 1 That originality in art does not ell.tail the disco"'el)' of novel, let alone unique, synmctical means or even archetypes is in.d.i.c ated by tl1e fact th~t a sequential gap-iill pattern very similar to the one Beethoven e.mploys here occurs in the first measures of Schuben's song, ' 'Das Wan.d.e rn'' (Exainple 79).


._..Moot

-

,,

..

·~

-

. ' '

"

t

•

'"'..., ,

1•

~

I

'

•

r.;,.

Il

'

•

,.• -

...

..

'

+ ... .. ..

'

~ -

J •

lo •

-

I

•

~

'

-

I •

I<

._

•

-.

•

'

'

'

Ll

LI Jll 111

(ft ~ .

•

•

t i-

. .. .Ii '

'

-

r

-8 u

•

a.. ~

~

•

'

• '

.

.

-

•r .

-·

~·

.• I

ill

ii

1 1'"

Material corn d1rc1tos autora1s

2-5 I

already been discussed. The third one, from C to Eb, implies descending morion to D (graph s ) and is panicularly important i1ot onl)r because Eb is a melodic and rl1ytlunic goal, but because it reir1fore.es, tl1ougl1 an octave highe.r , the irr1plicacions generated by the linear motion and deceptive cadence of the .first n.vo measures (graph 6) . These melodic in1plicacions are strongly support:ed by tl1e ht1r111011y. Because tlie Eb comes on a six-.four chord '\vhose appoggia:tura tones 11ormally resol\ e by step, the most probable note to follo·\v is D. The implied D could have folloi.ved direcd.y, as S:ho-vvn in Example 142, or could l1ave been realized in some 0th.e r v\ray. One of these Beetl1ove11 employs \\ hen the triadic motion is repeated a. tllird lligher in measures 8- 1i (Exa111ple 141 ). He n1akes the continuation seqt1ential, so the gap fro·m Eb to Gb is filled by the Fb in the ne>.i.: n1easure, and tl1en 1noves t:l'lrol.1gh Eb 1

1

to D (grapl1 4). 8 Botl1 the earlier tri~tdic motion (graphs 3 and 5) and the la.t er o,ne (graphs 3a and 4) con\rerge 011 the Eb and move together ro the D in measure r 2.

Descending motion does in fact follow tl1e Eb in m·easure 4 Bl1t ir is not by step. The D is conspicuo·us by· its absence. The gap pme.rnmg is emphasized agai11 by tl1e skip of a sixth at the begiru1ing of tl1e ecl1~11d again the Dis avoided~ The implied motion fro.m Eb to Dis not realized until the close of this section of the introdtlctiotl-in the mO\"Cinent from measure 1 i to 1 z (Example 141, graph 5). Not only are rl1e registral relationshipsthe ocn1v·e Eb-tl1e same in measure l t as il1 meas·ure 4, but the i11tertrallic relacionslups are sin1ilar: in both plac es Eb is a sixth a,b ove tl\e bass, first as part of a C-minor six-four chord and then as part of an Eb-minor triad in the first inversion. (graph 5a). The octave Eb in measure 11 is the result of s The earlier. less important gap fr-0r11 Bb to Eb is filled b}' the secondary voice sh.o~m in gnpl14,a.


I ~

~

'

( tU

,., '

'

•

~

~

I

@

•

""

~

I

I

~

•

' '

"""

·•I

I

I,

,..,.

~ .. ~

'

I

•

-

t.._,

·r~

'

,_,

'

1194

.... ~Wi" ... (frji. '

I

'' '• '

1,

I

I

_.....

·•

... 1·

~ ,... ...

l

J

~ ~t..

·'

-

~

~

·-

-

~

• ,._..•

~

~

'

'

~ ...

•

ut\ •

~

..

•

----

I

'

'

'

:~ '

'

~.

~

• 1

,f j '

1

"l ~·

'

[')

•

,!

...

· ~~

~

~

~

'

~I

.D

•

1r

'

'

(

-

l l'9

11

I•

-..1....

-ri~ 'ill.-

•

'

•

~

'

I

-

'

'

'

~

1 ~

·1

•

@

~I -

'

'

•~ •

t6

~

'

..... ,

\~

I.

'

I

'

•

I

'

. . •.. I• •

~

~,_ '

~

I ~ '

I

' .-i ~

I

'-

~

,.

'

-

'

rfi

II

I

•

•

' ' ~

'•

J ~ ~

'

'

Material con1 d1reitos autora1s

A SUMMARY EXAMPLE

2 53

the division of th·e melodic line \.Vhicl1 is created by the triadic motion in measures 2-4. The lo\ver Eb wluch closes the motto (and its varied repetition. in n1easures ·7 and 8) is not reall}" displaced into cl1e upper octave, bur persisrs_, moving co the D in n1easure 11 (graph 6) . Like the patterrlS '\vhich generate tl1em, iri1plicatiotlS are hierarciuc. For example, tl1e in1plications generated by the ga:p-fiJl patterns x and x'~ though not immedia,tely realized, are essentially low-le\1el-in the foreground (graphs 1 and i). TI1eir main pitches give rise to a higher level triadic structure (graph 3). 011 a still tugher str.u.ctural level, tl1e prolonged Eb and Gb, "\
parallel. fashion.. As it occurs at th·e beginning of the Allegrt>, the progression is equivocal. Usually it is understood as a progression f ron1 V1 to V in tl1e i11inor mode. Here it would be in C minor ( 1~) . Or less frequently. t'h.e first chord n1ight function as a Neapolit~n sixth. In tl1e latter c:aset the bass w·ould move t1p a whole step so that the f.o llowing chord (G major) "vould be in the sixfour position ( 144C). As it occurs it1 measure 17, however, the Ab-major chord has been so firmly established b)r preceding evencs as the subdominant of Eb major that a ''re~ding>' in C minor seems out of the question particularly since such an ai1alysis is not supPorted by \vl1at follows. ll

AC&,••

11"'

-

=-

- 5 -

·~

·1

rr

-· Exa.1nple 144 "


EXPLAINING l\if.USIC

This arnbigi1ity is Stlpported by t.h.e f acr that n1elodic and rhythnrlc processes are not co11gruent. T11e melodic c-0nn.ectio11 between tl1e t\VO cho.rds is stron.g -Il1e stepwise parallel motion is uru1listal{able and is reinforced t)y its subsequent conri:t1uation .in measure 1 8. In the foreground,. there seems to t>e f'l'arent linear motion from Ab (n1. 17) to D (m. 19), and bacll to Ab (m. 20). Bt1r the rhychmic con.t1ec.Lior1 bet\tlee11 tl1.e two chords is te11ous. For the fourth .beat of measure 17 is unquestionablyr tl1e beginning of ar1 anapest r.hytl1m \\'l1ich reaches rer11poraf}r closure in me41sure .19, as the analysis under Example I 44A s'h o\VS. I""iowe,rer, thot1gh the fourt:l"1 beat of meas11re 17 is 111arked. bec<1use it js the beginning of a rh ythn1ic group, its n1etric position is \Veale. Co.nseq\1e11cly, fro1n a melodic point of vie'v~ the bass (B) seems like a passit1g-to11e bet\veen ,C and the Bi, ar the beginning of m.easure 18. T he proximate co11Sequence of this set of relationships is, as we shall see~ that tl1e A b c·hord is p·r i111arily related not to the phrase '\ivhich im:mediately" follo·ws, bur to the dorninru.1t:-to11ic progression at th.e end of the tl1eme, in rne.asLtres z.o and i r. The more remote consequences are to be foi1nd at the beginning of the development sect.ion and it1 the coda. After a chord on the domiruu1t of Eb major, tl1e development section conrint1es witl1 the Ab-major to G-r11a)or relationship (Exarnple 145A) . Bt1t •

• +

Example 145


A SUMMARY EXAMP'L E

now the progression moves to;. and is understood as being in, C minor-which appropriately enough is the main key of the development. But here too, th.e rnocion is quite uoifor·m and continuous .rather than cade11tial. '\Vbat seems to ·be called for is a cade11ce in\rolving the Ab to G, chord. progression which can b.e un.equi1roca1Iy interpreted in the tonic~ Eb major; and ideally one which n1oves through a 1~:-V chord progression. And tl11s is preci9ely what occu.t'S at the crucial cadence of the coda (Example 1.4.; B). There, the G-major triacI is rreate~ as a passi11g cl1ord-a possibility iI1 the Allegro theme-\.vhich moves to 1: in Eb n1ajor. TI1e great imp<>rtance of the cadence is sho'\vn in the fact that follo,ving it, the ' 'Lebe-ivohl'' motto is presented f 01· the first tin1e i11 ics arc·h etypal, l1orn-fifth_s form ( meas11res r97-199). With this cadence ·che ''problem'' generated b:y the deceptive cadence in measure 2 is literall}r resolved~ 4) Tl1e develop111e11t section b:egins and ends i11 C mirior and in tllis respect is related to the deceptive close of the horn-fifths pattern in measure 2 (Example :141 ). But the seco11d deceptive cade11co ·V 1 to Vl in Eb-fronl

measlu:e 7 to 8 also has consequer1ces in th.e development. After a number o·f measures \'7'11ich are meJodically, rhythmically, and harmonically ambigt1ous, a h:armony which we realize (in retr,ospect) ~vas an augmented..Jsixch ·chordIV~l in Eb minor- leads to a clear do111inant-sevenrl1 cl1ord. in meas,ure 87

(E.xample 146).

--

--

.

~·

I

Here, for a brief moment, the relarionslup betw'een the deceptive cadence in minor and the l1alf-step progression of t'h.e Allegro seems to be n1ade

explicit. The motion of tl1e dominant of Eb minor to, Cb major in measure 90

is the same as that of the dec,epti\re cadence at measure 8 (Exarnple 147, graph 1 ) and the re,rerse motion,..f:rom the Cb-major triad to the Bb-major one is the same as th.e half-step progression of tl1e Allegro th.e,m.e (graph 2.). 1


EXPLAIN'ING l\.II'USIC

.

:r-Jl E:.-
3. T here is a c:o:ntemplacive, itnprovisatory quality abouc the slow intro(iuctio11. T l1e "quasi fan.tasian feeling is the .result of a. lacl<. of strongly processive relationships between successive foreground .events. (Note agai11 tl1e intimate . co11nection benveen ethetic and implicative relationsl1ips-and hiera.rchic ones as
1

, - "-'l '

-7

~

on-going (Example 148.A, level r) . T he group is mobile and implicative because of the powerful effect of the decepti,re cade11ce, ru.1d because at the beginning of the 1no\ren1e11t ·we i1.atuntll;r look forward to n1or.e mt1sic. ~;{any of the conseqt1ences of the deceptive cad.ence have already been

co11sidered. One more iwill concern us here. The deflection from a clearly implied tonic establishes a :particularly powerful harmonic goal - one regenerated in measures 7 and 8 by a second decept ive cadence: an unequivocal authentic cadence i 11 the tonic, Eb major. The ''need'' for such a cadence ac·ts Jike a magnetic pole, gi\ring direction to the Ada.gio even as it appears


A SUMMARY EXAMPLE

to be ctiriously inconclt1sive and hesitant. Ti1e tonic goal creates tl1e ambience within ' vhich the ensuing e,1.ents are underStood; the first twenty n1easures have an .authentic cadence in Eb as their primary and principal goal.9 Becat1se the patterning of tl1e motto is q11ite closed. and because no harmo.tuc, melodic. or rf-1ytl1nlic process connects it to the prolongation"' ·tl1e rela.t ionship between them is essentiall}' additive. The prolongation follows the 1notto bt1r is not implied b)T and does nor follow f1·om the motto. For these reasons, tl1e motto tends t·o be un.derstood as a discrete entity, connected not so much wicl1 tl1e measures '\\rrucl1 con1e directly after it as with its o~rn varied repetition in measures 7 and 8.

The first two 1neas·ures of th.e prolongation create a clearly end-ac·cented, ar1apest rhytl1mic group on the }}rimar)r (I) level (Exan1ple z48A). Both because the B is implied by the seco11d gap-fill patterning (gr·~iph 3} and because the dominant l1acnno11y is the goal of the preceding harmonic progression (C mino·r: V~/I\T-IV8-IV3#-I.1-V ) closure is stron.g·. But this point of relative stability and arrival is imn1ediately \veakened by tl1e ecl10 repetition of tl1e a.ccented part of the pattern; th.e masculine ending is made feminine (level r, measures 4-5).

•

J. ..

...

J

I l

pa :

g.

' llq _

.' I

.t+ 2

~

§

;

:~ ~

[C f

,,_ -"' •

• .

iioj

t

-

''... St

II

4_

,, , --·-''l ..,1 ,,' ". •

0 .

