developmental speech perception by janet feldma …
TRANSCRIPT
DEVELOPMENTAL SPEECH PERCEPTION
by
JANET FELDMAN WERKER
B.A,, Harvard U n i v e r s i t y , 1974
A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF ARTS
i n
THE FACULTY OF GRADUATE STUDIES (Department of Psychology)
We accept t h i s thesis as conforming to the required standard
THE UNIVERSITY OF BRITISH COLUMBIA December, 1978
@ Janet Feldman Werker, 1978
I n p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m ' e n t o f t h e r e q u i r e m e n t s f o r
an a d v a n c e d d e g r e e a t t h e U n i v e r s i t y o f B r i t i s h Columbia, I agree t h a t
t h e L i b r a r y s h a l l m a k e i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and stud y .
I f u r t h e r a g r e e t h a t ' p e r m i s s i o n - f o r . - e x t e n s i v e c o p y i n g o f t h i s t h e s i s
f o r s c h o l a r l y p u r p o s e s may b e g r a n t e d b y the H e a d o f my Department or
b y h i s r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n
o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l n o t b e a l l o w e d w i t h o u t my
w r i t t e n p e r m i s s i o n .
Psychology Department o f
T h e U n i v e r s i t y o f B r i t i s h C o l u m b i a 2075 Wesbrook P l a c e Vancouver, Canada V6T 1W5
Date
i i
ABSTRACT
Previous research has indicated that infants have the a b i l i t y to
c a t e g o r i c a l l y discriminate many of the d i s t i n c t i v e features of speech
sounds regardless of t h e i r exposure to a language i n which such d i s
t i n c t i o n s are important, whereas adults of one language group may have
d i f f i c u l t y d i s c r i m i n a t i n g l i n g u i s t i c features that are important i n a
foreign language. This suggests a decline during development i n l i n g u i s
t i c perceptual a b i l i t i e s , during which the a b i l i t y to discriminate non-
relevant features may be l o s t . This study was designed to be the f i r s t
i n a s e r i e s of tests of such a decline and involved comparing E n g l i s h -
speaking adults, Hindi speaking adults, and six-month o l d infants on
th e i r a b i l i t y to discriminate foreign and native speech contrasts.
Two p a i r s of Hindi sounds, and one p a i r of English sounds were i n
vestigated i n t h i s study, Infants were tested i n a " v i s u a l l y r e i n f o r c e d
i n f a n t speech d i s c r i m i n a t i o n " (VRISD) paradigm. This i s a di s c r i m i n a t i o n
paradigm i n which the infant i s conditioned to turn i t s head when there
i s a change i n the auditory stimulus. A va r i a t e of t h i s paradigm was
employed f o r the adult subjects using a button-press, rather than a head
turn, as the c r i t i c a l behavioral response.
This study y i e l d e d support f o r the notion that i n f a n t s have the
a b i l i t y to c a t e g o r i c a l l y discriminate d i s t i n c t i v e features of speech
sounds regardless of exposure to a language i n which such d i s t i n c t i o n s
are important. Some support was also given f o r the idea that there may
be a decline i n speech perceptual a b i l i t i e s with e i t h e r age/or l i n g u i s t
i c experience, but th i s f i n d i n g was only s i g n i f i c a n t with one of the two
Hindi sound p a i r s .
i i i
Two explanations are offered for these r e s u l t s . I t i s suggested
that the English population may have had someexperience with the non
s i g n i f i c a n t Hindi contrast. A l t e r n a t i v e l y , i t i s suggested that d i f f e r
e n t i a l perceptual distance may have accounted for the differences between
the two Hindi sound p a i r s . Two l e v e l s of processing, acoustic and l i n g u i s
t i c , are then invoked i n t h i s explanation.
i v
TABLE OF CONTENTS
Page
ABSTRACT l i
TABLE OF..CONTENTS ........ i v
LIST OF TABLES ........ ,v •
LIST OF APPENDIX A TABLES . . . v j
LIST OF FIGURES ........ v i i
LIST OF APPENDIX A FIGURES v i i i
ACKNOWLEDGEMENTS ..... . . . i x
INTRODUCTION 1
.METHOD . 19
RESULTS 29
DISCUSSION > 36
CONCLUSIONS 44
REFERENCES ........ 45
APPENDIX A . . . . , , 53
APPENDIX B . . . . . .i. . 58
V LIST OF-TABLES
Page
TABLE 1 Hindi Stop Consonants ....... 7
TABLE 2 F r i c a t i v e Discrimination by Age ....... 10
TABLE 3 Analysis of Proportions f o r Retror-flex/Dental Contrast' .,„.... 30
TABLE 4 Analysis of Proportions f o r Voiceless Aspirated/Breathy Voiced Contrast ' 30
TABLE 5 Mult i p l e Comparisons,on Retroflex/ Dental Contrast 31
TABLE 6 Multiple Comparisons on Voiceless Aspirated/Breathy Voiced Contrast ^
TABLE 7 Infant Mean Number of T r i a l s on Speech Contrasts 33
TABLE 8 C e l l Means and Standard Deviations For Number of T r i a l s to C r i t e r i o n ,. 35
TABLE 9 Source Table for Number of T r i a l s to C r i t e r i o n ; Analysis of Variance 35
v i i
LIST OF FIGURES
page
FIGURE 1 Mean Number of Sucks Per Minute ..... 17
FIGURE 2 Arrangement of the Experimental Site ..... 21
FIGURE 3 Infant During Control T r i a l 2 2
FIGURE 4 Infant During Experimental T r i a l ......
v i i i
LIST OF-APPENDIX A FIGURES
Page
FIGURE 1 Overlap of A r t i c u l a t o r y Features ...... 54
FIGURE 2 Abbreviated "d" Spectograms \ 54
ACKNOWLEDGEMENTS
I would l i k e to give s p e c i a l thanks to Dr, Richard Tees,
Dr. John G i l b e r t , and Keith Humphrey f o r giving so much assistance
and encouragement throughout'the project. Thanks are also due to
Dr.'Robert Fremder and Dr. Tannis MacBeth Williams for t h e i r w i l l r -
ing assistance during c r i t i c a l stages of t h i s project. 1
1
INTRODUCTION'
This study was designed to be a f i r s t step i n . t e s t i n g the following
hypothesis t
Humans are born with the a b i l i t y to discriminate the univ e r s a l set of d i s t i n c t i v e l i n g u i s t i c features, Through development, there i s a decline i n t h i s a b i l i t y as ,the perceptual space becomes organized to approximate the phonemic categories used i n the na^ t l v e language,
This idea (that some perceptual c a p a b i l i t i e s decline with age) con<-
f l i c t s with most approaches to perceptual development. According to
Gibson (1967) f o r example, perceptual capacity becomes i n c r e a s i n g l y d i f
f e r e n t i a t e d and ref±ned through ontogeny, Perceptual learning r e s u l t s i n
an Increase i n a b i l i t y to acquire information from the environment as a
r e s u l t of experience with the array of s t i m u l i provided by the environment.
Thus only those s t i m u l i present i n the environment lead to refinements
i n perceptual.capacity. A d d i t i o n a l l y , perceptual learning i s seen as an
increase, rather than a decrease, i n discriminatory a b i l i t y ,
Other contemporary•theories of perceptual development also support
the notion of increasing d i f f e r e n t i a t i o n as a r e s u l t of experience.' For
example Hebb and Piaget would both view perceptual development as r e f l e c t
ing the construction of "schemata" as organizations of the lelements of
sensory input. The schema i s then capable of being modified with further
experience and i s thought to influence the perception of objects by
"guiding" the i n d i v i d u a l to c e r t a i n d e t a i l s and by aiding i n c l a s s i f y i n g
and understanding the sensory input (e.g., c f . Tees, 1976), Somewhat
s i m i l a r l y , perception has been regarded as a process of category b u i l d i n g
2
and inference drawing (Bruner, 1957). As more categories are b u i l t ,
perception becomes l e s s dependent upon the d e t a i l s of the sensory input,
and more on the categories. Through development the categories become
more accessible so that fewer d e t a i l s are necessary for c l a s s i f y i n g an
input.
