Você está na página 1de 16
Toward a Theory of Timbre: Verbal Timbre and Musical Line in Purcell, Sessions, and Stravinsky

Toward a Theory of Timbre: Verbal Timbre and Musical Line in Purcell, Sessions, and Stravinsky Author(s): Robert Cogan Source: Perspectives of New Music, Vol. 8, No. 1 (Autumn - Winter, 1969), pp. 75-81 Published by: Perspectives of New Music

Stable URL: http://www.jstor.org/stable/832123 .

Accessed: 06/12/2013 04:29

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

.

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

.

information about JSTOR, please contact support@jstor.org. . Perspectives of New Music is collaborating with JSTOR to

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

http://www.jstor.org

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

TOWARD A THEORY

OF TIMBRE:

VERBAL TIMBRE AND MUSICAL LINE

IN PURCELL, SESSIONS, AND STRAVINSKY

ROBERT COGAN

TIMBRE,

of all the parameters

every

of music, is theone leastconsidered.

It lacks not only an

obscurity is in part notational, in partanalytical -

historical.While

fixing and notating the timbralfeaturesofa

cent (1750 is a convenient date, as a norm), the peculiarity of timbral

notation has not been

recognized. timbres, butratherthemeans- particularly theinstrumentalmeans-

adequate theory, but even an inadequate one. Its

musician

and in each

respect

knowsthatthe compositional act of

workis comparatively re-

Its usual notationindicates not

analogous

to luteand

instrumental fingerpositions. Can

in

manyrespects

relationsifall

pitch

used to achieve them. (This is

guitar notationof pitches based on

one

notationwere based on

the

followedcen-

turiesafter pitch and rhythm notation; ithas also remainedone step

furtherremoved fromitsessential

than those other notations (however questionable and inadequate theymaybe).

imagine

the

problems forthe theory of pitch

instrumental manipulation, ratherthan on

Thus timbralnotationnot

only

territory -

the natureof timbre -

resultingsounds?)

Timbral

analysis

is even more recent than timbralnotation.The

Quality

V

Chapters of Musical Tones" and

subsequent theoreticalcon-

by

essentialformulationsof Helmholtz date from1850-80.

and VI ("On the Differencesin the

"On the Apprehension of Qualities of Tone") of hisOn theSensations

of Tone have

siderationof timbre.These later studies have been undertaken

acousticians, engineers,linguists,physicists, and psychologists(among

others); rarely, if ever, by musicians. To be done with accuracy,

timbral analysisrequiressophisticatedtechnology of a kindwhichhas

existed forless than

remarkable for their period, contain inaccuracies due to technical

limitations. Helmholtz showed that timbre (which he called

dered as qualityby his translator)depends principallyupon the num-

provided the basis for all

fiftyyears.

Even the observationsof Helmholtz,

Klangfarbe, ren-

. 75-

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

PERSPECTIVES

OF NEW MUSIC

partials of a fundamental.

Thus timbral analysisrequires measurementand considerationofthe

spectrum of a sound. This is a matter requiring both tech-

overtone

nical and conceptual delicacy. In fact, the crude, specious use of

the overtone phenomena

Hindemith, by pure

admirable work), and by

more difficultthe proper use of these phenomena to illuminate

timbre. (Helmholtz

danger of inappropriate

generalization from the

insufficient, pains

ber and relative intensity of the

sounding

by composer-theorists from Rameau

to

theoristssuch as Schenker (despite much other

engineers

such as Olson has in itselfmade

was highly aware of the

overtone series and

took considerable,

if

to avoid it in his own work.)

significantly illumi-

a theory of timbreas a whole. It

does intendto showthat using timbralinformationnow (and forsome

years) available, certain aspects

nated-that it is possible to make a beginning toward the musical

analysis of timbre.