)

+

;1

0

Example 148 further, one migl1t r easonably suggest that the whol.e mo,rement has as its goal an authe:nric cadence \Titb. the Eb above middle C ·io the sop.rano, rea,ched on a clearly a.rticularod end-accented rh}'thn1 . An

EXPLAINING MUSIC

The echo ''opens u.p '' tl1e rhytl1mic strtJt ..'tt.tre of the prolongation, but

mobilit}r is a.clueved at the ,expense of i1nplicati\1e con1.'.leccion. Nlelodically, th·e echo separates tl1e end of ri1e tnain p{lrt: o.f tl1e prolongatio11 (measure 4) from its implied conri11nacio11 il1 111easu.res 6 and 7 (grapl1s 2 and 3) . Rhythntlc~tll y

and harn1onicaJly, tl1e echo is onl)r \\'eakly co11necced to wl1at foll.o\VS. R e,giste.r helps to relate rl1e t V\' O events~ as docs r:he i--rescetido vvhich Beetl1ovet1 calls for. · evertheless, tl1e patterr1 beginning on rl1e last cighth-rlote of i11easure 5 is not a goal of the 1)recct strot1gly coru1ecte(l \.vitl1, t'he h.igh Bb. The Cl'ld o.f i11eastll'e ,6 rela.tes back to the earlier gap~fill patterns and to tl'le :first stateme11t of the n1otto, connecting it "virl1 its repetition in 1neasure 7. Although the .n1elodic n1otion in 1r1eaSt1r.e 6 is coward the follo\vin.g G, tl1e r l1yth.111ic group is nor strongly anacrt1stic. For the crescendo to a11d the sforzando on the fi1-st beat O·f 111east1re 6 tend to tie the f.ollo\ving \\reak beats to tl1e accent\ rna1cing th.e group li1to a fused trocl1ee (level 1). It1 other Virords, melodic orga·nization and rhythmic sn~llCture are l10t congi'l.1e11t at this poirit. If this a1L'lly~is is correct, it vvould seem that, fron1 a n1elodi c poi11t of view» the l1igii. Bb is left l\~tn,ging .in n1id air- witllout conr1ectio11 \Aritl'l wl't;11: directl}r follows. Its con11ection is t1ot \\rith tl'te rnocion tl1rough A to G in the lower octave~ thot1gh tllis acts as a 1cincl of l;rovisional realiiatio.n. anc.i a clt1e to irs probable motion. Irs patterni11g is register-specific; it is cor1nected witl1 tf1e B in measttre ao (grapl14) 1 \\"l1ere it inoves to rl1e G, which

is even 111ore strongly implied b,y the b in measure 17. In the .first six n1easures, ,processive bonds appear, .for the most part, attentiated.. The nonin1,plicative con·n ectio11 l)et\veen tt1e n1-0tto ar1d tl~e prolonga.tion i~rluch co111es after it, the additive rel_atio1JShip berwee11 the cadence on G a:n.d its echo tl1e weak progressio,11 fro111 the ecl10 to the harn"1oriies ,:vhicl1 foll,O\'-' it, and the a111big11ity of n1otion back ro the nlottorhese are r:el~1ted to, and complemented by, wi1at migltt 'be c..-a.lled a countercu11J1ilative 11.ierarchic strt1ctttre. 111 a. cun1ulaci\re stru.crtire col1esive, srror1gl}' bonded, hierarchic stn1ctures are created wl1en t1nits of equal length ( e.g. 1 4 4, as in an anrececient-conseqt1ent phrase) are: co111b.ined, or \vl1er1 sh< ) rter groups precede a longer one ( c. g., 1 + l 2, as in a bar-forrn) . These 111eaSlires exl1ibit tl1e opposite l
1

+

+


A SU~L'\1ARY EXAMPLE

The second sL'< measures are similar to the first in basic phrase strtlcture. Yet n~r the end, th.ese ineasures are n1ucl1 more markedly 11rocessive tha11 their earlier counterparts, bec-ause of the strOl1g se11Se of goal-directed motion created by tl1e sequence iri ·1ne~-ures 9-1 l. rfl1e passage is s11btly cotnplex a,nd \VOrtl1 anal}rzing in sotne dett:til (Exar11ple 149). Harmonically, rhe main sequential motion, \Vhlch en1ploys a version of fourtl1-species ·c ount·erpoit1t, begi11s wich ti1e upbe~tt to t11easttre IO (gra:pl1 t ), but this progressi-011 ca.n be traced. back t:o measure 9, wl1ere the ap})lied domi11ant (V/ \ 71-'V I) relationship is fust presei1ted. .lelodicall.y, tl'le sequence contains tllfee in terrelated patterni.tlgS. The first 1 defined b)r the llal'monic and rhythmic srri.1cture, co11sists of three ''aried statements of a for~ground ·111o·tive (grapl1 2) . But this patter11 is bile,1el-macle up of the two linear strand.s sl10\.'\rn in grapl1 3 and 3a. At tl1e s.a1ne time, the upbeat skip of a third suggests the possibiiit)" of a gap-fill p-atterning (graph 4). Each of tliese patternings mov·es tl1ror1gh Eb to the do,~rnbeat of measure r 2, making the D an elnphatic point of arrival. Thottgh the seq11ence does 0011tir1t1e, in .another sense it ends wit11 t'he third eigh:}Il-note of measure r i, where a re\7 etsal takes place. Harmonically, the first inversion Eb-minor triad breaks tl1e previously establisl1ed patterneven though the b·a $ contin.u.es as before. l\'1elotiically, tl1ere is an i11teresring paradox: l:>ecause the intervallic relatiorlSlups are continued (as the Cb moves t·o Eb, at the end of 1neasure 11 ), "''e are confident that tl1e sequence is 0\7 er. For instead. of coming ()U an acce11t as it \.vould have done 11"t1d the pattern been rhythmically regular, the Eb con1es 0 11 cl1e last eigh.th-note of the measr.tre. 1\ s a result, the Cb is t1ndersroo(i in retrospect as an elision: it acts both as the end of the previous r11otive and as the beginning of a new ot1e. Because it ''shou1d1 ' l1ave bc·en an accented i1ote, tl1e Eb re.ceives special psycl10* logical en1ph.asis. Beetho\re11 ''aclcno\vled.gesn this change of placement and of function. by stipulati11g tl1at tl1e E . be played lot1der than the preceding a11d following 11o·res. Thus sr.ressed, the Eb is a particttlarl}r strong upbeat. Both for this reaso·n and be.cause it is t1'1e goal of n1uch. of the previous melodic patreini11g (also see Example 141, graphs 5 and 6), the D is a particularly he-J\'}' do'\>vnb.eat. T hough it is an en1phatic do\v'11. beat, the D in measure 1 .z is very 1nobile, for a numt>er of reaso11s,. l{hy rl1micaily, the motion from measure r r ~o I 2 is end-accented on the lo'\.vest level-as the nlorion frorn measure 6 1to 7 '\Vas nc>t. But 011 the next level ( 2), tl1e ia.ml)ic gr<>UJ>,. taken as a \;\rl1olet is ''realc: a high-level afterbeat. T o a considerable extent this is bec~1use the ac.c·ent is i1arn1onized by an unstable second-inversion dominant-seventh


160

,E XPLi\INING lV!USIC

chord, Finally, tl1ot1gl1 tl1e meloclic-rl1y·rl111uc organization of the ttpper ''oices is subtly, but sig11ificantly changed, the bass lir1e moves as before; and, since irs motion is quite unifortn (graph 5), co11tinuat.ion is in1plied- and

occ·u.rs.

I

'

1/fl·1

le

5 . .

J/J·1,

.•

!!l

..

'!l ,

__ -_

..

r·

------·· ""'l:T_.------IP.· - , ;~.

~] ~ --.,::-t-,_-_

\ .t . "' ' _ '----------------.-...---~'!!'!!' ~.-'

L.

'

l

I

I

2

M

"'''

u

~ •

w :

JL

I

:

,.

.••

,I

,.

t U' ~A~ ~ ...

-- _,t·:s.l£J , 1, l. ~ 0 "">' ~.. "l e

r

•

-J\V

\·~

.~/ _

\l'J\ \( 1£ ill - .! "

_:. - 'ii. TI 5 ;;._

•

It, __:.'-'r"-"----1 -

-~

,._.:!0;?,:.....:""::x-1~ . ·-~·:::•:w•-=·=§,,----z- \-.=:,'!; ,__,,..J,I

I '-

J

5

- .·. . : !

'~--~--------.--~~-~~~~~~·~·----/

-

Ex.a111ple 149

If the motio·t1 at tl1e e11d of chis seccion is 1nore 1)atently' processive tl1a11 that of the preceding one (n1easures r- 6)., ·the niotion at the beginning is, if a11ythlt1g, less so. Because of tl1e Bb upbeat in the bass in measure 7 the full domin.ant-se,ienth harmony, anc} the co·n trary motion in the ot1ter voices, 1


clomue at the end of rhis variant of the morto is considerably more f orcefuJ than it was in the first version of tl1e lLebe~rohl" motive. ~1o:roove.rt the separa.cion between tl1e n1-0tto and its prolo11gation is emphasized b:y parallel

skips i11 all \. 0ice.s.. Despite rhe dor11inant barn1ony \.vhich precedes it; t:lte second sra.t ernent of the n1otto has the aura o.f a fresh begittning-partly be7

cause the striking dimin.ished-se\1enrl1 cl1ord \\1 hicl1 har1nonizes the first note d<:>es not follo"\.v fron1 tr1e do111inant cl1ord, but is co11t1ccted with cl1e b-a.rn1ony at the beginning of measure 3. .As a res11lt 1 there is a feeling that the ''Lebewohl't n1otto re:rurns not becaltse it is i1nplied by the prolongation that ·precedes it, but because tl1e previous state.me11t of the motto was deflected fro1n its goal. It '"vas only a ,provisional realization. T t1e repeticioc1 in 111easures 7 and 8 is, so to speak, a second ''try'' at reaching a caden_ce in Eb; and

it too is aborciv e. Eve.n t1'1ougl1 tl1e prolo11gation u1 measures 3-6 intervenes, the second staten1ent of the rnotto is understood as being rel.ated to, even sublimlnall)' grouped '\vith, the first. It is this relatioi1ship of variecl repetition at the sa.me pitch-level i;:vhich gi~~es rise to \Vhar l earlier r:e ferred to as potential hierarch)l'· For, to tl1e extent tl1a:t tl1e tvvo stateinents of tl1e motto gro11p toget;h er- are felt to constitu.t e a si11gle ever1t-they imply the possibility ·Of a ba,r-form orga11izarion: :a +1 This ty~e ·Of structure suggesrs itse.lf because, as "''e 11ave seen tin1e and time again in the course of this study, like patterns tend to in1ply a more e>.."tensive event to which both can be related. 1

++

The possibility of such grouping is indicated by the dotted li11es of level i11 ti1e rhythn1ic :u1alysis of R'Yalnples 148 and i.49.

1

The rnro prolongations are also related to one another, so that tl1eir combined high-level motion, Eb to G0, implies the Bb so prominent in measures ;z, 1 and 12 (Example 14.3 ) . It als-0 follows from this analysis that the motto's main motion, doubled at the <)Ct:a\re by motion through the triad (Example i41 ), is to the passage beginning with the D in rneasure 12. And this hypothesis is suppo:rted by tl1e constructiot1 of t:l1e Allegro th.eme whicl\ combin~ the rtvo events. .i\ileasures 12- t 6 are quite straightforward fi,nd, in conrrast to ~rhat precedes: tl1e1n, n1arllfestly g oal-directed (Exainple 150). J\4elodically, th,e mid-

dle-le\rel motion is a diminished fifth, from D to Ab, '\Vhich implies the G r·eali.zed in measure 1 7 (graph 1). The foregroun.d scale-line implies co:i:1cinuation to Bb, because conjunct motion begun on a °"'eal{beat rends to conti11ue o.x1 to the next do\vnbea.t. Thti-s tl1e F,s in mefiliure 13 and 14 .are con-


EXPLAINING .M.USIC

rinuarions of, and appropriate goals for, the linear patterns ·wlucl1 leatio11 (gra1:>h 3) leads to tl1e implied Bb (graph .2 ). In additio.11 to tl1e convergence of these co11jt1r1c.'t pa.tternings, rl1e B.b ''left 11anging'• in n1easti.re 6 also '\:V~ls analyzed as impl}ting the Bb in n1easure 1(). rfhLlS tftree implicati,re pattern.U1gs C(>nverge 011 the penultimate no't e of th.e Allegro then1e. \

"" -- - - - - - - - ----- - -. .~ -_ -. .. _ _. . . . . . ._. . --

...,1

UdvC·f"~

. . . _ , ) , .- _ _ ___ ii_ _ _

, ~ -

. :-#-

',

'

'

-··

'

,,

-

•

-

~

. ..

~ . -

2 .

"'

2

. J

I

-__ •

,

'

..

~\L~~-=-==-----~--.~~~*~ J----~--·----~z*"·/ ,..._=· ------

Example

150

Rl1yth111ically measures r 2- 1 6 fu11ccion as a11 ar1acrusis to th·e Allegro~ botl1 on the ltigh~""t le,rel and, ii1 i:l1e last two n1eastires, on lo\\rer levels as '\iVell. Here tl1e patt:er11 of rep:e tition l~lte11t u1 the ''Lebe~~oltlsi n1orto ati.d its reSt.'ltcment seem to be made 111anlfest. A n vo-beat n1otive (x) which is al1nost a11 inversio11 of tl1e n1otto is repeated (x1') -implying that a lo11ger, fottr- be-at event will follo\V. Instead, however, a \1 aried \rersion of the moti\re (y) is state.d' a third higher and tl1en repeated (y'). T he two pairs combine to forrr1 a lrigl1er-level repetition, ( 1 + 1) :r + 1), vv:hi.cl1 te11ds to, itrrply a fourmeasure group. ''Tends," beca1ise 'vl1e11 t'he Ab's ir1 111eastire 16 con1e 011 weak beats, the pattern i ' prever1te;d from reachi11g sa·risfactory closure. Tl.1e '~'eak beats becor11e upbeats wJtich lead to t he Ab that begitis th.e Allegro. In short, like the repeticio11 of the motto, tl1ougl1 rnore ·explicitly, rhis passage is a pote11tial bar-fon11, and cf.1e co.rnpelling force of cl1e Allegro tl1e1ne arises in part out of the fact rll~lt it is the realization of the 11eed for a collesi" e fot1rn1easure 1nelodic-rl1)rtl1111ic-harmorric event.

+· (


A SUMMARY EXAi"'1PLE

4·

The driving energy of the ~.\llegro theme is also a result of rl1e compressior1 and unifi.catior1 of elements that \Vere pre iou ly only Loosely co1mected,. 11 Tl1c ' Lebe\vohl inotto tlnd the last sectio11 of tl1e Adagio are ~velded togct11cr to for1n a single evc11t-r:1.1e Altegro th.c n1e. As Example I 51 shows, the.re is a clear cc>nforma11t _relationship l>et\veer1 tl1e motto, together '\vith its exte11sio.n. to m.easure 12, an.d the d.escendir1g motio11 o.f t'he Allegro them.e. TJ1e sirnilarit·y bct'\vecn the end of the slo\iv ir1trodt1ccjon (111easures 12-16) a11d tl1e last part of the theme is no less suiJ
-

.. .'' . ~ ...

•

•

-

1

•

•

l •

'

©

~

]

•

.....___...

-

~

-

-

j

•

» i·~

M

J

·-

--

I

~

1 -

'

-

-

-

/(

-

'

-

'

.._, ~~ ·-• .

It

~ -

k

-

i

-JI

--

,

f·

•

'

'

•

11

[

-·''

.

-

-

@

@

.- .=i

I

-

!

~ ~

@

-

!.

-

® ~i

--

·t.·

~ ~ ..- ~

••

~~

:;i '

~

l

~ - ,..•·

~.,~ -

®

l.....

~

-

' ... •

'-: I

.......

•

•

I

@

~

•

......

f a\

·•'

·' •

-· •• '

I

-

J.

