Professional Documents
Culture Documents
REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/833096?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
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.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
Perspectives of New Music is collaborating with JSTOR to digitize, preserve and extend access to
Perspectives of New Music
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
COMPOSIT'IONAL CONTROL OF
PHONE'lIC/NONPHONEI'IC
PERCEPTION
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
Phonetic/Nonphonetic Perception 139
Marinetti, early sound poets such as Hugo Ball and Kurr Schwitters, Lettrists
such as Maurice Lemaitre, and many others. With the advent of musique con-
crete and electronic music, composers such as Stockhausen (e.g. Gesang der
Junglinge), Herbert Eimert (e.g. Epitaph fir Aikichi Kuboyama), and Luciano
Berio (e.g. Thenma-OmiggiaJoyce) began using recorded and processed speech
sounds (along with nonspeech materials) to focus listeners' attention on the
sounds of speech. Composers such as Luciano Berio (e.g. Circles), Gyorgy Ligeti
(e.g. Nouvelles Aventures), Kenneth Gaburo (e.g. Maledetto), and many others
wrote purely acoustic compositions to the same effect. More recently, com-
posers of computer music have brought digital technology to bear on the prob-
lem. These works include Charles Dodge's Speech Songs, Tod Machover's Soft
Morning, City!, John Chowning's Phone, Paul Lansky's Six Fantasies on a Poem by
Thomas Campion, and many others.
A variety of perceptual ambiguities can be found in these pieces. These
include ambiguities concerning:
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
140 Perspectives of New Music
VOICE-LIKENESS
The most heavily processed sections of Eimert's Epitaph fir Aikichi Kuboyama
convey the rhythms, rates, and directions offormant change which we associate
with speech. It is even possible in these sections to identify many syllable bound-
aries and even some classes of phones (sonic representation of phonemes). It is
possible, for example, to distinguish stops, as a class, from fricatives and vowels
as classes. And yet, while it is clear that these heavily processed sounds are voice-
like (or "speech-like"), no specific and intelligible phonetic message is conveyed.
We can generalize the distinction: "Voice-likeness" involves source recognition
and description; a sound can be voice-like without conveying a specific phonetic
message.
A compositionally useful definition of "voice-like" aspects of a sound might
be: those aspects which cue listeners' association with the human voice in a given
context. The subjective nature of the definition reflects the speculative nature of
composing for a diverse audience. It is more useful for the analysis of perceptual
effects to define "voice-likeness" in terms oflisteners' associations than in terms
of the actual source. Some extended vocal techniques or other unusual sonic
output of the human vocal tract heard in isolation might not be "voice-like"
under the above definition; sounds produced by electronic or acoustic instru-
ments other than the voice are often considered voice-like in some aspects. Also,
because a signal may be voice-like in one aspect (e.g. vibrato rate or intonation
contour) and unvoice-like in another (e.g. timbre of the glottal source etc.) it is
often more useful to refer to voice-like aspects of a sound than to voice-like sounds.
It is intuitively obvious that an association with the human voice can be pro-
duced-by clearing the throat or coughing, for example-without any phonetic
message involved. On the other hand, the presentation of a specific phonetic
message (whether by human voice or electronic or acoustic instruments) cer-
tainly cues an association with the human voice, and must be regarded as voice-
like in that aspect-even if the glottal source or other aspects of the sound are not
at all human-sounding or "voice-like."
But there is a much stronger correlation between voice-likeness and phonetic
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
Phonetic/Nonphonetic Perception 141
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
142 Perspectives of New Music
and at a rate which could not be approached by a human performer. For the
most part, the texts of these songs are intelligible-despite the "unvoice-like"
behavior of the fundamental frequency in the sections to which I refer. How-
ever, the nonhuman behavior and the "weighting" of the information (rapid
discrete pitch change on a single vowel) draws attention to the nonphonetic
aspects of the sound-sometimes to the extent that the vowel is temporarily lost.