On the other hand, behaviorists such as M i l l e r (1948) view perceptual
development as r e s u l t i n g from p o s i t i v e feedback from the environment. A
chance response to an appropriate stimulus r e s u l t s i n p o s i t i v e feedback,
and t h i s motivates the i n d i v i d u a l to discriminate that stimulus from others
a second time, etc.
Although these theories deal adequately with many facets of p'ercep<-
t i o n , they do not deal d i r e c t l y with d i s c r i m i n a t i o n . According to S-.S'.
Stevens (1939), discrimination i s the fundamental cognitive act. , The
a b i l i t y to discriminate the fundamental features of speech i s the necessary
f i r s t step i n language learning.
0n the basis of elementary discriminations, then, we make our f i r s t rudimentary classes and i n doing so we have the f i r s t step toward ge n e r a l i z a t i o n . (Sevens, 1939, p. 58).
One, perhaps unique aspect of speech p e r c e p t i o n i s the accumulation
of evidence suggesting that "categorization" of speech sounds i s an innate
propensity of the human organism, That i s , rather.than having t o . l e a f n to
d i f f e r e n t i a t e phoneiic features and then organize them into categories,
human infants seem to respond from the beginning, to speech sounds accord
ing to phonemic categories. This kind of a b i l i t y would enable the infant
to break the continuous auditory input of speech into d i s c r e t e segments,
thus s t a r t i n g the process of the eventual i d e n t i f i c a t i o n of meaningful seg
ments. For example, when given synthesized speech sounds varying along
71) Before proceeding with t h i s section i t i s necessary to understand the basic research i n experimental psychology concerning the decoding of speech sounds and the r e a l i t y of d i s t i n c t i v e features. Interested readers are r e f e r r e d to Appendix A.
3
a continuum, infants as young as 1 month of age show evidence of d i s c r i
minating the acoustic cues which d i f f e r e n t i a t e phonemes, but do not r e
spond to within category v a r i a t i o n s , ( e . g . , Elmas,Siqueland, Jusczyk, &
Vigo r i to ,1971).
In each natural language only a subset of the u n i v e r s a l set of
phonetic d i s t i n c t i o n s i s u t i l i z e d to d i f f e r e n t i a t e meaningful sound seg^
ments (phonemes). In an i n d i v i d u a l language, phoneme categories may
broaden to .include more than a single phonetic s p e c i f i c a t i o n (allophonic
v a r i a t i o n ) . For example, i n English, a s p i r a t i o n does not connote d i f f e r
ences i n meaning. Thus i f an English speaker pronounced " p i l l " with an
aspirated or an unaspirated v o i c e l e s s stop consonant, other English speak-,
ers-..would understand the word to be " p i l l " and would ignore the i r r e l e v a n t
phonetic v a r i a t i o n s , It can be seen then, that iri terms of learning one's
own language, perceptual development can be viewed as a process of groupr-
irig i n i t i a l l y d i s t i n c t speech sounds into broader categories, thereby esr-
t a b l i s h i n g the required set or class of phonemes relevant to one's native
language.
In addition to allophonic variations,, speakers of a , p a r t i c u l a r
language may omit a subset of phonetic vfeatures from use. There i s some
evidence that adult l i s t e n e r s have d i f f i c u l t y d i s c r i m i n a t i n g those features
when hearing them f o r the f i r s t time i n a l i n g u i s t i c context (Lisker &
Abramson, 1968).
The s i g n i f i c a n c e of the discovery that young in f a n t s can discriminate
some l i n g u i s t i c a l l y relevant d i s t i n c t i v e features of speech does not
nec e s s a r i l y imply a conscious recognition of speech sounds as language,
although Eimas (1975) has stated that t h i s discriminatory a b i l i t y "indicates
•: 4 •
that infants have some knowledge of the phonetic structure.of language,
s p e c i f i c a l l y , they must have knowledge of at l e a s t some of the phonetic
features." (p. '214). It i s unfortunate, however^ that at present there
i s no way t o . d i f f e r e n t i a t e s t r i c t l y acoustic from phonetic knowledge. In
any event demonstration of discriminatory a b i l i t i e s that r e f l e c t phonemic
features suggests that infants'have at l e a s t the necessary f i r s t response
c a p a b i l i t i e s to begin organizing t h e i r l i n g u i s t i c input into a f i n i t e set
of categories. Whether the infant has knowledge of the phonetic structure
becomes•irrelevant. Michael Laine, i n h i s Introduction to Structuralism
(1970) stated that when people use t h e i r own language
"...they c o n s i s t e n t l y and constantly apply i t s phonol o g i c a l laws ( i t s structure, i n other words) i n t h e i r speech. They w i l l not, unless they are versed i n l i n g u i s t i c s , be consciously aware of them. Nor, i f asked, would they be able to supply those laws....... What the observer sees i s not the structure, but simr-p l y the evidence and product of the structure, ..... There i s i n man an innate, g e n e t i c a l l y transmitted and determined mechanism that acts as a s t r u c t u r i n g force., Moreover, t h i s inherent q u a l i t y or capacity i s so designed as to l i m i t the possible range of ways of s t r u c turing, (Caine, 1970, p. 15)
Thus the response c a p a c i t i e s of the human infant would l i m i t the range of
categories used f o r l i n g u i s t i c input, .
The hypothesis that speech perceptual development may proceed from
the u n i v e r s a l to .a more l i m i t e d set of categories does not o r i g i n a t e with
t h i s study. Similar speculations have been made by other researchers
(Eimas, 1976; G i l b e r t , 1975; P i s o n i , 1977, Trehub, 1976). I t i s necessary,
however, to a c t u a l l y test whether or not infants can discriminate non-
native speech contrasts with greater ease than adults i n a s i n g l e , properly
designed study,
5
In order to examine t h i s hypothesis, infants and adults were tested
on two p a i r s of Hindi speech contrasts, JandLO'ne,pal'r of English speech^con-
t r a s t s i n two very s i m i l a r d i s c r i m i n a t i o n paradigms.
The English p a i r used was the common /ba//da/. These are both voiced
stop consonants d i f f e r i n g only In place of a r t i c u l a t i o n ( b i l a b i a l vs.
a p i c a l ) . This i s a r e l a t i v e l y common d i s t i n c t i o n across natural languages
and has been previously studied using i n f a n t s .
The f i r s t Hindi p a i r was the unvoiced, unaspirated r e t r o f l e x vs, dental
stop /Ta/'/.ta/ where place of a r t i c u l a t i o n i s the c r i t i c a l dimension. Retro-
f l e x consonants are produced by c u r l i n g the tongue back and placing i t pos^-
t e r i o r to the alv e o l a r ridge. In contrast, dentals are produced by plac i n g
the t i p or blade of the tongue against the back of the upper front t e e t h .
Dental i s a common place of production i n En g l i s h (and i n most natural
languages) whereas r e t r o f l e c t i o n does not carry phonemic information i n
English. A d d i t i o n a l l y , r e t r o f l e x consonants are uncommon across natural
languages and t h e i r contextual d i s t r i b u t i o n i s r e s t r i c t e d even i n those
languages i n which they have import.