This brief

paper will not supply

of music can be

I

Languages

are timbral systems of considerable complexity and

and some vowel-like consonants). The

Qualities of Tone"; even

Bell). Since language and its tim-

language

interrelationships between timbre and

subtlety.They consist of a variety of attacks (consonants) and su-

stained timbres (vowels

acoustical nature of language sounds has been studied more thor-

of timbres.A section of Helmholtz'

during histime

workingintensively in thisfield (including

Chapter

oughly than any other group

V is named "Vowel

a number of otherswere

Melville Bell, father of Alexander Graham

timbrehas been so greatly studied, and since

bres are components of many musical works, it is a useful starting

place for consideration of

other musical features.

The basis of the

timbral analyses in this paper are the

spectro-

graphic analyses of R. K. Potter, G. A. Kopp, and H. G. Kopp in

Visible

Example I reproduces spectrographicanalyses

English. (The analyses

partials of each vowel.Potterand the

widthof 300

cycles, details, but shows most clearly the

foreach vowel. Such analysis was done in the

3,500

between70 and

Speech (Dover Publications,

were obtained

originally published in 1947).

of the vowels of

by filtering and measuring

the

used a filterwitha band-

Kopps

whichresultsin the

suppression of some minute

concentrationsof vibratingpartials

range

bulk of vibration energy and variation

in structureare concentrated.) It is the second bar fromthe bottom in each spectrograph whichmost strongly characterizesvowel sounds,

due to its strength and variability of position. Potter and Kopps

cycles, where the great

S76

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

TOWARD A THEORY

OF TIMBRE

: 9 eve ~ t at ak oot .t. fath.r b.y. b.t
:
9
eve ~
t
at
ak
oot
.t.
fath.r
b.y. b.t

Schematic Representation of the Spectrographs

!i~i~iiiii~i ii r (it) a (hate) ? (met) to(at) an(s) i eve) . . .
!i~i~iiiii~i
ii
r (it)
a (hate)
? (met)
to(at)
an(s)
i eve)
.
.
.
.
.
?
:!~ ii~i~~iiiiiiii~ll
,
a(fa
ther(
)e)
(put)
i
iboot
4
a
4 (u)
a(bout)
(Genera (church) A
a) (lsltern) (chureh)

77 -

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

PERSPECTIVES

OF NEW MUSIC

referto the concentrationof partialsrepresentedby thatbar as the hub. It is clearfromthese analyses thateven whenthefundamental pitch

of vowelsis identical (as in these spectrographs), vibrationsare set up

which

upon the

emphasizes different partials and different regions

hub of each of thevariousvowels- their

is different. (An analogous situationobtainsforvarious

of course.) While

be modifiedsomewhatin variouscon-

texts, it does unvoiced, as nunciation

range, or pro-

vary in intensity in different frequencyregions, depending

specific vowel. The

overtone spectrum

of each vowel of partials. The

point

of

greatestintensity -

instruments,

a vowel

may not lose itsindividualstructureand identity, even when

in a whisper (nor do differencesof sex,

affectthese partial-concentrationpatternssignificantly).

arranged, then,

in a scale

according

to the relative

referto this, as Helmholtz did, as a

of brightness (the term

is irrelevantfor our purposes; it is the

whichis important).(See thevowelorder under "motionof

2, 3, and 4.)

The vowelscan be

height of their hub; one might

scale

ordering

the vowel hubs" in Exx.

Only the diphthongspresentambiguities; their patterns are chang-

any

anydiphthong is as de-

othervowel. They are no

ing ratherthan fixed.However, the change in

finedand constantas the structureof

more unanalyzable or arbitrary than othertimbres.

II

usage by composers, text settings for voices? One senses thatthe answer

is yes, yet this intuitionis never tested. Musicians have analyzed

for example, in

Does the nature of vowel timbreaffecttheir

textual settings in a variety of other ways. Reese showed that

rian chant

example he chose, "Jubilate Deo"; Music in theMiddle Ages,p.