I

Example

.•

J•

~

15 1

The patter11ing of cl1e ba.ss-lines is also sin1ilar. The relationship between the bass-ljne at the end of the slo\v introductior1 and the end of the Allegro then1c is ob,"iot1s (Example 15 1) . The co11formant relationship between the bass-li11e at the beginrilng of the Ad'agio and that at t:J1e beginni11g of the A.l.legro is son1ew ~1at less so, but, as Example 1 s2 s'ho\.vs, ir exists. It is not prir11aril)' that both ·pattcr11s are cltrornatic descending li n.cs, l)ut that in e'Jcl1 case tl1c m,ain pc>inc of strucn1ral a.rciculi'l'tio11 ( n1east1res 1i a11d 19) occurs on a do111ina11t-sevci1tl1 cl1ord it1 second invcrsio.r1, \vicl1 D in the soprano. The co1npacti11g synrl1esis is .not r11erel;r '.vithin the bass and soprano separately, but between tl1en1. In the Adagio the descendir1g ua. ~-li11e accompanied the

Mater al corvi dire tos autora s

EXPLAINING

Example

~tUSIC

152

}Jrolo11gation, 11ot the ·motto. In the Allegro theme, the bass-line derived f rotn the prolongation pattern is1 as it were, fused witl1, almost imposed upo~ tl1e melodic line of the mott·O. Tl1e i11tensity of t he opening of the Allegro is tl1ereby it1creased co11siclernbly. Rl'lythnucally tl1e Allegro theme is a single e\rent on the l1ighest sr.ruct:ural le\rel {Example r 53, le'\rel 3) . The dotced half-r1ore, Ab, is both the goal of the rhjrtlmtlc events '\vhicl1 inm1ediately precede it an.d. t11e accented beginnirig of the new patter11ing. It is se1)atated out, however, as an acce11t on rl1e highest level: that is. it is relat e.cl not to tl1e iambic group, J'J f J ,

'·--

\ -\ii I 3

_t

~ _,

z: a

which directly follows it, or e'ren to t he varied ' rersio.n of the motto, but ro the \vh.ole "' real\: group that unites it ro tl1e accented G at the e1Jd of the theme (me.asur.e 1 r ) .

--

-,

,, __,,I .%. \ - I -

'

'

.a

•

' - ,,_,_ ...., \.Yr

='' '~ \ ....I

\ ;_

v

'il

-

V -~ tz

1

4\ -

z

:::

.

,

V I

~

t.!'*' *'~I"-.., S te c 1l*i V J .

-

! .?

''

} 5

, •

I

Exat11ple 153


A SUMMARY EXA~iPLE

In th.e second rhythmic level (2) .. the next event is a cohesive, bar-form version of the ('Lebewohl· motto. But the clear artici1lation and closure pore11d-al at the end of the group does not rake place. The extra Stress of the sforzando on the downbeat Eb compels it to· function as ·the beginni11g of a rh.ythmic grot.1p as weU as the end of one. Beethovert 1nakes. tl1is relationship clear by putting a pf1rase mark from the Eb to tl1e Bb. The rising scale is anacruscic ro the G in measure 11 , and its fusion, througl1 the pivotal Eb and D, witl1 the earlier bar-forrn pattern mikes the whole middle part of the then1e functions as a weak group connecti1tg two higi1-level accents (level 3) . 1

This analysis of the rhythmjc struct11re o! the Allegro theme helps to 111al{e the larger ·melodic and harmonic patterning clear. It it1dica.tes that the over-all melodic motion is from the Ab in measure 17 to tl1e G in m.easi1re 21, and tl1at the basic t1arn1onic progressio11 is an aut:hentic cadence in the tonic (Example 154).

l .

Example 154 Since~,

as we have seen, the Allegro ·t heme is a synthesis of the main patter·nings of the Adagio, it is riot unre~sonable to consider the vvhole passage-the first twenty-one .meastu-es-as a. si1igle e,rent: more spe.cifically, as an extended pr·oloI1gation of tonic harmony with tl1e thir,d in rhe soprano voice. Though the prolongations in the Adagio ultimately lead the motto to tl1e D in n1easure 1 2 and are important because of the long-range im·plicarlon tvhich they generate, chey do 11ot play a declsi,re role in shaping the process of the highest level. Nor is rhe arrival o.f the D-either in measure 12 or in 111easure I9 an event on the higl1est level. Botl1 u1 the Adagio and the Allegro, whate,ver structural impona11ce the D inigbt ha\re had is undermined by the six-four position of the dominant cl1ord and by the 011-going

motion ·o f the bass-line. T ,he harmonic~melodic processes gener~ted in the Adagio are nor closed or even decisively articulated at the begi11ning of the Allegro. They 0011tinue to the cadenoe in i11easure 11- at which point a. ne\v impulse carrie.s the music for,vard. One u1dication tl1at these measures constitute a single process is


266

EXPLAINING MUSIC

that the bass-line m-0ves in an essentially contir1uot1s Li11ear tnar1ner fro1n tl1e C in nleasure 1, t:hr,ot1gh tl1e C ii1 111east1re .1 7, and to tl1e Bb in 1neaStire

zo---at: \ has 111erir, then, melodically, the first twent}1-one measures of the J11overne11t: m.aJ' be consideretf to be an extende.d neighltor-r1ote figure: G- Ab-G; and, harmonically, ttle}r consritute a rugl1-le·vel cade11ce lll Eb n1ajor (Example I 55).

r-1 E.'Xample 155 The first t\'-'Cllt)r-one 1neast1res of tltls 1n.oven1ent are a single process, bt1t tl1ey are differentiated ir1 tern1s of form. Fro111 a formal point of \ne\v, r.he sonara-forr11 structure hegi11s in n1eastire l 7. ~I~his begin11ing is rnade clear b}r th.e cl1ar1ge in r.e1n110, by t:he at)rupt forte on the repeated Ab, arid b}r the

co11trasti11g ch.aracter of tl1e Allegro tl1en1e, \Vhich,. thot1gh derived frotn tl1e patterns present.ed in the Adagio, .is significantly ne\v. There is, iu shortt a, biftrrcatio·n of form ~ind process \~1 hich can be diagra111n1ed as follov"\i'S:

PROCESS: (melody) G . . . . . . . . . . . . . . . . . . . . . , .... Ab- Ab~ G I\!C-v i: n·- - I. (' h".<.,..mo·n..y) 1· tl>l..

FORl\il:

--

•

•

•

••

,

.

.

.

... .

,

.

...... . . ... ,

••••••

I11troducrio.i1

Sonata forn1

a(1 - 6) - a'(7- 12 ) - b(13- 16)

A (17- 21) first theme group

·rhoug l1 cl1e imf)licative processes ge11erated ir1 tl1e Adagio tra11sce11d this fornlal articttlatioc1 a11d l1;;rve import.1nt cor1seqt1ences i11 the sonara-for111 proper, it is ne,1ertheless a distinct ancl St~}Jarate f orm'1l enrit)r as is sho"'rn by cl1e placernent of rhe repeat :r11arks. Beecl1cJ\ e11 does r1ot retr1r t1 to tl1e ten11.po and pace of tlle Adagio, inclt1ciing it as part of tl1e sonata-forn1 struc1


A SUMMARY EXAMPLE

ture, as he does in the first movement of the String Quarter in Bb Major,

Opus I JO. The for1nal articulation which rt'l;uks the begin.ni11g of tl1e Allegro '"vas attributed not 01tly to the inarked char1ge in d)rnat11ics at1d ternpo, but also to tlie ·e'\rident contrast be~vee11 the ethos of the Adagio and that o.f the first Allegro theme. TI1e ethos of rl1e Ada.gio has alre~dy been considered. Tl1e char~tcter and f eeli11g-tone of the ...t\ llegro the111e is not n1ere.ly tl1e result ·Of the faster tempo and louder dynarnic level, but of syntactic processes@mel·ody, rl1yth1n, and I1a.n11ony. Ir is not juSt ten1po \Vhicl1 change.s, pace cl1anges; and it is not just dy~na11ucs which change, but the le,rel ·of psychological intensity. And these, of course, complement one a11otl1er. M·elodically, the separate events of t11e Adagio are compacted, into a single ltineric gesn1re. Here there are 110 introspective pauses, no contemplative repecitio11s. Tl1e .cfuo111acic 111orio11 of the bass is enl'pl1acically goaldirected, and because each nore of the motto is accompanied by two harmonies insrei:d of one, pace and psychological intensity ·are increased" Moreo,rer·, beneath the ambigtUt}' of the chro111a,cie bass motion lies the the latent Possibility of a rapidly 1n-0ving cy~cle of fifths-a possibility that contributes something to the t1rge11cy of t he first llart of tl1e theme. Rhytlun, too, is imporcant. On the lo"-vest level ( 1) (Example 153 ) , rl1e end-accented groups empl1asized by cl1e inelodic-harn1onic repetition across the barline~are incisive. Their con1biuatioi,i, on the next level, creat es a potentially end-accented bar-form. .But t he final notes of this pattern function as the beginning of .a ne'-v rhytbnlic in1pulse as well.. This, restructuring not only prevents the relaxation of te11Sion wl1ich wo1tld have occt1rred had the phrase closed on D, b t1t the elision of th.e tv\'O patternings increases tl1e force of tl1:e melodic-rhytl1mic motion. The lcinetic character of the t hen1e, thus reinforced, is carried fon:vard by tl1e oontrai~y 1nocion between sopr·ano and bass~

Like otl1er reJationsl\ips-i111plicative, hierarchic, and oonformantgenerated during t he :first t'ir\' Cllt}r measures of the ''Les ...~diet1x>' Sonata, t11e contrast betwee11 tl1e conter11plative melar1cl1oly of the Adagio and the fervent: intensit}r of tl1e Allegro the1ne has ·conseq11ences rhr·ough the m.ovement. But the analysis mu.st ei1d here-not o·nly becat1se the reader's interest 111ust be flagging and his paticn.ce exhausted, b t1t because, as inentiotled ·earlier, the rigorot1s analysi~ of etl1etic relatio11sl1ips is beyo11d my kno\vledge and sl

EXPL.AINING MUSIC

M.uch 1vork needs to be do11e, both in this aspect of analysis aud in the study o,f implicative. processes and hierarchic strucrures. There will be disagreeme11t:s about these matters. So1i-1erin1es different exrplanations of a passage or co1nposicio.n \.vill really con.flier: for instance, '\vhen rhey Stem from fundamental differences in theory . At other tin1es, disagreement will be more apparent t1'1an real, and explanations will tJe con1plen1entary, not conflicting: for instance, wl1en diff:erent hierarchic levels or different parameters are made the chief focus of attention; or 1vl1en different kinds o·f relationships confor·mant, implicative, 11ierarci1ic, or ethetic 011es--Gre the basis for the analysis of a parcicl1lar " 'Ork. Wh.arever the reasons for su.ch disagreeme11ts, they should encourag·e, rathe.r than discourage, critical analysis~ not only bec ause tl1e task itself is challengitlg and fascinating, hr1r beca11.1se there is no escape. For our devotion to music ultimately stems from our delight in, and love for, parcict1lar compositions. 1\nd e\rerything \Ve do-all of ot1r study ai1d researcl1-seeks in the end to i11t1mi11ate as full)r as possible the S():Urce and basis of their po'ver co enga.ge and entrance us.


Index o Subjects and Names Abraham, Gerald, ~ n. Abstraction. See Patterning; Understand• ing Accompanirnent: absence of, permits ambiguity, ~ structure clarified by. 161 , r70, 175, 184

Actualization. See Realization Ad hoc h)'potheses. See H )'po theses A esch}rlus, 1..6. Aesthetic experience. See Experience Affective experience: depends on cognition, ~ 113; deviation and, ll J ; intellectual experience and, Qi 1 r 3, z 13. See also Ethos; Kinaesthetic experience Alternatives: awareness of, affects understanding, !_Cl,_ 111-113 ; composer,s~ considered by critic, 18 f. (see also Composer s) ; awareness of, depends on patterning, 111 f. S ee also Criticism; Implicative relationships Analysis. See Criticism; Style analysis Anomaly: discussion and explanation of, !Qi 1!!t2z3, 247; possibility of, w f. Ancecedenr-co11sequer1t parrerns ( A-A'):

discussion of, to-1 i, !.12 !.2 f. [Ex. !1 lL ~ ~ 2i.i !.H [Ex. !11 E f . [Ex. ~ ~ ~ ~ f ., ~ ~ [Ex. 4Z1 ilU f. [E.~. ii1 1031 !..41 [Ex. 56] , 142- 144 [Ex. 2£1

~

211

l.8.1 f. [Ex. ~ 218-130 I Ex. 129] , 210-134 [Ex. 130-131], 240 f. [Ex.

[E."<. 134] i equivocal case of, 132 f .; hjgh-level, ~ rhythm does not structure, 86. Archet}'·pal scl1cn1ata. See Schemata Aristotle, 2.6. Avant-garde, music of, 66 Axial n1elodies, 2!b 183- 191: additive and f orn1al. 2!i djscussion of, 24.t !1i [Ex. 51] , 184- 187 [Ex. 221 187-191 [Ex. roo-101 ] , I..8fi f. [Ex. 1o6]

Bach, J. S., ~ 2L J 10, 205 •

!.!_,

u f., !.h 1!i ~ ~ ~

Bar-form patterns (a-a-b): anapest rhythms cquivaler1t to, ~ (see also Rhythtn ); discussion of, l2 f. [Ex. !11 85-98 passi1n [Ex. 47] , 1..8.8. [Ex. 100] , !12 [Ex. t 33] , 16..1 [Ex. I 48-149] , z..61 [Ex. 150) , 165- 267 [Ex. c53 ] ; other instances, but not discussed as bar forms, Examples l.L !2J. 63, 66, 85- 87, r 34; sf>nata-form not, ~

Baroque music, 209: incomplete rhythms • uncommon in, 205 Barth, John, 1 Beardsley, M .. C., ll n.

• References to particular composicior1s are give11 in the Index of Music.

Material com dire tos autora1s

lNDEX OF SUBJECTS AND }JAJ\..1ES

270

Ileetho,·en, L. van, 8. u_, !!:L 5.L 67, ~ 110; lace st)·le of, E f. Beginnings, in11>onancc of, !...!.z Bci11t1n1, Eduard ,·an, £_ Berg, A ., i1 Berlioz, H .,. ~ i1 Ber11stein 1 La\\'re nee, ~ n. Ilifurcation of f<>rn1 and process. See Forn1 and process Binary for1tl ( i\ - B): discussed,

16] . See also Rou11ded Bot1lez, Pierr e, ~ Brahms, J., !Q. 55 , 11

~

[Ex.

l'inar~· f orn1

Broder, Nathan, ~ Bro\vn, Earl, 6 Bro\vn, !\ onnan 0 ., l Bruclmer, A., l1 Bunge, ~ 1 ario, 8 n..