The second Hindi p a i r was the unvoiced aspirated, dental stop vs. the
murmured dental stop /t*V/d*V i n which a differ e n c e i n voice onset time i s
the c r i t i c a l d i s t i n c t i o n . Voice onset time i s a combination of the state
of the g l o t t i s during an a r t i c u l a t i o n and the presence or absence of a
period of voicelessness during and a f t e r the release of an a r t i c u l a t i o n .
In English only two categories of voice onset time a r e ' d i f f e r e n t i a t e d ; —
voiced (the vocal words are nearly together so that they vibrate) and
voi c e l e s s (vocal cords are so f a r apart that they cannot vibrate at a l l ) ,
There are two a d d i t i o n a l categories i n Hindi; v o i c e l e s s unaspirated and
6
brea thy v o i c e d (murmured) s tops , (See Table 1 ) ,
As can be seen, bo th p a i r s of H i n d i sounds i n v o l v e d i s t i n c t i o n s b e
tween ca t ego r i e s that do not e x i s t at a l l i n E n g l i s h and are r e l a t i v e l y u n
common to n a t u r a l languages. Evidence has been p rov ided (Singh &"Black ,
1966) tha t these p a i r s would thus be among the hardest f o r E n g l i s h speakers
to d i s c r i m i n a t e . A d d i t i o n a l l y , S . E . B lums te in suggested tha t i n ga the r ing
p i l o t data f o r the s t u d y . o f r e t r o f l e x stop consonants (Stevens & B l u m s t e i n ,
1975), E n g l i s h - s p e a k i n g subjec ts d i d seem todemonstrate d i f f i c u l t y i n d i s ^
c r i m i n a t i n g the unvoicedr-unaspirated d e n t a l v s , r e t r o f l e x stop consonants
(personal communication, February , 1978),
The d i s c r i m i n a t i o n paradigm used f o r the i n f a n t s i s c a l l e d " v i s u a l l y A
r e i n f o r c e d speech d i s c r i m i n a t i o n " (VRTSD).• A v a r i a t e of t h i s paradigm was
used f o r the a d u l t s . VRISD was f i r s t developed as a v a r i a n t of the c l a s s i c
"p l ay audiometry" o f D i x and H a l l p i k e (1947) , i n which a c h i l d was condir^
t i o n e d through a s s o c i a t i o n to, push a bu t ton i n response to a sound source
to e l i c i t the p r e s e n t a t i o n o f a p leasan t p i c t u r e . Suzuk i and Ogiba (1960)
modi f i ed the response to a s imple headr-turn i n order to t e s t the hea r ing
a b i l i t y of c h i l d r e n under 3 , c a l l i n g i t v i s u a l re inforcement audiometry
(VRA). Al though Suzuki and Ogiba (1961) repor ted low success r a t e s f o r
i n f a n t s one year and younger, more, recent work ( W i l s o n , L e e , Owen & Moore,
unpubl i shed manuscr ip t ) i n d i c a t e d tha t 90% of i n f a n t s as young as age 5r- l /2
months can be t e s t ed w i t h t h i s paradigm. More r e c e n t l y researchers have
begun us ing t h i s paradigm i n speech d i s c r i m i n a t i o n s t u d i e s ( E i l e r s &
M i n i f i e , 1975; E i l e r s , W i l s o n & Moore, 1977; H i l l e n b r a n d , M i n i f i e & Edwards,
1977; K u h l , 1976, 1977; M i n i f i e , 1976) . In these s t u d i e s the i n f a n t was.
cond i t i oned to t u rn h i s / h e r head toward a sound source when he/she de tec ted
7
Table 1
Hindi Stop Consonants
Voiceless unaspirated
Voiceless aspirated
Voiced Breathy voiced .
b i l a b i a l
dental
r e t r o f l e x
post-alveolar a f f r i c a t e
v e l a r
p a l (take care of)
* TA f (mode of
singing)
t a l (postpone) t f a l (gS)
kan (ear)
p a l (edge of knife)
h ** £ an ( r o l l of cloth)
t h a l (place f o r buying wood)
t f a l (deceit)
k an (mine)
bal (hair)
J? hal (forehead)
h ** (charity) (paddy)
dal d h a l (branch) (shield)
d 3 a l (water)
gan (song)
- h , a l ,
(glimmer) h
g an (Kind of bundle)
= comparison Retroflex/Dental **
= comparison Voiceless Asp/Breathy Voiced
8
a change i n the s t i m u l u s , and was rewarded w i t h the appearance of a f l a s h
i n g , no i s e -p roduc ing toy a n i m a l . Not o n l y cou ld i n f a n t s as young as f i v e
months perform s u c c e s s f u l l y i n t h i s paradigm, .but they a l s o show the a b i l i t y
to t r a n s f e r t h i s " l e a r n i n g , r s e t " to new s t imulus p a i r s ( K u h l , p e r s o n a l
communicat ion) .
In fan t Research
Al though speech researchers are u n i t e d i r i t h e i r b e l i e f tha t by age 2
months i n f a n t s demonstrate a s o p h i s t i c a t e d a b i l i t y to make phone t ic d i s c r i
mina t ions , , (Eimas , 1974, 1975a,b; E i l e r s & M i n i f i e , 1975; K u h l , 1976; M o f f i t ,
1971; Trehub, 1973, 1976; W i l l i a m s , , i n p r e p a r a t i o n ) , the meaning of t h i s
d i s c r i m i n a t o r y a b i l i t y i s debated. Accord ing to the a c o u s t i c cue h y p o t h e r -
s i s , these .apparent "phonet ic d i s c r i m i n a t i o n s " , are based on lower, order
a u d i t o r y d i s c r i m i n a t i o n s . a n d are viewed as no d i f f e r e n t from non^-speech
a u d i t o r y p e r c e p t i o n (Stevens & K l a t t , 1974) . In c o n t r a s t , proponents o f
the l i n g u i s t i c cue hypothes i s c l a i m tha t speech sounds are heardo. s p e c i f i c
c a l l y as speech (Eimas ,• 1974; G i l b e r t , 1971; M o f f i t , 1971; Morse , 1976;
Trehub, 1973), and tha t speech d i s c r i m i n a t i o n i s an inna te b i o l o g i c a l a b i l i ^ -
t y important i n the process o f language a c q u i s i t i o n , (Eimas, S i q u e l a n d ,
Jusczyk & V i g o r i t o , 1971). A d d i t i o n a l elements of the l i n g u i s t i c cue hypo
t h e s i s i n c l u d e the f o l l o w i n g : (1) p e r c e p t i o n o f consonantr-vowel sounds i s
c a t e g o r i c a l and l i n g u i s t i c a l l y r e l e v a n t , (2) speech p e r c e p t i o n i s non^-
developmenta l , i . e . , i n f a n t s respond to phone t ic s t i m u l i i n the same way
as a d u l t s , (3) non-speech sounds,are not pe rce ived c a t e g o r i c a l l y and (4) hu^
mans are the on ly spec ies tha t can pe rce ive speech sounds c a t e g o r i c a l l y .
Whi le there i s some evidence to suggest tha t i n f a n t s and a d u l t s can
p e r c e i v e some nonspeech sounds c a t e g o r i c a l l y ( Juscyk, Rosner & C u t t i n g ,
9
1977), that other animals can show c a t e g o r i c a l perception f o r phonetic
features (Kuhl & M i l l e r , 1975), and that within category perception i s C2)
possible under, extreme conditions (Samuel, 1977), the c r i t i c a l issues
for the present study are whether or not perception of consonant-vowel
sounds i s organized according to phonemic categories and whether infants .
can respond to phonetic s t i m u l i i n the same way as adults.