It has oftenbeen suggested thatwords of semantic importance are

placed prominently, at high, or occasionally low, points. Symbolic analysis is not unknown.Of these only the firsthas to do withthe na-

ture of

structureof words ratherthan theirtimbralstructure.

even that has to do with the

Grego-

the

frequently risesforan accented syllable(particularly in

166).

sound, yet

rhythmic-dynamic

It is perhaps ironic that in the art of "sound and time," in Stra-

to sound. For the most

analysis" ofvocal music,

vinsky'sphrase, so littlenoticehas been given

part

the

tion does not exist.

one has been contentwith "pure musical

purityderiving in part froman assumption

thatthetimbral ques-

Examples 2, 3, and 4 constitutethebulkofthis

study.They present

of "When I Am Laid in Earth" fromPur-

phrase-by-phraseanalyses cell'sDido andAeneas,portions ofSessions'"On theBeach at Fontana,"

- 78.

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

TOWARD A THEORY

OF TIMBRE

and Stravinsky's "Full Fadom Five" fromthe

Shakespeare. The analyses consistin each case of the vowel

the texts, fromwhichis extracted the line of hubs. This line of hubs

represents the essentialacousticalmotionof thetext per se, a motion

which, according to the successionof vowels,

ous

vide

crucial.) The line of hubs is then

the voice

are detailed in

Examples

only

progressesthrough vari-

Three Songsof

William

spectra of

frequencyregions,high

only

and low, as charted. (The consonants pro-

less

brief interruptions and articulations; as such they are

compared

commentary in the

withthelinearmotionof

melody; particularlyinterestingaspects of this comparison

graphs. The following serves

boxes below the linear

2-4 should be considered in detail.

as a summary of whatis revealed in them.

PURCELL.

The

phraseapexes

ofits risingprincipal linecoincidewith

of

"laid" and "may," the "i" 's

motionsof the melodic line into

vowels:

high-spectrum(bright) vowels: the "e"

of "re-(member) me." Descending

theinnervoice lead intoor

the "3" of "earth"; the "6" of "wrongs";

ble"; the "a" of "ah"; the "o" of

piece is on the prolonged,

spectrum in English.

timbral-linear correspondences: the

"laid" (m. 3) and the later

"re-(member) me" are crucial in

timbral-linear structure of the piece.

voice descent, mm.6-8 ofthevoice

tion of the low-spectrum "no trou-ble," just

descent depends upon the repetition(and prolongation!)

spectrum "ah for-(get)."

coincidewith low-spectrum(dark)

the "A" and "a" of "trou-

apex of the entire

"me," the highest vowel

Textual repetitions reiterateand intensify these

"for-(get)." The

stressed "i" of

repetition of the high-spectrum

multiplerepetitions of the high-spectrum

creating the apex of

principal

Likewise, the contrasting inner-

the

melody,depends upon the repeti-

as the

concluding

linear

of thelow-

verbal timbral design is characterized by prevailing

high-spectrum sounds contrastedwithan accented end-timbre (in 11.

2

structureof the melodicline

tailed correspondences

withthe

ting

the detailsof timbreand line

omitting unstressed syllables,

SESSIONS.

The

and 3) of marked low-spectrumquality:"groan" and "stone." The

is parallel.

In addition, innumerablede-

are to be found: the linear

apex coinciding

the set-

highest

vowel

spectrum, the "i" 's of "sen-(ile)sea";

fact,

of "numbers each," etc. In

agree almost entirely.

STRAVINSKY.