Cadential 1>attcrn.s. See Gesture Cage, Joh11, h ~ Ho Changing-note 111elodies, 191- 1<}6: disctission of, ~ f. [Ex. 102] 1 192- 194 [Ex. !_Ql] Character, r11usical. See Ethos Characteristic sr~1 lc: a basis for unit)•, 6B. £.;objectivity of, ~ Choice: i111plicatior1 guides, 112; necessity of, r 1 r f.; llasecl 0 11 patterning. 8.. 11 r. See also Composers; Criticisn1 Classical r>criod, n1t1sic of: axial 111clodics unco11n11on in, 184; cadcr1rial patterns in, 152, 167, 211, 2171 241 ; changing-note melodics occur in, !..2.! n.; inconiplcte rhyth111s occur in, 205 i ar1d 11ineteenthcentur)1 1nusic contrasted, !2.i n.; schcn1ata i11, !..!..2 f.; sin1plici~· of dcccptiYc, 163, 176, !..H.b 191, !J I

Classification: not the goal of criticisn1, 67.J !..4i St)· le anal~·sis involves, L. ~ 18; under~-randing depends on, t.8 f., 21 1 Closure : additi\rc for1ns lack, 9~-:¢, 181 f.; l>asis for, ~ cr>da enhances, 1 38; convergence enhances. 157, t6j, 18~, 198, 224; degrees of, R& f.; elision \VCakcns, w f.; e11d-accented rhyrhn1s cr eate, 18, ~ u.8 (see also Rh)·thrn); extension enhances ( see Extension ); f orn1al e\'Cnts

defined t>y, ~ 90; structural gap creates ( see Gap, stn1crural); harmonic (see Harr11on\') ; hierarchic c hancrcr • of, 89; hierarchies ~'tru cturcd b)' , ChaJ>rcr l\T pnssi111; a11d iniplication, Chapters \ TI-VIII pa~·si111; interruption heightens, 191 ; 111clodic ( see i\•1 clody); r1onc.:origrucncc affects, H.b ~ f., 88, U4l based 011 11arreming, ~ para111ctric congruence creates, !!2i performance enhances, 1 38; takes prccedc11cc over i111l1licarion, 1 38i prolongations dela)1, 1 30, !.1.! ( see ,1/so Extension); rcpcticion and, 52, 84-86, 88, 101; return e11hances, 50, ~ 88, 138, 152, 155; re\·ersal articulates, !...!.Q (see also Rc\'ersal); rh~·thmic (see Rh~1 chn 1); structure inferred fron1, 89; st~·Jistic experience and, 138; to11al departure needed for (see Insrabiliry); uniforn1icv \vcakens, 16. !.!2 (see also • U niforn1it\'• ) Cohen, l\ lorris R., I.8 Co111plcn1cnrar)' n1elodics, 175-183: cot1forn1a11ce in, 175, 177, !.l!! f., 181 ; con\·crgcnt and divergent, r..8a f.; dcfi11irion of, 175; discussion of, !1.! f ., !.ll [Ex. 91 l, W f. [Ex. 94], 176-178 [ l(x. 2i1 178- 180 [Ex. Q6:91] , 181-183 [Ex.. 98) c()ll1J')lexit)': f
Material com d1r rtos autora1s

INDEX OF SUBJECTS AND NAl\IIES

hie, 26-28; intention and (see Inrencionalit}' ); masking of, 45-48, ill memOI)' aided b)1, ~ 70, 2Zi n1cthodology and ( see l\1cthod, exigctic; Reci, R.); morivic indi,1 iduality enhances, ~ f.; in nineteench and t1.ventieth centuries, 55l.2.i parameters and (see Parameters); proccssivc, ~ f., iJ. f., 61 ; ps)rchic 1)arsin1ony and ( see Parsimon}' ) ; in prolongations, -z.27, 241; Reti's anal~'Sis of, 59~ 7.!!. f. ( see aiso Rcti, R .); rhyth111 ar1d ( see Rh)rthm ) ; significance of, bec\veen n1ovements, 5.1 f ., 70-72. 2Zi significance of, between works, 7 1-7 3; strength of, 46-49, 2.1! Zi f.; style and, 74-75; thematic transforo1ation and, 55- 59; unity and, 64-67 ( see also Unity) Cor1junct patterns, i 31-144. See also Linear melodies Contextual discrepancy, 1 21, 196-11 3; basis for implication, !!&i ? 18, z44. See also Potentiality Continuation, principle of, lh 130 Convergence: discussions of, 100 [Ex. fil ~ fEx. 68]. 142-1 44 [Ex. 70-74) , !12 f. [Ex. 76] , !J2 [Ex. zz1 ill [£.'<. 82]. ~ f., ~ [Ex. 86] , !1Q [Ex. 90], w_ [Ex. 2.!.1 um [Ex. ¢-97] , ~ [Ex. 105], 214 [ E.,. 114] , 16.z [Ex. 150] Cooke, Deryck, ~ n.. Counterfactuals: ps)rcl1ological sig11ificar1ce of, r r 2 ; understanding and, ill. Counterpoi111:: fourth species, !.i.! f., 168, l 59i gesture and, :ut8 f.; probabilistic, 1.i pure process and, 2! n.; sixtcenth-centur)1, 8 Critical anal)rsis. See Criticism Criticisn1: ad !Joe h)' potheses used i11, 12!..4 (see also 11ypoc)1cscs) ; affective experience and, ~ alternatives considered in, l8 f., 116; composer's choices and, LB f .. u f.; conscientious, lii description and, 2? disagreements about, 121, 268; docucnentation and, 2..1 f.; not exhau~-ci,•e or definitive, 14 f., 105; explar1ation and, ~
271

~ f.,

r 10; n1ethods used in,

Chapter \ TI passi111,

u f.,

~

n. (see also .~v1ethodology; Methods) ; need for, !.ii 168; objections to, 4-6; particulars explained b)', 6 f., 12-1 5 (see also Hy231

potl1cses ) ; a11d perforn1ance (see Performance ) ; relevance of sk'e tches for, ~ 77- 79; style analysis and (see Style anal)rsis) ; theory and, 2! !b ~ views of composers and theorists and1 2.1 f.; indi vidual responses r1ot explained by, 1 Deflections: alternative goals created by, 119, 130; definition,

u.8 f.; discussion,

uu rEx. 54], u.8 (Ex. i21 !.!i [Ex~ hl !.ii rEx. 80], 162 (Ex. hl !II. [Ex. 25.1 gap creates, !l.:i: (see also Gaps, structural ) ; reversal a special case of, ~ ( see ai;-o Reversal) Deccr111inism, mistaken applied to music, Ul

Dialectical change, music as , 56-59, ~ f. Digression: need for, 10=1:z. [Ex. 1 & 41 !h 241 . See also ln!l'tabilit)' Disjunct patterns, 144-174. See also GapfiJl n1elodies; Gaps, structural; Triadic rnelodies Dra)' , Willian1, !.2 J))•nan1ics: ethos and, lli 4£ f ., 128-130, 166, 11 1, 243 ; ps)rchological intensity and, 267; structure clarified b)r, 1oz, !.ti! 155, 2231 167; S}'ntaccic function of, l i f. ( see also Stress) Echo: n1orphology nor111alized by, 104 [Ex. 57]; potential realized by, w [Ex. Z21 stability \.V·eakened b)' • !iZ f. [Ex.

~ Eliot, T. S., 6 Elision: bifurcario·n of form and process creates, 100; eo1phasis created by, !..12 fEx. 149]; morphol
\•iarion and, 119, 221 . 121 , 143 ; implication inseparable frorn, 218. 223,


INDEX OF SUBJECTS AND NAMES

272

Ethos ( conti11ucd) 243-245, ~ f. ; kinaesthetic experience of, 113 , '12 f., !!! (see .1lso Kinaesthetic

experience); rr1asks conforn1ance, 48; parameters a11d syntax define, ll:i ~ f ., 140, L.66. f., ~ f~ 130, 243, 256, 267; perfor111ance aod, ~ f., r 18-130; theory of, needed, 246j \\lords inadequately express, 243, 246 Events, kinds of. See Deflection; Ger1erative evcnrs; Medial cver1ts; Prolonga-

ll.1 Chapter I\ T

passi11z f Ex. 47]; i.nf errcd f ron\ closure, ~ kinds of, 9521.!. naming of, 2.!.i ~ nonco11gruence and (see Parameters, noncongruence of); process and, djstinguisl1ed, ~ 90~ ( see also Form and process ); processive ( see Processive for111s); relationsl1ip bet\veen movements and, 21 (see also Con formant relationships). See also Antecedcnc-consequcnt patterns; Bar-

tions; Reversal; Tern1inal events Evolution, co ncept of in music, Q± f., 2..! Expectation: concept of in1plication pref erred to, !_!in. Experience : aesthetic, ~ !.2.i 168; affecti\•e ( see Affective experien.ce ) ; intelligible, 3-6, !.2J 11 li kinaesthetic (see Kinaesthetic expcrie11ce); n1usical (see Understanding; Listener ) ; n1ystical, ii na-

for1TI patterns; Binary form; Rounded binary form; Sonata-fom1; Ternary for1n Form and process: bifurcation of, defined and discussed, ~ (Ex. fil 98-100 [Ex. fil Ull [Ex. 54], ~ [Ex. 511 !.2l [Ex. 103], !!2. [Ex. !.!2Ji !_li fEx. !.1!1 265267 [ E.x. ~ & 150]; hierarchic alternation of. 90; mixed in 11·1ost \vorks, 2!t 22 Frar1ck, C., ffi iZ.i 2§.

i vc, l! ur1ique, ~ f. Explanation: aesthetic and generic distin-

Gafurius, F., u

guished, 77- 79; complete, n ot 11ecessar y,

Gap-fill rnclodics, 145-1 57: cornn1on in

!.4J ~ f.; conceptualization needed for, 16, 113, 116i description not, 2.i goal of criticisr1\, !..Ii haza rds of, 25; laws and,

tonal music, 101; described, 145; gap and fill separable in, uu f ., 109; highlcvel, 100=-102 LEx. fil 150-r 51, ~ f. [Ex. z!1 111i.ddle-level, 151-154 [Ex. 12.1 160 [Ex. hl L6..l rEx. ~ lowlevcl, uu f. [Ex. il1 103, !.il [Ex. ~ ~ f. fEx. hl !21 f. [Ex. hl W £. [Ex. liL !.iZ f. [Ex. £ ~ f. [Ex. 7zJ. ~ [Ex. 86] . 166 ([Ex. hl !.2.2 f. [Ex. 106] , :w:8 f. [Ex. tOQ=l to], 2 c4-116

8-14; le11grh of, !.!i f.; 111ctl1ods used in (see ~I eth<>d<>logy; 1'1cth<,ds, analy tic);

reasons en1ployed in, ~ S ee also Understanding Extensions: c losure enhanced b)', 2.ii 130, 138 [Ex. !11 150-15 2 [Ex. 78] , l.6o

[£ x.

~

~

[Ex. 2.21.i rnomen.t um f. [Ex. 13 1)j n1orpho-

slowed by, w Jogical length stretched l>)', Ifil.

hl IM I Ex. hl m

rE x.

t.Ex. 2.!.1 w f. [E.x.

131]; '"ithin phrase, in1plicati,,.e, 136-

212 [F.x. U1] Faulk11e.r, , ,\ fm., l F'erguson, Dor1ald N., ~ n.. Forn1: additive, Bae hierarchies, 2! !!!.1 93~ (se<: also Axial rnelodies; O sth1ato ; Strophic for111; Ternary for111 ); co11f ormance articulates (see Conforr1\ant relationships ); evenc.r; in, defined by closure, ~ (see also Closure ); l1ierarchic structuring of, 38-40 [Ex. !J.1. 44 f. [Ex.

[Ex. 141

117]. 111

[Ex.

130], 24C):=253

LEx.

l

Gaps, structural : ancillat)' , 1 39, 146 f., 199; closure e11hanced by, 141, i ~ f., 170 fEx. 221.i !.2J [Ex. 25.1 deflection crea red by, !l..4 [Ex. hl t.00 [ E.x. hl d escending, unco1nmon, 1 34; explanation of, ~ filled by proxy, I..66 [ Ex. ~ perfect fourths as, ro3 n., 145 1 1.6.I n., 166; imply fill, ~ 124 f. [Ex. hl !..J.i [Ex. ~ 1441 ~ [Ex. 86], ~ [Ex. 101] ,

!..21.

[Ex.

108],

12 3], lll

LEx. uB. LEx.