This restates the point made in the preceding section: unvoice-like behaviors
may draw listeners' attention away from phonetic information available in a sig-
nal. Moreover, the reverse can also apply: voice-like behaviors may draw lis-
teners' attention toward a phonetic interpretation of an ambiguous signal. (For a
related experiment, see Tsunoda 1971). Both of these influences play a role in the
strategies outlined below.
PHONETIC/NONPHONETIC PERCEPTION
I have distinguished above between voice-like sounds and sounds which cue
phonetic interpretation. I have also pointed out that any utterance which con-
tains phonetic information also contains nonphonetic information. It is possible
to go one step further to say that those aspects of a sound which determine
phonetic content (e.g. formant transitions in stop consonant/ vowel syllables)
can, to an extent, also be discriminated on a purely sonic basis. That is to say,
listeners can, within limits, discriminate between two signals which they identify
as being the same phonetically (both identified as the syllable /bi/, for example)
and which are identical in every aspect except in the exact structure of their for-
mant transitions. Moreover, under certain circumstances, a signal can be per-
ceived-and discriminated-either alternatively or simultaneously as speech and
nonspeech (see, for example, Bailey, et al. 1977). When sounds are heard as both
speech and nonspeech simultaneously, the phenomenon is called "duplex per-
ception" (Rand 1974; Isenberg and Liberman 1978; Liberman 1979).
Thus it is possible to derive two different types of information, sonic and
phonetic, from the same aspect of the same signal. This is so because of the con-
trasting nature of the two perceptual processes the listener brings to bear.
Phonetic perceptual processing involves attending to a more or less continu-
ous stream of sound as a carrier of separate and serially ordered phonetic seg-
ments. This is not to say that the sonic signal is itself "spliceable" into separate
phonetic representations. It is not. Adjacent phones of any given syllable are co-
articulated by the speaker and "interleaved" in the resultant acoustic signal. (See
Liberman and Studdert-Kennedy 1977.) The acoustic signal is decoded into sep-
arate and serially ordered segments, however, by the lister.
Sonic perceptual processing, on the other hand, does not usually involve the
discrimination and "labeling" ofsegments. Although we can discriminate sepa-
rate events in the signal, and can recognize the order of those events, we often
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
Phonetic/Nonphonetic Perception 143
COMPOSITIONAL STRATEGIES
* AWAY from the phonetic information in the signal and TOWARD the
sonic information (in the case of sounds which lend themselves easily to
phonetic interpretation), or...
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
144 Perspectives of New Music
The first class of strategies has its effect by de-emphasizing or destroying the
relationship of the sounds to a hierarchical linguistic structure and by drawing
the attention instead to the purely sonic information. By themselves, these strat-
egies are often less compelling than the class two strategies, discussed below,
and, if listeners allow their attention to remain on the elementary linuistic infor-
mation in the signal, they may find these approaches unrewarding. Each strategy
may be more effective-as in most of the examples cited-when combined with
other strategies.
Class one strategies include uses of:
Some of the earliest examples of this approach taken in this century are the
nonsense poems of the dadaists and the Italian futurists. Many of them organ-
ized their nonsense poems sonically by working with a highly restricted phonetic
vocabulary. More recent examples of this technique include Ligeti's Nouvelles
Aventures for three vocalists and seven instrumentalists (Example 1).
I also include in this category text "deconstructions" such as Daniel Lentz'
Songs of the Sirens and Cage's 62 Mesostics Re Merce Cunningham because the
texts-in their fragmented form-draw the attention away from the syntactic
level specifically by means of their lack of morphemic and syntactic content.
It should be noted at this point that, just as it is possible for a text to be syntac-
tically nonsensical or morphemically nonsensical, a text may also conform to or
violate the phonological conventions of a given language. Thus the nonsense
word "shtimp" would be "phonotactically well-formed" in German but not in
English-"sh" does not precede "t" as an initial sound in English. Moreover,
the set of phonetic sounds itself varies from language to language (with large
areas of overlap). Tongue clicks, for example, are phonetic in some African lan-
guages but not in European languages. Thus even the question of whether lis-
teners hear phonetic nonsense or nonphonetic vocal sound in a given text may
depend to some extent on the languages with which they are familiar. A non-
sense text may thus say a great deal about itself even without involving mor-
phemically or syntactically intelligible information.