The line

reproduces not onlymanyspecific details

general angularity.Particularly

usually

taken

of the timbralcontour, but also its

interesting is the treatmentof the word "his." What is

as

to be an exact correspondence

of the diphthong

the beginning of the diphthong, he considers it high; when it de-

"ai" is interesting: when the verbaltimbrerisesto

Stravinsky's mannered and idiosyncraticway witha textturnsout

of setting withtimbre.His treatment

. 79

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

PERSPECTIVES

OF NEW MUSIC

scends to its beginning,

its

the natureof the textand

he considers it low. This

exaggeration

of

ambiguity

adds

further angularity(completely

its setting).

given

to

consistentwith

Special attentionmust be

the last fragment,"ding-dong onomatopoeia-is everything;

musical

bell." Here the timbre-the verbal

meaning (in the semantic sense) and syntaxdisappear. The

setting is a setting of

semantic-syntacticaspects. The

cise timbral correspondence. But thereis yet

ence. Example

almost pure verbal timbre, drained of itsusual

melodic design, indeed, shows pre-

a further correspond-

In each

spectrographs of "ing" and "ong."

A shows

ing ong
ing
ong

Ex. A

case

of its

strikingly with Stravinsky'ssetting:

the ending "if" (ng)

preceding vowel,

has the effectof

continuing the structure

butin greatly diminished intensity. This agrees

III

ding

dong

have not been presented to settleand close theissue

Examples 2-4 of timbreand its

to

numberof

large

be viewedin this way. The

portantenough

and research.

questioning It is obviousthateven whereverbaltimbreis an

the determinationof the totalmusicalstructure (as it seems to be in

important factorin

relationship to othermusical parameters, butrather

open it. These examples have not been specially selected froma

samples; they are, rather,

almostthe firstworksto

correspondences seem consistentand im-

to indicatea possible fruitfuldirectionof intensive

these examples), itis nota sole determinant.There willbe cases where

timbreis slightly relevant, and those where color derivesfrom

posing the purely

tered forces," wrote

(as another paper will

Debussy, one of the great timbral composers

timbral implications. "Musicis thesumoftotalscat-

op-

show). The role of

analysis is technically

(and accurately) to reckonthat sum; thiscan only be achieved by the

. 80.

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

TOWARD A THEORY

OF TIMBRE

the timbral.Withthe

innovationsof

through the

sound barrier.It is leftfor theory to follow. The organization of verbal timbreadds a thirddimensionto lan-

tronic

Debussy, SchoenbergianKlangfarbenmelodie, and elec-

precise appraisal

of all of the forces,

including

freedom-of-synthesis,composition

structure, already

has moved

defined in termsof semantic meaning and

questions

structuring of timbre, and the relationships

both language

analysis (and

stage,immenselysuggestive.

guage

verbal rhythm. The aestheticuses of these timbral

hardly

a point

raised (the

between timbralstructureand other dimensionsof

and music) point to a new crucial and deep level of

composition!), one whose implications and consequences are hardly

foreseeable, yet, even at this

New England Conservatory of Music

possibilities are

known. Even more than with rhythm, the timbreof words is

where language and musicbecome almostone. The

scaling

and

.81

.

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

Vowel spectra:

Vowels: (

e

Text: {

When

i-

A5-

a-

2

I

j"

U-

2-

G-

F- F-

Wa

vl_-

al

I

(The sectionalrepetitionsareomitted.)

When

I

am

a

e

a

e

am

laid,

am

laid

laid,

Thehigheststressedvoweloftheverbalphraseinte s

ofspectraandhub,the"e" of "laid", coincideswiththe

stressedapexes,B6andC,ofthemusicalphrase.

co,,

am

laid

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

Vowel spectra:

Vowels: (

Text: (

e

-

may

a0

my

wrongs

i

cre

0-

U-

U-

S

SD

G-

F-

-

C-

E F-

"0

G -

r-

Thenewmelodic apex,D, coincides

withthereturnto

vowel quality, the "e" of "may".