105] , 203 [Ex. 107], 100 [Ex. 120] . 220 [Ex. 122-

!.!2. f. [Ex. 129],

t 14].

w

gs [Ex. !.!5. BJ, [Ex. 130]; fill of,

Mat nal com d1r 1tos autora1s

273


pro·vis1ona1,

reversal created by (see R eversal ); size of, and irnplication, 145; triads as, 102, l IQ, 125-117, 147, 2 r8 (see also Triadic melodies) Generative events, 2-!.l J 18, 141, !A'.i Gesture, 1o6-213 : basis for, 207 f .; cadencial, 110 [Ex. 59A], !_£ [Ex. ~ & 78], !..ZQ rEx. ~ 111 [Ex. !..!i & 132], 2 14m [Ex. 117] , !..!1 [Ex. I 19], !4:.! [Ex. 134) ; arid Characteristic sryle, 207 Q.:.i coda, ~ [Ex. !Ii & 116], !i3 f. [Ex. !..ll & r 36]; conventionalit}r of, 107-113 passi111, 243- 245 passi11z; discrepancy of context and, 207-116 passi111, 246 f.; fugal, .l.Q8 f. fEx. 102:"1 to]; kinds of, 206'f.; n1obile, 118-221 (Ex. 121-124]; potentiality and , 2o6-226, :z.43- 247; t};polog)' of, needed, 207. See also Sche1nata Gombrich, E. H ., ~ f . Grout, D onaJd J., u Grove, Sir George, :u8. Gushee, Lawrence, u n.. 111;

Harmo•'l}' : an1biguous, 140, ~ w f.; closure and, Chapters IV, Vl-VTII ,passi:nz; conformance and (see Conformant relationships); ethos affected by, 140, 2191 230, !.iJ f., 267.; gesture and (see Gesture, cadential); hierarchic disconrinuity of, ~ f ..; implication and, Chapters Vl-Vlll passi111; n1obility of, ~ 86, 1 ~ f ., ~ 1981 266j noncongruent in deceptive cadence, 81, and se111icadence, 85 ; probabilistic, 2i rhythm clarifies, 265; structure articulated by,. Chapters I\T, VJ- Vlll passi111; style and, lli ~

Hegelianisn1, exigetic m ethod based on, ~ Hierarchic equivalence: la,v of, !1ii particular im"tances discussed, ~ [Ex. 75], !.i.2. rEx. zz1 !J£ [Ex. illi !ll [Ex. Z21

f. [Ex. 2.!.1 ~ [Ex. 2.5.1 ~ [Ex. 101] , 200 [Ex. 1o6] ; provisional realiza-

!._TI

tion and, !_!! [Ex. ~ 134-136 [ Ex. 66], 142-144 [Ex. 2Q & 73-74]. See also Structural tones, selection of

Hierarchic structures: analyzed, 1,8-40, 83-89, C hapters \TJI- Vlll passi111 (see also 1: om1, hierarchic ~'tructuring of; ,l\1f elody, hierarchic structuring of; Rh ~rthn1, hierarcl1ic structuring of); arched, 21.! braided, 2Zi conformance and, define forr11, 2!..!. conjuncci,rc flat,

2b countercumulacive, 2 58; criticism analyses, !1.i different disciplines study, ios; discontinuity of, ~ f.; flat, ~ !b 93=96, ~ f. (see also SttOJ)hic form); form a11d process alternate in, 90, 120. 2 z.Sj i111f)lication a11d, 120, 253; in1portance of, 66 f., ~ inferred fron1 closure, ~ levels of, and forms, 81. ~ f., 131 ~ n1clodic, 111, !.2il objecti\rity of, 41 parametric closure articulace.s, 81=98, (see also Closure); patterning basis fo r, 8J. f.; potential, 256, z..6.l [Ex. 148150] i strongly shaped, 194; unity dependent on, ~ ill 66 f. H}rpotheses: ad hoc, expJajn particulars, 11-14 r8, ~ 118, 111, 158, !.HP.• 187; • • • • cor11J11on-sense, Q-11; cr1nc1Sm requrres, 21 !.2.i 116; from other disciplines, 14; i1nplicative inferences as, 113, J 16; sryle anal}rsis provides, !1.? l.8 f., 118; theory of music provides, 21 !.it 18, !.i Jconology of 1nusic: possibility of, ~ f., !£ Implicative relationships: affect and, 113; alternative, !.2 f., 11 r f., Chapters V l-

Vlll p.assi111; ar11biguity of,

1 11,

ij9_ f.,

~

f .; a11ornal}' and, 111 f., J 18, :i4oi awareness of, 1 ~ f., 161 , 237; axial melodies, weak, ~ 185-186; bases for, 113; choosing based or1, r 1 r f .; closure ta lees precedence over, 138, 156 Thi con1plexicy of, 113, 119, 197, u8 f.; conforn1ance 111askcd by, il.i conjunct patterns generate, 1 31-144 (see also Linear melodies); context affects, i! f., 196-213 ( see also Potentialit)' ) ; convention enhances, :07 f., 217, 119; criticism explains, !L 116, J 30; defined and discussed, 110-113; deflections change, 118, 130 (see also D eflections); dela}' enhances, ~ !1!i 161, Chapter VII pas-


INDEX OF SURJF.CTS AND NAMES

Implicative relationships ( co·11ti11tJed ) si111; clisjur1cc j)atter1ls generate, 144- 174 (see also Gap-fill melodies; Gaps. scrucrural; Triadic 111elodies); and e[hos, 140. 156; expcccacio11 a11d, r 1 ~ n.; extensic>ns and ( ;·ee Ex[ensions ) ; gaps create (see Ca p-fi II 111clodies; Gaps, strucrura1) ; gestures c reate, 206-2 r 3, 1 18. !.!.J (see ,1/so Gestures ) ; har1non}' and, 84, 85 1 I 00, Cha peers Vil-vr1 I passi11i; are hierarchic, 1 ?O f ., ' 3 1, 2 53; l1icrarchic equivalence governs, !1!l (see also Hierarchic cqt1ivalcnce); as ll}'))Otheses, I 1 3, 116; incomplereness and instabiliC)' basis fo.r, 118, i:z.~. Chapters Vll- V lll pnssi1t1 (see also I11co111pleceness); inferred f ro111 patterns, 1 Io, 11 l ; ki11aesthetic aspects of, 113. 119; la ~,· of

conr1r1uario1l governs, I 30; n1elodic, ~ 22.i Chaprcrs \ ! I-\Tf ll passi111 ( see also J\•l elod!. ) ; 1nobilit)' and ( see J\lobility); patcerning basis for, !h. f!.h 1 1 o, 113 f ., 121 , 1 ~o. Chapters Vll- \ 1 111 passi1Jli pocencialic)' creates, 222 113, 1¢=11 3 (see also Potcntialit)' ) ; probabilit)' and ( see Prol)ability) ; prolongation and, 119, 151, 126-1-lt, Z.f.Q, 261, 165; realization of ( see R ealization ) ; repetition and , i! f., io6, 2 29, 2 31 ; retUin regeneraces, u8. f ., c26, !±1 f ., 180, 182, 156i rctrospectiv·c understa11ding of, r 11, r 13, ~ r r8, Chapters VII- Vlll passi111; reverberation of,

reversal and ( l'ee R e,·crsal) ; rh!rth_m and (see Rf1!1 th111 ); schc111ata and ( see Schema[a ); as sigrts, 1 11 ; 5111all differences affect, 3J f., !.!l f., 195, !!2 f., !..l! f. (see also wfechods, anal}·tic) ; specificiC)' of, ~ 11 7, 137, 143, 181, 212 , 246, 148; sy111n1etrica l patterns generate, 1 74- 196 ( see 11lso A.xial melodies; Cha11gi11g- nocc n1elodies ; Cornplc111entary 1neJodies); tonal te11dencies gcr1crace, r 2 3- 126. Cha peers \ Tfl- VIII passi11z; unancicipated eve11cs and, 20, !..!.1 f.; undersca11ding, 28, r 10-1l3, 117; uniforn1it}' generates ( see UnifonniC)·) ; '~·eak ening of shape creates ( see Shape ) l11con1pleccncss: i1l1plication generaced by, t 1 71

1? 1;

proccssiYc and tonal djscinguishetl, 113; ~>ro cess ivc and to nal discussed, 1 2 3!.!L Chapters Vll- \ ' 111 passi111; rh}1thn1ic, 203-2o6 (see also Rhych111 ) India, 1nusic of, iii 66 D' lnd)' , V., iZ lnevitalJilit}· in music, :w f ., 56-58 lnscabiliC)'; required for closure, 88, 136~ 138, !.22. f., 191 , 200. See also J\ lobility ; U tliforrnicv • lr1ccntional it)f: cunforn 1ant relationships and, 73-75; ini probabiliC)' ir1dicaces, Z4i !..42 n. 113;

Jau1es, \ Vrn. , L 6 Ja1)an, 111usic of, 109 Java, 111usic of, ~ 109 Jo11es, David E. H., !..5. f cr1l1a11, josepll, Ull ff. Kinaesthetic experience: affect and, 11 3, ~ f.; association arld, ~ f. ; i11elody and, f. ; timbre and, 112 f . Kirb~', F. E ., 68. f., 207 n. l(ocsclcr, .A..., !.1 f ., ~ !b ~ Kohler, L,. a11d Ruchardt, A., phra.c;i11g of i\ lozart Th e111e, Chapter 11 passn>J

w

Kurth, Err15t, u

See F.xJ1lanation; H)·pothcscs Learning: educacjon and, !1i through 111usical experic11ce, !b 16, Zii ~ See also I .. i~-rencr; Understanding Lehrer, To111, i Lenr1ebcrg, I Jans, 2 12 n. Lic htentha I, Peter, 6B: Linear 1T1elodies 1 i 31 - 144 : bilin.car c o11verging and di,·crging disti11guisl1ed, CO.f. f. ; bilinear converging, discussed, !.;1Q f. [Ex. 68] , 1-11- i44 I Ex. 70-74] , ~ fEx. 86] , 178- 181 [Ex. 96-97], u.8 f . [ Ex. 129] ; bilinear diYergi11g, discussed, l 31-133 [E x. 64-65] , ~ [Ex. 76] ; bili1tc
z

Mat nal com dtr 1tos autora1s

INDEX OF· SUBJECT'S AND NAi\1ES

66] , ~ [Ex. ~ f.

IE.x.

240

[Ex.

fil

~

2 75

fil

111J,

I 34];

VJ- \ 7111 passi111; linear (see Linear r11clodics); motor bc-

f. [Ex.

h.avior ( kinacsthesis ) and, f. ; 11eighbor-note pattern, 265 f. [Ex. l 55]; of nineteenth century (see Nineteenth-centllr\1 ' r11usic); present stud\' of, lin1itcd, 1 ro; sin1jlar, co111pared (see Methods,

l'niddle-lcvel,

Ul2.

[Ex. hl 210 [Ex. 112-113], ll.J f. [Ex. !..!41 224 f. [Ex. !..!ili lowlcvel, 83-87 [Ex. 47] , 2! f. LEx. ~ 11 s.

i21

w

hl

[Ex. 6o] , lli f. [Ex. hl u8 [Ex.

[Ex.

z.z1 !.iQ

u.8 [Ex.

ill

[E.x.

~ ~

f. (E..x.

hl

~

fEx. hl 1.00

[Ex. ~ I..6.r f. [Ex. hl !§3 f. [Ex. 86], L.6.6. f. [Ex. 88] , w [Ex. 211 ~ [Ex. 25.L t.8ti [Ex. 221 188- 190 [Ex. 100101 ] , ~ [Ex. !.Q11 !.!.l f. [Ex. ~ & 126], 236, [Ex. 133) , 240 [Ex. 134) 1 !!! f. f Ex. 141], !.i! f. [Ex. !.i! & 145B] Listener: COlt1petcnce of, based on expcrie11ce, 1 1 o, I 30 1 207; conceptualization unnecessar}r for , 15-1 7, r 16; gestures and schen1ata understood by, 107, 1 14, 213 ; irl1plications understood b)', 1 1 o, 113, inferences made by, 110, 1 r3; kinaestl1ctic experience of, 141; memor)r in1portant for, 8.o (see also l\lemory); probabilities understood U)', 2 8. 130; pro\•isio11al realization reinforces, 11 7; ref erenee to not necessar}', 1 1 o, ~ n. See also Understanding Liszt, F., ill ~ l.Z l.-ord, Albert B., Iii n.. Lo renz, Alfred, ~ 116;

l\1[acha\lt, G. de, ~ !..QQ

l\1larthesot1, Jol1anr1, 11 2, !J2 Mcdja} c\rent, 2b 1!9. f. ~lelod)' :

an1biguit}' of, 169; ascending vs descending, w f.; axial (see AxiaJ n1elodies ); changi11g-note (see C hanging-note melodics); choice of exa1nples, 11 o; of Classical period (see Classical period ); classification of, 131 , 1.06. f.; closure articulated b)' , 83- 891 2..! f ., 98-1042 Chapters \ 1J- \ TIII pa1si111; cornplcmcntar>'

(see Con1plen1entary nlelodies); conjunct and d isj unct intenrals basis for,

3 1; gap-fill (see Gap-fill 111elodics; GaJ">S, structu ral ) ; gesture and (see Gesture ); hierarchjc structuring of, 131, 81-88 LEx. fil 28-100 (Ex. .u1 100i

H1 C hapters

!_2.4 f. [Ex. 103-104], 231- 233 LEx. r30-11.!1

lD..2

[Ex.

w

;

anal}itic) ; adequate theory of, lacking, 109; triadic (see T riadic rnelodics) J\•1en1or~' : axial n1elodics ser\'C, ~!bi conforn1ant relationships aid, ~ 58, 70, ~ 2.Ii l1ierarchic struct:ure facilitates, 80; individuality and, ~ schematic aid, ~ l\ Jeter: cross, weakens shape, 200 f .; disturbance of, 2 36; ethos and, u8 f .; fast t e111po empha~izes, i!..! 83 , u8 ~1lethod, cxigctic: circularity of, 65; defi11cd §.4. f.; use of, ~ f. , 1.!

i\·i ethodology, t.16-r 23 : anomaly and, 1 18; 111ental experin1ent part of, 116; musical-

ity gujdes, !.fu normalization, !..Z (see also Jv tethods, anal)rtic); objectivity of, 18, 64, J 16; octave transfers and, !31 ( see also Occave); Rcti's questioned (see Reti); retrospective r1ccessary, 1 £3 , 116; scope of, ~ Hi specification of implication needed, I 17; strict rules irnpossible, ~ I 2 t; and selection of structural tones (see Struc tural tones, selection of); unrealized impljcation and, 1 1 7. See also Criticisrn; Methods, analytic i\1cchods, anal)•cic: b)~ n1odification or r1orn1alization, u [Ex. 1- 3 compared wich l & tli 31-34 [Ex. 8-10 w.irh §.1 3.i [Ex. I ID ,,rith ~ !} f. [Ex. 1i \Vlth !!.1 ~ fEx. z 3C '"'ith z 3BJ, ~ [Ex. U \\rich ill.,. 50-5 r [Ex. ?8A & B \\•ith 28C

411.i.