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
Phonetic/Nonphonetic Perception 145
4 X ABRUPTO:
[Fl.utoe: *mtot. _ . p Fl. cclo]
rpre) r 5 r- o------------.
.8A"120
P. C -S-^1
me*l,l"^t,ur .rt_.I- ._ i----1 L ,
I'S
?ki K T Ji
.~,.,.. ' we,o,,. . , iT;.
I !ts!b! r!gE! c:A riLdr! ka! tjz!
Alt' o.~s* ^ ot .,.o,b;ots (Ite*FOOP Alt:tB I..... '
lt , ^YO l. sempfe
Haibti..oc, T 4, c4Lt=. 2 lMo.t
Avo so .
,, t4t.
t.* AooA5W
rate JfY\)
k ?-3e$| r At;t-
~ r 4 ^ r
'/ f - - St
;bJsIt - r>
,e 1 *
if je` k/ . Y ' '
jP. W ; , t e pW t^ - 1 - - -'.1 - -
B 7; ^fd^? ^7 1Th' ^ f?Syi K r
P -3-
;!t
<L'k
V?! kv! thae! ra! 6k.i o! t' s-- 'y!
?,! co!
(-
JA,ffi Wr^ '
[": ?-J
q
OLB tko- I - y
---- 'l _
b4l;B"F - - iiQ P-^K
if__e,
v -."-ooasroE,_
'~"~"~~.d.- - ":' ?et,
EXAMPLE 1
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
146 Perspectives of New Music
Whereas the first class of strategies may involve a single source continuously
presenting phonetic materials, the second class makes use of OTHER
SOURCES and/or INTERMITTENTLY INTELLIGIBLE phonetic materials.
These latter strategies draw attention to sonic information by means of:
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
Phonetic/Nonphonetic Perception 147
pay conscious attention to the sound of phonetic cues in their attempt to decode
an ambiguous message.
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
148 Perspectives of New Music
+ tf--- .t
. i i ;.
I I I
I~~~a b
-
. I""' b
I I --I i
I I _ j
Luciano Berio-CIRCLES
? Copyright 1961 by Universal Edition (London) Ltd., London
All Rights Reserved
Used by permission of European American Music Distributors
Corporation, sole U.S. agent for Universal Edition
EXAMPLE 2
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
Phonetic/Nonphonetic Perception 149
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
150 Perspectives of New Music
6 (
SI
S2
A I
A2
TI
T2
BI
B2
PIANO I
PIANO 2
PERC I
PERC 2
ORGAN
EXAMPLE 3
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
Phonetic/Nonphonetic Perception 151
SOURCE/FILTER EFFECTS
As I pointed out above, any sound which can be interpreted phonetically also
has a nonphonetic component-sonic information which is not directly relevant
to the phonetic code. In many languages, changes in voice quality (e.g. raspy,
harsh, whispered, etc.) and pitch do not change the phonetic message itself.
By employing glottal waveforms which display attack and decay patterns, par-
tial frequencies, and behaviors of pitch and timbre change which are unusual or
impossible for the human glottis to produce, composers have drawn listeners to
focus on information in the speech signal which is not directly relevant to the
phonetic code. Moreover, this "nonphonetic" focus seems to transfer, to some
extent, to a nonphonetic focus on formant resonances and other aspects of the
sound which determine the phonetic message. This transfer of attention does
not preclude the possibility of phonetic and syntactic intelligibility, but it leaves
less attentional capacity available to this task.