Several studies have already been undertaken to examine these two
c r i t i c a l Issues. Two measures have been used extensively i n these studies;
heart rate deceleration (HRD) and high amplitude sucking (HAS) (both have
been discussed extensively by Morse (1974), HRD i s based on the assumption
that heart rate decelerates i n response to a novel stimulus. In contrast
HAS involves a behavioral response by the infant (sucking), Both HRD and
HAS are used i n a habxtuation^dishabituation design, contingent upon rein<-
f o r c i n g properties of the speech s t i m u l i . Both have the following weaknesses.
(1) Tt i s d i f f i c u l t to i n t e r p r e t negative .results. I t i s never known whether
th e , c h i l d was unable to make a dis c r i m i n a t i o n or did not f i n d the s t i m u l i
i n t r i n s i c a l l y r e i n f o r c i n g . (2) I t i s impossible tocompare data across
i n d i v i d u a l s . (3) There i s about a 65% rate of incompletion, that i s , data
can only be reported on about 35% of the sample. In addition, the HAS
paradigm can only be used with infants up to 12 months of age. The HRD can
be further c r i t i c i z e d f o r i t s s e n s i t i v i t y to both the state of the infant
and the c h a r a c t e r i s t i c s of the speech s t i m u l i (Morse, 1974).
In s p i t e of the problems Eimas et a l . (197.1), using a HAS paradigm,
were able to provide answers r e l a t i n g to the fundamental questions. For
example, i n an experiment i n which voice onset time (VOT) was varied by (2) ~ T" "~ ~
'These findings are a l l taken as evidence that there i s both a l i n g u i s t i c and an acoustic stage i n processing, but do not negate.the psychological r e a l i t y of c a t e g o r i c a l perception i n a l i n g u i s t i c context.
10
I n t e r v a l s o f 20 m s e c , i n f a n t s aged 1-4 months were shown to d i s c r i m i n a t e
between the v o i c e d and v o i c e l e s s stop consonants / b / and / p / a long the
same phonemic boundary as a d u l t s ; tha t i s , they showed bo th ' c a t e g o r i c a l
d i s c r i m i n a t i o n and l i n g u i s t i c a l l y r e l evan t p e r c e p t i o n . I n a d d i t i o n to
c a t e g o r i c a l d i s c r i m i n a t i o n along the VOT continuum, c a r e f u l l y designed
fo l low-up s t u d i e s have demonstrated tha t i n f a n t s can c a t e g o r i c a l l y d i s c r i
minate the / r / - / l / d i s t i n c t i o n (Eimas,' 1975b), stop consonants i n f i n a l
p o s i t i o n (Eimas, 1 1974) ,• p l ace o f a r t i c u l a t i o n ( J i i sczuk , 1977a) , most o f
the f r i c a t i v e s ( E i l e r s & M i n i f i e , 1975) , some vowels (Trehub, 1973) ,
g l i d e s i n i n i t i a l and media l p o s i t i o n ( Jusczyk , Copari & Thompson, 1977) ,
and n e a r l y every o the r fea ture on which they have been t e s t e d .
Not a l l the a v a i l a b l e evidence i s suppor t ive of the c e n t r a l hypothe^
s i s o f t h i s s tudy . TMrre -eisl.£Someiesugge_s:t-io:nmthat f r i c a t i v e p a i r s may be
d i f f i c u l t fo r young i n f a n t s to d i s c r i m i n a t e ( E i l e r s & M i n i f i e , 1975;
E i l e r s , W i l s o n & Moore, 1977) . Us ing v i s u a l l y r e i n f o r c e d i n f a n t speech
d i s c r i m i n a t i o n (VRISD) , i n f a n t s aged 6"8 and 12r-14 months o f age were
compared on t h e i r a b i l i t y to d i s c r i m i n a t e n a t u r a l p a i r s o f both F i / 9 i and
F u / 8 a . T h e i r r e s u l t s are summarized i n Table 2, A s i g n i f i e s the i n ^
fan ts cou ld make the d i s c r i m i n a t i o n , and a " - " s i g n i f i e s tha t they cou ld
n o t .
Table 2
F r i c a t i v e D i s c r i m i n a t i o n by Age
6-8 mos, 12-14 mos.
F i / 0 i r-.. +
Fu/0a
11
E i l e r s suggests tha t these f i n d i n g s show tha t "improvement" does occur
w i t h age i n f r i c a t i v e d i s c r i m i n a t i o n because the 6-8 month i n f a n t s f a i l e d
both p a i r s , w h i l e the 12-14 month i n f a n t s passed the F i / 9 i d i s c r i m i n a t i o n .
E i l e r s et a l . (1977) concluded tha t t h e i r f i n d i n g s c a l l i n t o q u e s t i o n
the u n i v e r s a l i t y o f phonemic d i s c r i m i n a t o r y a b i l i t i e s i n young i n f a n t s and
suggest i n s t e a d a more complex p i c t u r e of p e r c e p t u a l development i n c l u d i n g
bo th a u d i t o r y c a p a c i t y and l e a r n i n g ,
T h i s " n e g a t i v e " evidence i s not uncha l l enged , however, Us ing the same
paradigm (VR1SD), Holmberg, Morgan and K u h l (1977) have found evidence of
Fa /0a d i s c r i m i n a t i o n i n a study o f 6-month o l d i n f a n t s , Holmberg et a l .
suggest exper imenta l c o n d i t i o n s may»account f o r the d i f f e r e n c e between
t h e i r r e s u l t s and those obta ined b y . E i l e r s et_ al_, (1977) i n tha t the l a t t e r
r e q u i r e d tha t i n f a n t s show evidence of d i s c r i m i n a t i o n w i t h i n o n l y s i x t r i
a l s . F a i l u r e to do so was i n t e r p r e t e d as i n a b i l i t y to make the d i s c r i m i n a
t i o n . Holmberg et a l . c l a i m tha t s t a t e f l u c t u a t i o n and v a r i a t i o n s 1 i n
a t t en t i venes s may make the s i x t r i a l requirement too s t r i n g e n t f o r younger
i n f a n t s . In any event , i t should a l so be po in ted out tha t f r i c a t i v e s have
been shown to be d i f f i c u l t to d i s c r i m i n a t e whether i n n a t u r a l or syn the
s i z e d forms i n the case o f the a d u l t s ( M i l l e r & N i c e l y , 1955) and c h i l d r e n
(Abbs & M i n i f i e , 1969),
Proponents of the a c o u s t i c cue hypo thes i s i n t e r p r e t f i n d i n g s such as
those presented by E i l e r s e_t al_ (1977) as evidence tha t d i s c r i m i n a t i o n i s
p u r e l y a u d i t o r y , not l i n g u i s t i c a l l y r e l e v a n t and thus sub jec t to l e a r n i n g .
They suggest tha t the c a t e g o r i c a l d i s c r i m i n a t i o n of consonants i s an a r t i
f ac t of the p a r t i c u l a r s t i m u l i used . For example, i n VOT s t u d i e s , i t has
been suggested tha t the c r u c i a l cue i s the presence o r absence of the f i r s t
12
formant t r a n s i t i o n r a t h e r than the r e l a t i v e amount of v o i c i n g ( e . g . ,
Stevens & K l a t t , 1974), However, M o f f i t (unpublished) i n v e s t i g a t e d t h i s
s p e c i f i c ques t i on i n astudy o f c a t e g o r i c a l p e r c e p t i o n of b i l a b i a l stop c o n
sonants . Us ing f o u r , p a i r s of s y n t h e t i c a l l y produced s t i m u l i , two p a i r s
d i f f e r i n g m u l t i ^ d i m e n s i o n a l l y and two p a i r s d i f f e r i n g i n v o i c i n g o n l y ,
M o f f i t found that i n f a n t s aged 30-60 days were able to make l i n g u i s t i c a l l y
r e l e v a n t d i s c r i m i n a t i o n s wi thou t the a c o u s t i c cue of presence or absence
of the f i r s t formant t r a n s i t i o n .