& DJ, ~ [Ex. 42 \vith t20 [Ex. 59A \Vith iQ1 124- r25 fEx. 6 1A & B \V.ith

!..i.4 [Ex. 8a. with 211 !.iZ [Ex. hl w [Ex. 2!1. !..25 [Ex. 103], u6 [Ex. x17B & B'with 117], ?18 [ Ex. 120F \vich A and BJ, 222 f. [Ex. 124] , ~ f. [Ex. ~ 236 f. [ Ex. 133B \vith A ], !J..2 IEx. 133C with A] , ~ LEx. 138 -..'.\rith 139141] , !.i! [Ex. !..i.! \\•ith 14 1]; b)' paired comparison, Chapter Tl passim [Ex. 6. \Vi th i1. £ [Ex. lil \Vi th ~ 86 f. [Ex.

filL


INDEX OF SUBJECTS ANO NAMES

(\'lechods ( co11ti111,ed ) iZ \vith !11 I.Ill. f. [ Ex. 55-56 \Vith H1 !.!1 f. [Ex. ~ \vith hl Chapters \'11Vlll passi111 J\ f itch ell, \.\f111., us ~lobilit)' : acco11tpani111ent enhances, 139; echo creates, ill f .; harmonic, ~ 100,

Chaprcrs \ .'JI- VIII passi11z; rnelodic, Chapters IV, Vl- \ TJII passi111; nonco11gruence creates, ~ ( see also Paramercrs, non congrucnce of); and c losure depend on paratl letric inneraction, 81, 88, C hapters Vl-\TJI( ptrssi111; prior orga nization and, 3.ii rltythn1ic, 28, 32-34, Th ~ ~ 100, 12 7, Chapters Vll- Vlll p.1ssi1J1; unifo r r11it}" creates, 1.6 ( see also U 11iforn1it)' ) J\lu111entun1: lengths extended b)', ~ 161, 233 . See also U niformit)' !\lo rphological le ngth : co111plcte before e xtension, 2 3 3; elision disturbs, 2 33; establish1nenr o f, 114, '5'5i equivocal, 132i exte nsio n o f ( see Extension ); nor111alization of ( see Prolongations, no rn1ali7.ing) ; patterning c hanges, ~ [Ex. ll1.i stretc hed ( see Extensio ns) i\fozart, \V. A., !.Q.2 117,

Narriicl ur, E::ug ene. ~ n.. Near Eas t, music of, 6.6 N eigl1hor-n o tc structure, ~ fEx. !iil Ninetecnth-cenc:ur\.r music: axial m elodies • occur in , ~ !!:..i changing-note n1clodies nut co111mon in, ~ n.; conformant relatio nsl1ips in, 55- 58, ZQi c ultural influences on, 56-58; hierarc hic sin1plicit}' cif, 163, 166, !.2.i n.; i11complece rhy thms in , 205; size of co ni posicions in, !l:2l ~ ~ ~ n.;, variations beco me processive . in, ~

Norn1alizarir>n. See Methods, anal~~ti c Notation, li111irations of, u f . O c ca\•e : fun ctions as goal, 102, 197; range defined b~·, 103, 118, !..4Q f ., 149, 1 58, 261, u6 !b 240; Stability enhanced b)r, Cl1ai)tcrs \ 7 1l- Vl ll passi111; rransfer, lcgirimac)' of, r 37 1 t.8.o. f., !21 f., 200, 206 n..

Orchestration : syntactic use of, Th~ L.8.6 f. , 202 f., 20j, ill Originalit}' : nature of, !.2. f., 213,

12?1

126 1

~ n.

O stinato: additive, .tlac hierarch)' , 2Q !b 23 f ., 106; change not implied by, i!

Palestrina. P. da, 8 Paranieters: closure a11d mobility· created uy ( Chapters IV, VI-Vlll passi11z ( see also Clos11re; Tn1pl icat1vc relationships; 1\·t obilit)' ); conforn1ance d efined b)' , 44~ confo rn1a11cc 111askcd by secondaJ.J·, iZ f.; congruen ce of, creares closure, 86, ~ congruence and 11011congruence distinguished, 81 ; hierarchic suucn1riog of, ~ f .;

n1cthodology and noncongrucnc, !L c2 r; noncongr uence of, ~ f. [ Ex. 47], 154£. [E.x. 221 r68f. [Ex. 901, 1;4 [Ex. 144] , 258 r£x. 148] ; noncongruent in deceptive cadence, 81 . and in semicarin1ar~' and secondary d iscjnguished, ~ ~ 4.t 88; sin1ultaneous ,·ariation of, i:1 f.; scy•le and, ii! S)'ntax of, and hierarchy , ful f . Parentl1 ~cs: concinllit)r interrupted by, U.2 f.; discussed, 139-241 rEx. I 34] ; in1plicatio n not: a ffec ted by, 119, 139; lack of conforn1ance in, !41 ; need for, explained , ~

P arsin1ony, ps) chic: conforn1anc e creates, ~ f.; pleasure of, ~ ~ 1 50 n. Patterning: involves abstraction, l f.; affect depends on, fu archety pal (see Schen1ata ) ; basis for, 16-18, !!Ji 204; cadential (see Gesture ) ; confor1nance and, 42.!_ Cha1)ter 111 passi111. (see also Conforrnant r elatio nships); conjunc t ( sec Linear melo dies) ; continuation of, 130; critical method and, 116; disj unct (see Gapfilled n1clodies; Gaps , structural; Triadic n1elodies ) ; d)·namics affect, ll ( see also Stress); hcterogcncit)' precludes, 26, ~ hier:irc hic struc ru ring of, Chapter JV passi11z (see also Hierarc hic structures) ; i111plication based on, !L 83, 113, 114124, r 30; incon1 piece, 118, !..!J (see also Incompleteness) ; learned vs. innate, 2 I 4-



218; not aJl, implicative, !..!..2.i performance niakcs clear, 29; potential (see Potentiality); proxit1lity and disj unctjon articulate, 1! f., 83, ~ 202; repetition of (see R epetition); rhythmic, ~ f., 204 (see also Rh}1 thm ) ; as signs, 111; similarin1 and difference needed for, ~ " ~ 204; symmetrical (see Axial rnelodics; Changing-note r11clodies; Con1ple1nentary 111elodies); u11derstar1ding based on (see Understanding) ; unifor111ity precludes, ~ ~ 104 Pavlov, l P ., !2 Perforrr1ance: alternative passible, ~ f.; closure e1nphasized by, x J8; criticis111 suggests, !2! Chapter II passi111; ethos and, ~ f., 118-130; genre and, 1!1 hierarchic structure and, 38-40, !.i2 f.; noncongruence articulated in1 ~ tempo and dynan1ics in, 40-42, 128-130, 138 f . Polanyi, Micl1ael, !.1.i !..2

s-..

Potentiality: basis for in1plication, 12 3, 196; hierarchic (see I-Iierarchic Structures) ; iotervallic, 11 f. [Ex. 14-1 5] ; melodic gesture and (see Gesture ) i pitch and (see Potential structural tones); rhyth1n and (see Potential rhythms) Potential rhythn1s: definition and discussion of, 32- 35, :u [Ex. 6-11], ~ [Ex.

.411

!.ll [Ex. 'ML 201-203 [Ex. ~ orchestration emphasizes, 201-103; perform ance of, lb i.! f. See also Rhythm, incon1plete Potential structural tones: definition and discussion of, 2.! fEx. ~ ~ f . [Ex. u1 !J_i [Ex. 66], ~ [Ex. 2i1 i.B6 [Ex. ~ 188- 190 [Ex. JOO-tor], 196-198 [Ex. ~ 198-100 [Ex. 106], 203 [Ex.

!.£21

[Ex. 124] Pousseur, Henri, 6 ll.l

Probability: in1plicarion and, lO, t 10-113, 11 6 f., 130, Chapters Vll- VIlI passrm; influence of frequency on, Ill f .; intentionalit}' and (see Intentionality); relati,rity of, 1..6o n.; statistics and, ~ 116 f.; style ar1alysis co ncerned \Vith, 2.i use of term, 28; theor y and,

z

Process: closed, not implicacive, 2£.t !.!2i 228; confor111ance a nd (see C-0nformant relatior1shi1)s); for111 ar1d (see Form; Forn1 and process); pure, 2! f. ( see also Processive fonns); r epetition and ( see Repetition); uniform, and closure ( see R eversal; Uniformity); ~reakcning of, 256-258 Processive forn1s ( 011e-part): 91:93 [Ex. 50] . uu f. rEx. il1 108-209 (Ex. 1091 !.Q] P rolongations, 126-2 37: before phrase, 1 JOj closure delayed b)' , 130; declarative, !!.2.f. [Ex. i21 m f. [Ex. ~ 88 & 128]; effect of, and context, 2 30 ff., 2 37; exte11sio115 (see Extensio11s); implicatio n within a comp lcx1 249-16s passim [Ex. 141, !.iJ & 148-150] ; kinds of, 116; implicative latency of, 2 28; normalizing, ~ [Ex. 57], ~ [Ex. ~ W fEx. 79], !!_2 f . (Ex. 119]. i l l [Ex. c30]. 236-139 [Ex. UJ1 See also Parentl1escs Provisional realization. See R ealization R atner, L eonard, :.u R caJization of implication : alternative, u.B f., 12 3, ill f., Chapters VII- VIII passi111; not achjeved, 117; closure and, ~ ~ r. !.2 f., Chapters \711-VIII passi1Jl ( see also Closure); convergence of ( see Convergence); deflections dela)' , uB (see also Deflections); dcla)' of, 116 f.; harmony assures, 181; hierarchic equivalence needed for, !..M (see also Hier archic equival ence); need for regular, l • 7; orcl1esrration. cm1 )hasizes, 133, 144. 186, 2 2 5; pro,risional, dcfi11ed, 1 17; provisional~ discussed, ~ [ Ex. hl m

[Ex. hl ~ [Ex. 66]. ~ [Ex. ~ zoo f. [Ex. 1o6), 2.06 [Ex. 108] , uJi [Ex. !..!11 !!.2 LE."<. 111B] , 222 - 124 [Ex. 114], !i1 f. [Ex. 138-139] , i.6.I [Ex. r48] ; proxi1nate, defined, 119; proxi111ate, dis~ cussed, u6 [Ex. hl Chapters VII- VllI passi111; by proJ1.-y, 1.66 [Ex. hl remote, defined, 119; rcn1ote, discus-scd, !..!1 (Ex.

hl W (Ex. hl ill [Ex.

221

142-144

t6o. [Ex.

[Ex. 70-74) ,

hl

~

[Ex.


INDEX OF SUBJECTS AND N AlVlES

Rcali1~tion

fil

of in1plication ( conti11ued) t.8.a f. [Ex. 96--Q7] , !.2:1. [Ex. ~

1()8-201 rEx. ~ & 10-6], 206 rEx. 1o8] , u6 [ Ex. 11 7], 224 - 126 [F.x. 115-1 16], 248 [Ex. ~ 253-255 rEx. !.fil :fill. [Ex. 150]; rules of proxin1ate and ren1otc, 130; specificit}' of, 117, Chapters \ 'l-Vlll passi111; transcends form (see F'orn1 and process, bifurcation of) R easons: rule a11d strategy, djstinguished, 11 f.; rule, and style, ~ 2 r 3; rule, and t heory, !.ii use o f, iJ1 criticisn1, !!!.:. See also H }rpotheses; Theory R eger, /\lax, i l f. R egister : in1pJicativc, !_N f.; spccificiC)' of, It z, Chapters Vil- V III passi11z R elationships. See Conformant relationships; Ethos; H iera rc hic structures; l mplicative relationships RcflCticior1: cl1angc i1l1plicd by, i! f.; closure and (see Clost1re); ditferences empl1asizcd by, li.i. differentiation i.n1pJ ied by. 85, 204i effect of, depends on conrcxc, i.! f.; forni and (see Strophic form; T ernary forn1 ); forn1 articulat.e d b)r, ) l f., 65 ; patterning not created b )r, 26, ~ 205 ; 1}roccss \VC3kened by, 2li not j)S_vch olc>gic:al, iii i!..i. rerurn dist inguished from, iii 42 f., ~ rh}'· th111ic (see Rh~·rh111 ) ; separation created by, 51-53, 67, 101; tension released by, 103; unity not t he result of, ~ (see also Unit)'). See also Co11formar1c r elation hips; R eturn Rcti, R udo lph: n1cth.o dolog)' of, d iscussed, 6o, 62-65; thec>ries of, questioned, 64-67, ~ f. Si.:e also C...011forn111 nt relationships Return: closlirc a11d for1n articulated b)r, 42 f., ~ !IB, Cha})tcrs \ l ll-Vlll passi11z; r epericior1 disti ngt1ished from ( see R epetition); si n1ilaricies emphasized by. il: See also Closure; lrnplicarion R everbcracio11. See l1nplication R eversal: c losure articulated b~·, 86 [Ex. !..ZL !.!i rLx. hl !J1 f. rEx. 66], 170 rEx. 29.L !.i9 LEx. 149] ; defined, lIQj !f-lf' creates, 110 [Ex. 59A], ~ [Ex. 68]. !il [Ex. zfili I.6i [Ex. hl !.li

fEx. 2!L

1781 !.22 [Ex. 106] . z 36 n. [Ex. I 33]; lack of. implies continuation, M..~

Rhythn1: an1biguous, ~ f.; anapest, as bar-form, 87, 1.8.B (see ti/so Bar-forn1 patterns); anapest groups discussed, !!z [Ex. ill~ f. fEx . .u1 uB [Ex. 63]. !32 [Ex. 66], ~ [Ex. z.21. ~ [Ex. ll1 1.6.:.

[Ex. hl ~ [Ex. 86] , t.66 I Ex. 87], !.15! [Ex, 90] , r..8.8 [Ex. 1 oo] , ill [Ex. ~ !.!! [Ex. !.14 !i! [Ex. 141] , !H I Ex. 144). ill [Ex. !.ifil.? augmenratior1 of, z 16; closed groui)s, M, 83, ~ f., 87, 128. !.5:1 f ., 205, 1291 2 3 1; closure articulated by, 83-89, 2-!J 98-104, Chapters Vl- \TJII p11ssi111; conformance and, ~ f., 41 f., 6o, !:.12! 182 ; context affecrs, 131, 137; d)'nan1ics affect (see Stress) i ethos and, .u.8 f., 130; gcsrure and, zo7 f .; groupings of, 28, .H.i hierarchic structuring of, 38- 40 [Ex. !11. :H_ f . [Ex. ~ 81- 89 f Ex . .iZ1 JOO [Ex. H1 ~ [Ex. u1 !:!! [Ex. ztl 1 56 [Ex. Z2 & hl !1.2 [Ex. ~ 1.8.l [Ex. ~ ~ LEx. 106],

[Ex. 11RB], !.14 LEx. 131], 236-239

228

[Ex.