Melodicdescentcoincideswithlow-

high-spectrum spectrumvowel, "o" of "wrong".

may

may

mywrongs

my wrongs -

cre- ate

cre-

ate

 

e

o

A

a

-

ate

no

trou

-

ble,

The repetition of"not of low-spectrumvowels largely abandonsits upp apexes BI-C-Dfora low terminology)region-A magnifies thevowel con

ous high-spectrum) andl

linevs.inner-voice line.

no

no

trou

trou

ble,

ble,

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

Vowel spectra:

Vowels: (

Text: (

R'

QG

0

,a

"'

E:

o *

-

a-

U u- -

D

-

F-

c-1

C

-

F-

i

Re

-

mmB

mem - ber

i

me,

i

E

Re - mem-ber

"Rememberme"repeatsthehighestvowel timbre,"i", at

end.Thishightimbrecoincideswiththeabruptreturnto theupper-lineD.

itsbeginningand

The repeatedD'sparalleland intensifytherepeated"i's".

Re - mem

-

berme,

Re -

mem

-

ber me,

i

me,

Againaninne

to low-spect

But

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

Vowel spectra:

"

0

Vowels: (

Text: (

-

ai

my

A- D-

G-

F-

0 F

G4

my

fate,

e

fate,

i

e

a

Re - mem- ber

i

me,

The high-point oftheentiremelodic

line,G, coincideswiththestressof

the highestpossible vowel spectrum, "i" ofme".

Re - mem -

ber me,

A

but

a

ah!

The finalfilled-inoctavedescentis largelyjoined tothe prolonged low- spectrum vowels "A,a,o".

but

ah!

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

mm. 1-11

Vowel spectra: 1

Vowels: {

Text: (

i-

C-

A -

G- 0-

U-

G-

oF-

Dl-

Ce

-

I

Wind

aI

whines

and

a

whines

aa

the

I

shin -

gle,

The single vowel "I" of "w'ind" and "shin-(gle)", and the pitch C join

pitch-timbralplateau at both ends ofthe phrase. The mid-phrasespectrum descent is paralleled in a generalizedway.

to fixa

Wind

Wind

whines

whines

and

and

whines

whines

the

shin

the shin t

ogle,

a

The

e

cra

The abruptdrop of vow the melodic line. Not on match;many details-for vowels"e" ("crazy") and

The

cra

-

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

Vowel spectra:

0

o

Vowels:{

Text: (

i-

a

A

0

aG-

-

U

G-

u -

G-

F-

c

6

i

se

-

a

nile

i

sea

a

3

i

num - bers each

E6-

C,-

Abruptrisetodouble-stressofthehighestvowel-

spectrum, "i"

is paralleledby thewide leap to the melodyhigh-

point, Gb-F, whichis

soundings of"i".

of"each"attheendofthe measure.)

of "se-(nile) sea".Thevowelaccent

maintained throughout the

(Thehigh F returnswiththe"i"

A

r. ----2,

se

-

nile

sea_

num-bers each

sin

I

sin -

-

a

gle

gle

aI

slime

slime -

 

I

-

sil

vered

 

This ph

thispa

point;i

ofthel

high-s

sil

-

vered-

 

sto

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

m.1

Vowelspectra:

_

Vowels: {

Text: {

v

Full

a

fa - dom

"

S

1)-

a-

u-

C-Db

A

Ab-

FZ.

E6-

aI

five

a

thy

a

Fa

 

a

a

-

ther

lies,

The parallelisms ofvowel spectrum andmelodic line aremostexact

if one

Fa-(ther),lies". The dipthong"al" is ambiguous-itssetting in

corresponds to its higherhalf, in

"five"

regards thevowelsoftheaccented syllables:"fa-(dom),five,

"lies"to itslowerhalf.

ull fI - om fi

Full

fa - dom five

th

thy

F,

Fa

-

te

ther

ls

lies,

mm.2

i"

Of

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions

Vowel spectra: i

Vowels:

Text: {

o

Those

o CA-

0s0

a-

a-

U-

DI,-

.4 FF

Those

a

are

are

pearles

a

pearles

a

a

I

that

were

his

that

were

his

eies,

eies,

This content downloaded from 78.104.70.246 on Fri, 6 Dec 2013 04:29:37 AM All use subject to JSTOR Terms and Conditions