Kuh l (persona l communication^ i s i n v e s t i g a t i n g the s p e c i f i c dimensions
of speech d i s c r i m i n a t i v e a b i l i t i e s i n i n f a n t s u s i n g a paradigm which i s r e
l a t i v e l y . n e w to i n f a n t speech p e r c e p t i o n r e sea r ch , the p e r c e p t u a l constancy
paradigm ( s i m i l a r to VRXSD), v i s u a l l y r e i n f o r c e d i n f a n t speech d i s c r i m i n a
t i o n . She has demonstrated that i n f a n t s can more q u i c k l y l e a r n to organize '
t h e i r l i n g u i s t i c c a t ego r i e s acco rd ing to phone t ic fea tures w h i l e d i s r e g a r d
i n g v a r i a t i o n s i n bo th speaker and i n t o n a t i o n than they can l e a r n to o rgan
i z e c a t ego r i e s acco rd ing to these l a t t e r dimensions d i s r e g a r d i n g phone t ic
f ea tu re s .
Fu r the r support fo r the two c r i t i c a l i s sues of the l i n g u i s t i c hypothe
s i s (as o u t l i n e d p r e v i o u s l y ) i s found i n c r o s s - c u l t u r a l s t u d i e s . Babies of
K i k u y u , Spanish and E n g l i s h speaking parents have been shown to d i s c r i m i n a t e
l i n g u i s t i c fea tures tha t are not r e l evan t i n t h e i r n a t i v e languages (Lasky ,
Lasky & K l e i n , 1975; S t r e e t e r , 1976; Trehub, 1973, 1976) .
A d u l t Research
In con t r a s t to the r e s u l t s of r esea rch w i t h i n f a n t s , s t u d i e s o f adu l t
speech d i s c r i m i n a t i o n have r e s u l t e d i n obse rva t ions tha t adu l t s have d i f f i
c u l t y d i s c r i m i n a t i n g phonemes tha t are not used i n t h e i r n a t i v e language.
13
Japanese adults have been shown to have d i f f i c u l t y d i s c r i m i n a t i n g the
English / r / - / l / . In a study by Miyawaki, Strange, Verberge, Liberman &
Fujimura (1975), 21 Japanese and 39 United States adults were tested on a
d i s c r i m i n a t i o n task using synthesized versions of /ra/ and / l a / i n which
the c r i t i c a l perceptual cue was s t a r t i n g frequency and t r a n s i t i o n of the
t h i r d formant (F3). The r e s u l t s f o r i d e n t i f i c a t i o n and d i s c r i m i n a t i o n
tasks with mature English speaking subjects from the United States, showed
t h e i r perception to be nearly c a t e g o r i c a l . On the other hand, native-.
Japanese speaking adults were able to discriminate only s l i g h t l y above the
l e v e l of chance. A comparison of these same two groups on discrimination
of non-speech counterparts of /ra/ and / l a / (F3) same as i n speech token,
but F l and F2 amplitudes set at zero) yielded a d i f f e r e n t r e s u l t . Both
groups showed equally successful discrimination of a l l comparison p a i r s .
Miyawaki et a l , concluded that experience with language a f f e c t s l i n g u i s t i c ,
but not acoustic perceptual a b i l i t i e s .
Studies with English speaking adult subjects have shown that they.ex
perience d i f f i c u l t y i n discriminating test tokens that span a non-English
lead boundary i n VOT (Lisker & Abramson, 1968). Discrimination data for
adult l i s t e n e r s (Lisker & Abramson, 1970) shows that Spanish speakers can
only d i s t i n g u i s h between two categories of v o i c i n g . In contrast, data pre
sented by Lasky, Lasky and K l e i n (1975)^ ind i c a t e that infants being reared
i n Spanish speaking environments can discriminate three v o i c i n g d i s t i n c t i o n s .
A d d i t i o n a l l y , Trehub (1976) has suggested that adults have d i f f i c u l t y disv
criminating the d i s t i n c t i v e feature, + stridency, as exemplified i n , f o r
example, the Czech /za/ /ra/.-
14
I t has been suggested that since English speakers have only, two cate
gories for prevocalic stop consonants they may lose the a b i l i t y to d i s c r i
minate sounds inv o l v i n g three and four categories. (Ladefoged, 1975), For-
example, the Eastern Armenian;- language has three categories for prevocalic
stops, and Hindi has four categories, Ladefoged's suggestion has been sup
ported i n the l i t e r a t u r e . For example, 'Singh and Black (1966) tested adult .
Japanese, Hindi> English and Arabic speakers le s s than a month a f t e r they
arri v e d iri the United States, A f t e r being traiiied for one hour'on a serie s
of consonant-vowel sounds, the subjects 'were asked to i d e n t i f y those same
sounds i n w r i t i n g . A l l l i s t e n e r s i d e n t i f i e d the sounds o f ' t h e i r native
language best, The Hindi s t i m u l i used i n the present study were p a r t i c u
l a r l y d i f f i c u l t f o r non-Hindi speakers to i d e n t i f y . ,
It should be noted, however, .that(jmemoryj requirements may confound
these r e s u l t s . I d e n t i f i c a t i o n paradigms (such as used by Singh and Black)
involve the use of short-term memory f o r l a b e l s and are thus inadequate
fo r t e s t i n g pure di s c r i m i n a t i o n a b i l i t i e s . Since many of the cross-language
adult findings are based on i d e n t i f i c a t i o n studies they could be misleading. .
In the studies reported by Trehub (1976) and Miyawaki et _al. (1975) d i s c r i
mination paradigms were used, and the adults s t i l l demonstrated d i f f i c u l t y
with the non-English discriminations.
It can be argued that r e s u l t s of tests of perceptual a b i l i t i e s can be
misleading under conditions of low motivation because subjects w i l l be i n
c l i n e d to use "everyday" categories. With greater motivation, however, sub
j e c t s may be able to make f i n e r d i s c r i m i n a t i o n (see Brown & Lenneberg,
1958), Although many studies i n adult speech perception have ignored t h i s
problem, research i n psychophysical acoustics suggests that motivation i s
15
c r i t i c a l . Samuel (1977) has shown that w i t h t r a i n i n g (and thus feedback) ,
a d u l t s may be able to d i s c r i m i n a t e w i t h i n category speech s t i m u l i w i t h
shor t onset t ime . Al though he c i t e s t h i s as e v i d e n c e ' f o r p e r c e p t i o n i n bo th
a phone t i c ( c a t e g o r i c a l ) and a c o u s t i c (cont inuous) mode, i t may be i n t e r p r e
ted as evidence for . the importance of .mot iva t ion .
i n summary, a rev iew o f the l i t e r a t u r e suggests that i n f a n t s may have
g rea te r d i s c r i m i n a t i v e speech a b i l i t i e s than adu l t s and tha t exper ience
w i t h language may narrow d i s c r i m i n a t i o n a b i l i t i e s . To da te , however, t h i s
ques t i on has not been examined p r o p e r l y i n a s i n g l e s t u d y . w i t h i d e n t i c a l
tokens and w i t h procedures that a l l o w f o r comparable i n d i v i d u a l da t a .
I n f a n t / A d u l t Research ,
Trehub (1976) desc r ibed a s e r i e s of ..four d i f f e r e n t experiments that
were grouped fo r a n a l y s i s that most c l o s e l y address the hypo thes i s of t h i s
study i n tha t she t r i e d to compare r e su l t s , of i n f a n t and adu l t exper iments .