13], 256-261 [Ex. 148-150] , 26.J~ [Ex. r53] ; ir11plical'ion and, 117, • 16 f., Chapcers VIl- Vlll passi111; i11cornplctc, an:tpcsts as. 204 n.; incomplel'e, discussed, !.1Z [Ex. 66]. 203- 2o6 L l~x. 10H], !A.! [Ex. 134]; meter an.ti, !:.!.I f., 169, 200; r11obiliry of (see iV1obilic)r); or.her para1nctcrs nonco1tg ruenc \vi1:}1 ( see Para111eters1 11oncongr t1cnce of); perforn1ance and phrasir1g o f, Chapter If passir1l, u8 f., 169 f .; pitch proxi111ity and, 11 £., 84; i1otcntial , related to incomplete, 203 f. ( see also Pote ntial rh)ithms); a primar~' para1netcr, ~ ~ 88i prior organization affects, 13 f.; repetition \\reakens, 5..!.: r 37, 2 57; s1nall differences affect (see Implication, sn1all differences affect); tempo influe nces, 40:41, 83 , u8 f. Rien1ann, Hugo, 41 Robbe-Gril ler, Alai11 1 1 Romantic n1usic. See Ninctecnth-cenl'ury I

•

lllU.'>IC


INDEX OF SU:BJECTS AND NAMES

2 79

Rothko, Tvlark1 66. R ounded binar}' fonn ( :A: :BA: ): comn1on in tonal music, 2£;. discussed, 2i [Ex. !11 f2 fEx. ill ~ 2..! [Ex. 411 ~ [Ex. 66] , i.8n [Ex. 96:-91]; hierarchic srructu.r ing of1 ~ in1plications realized in second part of, 201, 2 I Qi repetition in, and implication, 138; sonata-form is con1plex, 2.! Russell, Bertrand, r6. l.8. Rurf1ardt, Adolf. See Kohler, L.

Schemata: aesthetic experience and, l 13; cadential ( see Gesture, cadential ); con1pletcness and, 1 30; discussion of, 11. [Ex. 41- 43), ti.£ [Ex. 111 ~ [Ex. 87-88). ~

[Ex. ~ ~ [Ex. 7JJ. ~ & !.Q2li

ethos arid use of, 221 , 143 ; gestures as, 207 (se e also Gesture); implicatio11 and, 2 14-217

[Ex.

117), 211- 226

[Ex.

124-

116] , Chapter Vlll passi111; indi\l'idualization of, ~ f.; learning of, 114, 117; n1emory and ( see A-1eo1ory ) ; nature of, z 1 3 f.; norn1alization of, !L z 16 (see also Methods, analytic); originality and, 213 , 2 26; particularizarion of, t8. f., l I 3; style anal)1 Sis concerned \Vith, f., ~ 1.!i 13oi understanding depends on, I.8 f .• ~ Z!i. ~ f., 219; understanding of, atemporal,

z

!!.2. n. Schenker, Heinrich~ 12, !:h ~ ~ Schonberg, A., lb. ~ Schumann, Roben, lli ili ~ !...!I Semicadence. defined, ~ Serial music: conformance i11, ili 66, ~ Z!! constraints of, 76; r11elodic style of, 8 n. Sforzm1do. See Stress Shakespeare, vVm., ~ ~ Shape: strong, 194; \\1cakening of, implicative, 132. 168-169, 201, zos Signs: n1usical patterns as, i.! f., il !b. u l Simon, Herben A., ~ !b Bo, 2..! !b 25 Sin1plicity: coordinate implications create, 121; deceptive in Classical rnusic (see Classical period) Sketches of composers: significance of. ~

77-79; theor}r required to explain,

u

f .,

~ ''·

Srr1ith, Barbara H., 8..I n. Sonata-form: both formal and processive, 2..!...! 21i hierarchic relationships central in, 66; 111otivic constancy in, i4i a rounded binary form, 2.!..t ~ Sonorit)' : etl1os and, 243; relationships clarified b)1 , 131. t44 rzo. 186, 223, !.!5 f. (See also Texture ) Stability: ec ho \Vcakens, 257; elision weake ns, 2 34i and insrabilit)' articulate form and process, Chapters IV, VJ- \' lll passi111; oc tave enhances (see Oc..-cavc) ; psychological need for, ~ uniformity does 11ot create, 2..6 Strauss, Richard, il

Stravinsk)', 6 211 Stress: function of, in tonal n1usic, 35 n.; rhythm affected by, 11, 35-37, !.!.I 100, l.81. ll!.t. ~ a. , z58, 26'; perforn1ance and, 3.L i!.? t 28-r 30i potential actualized by , 11.i te1n po affects, i!.i uB Strophic for111 ( A-A' . . . An): additive and fornlaJ, 93::¢, 2Z.! axial ( see Axial melodies); closure of, weak, 1 56; conforn1ance basis of, 66. 21.i hierarchically flat, ~ !b 2S f.; length of, variable, 2ii occur on lo\'1-levels, ~ !h !!J. f.; ostioaro a kir1d of, 2J ( see also O stinato); therne and variations as, l!.i 4.h .i!i .2!i 25. f.; trended changes in, 2i Structural gap. See Gap, strucrural Strucrural tone, realization of. See Hierarc hic equivalence Struc tural tones, selection of: ad hoc h)rpothescs used in, 64, J 1 r; cases invol\~ng, discussed, Chapters Ill-IV, VI\ TIII; difficult, 12 1; guides used in, 12112 3i musicality needed for, 63 f., 1z1; r1ecessit)r for, !1 f., 63; rules needed for, ~ f ., t 2 I; theory may affect, 18, 61, ~ f., i l l Strtlnk, Oliver, 56 Ila St)rle: baroque ( see Baroque music); characteristic ( see Characteristic style); classical (see Classical period, tnusic of); composer \vorks \vithin, 10; conformant ~



280

St\: le ( co11ti1zued) • relationships and, '[}:_ f .; c.riticisn1 and (see St)Tle anal)rsis); culture a11d, 56-58; explanation and, 1 17; gesture and, 207 f. ; hierarchic relationships and, 66. f.; intencionalit)· and, w learning and, zn., 16, L8 f.,

~ (see

1 38;

of nineteenth-century

nlu.-;ic N inereer1rf1-cenrur,, n1usic ); • paranlctric in1ponance and, 5 5, 89; patterning a11 asp ect of, !.Ii probability and , fi ~ lli relationships en1phasized var>' with, 66 f. ; schemata an aspect of, f...!J f.; S}'lltax a11d, !.12 1 30; theories specific ro, 109; understanding dep ends 011 knO\Ving, Zi.i I 38, 1 06, !.!j . St)•lc anal)·sis: c lassification and (see Classificacion); criticism depends on, 2i ~ tB f.; criticis1n disringuisl1cti f ro111, 6-z, !.4 f .; heuristic, !..2.i norms and schemata descril1ed by, L !..i? LIL lli p robabilities defined b)', Zi schemata and (see Sc hemara); current stare of, z 5, 1 17; statistical 1rlerhods appropriate to, LJ. 1ft !!!l cheor)' discinguished from, f ., 25; concerned w ith typolog)', L !.i! 207 S)'tn n1cu.ical patterns, 174- 195 ; conformance in, IH_ f., 177, I~ f.; foreground

z

n1asks, r 79 ; retrograde relationship in,

U2.: See also l\.xial n1clodies; Cnangingnote melodies; Complemencar)' melodies

construction and confirmation of, 2 2; criticisn1 based on, 2.i !.4.t !.L 113; criticism forced co fie, 18 1 21; of ethetic 1

w

relationships, 246; generality of, hazards of, =4 f., ili of hierarchic srructure, C l1apcer IV passi111 (see also Hierarchic ::;tructures ) ; and history of theory, discinguished, 2 :z; of implicacion, Chapters V- Vl fl passi111 (see also Implicative relationships; l\ilelod)r) ; not inclusi1le, 109; la \\·s de,•cloped by, 1--9, !.! (see also H )1potheses ) ; musicality guides, 18; objective use of, ~ .1...1.6 (see also JVtechodolog)1; Alethods, anal)rric; Structural tones, selection of); refinement of, needed, !:..b 15; relevanee of past, u f.; Schenker's (see Scl1enkcr, Heinrich );

sketches explained by (see Sketches of co111posers); style a11al}'Sis distinguished fro111, 1 f.; under:standit1g not based .on, !.§... '!Ji of 1nusical unity (see Unity) Thon1 so11~ \Vm., Tl n.. Tinlhrc: i1npli.cation and !.12: See also Orc hestration Tirro, Frank, ~ n. Tonal n1usic: fan1iliaricy with, assumed, !1 (see also Learning; Listener ); S)' ntax of, c hanges '\Vith hierarc hic levcJ, fu2 f. To1.:e)1, Donald F., ~ ~

Tradicion.: objecti\'it}' of,

~

po\ver of,

207

T empo: dJ'namics and, !.! (see also Stress); cthc>s and, ~ f., z 2 8-1 30 139, 166, !.iJ f., 267; pace distinguished from, 267; rhythm and (see Rhy chrn ) ; scruccure articulated by change of, 102 1 267 T 'er1ninal e\'cncs: not implicative, r 10. See 1

11/so Gestur e, cadential T ernary forn1s ( .r\-B-A ), 8..t.:. not processi~·e, ~

hl linked,

f.

T exture: contra1)uotal gesture and, structure articu lated b)·, 100. 101, 170, 1.86

Transcendentalisn1: music of, fu views of, l Triadic melodies, 157- 174: complementary, !!E. [Ex. 103]; co11tinuous, 167-170 [Ex. 90), r 70-i72 rEx. cl 209 f. [Ex. 11 r], 116 rEx. 113] ; as gaps, 100 f. rEx. 54), 103, !.41 [Ex. s.Q1 !.!Z [.E x. iQ1 ~ f. [Ex. zz], ~ [Ex. ~ 162 f. [Ex.

1.Q8. f.; !].!

f.,

Thcmacic transformation, 55-59. See also Co nformant relationshi ps Theor}': ad }Joe h}' p<>theses and, 11 -1 4 (see also I-T ypotheses) ; of conformant relationships ( see Reci, theories of );

f. [Ex. hl 1.66 f. [Ex. 88]; high-level, ~ [Ex. 411 !..!Z [Ex. 62] , Lfil f. [ J~x. 85], 163- 165 rEx. 86], 190 f. [£ x. IOI], 194 f. [Ex. 104], 111 f. [Ex. 114A] ; n1iddle-lc\rel, ~ [Ex. cl 1r5, 118, 120 [Ex. i21. 124f. [ Ex. §1..L 158-160 [Ex. hl r.6.I f. [Ex. hl 1.6.6 [Ex. hl !Qi [ Ex. 176-178 [Ex. 25.1 209 [Ex. 111 ] , 210 [Ex. 121], !3_! ~

m


INDEX OF SUBJECTS AND NA1\•1ES

rEx. 130] , 140 rEx. !...li1!il rEx. 143]; low-level, L.8..2. f. [Ex. ~ 1..86 l E..x. 221 z.00 [Ex. ro8], !.!.2_ [Ex. 121] , 111-213 [Ex. 124], ~ [Ex. 115B] , !fl_ f. [Ex. 12s·1, !.42 [Ex. !A:.!.l (see also abo'Ue con1plcn1el1tary; conrir1uous) Triadic patterns: con1ple111e11tary, u..! f., 191 i continuation and fill in1plied b)1, 100-102,

l2j, 1451 147, 2I8, 240, l49i

continuation implied by, 120, 124, 15816o, 161, ~ f.; continuous, 167-170; c111bedded in linear morio11, !.4! f ., I 50; extent of, depends on patterning, 157-161,

Chapter Vll passi11i; harn1ony affects, 121,

240 &;

in1plication and, Chapters

VII- Vlll passi111; linear patterns based on, 168, 170- 171 1 209 f.; linked, 165- 167; norms in tonal music, 157; PotentiaJ1y endless, 169; recapitulation makes n1ani-

fe!t"t, Lz.4: See also Triadic melodies

depends on 111en1clry, 8-0; 1)atterniog necessar)' for, 3-6, 1 1 o, 130, 21 3; probabilit)' a11d, 18. I 16 f. ; p rospective, 18, 111; provisional realjzation and, 1 tZi 216; retrospective, 28, 11 1; lll f., !!l f., 1..6o fh.i schen1ata and (see Schemata) U niforn1it}r: closure requires br eak in, I!_2 f., 1:20, 115, 147,. 162, 1702 ~ f., z36 n., 2 392 266; continuation implied b)•, 136 f., 162, r8r, 198, 260; harmonic a111 biguit)~ and, ~ mornenrum created b)·, 125, 137, 148; precludes patterning, ~ ~ 104

f. Unil)' in music: basis for, 66-70; conforn1ance not sufficient for, 6.6 f.; confom1ance said co create, ~ 64; dialectic \.riew of, ~ f .; definition of, difficult, 66; hierarchic structure and 66 f.; of multi1t1oven1ent works, 68-70; various relationships create, 66; sryle and, 66, 6.8 f ., value and, ~

Understanding: i11volves abstraction, l f.;

of alternati\•es,

~;

complexity of, t 13; context affects, 21 (see also Conte>..'tual d_iscrcpancy); criticisln depends on, i zf., I 16; education enhances, !_1j 111 -11

explanation distinguished from , I 5-17; hierarchic structure facilitates, So; implications, 110-113; in1plication changes, knowledge, but experience, basis for, fb 15-17, 1..5..i 138, 107, 213; 1r 1;

11ot

z

\ i\fagner, R., 12, iL iL ~ I 10 ' W cakening of shape. See Shape \.-Veber11, A., 8 ,! b ~ 109; views of, ~ f. \.Vell-Ten1pered Cla...rier, 21 \Vllite}lead, A. N., ~ Win1satt, W . K., 13 n. \Volfe, Ton1, 1 \Vords\vorth, Wrn., J


Index o Music -------------.---~--------------

Bacl1, J. S.: Art of che Fugue, 91. n. B i\ li11or l\ lass, "Qui sedes ad dcxrra1n patris": ~ !Ex. ~ 3 B ] Parrica No. l fo r H arpsichord, Burlesca: 196- 198 [Ex. 105] Partita No. i. for Harpsichord, Giguc: 1 2 0 fEx. 113) Prelude and Fugue in G. i\l inor for Organ 5-f2 ): !..±!. n. \Vell-T cn1pered Clc1\· icr, Rook I Fugue in c: ,l\1ajor: ~ f., ~ n. [Ex. ~ r:ugue in Cl J\ linor: !..2..! f. [Ex. 102] Fugue in o:; l\1l inor: .u:u f. LEx. ill f.'uguc ir1 Jo' i\ l inor: 21 f., 98, ~IL

(B'''' '

[ Ex. ~

Prelude in C i\ lajor: 2! f., ~ L£.x. 128A] Prelude in C l\linor: 2.! f. ':\'ell-Tempered Clavier, Book l'I Fugue in C l\ lajor: lilB f. [Ex. 11 0A] Fugue in .D l\1int1r: r48- 150, 151, 174, ~ LEx.

nJ

f.'ugue in. F i\1inor: 98- 1001 ~ ~ n... [Ex. u.1 Bartc>k, B.: String Quartet "No. SJ V ~: l !i f., 1 i.8 f ., 110 f., 160 n. LEx. i2J

Sering Quartet No. 6: 57 Beechove11, L . van: Piar10 Sonata, Opus ~ {uPathetiquc"): 68 P iano Sonata, Opus 81a ( "Les Adieux"), I : ~ ZJ f. I Ex. fill Chapter VIll [Ex. 136-155 ) Sonata for Violin and .P iano, Opus !:ii

111 : ~ f. [Ex. 63) Sering Quartet, Opu~ LB Nu. !..i l : z!! n. String Quartet, Opus i2 Ne). h 111: ~ String Quartet, Opus 127, I : 100-10 1, ~ LEx.