In Still, for example, Roger Reynolds uses "vocal fry" extensively (and slows
the rate at which the text is presented to well below normal speaking rates). In
SoftMorning, City., Tod Machover uses the sound of a contrabass as the glottal
source (by means of a "cross synthesis" technique) for some fragments of the
text in the tape part. In The Stoy ofourLives, Charles Dodge uses a wide band of
converging and diverging sine waves to represent the "book voice" in his syn-
thesized-speech setting of Mark Strand's poem. Bengt Emil Johnson's 3/1970;
(bland) III ends with electronically processed vocal sounds most of which retain
their formant structure (and hence their phonetic identity) while the timbre of
the source is significantly altered. Thus vocal, instrumental, and electronic
sources have all been used for this effect. In each case, the marginal intelligibility
of the text is due partly to the above-cited transfer of the listener's attention to
nonphonetic aspects of the speech signal.
CONCLUDING THOUGHTS
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
152 Perspectives of New Music
* to construct the composition parallel to (but with little regard for) the
organization ofthephonetic segments in the text.
There exist formidable pieces which take each of these approaches. However, a
discussion of the ways in which the perceptual issues discussed above relate to
these pieces and to these general approaches requires a separate exposition.
The word lives a double life. At times it grows like a plant and produces a
cluster of sonorous crystals, then the beginning of the sound takes on its
own life, while that part of reason which we call "The Word" remains in
shadow; at other times the word places itself at the service of reason; the
sound ceases to be "omnipotent" and absolute-it becomes "NAME"
and weakly carries out reason's orders. Thus, now reason obeys sound,
now pure sound obeys pure reason. It is a struggle between two worlds, a
struggle of two powers which is constantly carried on in the heart of the
word, giving a double meaning to language: two circles of shooting stars.
(Hausmann, 1969, 53)
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
Phonetic/Nonphonetic Perception 153
NOTES
REFERENCES
ARTICLES
Bailey, Peter J., Quentin Summerfield, and Michael Dorman. 1977. "O
Identification of Sine-Wave Analogues of Certain Speech Sounds." Haskin
ortories Status Report on Speech Research SR-51/52: 1-25.
House, A.S., K.N. Stevens, T.T. Sandel, and J.B. Arnold. 1962. "O
Learning of Speech-like Vocabularies. " Journal of Verbal Learning and Beh
1:133-43.
Isenberg, D., and A.M. Liberman. 1962. "Speech and Nonspeech Percepts
from the Same Sound. "Journalofthe AcusticalSociety ofAmerica 64, Suppl. No.
1:J20.
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
154 Perspectives of New Music
Rand, TC. 1974. "Dichotic Release from Masking for Speech." Journal of the
AcousticalSociety ofAmerica 64, Suppl. No. 1: J20.
Remez, Robert E., Philip E. Rubin, David B. Pisoni, and Thomas D. Carrell.
1981. "Speech Perception without Traditional Speech Cues. " Scince 212(4497):
947-50.
Stockhausen, Karlheinz. 1964. "Music and Speech. " Die Reihe 6:47-56.
Tsunoda, Tadanobu. 1971. "The Difference of the Cerebral Dominance
Vowel Sounds among Different Languages." The Journal of Auditory Researc
11:305-14.
SCORES
Berio, Luciano. Circles, for female voice, harp, and two percussion
London: Universal Edition, 1961.
Machover, Tod. Soft Morning, City! for soprano, double bass, and co
generated tape. Paris: Ricordi Press, 1980.
SOUND RECORDINGS
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms
Phonetic/Nonphonetic Perception 155
Bodin, Lars-Gunnar. For Jon III (They Extricated their Extremities plus for John)
(Phono-disc) Fylkingen Records FYLP 1029.
Johnson, Bengt Emil. 3/1970; (bland) ILl. (Phono-disc) Sveriges Radio 20-1.
Lentz, Daniel. Songs of the Sirens. (Phono-disc) ABC Command COMS 9005.
This content downloaded from 185.38.254.12 on Thu, 11 May 2017 11:16:41 UTC
All use subject to http://about.jstor.org/terms