In experiments I and I I , Eng l i sh -Canad ian i n f a n t s aged 5-17 weeks were
t e s t e d i n a HAS paradigm f o r t h e i r a b i l i t y to d i s c r i m i n a t e the o r a l n a s a l
vowel d i s t i n c t i o n which occurs on ly i n French and P o l i s h /pa//f>a/, and the
d i s t i n c t i v e fea ture o f s t r i d e n c y e x e m p l i f i e d by the Czech / z a / / r a / . The
group r e s u l t s suggested that i n f a n t s cou ld d i s c r i m i n a t e these fea tures
(see F igu re I ) . Trehub then employed a change-no change paradigm to d e t e r
mine whether E n g l i s h - s p e a k i n g adu l t s cou ld d i s c r i m i n a t e the Czech p a i r . A
s i g n a l d e t e c t i o n a n a l y s i s i m p l i e d s u b s t a n t i a l confus ion f o r the two Czech
sounds (d '=,83 compared wfth d '=1.00 fo r the common E n g l i s h p a i r / b a / / d a / ) .
Trehub then compared the adu l t da ta to the i n f a n t data o f Experiment I I ,
and to, some 1972 / b a / / d a / i n f a n t da t a . The two se t s of i n f a n t da ta were
analyzed by an a n a l y s i s o f v a r i a n c e (group x the suck ing r a t e i n the 5 p o s t
decrement minutes x language),. There was no s i g n i f i c a n t i n t e r a c t i o n between
16
experimental and control groups and language suggesting equal discrimina-.
t i o n a b i l i t i e s f o r English and Czech contrasts. When compared to the
adult data t h i s suggests English infants have greater d i s c r i m i n a t i o n a b i l i
t i e s than the adults. As Trehub h e r s e l f points out, however, the use of
d i f f e r e n t paradigms i n d i f f e r e n t labs makes such comparisons only specula
t i v e . A d d i t i o n a l l y , the infant findings were based on group (HAS) data;
responses had to be averaged over the postdecrement time period whereas
adult r e s u l t s were based on di s c r e t e i n d i v i d u a l data (although grouped f o r
an a l y s i s ) .
P o t e n t i a l l y relevant findings have also been reported very b r i e f l y by .
Bower (1977). When l i s t e n i n g to the speech of t h e i r own language, adults
perform a "dance" of subtle body movement which r e f l e c t s the d i s c r e t e units
of the continuous speech input (Condon & Ogsten, 1971). Without giving
any.experimental d e t a i l s , Bower, claims that while adults perform t h i s
"dance".onlyhinsresponse to t h e i r native language^ i n f a n t s "dance" to the
speech of any language.
Indirect evidence supporting the hypothesis that i n f a n t s may be better
able to discriminate non-native sounds than adults i s provided by studies
of second-language learning i n c h i l d r e n . The b i o l o g i c a l argument of the
s t r i c t - c r i t i c a l - p e r i o d hypothesis as put f o r t h by P e n f i e l d and Roberts
(1959) and by Lenneberg (1967) suggests that the c r i t i c a l period for language
a c q u i s i t i o n l a s t s from about age two u n t i l puberty, and is.due to the lack
of complete hemispheric s p e c i a l i z a t i o n . Although t h i s hypothesis has been
c r i t i c i z e d from the point of view of age of l a t e r a l i z a t i o n (Kimura, 1967)
and on the basis of B u r s t a l l ' s hypothesis of a l a t e r optimum age for
language learning ( i n McLaughlin, 1977), the evidence for a c r i t i c a l period
17
F i g . 1 Mean number of sucks per minute, as a percentage of the maximum predecrement sucking rate, for 5 minutes before and a f t e r the decrement c r i t e r i o n .
18
with respect to accent i s strong. I t has been found that the younger the '
c h i l d , the more perfect the prdriounciation i n learning a second language
(Asher & Garcia, 1969). A d d i t i o n a l l y , .case studies of b i l i n g u a l c h i l d r e n
reveal that younger ch i l d r e n seem to do better on the acquistion of phono
l o g i c a l features (McLaughlin, 1977), demonstrating a greater perceptual
f l e x i b i l i t y i n younger than i n older c h i l d r e n . F i n a l l y , recent speech per
ception data of adult b i l i n g u a l s suggests t h e i r perception of acoustic con-
tinua r e f l e c t s the phonetic categories of both languages ' (Carramazza, Yeni-
Komshian, Z u r i t & Carbone, 1973; Williams,.1975).
The hypothesis examined i n t h i s study i s a c t u a l l y intermediate to the
l i n g u i s t i c and the acoustic cue hypothesis, According to the present hypo
th e s i s , the infant has the a b i l i t y to respond to the universal set of
phonemic d i s t i n c t i o n s at b i r t h . Whether t h i s a b i l i t y i s purely acoustic
or l i n g u i s t i c i s not important. What i s important i s that since the i n
fants can discriminate the u n i v e r s a l set of l i n g u i s t i c features, they are
able to break the continuous flow of speech input into i t s d i s c r e t e fcap
tures, and thus begin segmenting the meaningful aspects of l i n g u i s t i c i n r -
put.. This predisposes infants to acquire t h e i r native language'with
greater ease than i f they.had to learn to discriminate the l i n g u i s t i c fea^
tures. The purpose of t h i s study was to test the hypothesis that through
development the individualsbbegin t o o r g a n i z e t h e i r speech perception cate
gories to moire and more c l o s e l y approximate, the phonetic categories used i n
t h e i r own language. This a b i l i t y may be purely auditory at b i r t h and be
come encoded l i n g u i s t i c a l l y through ontogeny, or i t may,be a s p e c i f i c
language a b i l i t y at b i r t h , as some neuropsychological ( G l a n v i l l e , Levenson
& Best, 1977; Kimura, 1967) and anatomical (Witelson & P a i l l i e , 1973) e v i
dence suggests.
19
METHOD
Subjects
Four groups of volunteer- subjects were tested i n a d i s c r i m i n a t i o n
paradigm (described i n the next section) on two pairs of Hindi speech con
trasts .
Group T consisted of f i v e Hindi speaking adults (three, males, two
females) aged 22-35. The subjects were rec r u i t e d through advertising and
word-of-mouth from the Vancouver community.
Group IX consisted of f i f t e e n infants ( s i x males, nine females) rang
ing in,age from 6 months,' 7 days to 1 months, 23 days,.with an average age
of 6 months, 28 days. The infants were re c r u i t e d from the community by
adv e r t i s i n g i n newspapers,, at well-baby c l i n i c s , and by telephoning people.
l i s t e d i n b i r t h announcements. Although English was the p r i n c i p l e language
spoken i n a l l the infant homes, ad d i t i o n a l languages were spoken i n four of
the homes.
Groups I I I and TV each consisted of ten English-speaking adults ( s i x
males, four females i n each group) aged 22-35 re c r u i t e d from the. Uni v e r s i t y
of B r i t i s h Columbia campus, As i t w a s ' d i f f i c u l t to f i n d adults with no
foreign language t r a i n i n g , notes were taken on formal t r a i n i n g and on i n f o r
mal exposure for each adult. Group ITT consisted of "naive" adults, whereas
Group TV was given l i m i t e d feedback i n the di s c r i m i n a t i o n paradigm to make
t h e i r task more comparable to^thec.ihfantatask.
Procedure and Apparatus
The procedure and apparatus for the t e s t i n g of the infants i s outlined
i n some d e t a i l as i t i s not widely available i n the l i t e r a t u r e . This pro
cedure i s c a l l e d v i s u a l l y reinforced infant.speech d i s c r i m i n a t i o n (VRISD).
20
As the adult procedure was designed to approximate t h i s infant procedure
as nearly as p o s s i b l e , i t w i l l only be described b r i e f l y . .