.HJ

String Quartet, Opus 130, 1: 267; II: 81 filh ~ f., ~ f., ~ LEx. fZ1 V: !..!1 f.

[Ex. 1198] Sering Quartet, Opus: 1 31, I: Ve f. [Ex. 42A]

String Quartet, Opus

132,

I:

v

f. [Ex.

~

f. [Ex.

-JiBJ String Quartet, Opus 133:

.Ji Cl S)•n1pho11)' No. Jj 69; 1: 205, ~ f. [Ex. 128RJ; II: li f. LEx. 31 ] ; III : ~ n., ll f. I.Ex. 45- 46] S)•n1phon)r ~·o. :b I\T: 218- 126, 14on., [Ex. 120, 11-1- 126] S)rllif>hony No. i! l & Ill : ~ f. [Ex. 1 s]; I\T: 3i ff. I Ex. uJ

• Ron1an nun1crals designate mCJ\remc11ts


INDEX OF MUSIC

Syn1phor1y No. 6 ("Pastoraln): ~ Syn1phony No. 7.J I: 142- 144 I.Ho [Ex. 70-74] Syn1phon)r No. ~ l.; H Sy n1phon)r No. 2..: 5L. ~ Berlioz, H....:. Syn1phonic Fantastique: 2Q. Borodi11, A.: String Quartet in D 7\1ajor, Ill: 112 f. [ Ex. ~

Brah1ns, J.: Quinte t for Piano and Strings, Opus ill I: 208 Sonata for \ 1 iolin an.
2tli IV:

1.! (Ex. ~A] Sy1nphony i o. !., I: 5g-63, 70-71 32- 37 & 19:40]; III: 19. [Ex. IV: 1.£ f., ll [Ex. l2J S)'t111>hony No. l! JV: Z! [Ex. 4 r BJ S]rmphony No. !! I: 170-174, ill 91=91]; II: 2! I Ex. fil lV: I!

[Ex. 40];

[Ex. [Ex.

Bruck ner, A .: S)·n1pho11y No. 'b Ill: 205 f. [Ex. 108] Chopin, F.: Prelude in E iv1inor, Opus 1B No. i: ~

75B]

Prel1tde d l'Apres-1\tlidi d )1111 Fa111ze, 10.B Dvorak, A .: Quintet for Piano and Strings, Opus ~ II: ~ f., I 20 ~ [Ex. .uJ Syn1phony J\~o . .i ("Ne\.v W orld"), IV : 184-187, tB.8 [Ex. 22] Folk-songs: 44- 46 [E.x. ~ ~ [Ex. 26] Franck, C.: Quintet for Piano and Strings, I : 18719 c, ~ !!:..i 1 04 n. [Ex. too-101] Sonata in A J\i1ajor for Violin and Piano, IV : 2Z n. Gemin iani, F.: Concerto Grosso in E J\•l inor 1 Opus l No. 1.t 1: W f . [Ex. 75 A] 1

Handel, G . F.: Concerto Grosso in G ~linor, Opus c2 No. 6, J[ : 140 f., ~ [Ex. 68]

166-167,

Haydn, F. J.: String Qua.net in Eb ~fajor, Opus ~ N o. h I: 112 [Ex. 6o] ; I\T: 240 f.

[Ex.

u.iJ

String Q uarter in Bb Major , Opus il No. 3, I: 181-183 [Ex. 2fil String Quartet in Bb i\1ajor, Opus ~ No. ~ l : 119 [Ex. 1..1J A] , u.6 (Ex.

!fl] S}'mphotl}' N o. 6a (" il Distratto" ): 68 S)•111pho11y No. 2i ("Surprise" ) , I: iZ £., 68 [ E x. !i] S)7 11lphony No. 21 in c rvta jor' I: 212 f .• ~

[Ex.

115- 116]; II: 163-165, 204 n. [Ex. 86] S)·inphony J\'o. 100 ("Military"), IV :

n n.

S)·mphon~'

41 CJ

~ [Ex. D ebUSS)', C.:

lvf essitrf,, Hallelujah Chorus: !:14 [Ex. 87)

No. w1 ("The Qock"), I :

E f. rEx. ~ Sy mphon) No. Ull in Bb l\·1 ajor, IV: !.2 S}·t11plloO}' N o . 104 ("London"), lll: t rs, 152, ~ [Ex. ~ & 106] ; IV : 20 1-201 [Ex. !.QZ] J\1al1ler, G .: S>rn1phony J\ro. it l; t.66 f., :u8 [Ex. ~ 1

J\•l ozart, \ \' . A .: Divcrtin1ento in Bb Major (K. 287. new 217b) , I: 131-133, 140~ !14 [Ex. 6465] L e N ozze di Figaro, "Se vuol ballare": 150-152, ~ l.6o [Ex. zfil Piano Sonata in C A1ajor ( K. 279), I : 120, ll1

LEx.

122]

Piano Sonata in A l\•l ajor ( K. 331), I : C hapter ll, M.i 2i f ., ~ [Ex. 5-10, l l & 14- 19) So11ata fo r Violin and Piano in E ~·1 inor ( K. 304), 1: .tQ.8 Sonata for Violin ac1d Piano in A Major ( K . 30s ) , 1: 16z=1 70 [ Ex. ~ Flute Quarter in A /\1ajor (K . 298), Ill:

!.il rEx. i§.1 Oboe Quartet in F l\1ajor ( K. 307 ), I : 191-196 [Ex. 103-104] 103.

String Quartet in D 1\i1i11or ( K. 42 1) ,

111: ~ f. [Ex. 76]


INDEX OF MUSIC

String Quartet in Bb 1\1:ajor ( K . 485 ) ,

III: !!2 [Ex. 119Al String Quartet in A J\!aj or ( l{. 464) , II: r78-181 [Ex. ¢--97] String Quartet in D 1\1ajor ( K. 575 ) , II: 212, z30-235, u..z [Ex. 130-132 ] Quintet for Clarinet a11d Strings ( K. 58r ), IV: ~ f., 50-52 [ Ex. !.l & 27-

W ''Haffner" S}1111phorl)' in D i\lajor ( K. 385 ) , II: t67, 112 [Ex. 89] "Linz" S)rniph.on}' in C 1\1a jor ( K. 42 5), I: 176-178, 180, ~ n. [Ex. 2.i] S)·mphon)' No. 12 in Eb ~·lajor ( K. 543) 1 IV: 10-12. !.il !.2 f., ill f. [Ex. h 4 & 135] S}·n1phon}' No. ~ in G i\ f inor ( K. 550) , Ill: 2Q Purcell, Ii; Dido trfld A ene.-is, 2Z Ravel, J\•I.: Rl1apsodie Espagnole; 23 f. [Ex. 5 J] Reger, J\•1.: \ Tariarions aod Ft1gue on a Theme of Ailozart: £ [Ex. 10] Schubert, F.:

~0. 1 2 1 : 111 2 24on. [ Ex. 124B]

String Quartet in Eb ;\t1ajor, Opus 115, II: 134-139, r45 n. 1 !.iZ.t 174 1 196 n. (E,x.. ~ Trio for Pi~no and St:rings, Opus 221 fl : 228-230 [Ex. !1.2] Sct1un1ann, R.: Alb11111 for tl;e Y 01,ng, "Soldier's Marchu: 121 , 125-130, 140, !..!.i !Ex. ~ Quartet for Piano and Strings, Opus .4L I \ T: 108-212, !..!..4 [Ex. 109 & 111-114] Quintet for Piano a11d Strings, Opus ~ ll: i i f. [Ex. 1.!J Sn1etana, F.:. Vlta'1...·n (The 1\1Ioldau ): L.6a f., 166, u8 [Ex. ~

Sr.rauss, R.: Till E11lrmspiegel: 123, 16r-163, 174, [Ex.

W

Stravinsk}1 ,

~

L.:

O ctet for "'' inds, 1: 209 [ Ex. I roB] Petroz1cbka: 123-125 rEx. 61 ] Telemann, G. P.: Suite in A i\•1 inor for Flute and Strings,

11: 158-160 [Ex.

W

O ctet for Strings and \iVinds, Opus I 66, l\T: 219 [Ex. 121B]

T chaiko,\•sk\•, , P. L

Die Scho11e Aifiilleri1z, "Oas \i\!andcm": 152-157, 204 !b 149n. [Ex. 7.9.. & 81-

Wagner, R.: Die 1\leistersinger, Prelude: 68 [Ex. 38] Trista1i u11d Isolde, Prelude: 135- 139 [Ex. 133]

W Sonata for Violin and Pia110 1 Opus !.11

Symphony No. il l\' : 'E [Ex. .+3A]


•

f'

cll.lj)tcr cc>11ccrns the significance of r110ti\ ic si111ilarirics and, more s1Jccificall)·.

rl1c 'iL '' J)<>i1lr <>f critics such as Rt1lf'h llcri. J he f1,t1rrh cl1ar>rer cc>11sidcr<; the 11arurc c>f hicrarcl1ic srruccures-rl1c ,,.rn1 a11<;urc, i11rcracr. ·1 he \CC<> n<.I, an<.I n1c>sr a111biric>t1s, })ttrt <>f the L><><>l.- accc111prs ro establish a fc>un tlacici11 f<>r criticis111 <>f t<)nal n1clc>tlic: rclaric>nships as f)roccsscs i11 \\ hicl1 earlier ~Jatccr11i1lgs in11)ly 1>r<>bal>lc ccn1rinuari1 >11s a11d pc>ssiblc l<:i11ds of clc>surc. f 111plicat ic>11s 111a)· IJc realized at ()llCC Clr lll,1)' l1c tlcla)·cd: the (JrC)CCSSCS itllJ>licir i11, Sa)', rhc <>pc11i11g niclotl}· of a "ic>naca-fc>rr11 111<>\ c111e11t 11ta\· nor be saris• fact11ril) fc>r111t1larccl t111ril rl1c rcca1>ul,1ric>1l <,r c\·c11 chc coda. ·rhc cliscussi<>Il~, copiot1sl~· illt1srratclJ, a~~t1111c sc>n1c fan1iliarit}' '' irl1 111l1~it·al 11<>raric>n. Gi,·c11 such fan1iliarit)', a 'vicJc r.111gc <>f rc:1der4' interested i11 111t1sic, ac~1 l1cc ics, <>r criricis111 sl1oulc.l fi 11LI ,\ 1r. \ ICj'<:r's itlcas a11
I ~•·<>~ 1\Rt> l~. \ I L\ ' Lil, l)1r\·t.1.I!-i f' ~) l l~ 11 r1>fCS'i<>r i11 the 1-I u111a11irics at

fl1c Uiil\ cr~if~- c>f Ckicago, is cl1c aut-l1or TIJe l?l1J·1l1111ic .~'tr11£·t11re of \l L1sic (\\'itl1 (; , \ \ '. c:()() ~>cr), a11d .\111sic, t!Je .4rts, t17J cf c>f

£111ot1011 tllltf

f tie!11f.

.ll c:1111i11_r4

i11 ,\ J11sic,

•

S~l~J

,

\T CC)\ll)()SJ J J()

/,

I

I

I

I

\ f 11

I

I

.,

lJ

} ~

1

,11 J f

Ir l

r 11

J3,r . GEORGE IJERLE " I l1c 111ajc)r 111c>11cJgra1>h t) n a pcrci 11c11r subjccr . . . highl)r rec1J111111c11Je all seric>us 111usic sruJc nts. '-C: l1oice 1

"Cl e~ rl )'

'''rirrc11 a11d C<>j)i<>t1sl)' illuscrarcJ ,,·ich pertit1c11c n1usical cxa111~>l cs, rhi.s i~ tile lJc.sr gc.:11cral intr<>ducti<>11 t<> serial c1..>Jllpc>siri<1n <.:u rren rl_,. a\'ai !able.''-. \111.ric f r111r11al Thud rt:' iscJ a11d c11largcd cJi uo11 $9. 5 o

I~ \

I)I I

\

I)

I

I I I. C) PI ll \ SI Rl \

13)' \'' I 1"'0N DE:\ N cxtra<>r111 rhac c>nl)· lc>11g, '' ilic fa111iliarit)' ,,·irh 1 Ia11ticl's opera~ cf .such a pcrfor111ancc, <>nc \\·a11ts rc> ~hl1ut '/3ra1.Jo!·" "~\ 11

•

--S11t11t£iay

Re1.'iew

"1~11c fLrst

con1prchc11Si\•c srupcras. . . . ,. \ ''ro k.." -1\ 1o t t.'.\'

I~ I / /1 I

J \ \ 11 f) t.

I~ Il I

l II

•

By Ell fC \\'t\J ~T l~R \ \ rf-llTE :\ Ir. \ \ 'hire's IJuol\ has lo ng l.>c<.:11 regc. r aJ1d 111ost cr>n1i1rcl1c11sivc ~r1tU}' c1f 13enja111 i11 fJri rre11. I le l1as ral i11cludc anal)'Scs c>f ;111 of J3ritrcn's 111ajor <>r>cras as ,,·ell ~ discussit>ns c>f Brirte11'.s n1<>.sr recent n1u~ical-rl
rakes irs place as rhc 111osr cc1n1 prcl1c11si \' C of pu blicatio11s <>11 [he life a11d \\"<>rl{s c>f the grear E.nglish co111poser.''

-Opera Ca·1Ul£ia R c,·isc(I a11n $ r o.oo

T T

Bcr){clc)' 94720

MEYER, Leonard B. - Explaining Music - Essays and explorations.pdf

Recommend Documents