In the present study, the experimental set-up consisted of a sound-
attenuated room with a one-way observationwindow adjoining the control room
i n which the experimenter arid l o g i c system were situated. The sound attenur-
ated room contained one'small table, i n the middle foreground of the room.,
A chair was d i r e c t l y behind the table on which the parent and infant were
seated with t h e i r backs to the experimenter. A second chair was located
across the table and s l i g h t l y to the l e f t from the parent/infant on which
the a s s i stant was seated. A speaker,was located i n the back r i g h t corner
of the room,.and a v i s u a l r e i n f o r c e r was located at a 45 degree angle from
the c h i l d ' s l e f t side (See Figure 2), To avoid i n f l u e n c i n g the i n f a n t , the
assistant and parent both wore sound attenuating earphones (see Figures 3
& 4).
The v i s u a l r e i n f o r c e r was an e l e c t r i c a l l y activated toy animal conv
tained iri a smoked p l e x i g l a s s box. The smoked glass made i t po s s i b l e to
see the animal orily when the v i s u a l r e i n f o r c e r was activated and l i g h t s i n
side the box came on. A c t i v a t i o n of the r e i n f o r c e r also made the animal
s t a r t moving and producing noise .(toy bear drums, chimpanzee claps cymbals).
In the VRJSD system, the experimenter presented a sound which did or
did not change. The experimenter and assistant both voted as to whether the
infant responded to a change. If both voted that the i n f a n t had resporided
within the required i n t e r v a l , a r e i n f o r c e r was presented to the i n f a n t .
More s p e c i f i c a l l y , , the e n t i r e VRISD system was c o n t r o l l e d by a l o g i c -
system to ensure maximum experimental control i n the test paradigm throughout
the test period. Sound I (e.g., ba) was f i r s t played through the speaker at
2 sec. i n t e r v a l s . When the experimenter activated the l o g i c system, the
FIGURE 2 Arrangement of the Experimental S i t e
A
TABLE
I
P
OBSERVATION WINDOW -
EQUIPMENT,
E
A = Assistant S = Speaker
VR = V i s u a l Reinforcer I = Infant P = Parent E = Experimenter
24
vote button held by the assistant l i t up to indic a t e that a response i n
t e r v a l was.beginning. The experimenter then selected e i t h e r track.A (con
t r o l track, no change i n stimuli) or t r a c t B (experimental track, i n which
the s t i m u l i changes, e.g.,1 to 4 tokens of da) according to a predetermined
randomized schedule. An adjustable timerwas set at 4-1/2 sec and activated.
During t h i s 4-1/2 sec i n t e r v a l , i f the infant turned' i t s head toward the
sound source, and i f both the experimenter and the ass i s t a n t independently
pressed t h e i r vote buttons the di s c r i m i n a t i o n was assessed to have been made,
arid the v i s u a l r e i n f o r c e s was activated f o r 4 sec. I t was necessary that
both the experimenter and the assistant vote within the s p e c i f i e d time i n
t e r v a l , and that track B had been chosen for the v i s u a l r e i n f o r c e r to be
activated. I f any of those conditions were not met, the r e i n f o r c e r did not
come on, (A manual override was also included i n the system which made i t
possible to activa t e the v i s u a l r e i n f o r c e r at any time to allow for f l e x i
b i l i t y during the conditioning stages of the VRA paradigm),
The VR1SD procedure began with a request to the mother to s i t behind
the table with the infant on her lap. The mother was given headphones de
l i v e r i n g music to prevent her from hearing the speech s t i m u l i . The a s s i s t
ant (who was also wearing headphones) sat across the table to the c h i l d ' s
r i g h t side and showed the c h i l d a s e r i e s of toys to keep him/her happy and
occupied during the t e s t i n g session. The assistan t ' s r o l e was extremely
important to the paradigm, as i t was necessary to,keep the c h i l d passively
interested i n the toys, but not so interested that he/she would disregard
changes i n the speech s t i m u l i . When the infant was atte n t i v e and calm,
with his head facing the toys, the experimenter began a t r i a l by se l e c t i n g
e i t h e r " s t a r t " track A (control) or track B (change) according to a prede-
25
termined random schedule.
The conditioning portion of the paradigm proceeded as follows. A
serie s of one s y l l a b l e sounds ( i . e . , Ba) was played over the speaker. When
the assistant indicated the infant was i n a state of readiness (by pressing
a s i l e n t foot button to activate a l i g h t that the experimenter' could see),
the experimenter changed to Track 2, and Da was played over the speaker.
Immediately following the f i r s t token of Da, the toy animal was activated.
Upon.activation, l i g h t s came on i n the p l e x i g l a s box, and the toy animal
started performing. The noise of the animal attracted the infant's atten
t i o n , and a head turn response was made to see the toy animal (see Figures
3 and 4).' This procedure of presenting the sound stimulus and a c t i v a t i n g
the r e i n f o r c e r was repeated for 2 to 3 more t r i a l s ; The experimenter then
waited u n t i l a f t e r the second token of "Da" before a c t i v a t i n g the r e i n f o r c e r '
to give the infant the a d d i t i o n a l seconds to respond. I f no head turn r e
sponse:, was made, the en t i r e procedure was repeated. As soon as one head
turn response was made upon presentation of-the sound change before a c t i v a
t i o n of the r e i n f o r c e r , a c t i v a t i o n of the toy animal became contingent upon
the i n f a n t making a head turn within 4-1/2 seconds a f t e r a changed stimulus,
and the paradigm was taken over by the l o g i c system. Sixty-eight percent of
the infants formed the association within an average of. nine conditioning
t r i a l s .
Two to four sessions were required to complete the t e s t i n g of each i n
fant. Parents were requested to bring t h e i r infants on days when they had.
no evidence of colds or ear i n f e c t i o n s . Care was taken to ensure that each
infant was comfortable i n the experimental room before t e s t i n g began, and
observations were made on the. infants a b i l i t y to respond to a sudden onset
of sound (they a l l d i d ) . On the f i r s t day of restin g the infant was condi-
26
tioned to one or the other of the Hindi sound pa i r s chosen on a random-
• 3 -U - (3)
lzed b a s i s .
Testing on the other Hindi sound p a i r was then completed on day 1,
or one of the next three testing sessions. Infants that did not condition
on the f i r s t day were given a second.day of t r a i n i n g . I f they did not be
gin to condition within the f i r s t f i v e minutes of t e s t i n g on' the second
day, they were switched to the common English p a i r /Ba//Da/, I f they did
not learn with that p a i r during the session, they were not continued iri
the study. (Mothers would not continue bringing a baby out to an experi
ment t h e i r baby would not succeed a t ) , F a i l u r e to respond to a new sound
change was followed i n the,same session by r e t e s t i n g the infant on a pre
v i o u s l y tested contrast. This was done to tr y to determine whether the
sound p a i r , or the state of the i n f a n t , was responsible for the negative
r e s u l t s .
Procedures as s i m i l a r as possible to these were used with the adult
subjects. The,experimental set-up was nearly i d e n t i c a l , with the subject
s i t t i n g at the table ^'facing the loud speaker. When the subject indicated
readiness, the experimenter would begin the t e s t i n g session. The subject's
task was to push the vote button on a "vote" box when a change i n stimulus
was detected. The-visual r e i n f o r c e r contained the toy animal, and was ac
tiva t e d upon a correct vote i n an experimental t r i a l .
The c r i t e r i o n f or d i s c r i m i n a t i o n was set at 8 out of 10 correct r e
sponses for change t r i a l s ; The mode of presenting c o n t r o l t r i a l s v aried -?ln t h e - l n l t r a l ^desigriZsre• had« planned/to 'cohditioh^infants on /Ba//Da/ and then trans f e r them to the Hin d i . p a i r s . Several problems arose. Most importantly, our mothers, t i r e d of making the long t r i p to.UBC would make up reasons not to return a f t e r the 3rd or 4th session. As i t was pr i m a r i l y important to have a within group comparison ori the 2 p a i r s of Hindi speech sounds, data was therefore c o l l e c t e d on /Ba//Da/ only when po s s i b l e .
s l i g h t l y between the infants and the adults. As i n d i c a t e d e a r l i e r , the
assistant indicated the infant's readiness by s i l e n t l y ' a c t i v a t i n g a l i g h t
for the experimenter to see, During the period the l i g h t was activated,
a l l head turn,responses (including f a l s e p o s i t i v e s ) were recorded. At
l e a s t four times during each seri e s of ten change t r i a l s , the experimenter
would not change the s i g n a l . During the other times the s i g n a l would be
changed between 1 and 8 sec.after a c t i v a t i o n of the "readiness" l i g h t i n
a random fashion. The a s s i s t a n t never knew i f a change t r i a l would occur
and would never know when i n the observation period that change would occur.
The actual c r i t e r i o n was. therefore at l e a s t 12 out of 14.
For adults, change t r i a l s occurred i n an i r r e g u l a r fashion ranging
from 8 to 30 sees. A l l f a l s e p o s i t i v e s were counted, so again the 8 out
of 10 c r i t e r i o n i s not an accurate i n d i c a t o r of the l e v e l of performance
since guessing could have occurred a f t e r every token (every two sec).
The c r i t e r i o n for infants and adults i s defined i n terms of experimenter
delimited observation periods, rather than i n absolute performance.
A second group of English adults was tested with l i m i t e d feedback to
provide a procedure more comparable to that of the i n f a n t s . In t h i s group,
subjects were runii.as i n a conditioning paradigm, with the sound change
paired with a c t i v a t i o n of the r e i n f o r c e r f o r the f i r s t X t r i a l s (where X
was the average number of t r i a l s to c r i t e r i o n fori:the i n f a n t s ) . A l l
adults were tested i n one session.
Stimuli
Three stimulus p a i r s , one English /Ba//Da/, and two Hindi /ta//Ta/ h h
and/t //d /, were used. The vowel (a) was used for a l l s t i m u l i as i t i s
common i n both Hindi and English and i s one of the most frequent vowels to
28
appear with r e t r o f l e x consonants (Stevens & Blumsteln, 1974). ; A l l tones
were made inthe Phonetics laboratory at the Uni v e r s i t y of B r i t i s h Columbia.
Each tape contained 8 natural exemplars of each sound ', This was to ensure
that v a r i a t i o n s i n duration, fundamental frequency and information would
be randomized both within and between categories.
29
RESULTS
Each i n d i v i d u a l i n each group was recorded as having e i t h e r reached
or not reached the 8 out of 10 c r i t e r i o n on the two Hindi p a i r s . An
analysis of proportions was then performed on this data (see Tables 3 and
4). This analysis i s based on the Scheffe theorem (Marascuillo, 1966).
I t i s used to compare', discrete, data i n which the proportions of i n d i v i d u a l s
f a l l i n g into given categories is- compared.
The n u l l hypothesis was that there would be "no s i g n i f i c a n t difference
among the proportion of i n d i v i d u a l s reaching c r i t e r i o n i n the four groups."
The r e j e c t i o n l e v e l f o r this•hypothesis was set at p = ,05.
For the f i r s t comparison (the Hindi'contrast Retroflex/Dental) the
o v e r a l l Chi-square obtained had a p r o b a b i l i t y of le s s than ,05 (p = .0000016)
and thus enabled r e j e c t i o n of the n u l l hypothesis. A s e r i e s of multiple
comparisons between each p a i r of groups yielded the r e s u l t s outlined i n
Table 5.
There was no s i g n i f i c a n t difference between the proportion of Hindi
adults and infants reaching c r i t e r i o n , nor was there a s i g n i f i c a n t d i f f e r e n c e
between the naive and the trained English speaking adults. A l l other com
parisons ' (Hindi vs. each English group and infants vs. each Adult English
group) were s i g n i f i c a n t .
Eor the second comparison (the Hindi contrast unvoiced, aspirated
dental vs. breathy-voiced dental), the o v e r a l l chi-square obtained was not
s i g n i f i c a n t (p .05, p = .0580), so r e j e c t i o n of the n u l l hypothesis was not
possible. However, since the p_ value was so close to .05, a s e r i e s of
multiple comparisons was performed to comb the data for trends. The r e
su l t s are summarized i n Table 6.
30
Table 3
A n a l y s i s of P r o p o r t i o n s f o r
R e t r o f l e x / D e n t a l Cont ras t
Group 1 Group I I Group I I I Group IV ( H i n d i A d u l t s ) ( In fan ts ) (Naive Eng. (Tra ined Eng.
A d u l t s ) adu l t s )
Reached c r i t e r i o n 5
Did not reach cfi€?^rXy:> A ttribif
T o t a l N 5
11
1
*12
1
9
10
0
10
10
Table 4
A n a l y s i s of P r o p o r t i o n s f o r V o i c e l e s s a s p i r a t e d /
Breathy v o i c e d Cont ras t
Group I Group• I I Group 111 Group IV
Reached c r i t e r i o n
D i d not reach c r i t e r i o n
T o t a l N
10
*12 10 10
A l t o g e t h e r , 15 i n f a n t s were ^ t e s t e d i n a l l , b u t 3 would not c o n d i t i o n to the paradigm, even w i t h the E n g l i s h Ba/ba sound p a i r . The remaining 12 cond i t i oned and 9 passed both speech sound p a i r s .
31
Table 5
M u l t i p l e Comparisons on R e t r o f l e x / D e n t a l Cont ras t
Comparison Confidence I n t e r v a l , P rob . of No D i f f e r e n c e
2 - 1 -0.306 to 0,140 .779
3- 1 -1 .165 to -0 .635 *..000 3- 2 -1 .163 to -0 .470 *.000
4- 1 -1 .163 to -0 .470 *.000
4-2 -1 .140 to -0.694 *,000
4-3 -0.365 to 0.165 ,774
Table 6 '
M u l t i p l e Comparisons on V o i c e l e s s A s p i r a t e d / '
Breathy Voiced C o n t r a s t '
Comparison Confidence I n t e r v a l Prob , o f No D i f f e r e n c e x
1 - - v V ' • i. v .
2- 1 -0 .436 to 0.102 .494
3-1 -0 .987 to 0.213 *,002
3- 2 -0 ,905 to 0.038 .153
4- 1 -0 .662 to 0,062 .232
4-2 -0 .585 to 0.318 .909
4-3 -0 .230 to 0,830 ,572
32
The only s i g n i f i c a n t d i f f e r e n c e occurred between the Hindi adults and
the naive English adults. The comparison between the infants and the naive
English adults was the next closest to being a s i g n i f i c a n t d i f f e r e n c e but
s t i l l had a 15% p r o b a b i l i t y of being due e n t i r e l y to chance.'"
Since the VRISD paradigm i s a conditioning paradigm, i t can be c r i t i
cized f o r allowing the infants a chance to "learn 1' the di s c r i m i n a t i o n during
the t e s t i n g session. As outlined i n the method section, Group XV (adults
with training) was included to address t h i s type of c r i t i c i s m . A d d i t i o n a l l y ,
two other sources of data were c o l l e c t e d to t r y to understand the r e l a t i v e
contribution of "learning" within the t e s t i n g session. The f i r s t source was
the i n c l u s i o n of the common.English /Ba/ vs. /Da/ sound p a i r whenever p o s s i
b l e i n infant t e s t i n g . The second source was a comparison of the number
of t r i a l s to e i t h e r c r i t e r i o n or stopping f o r each of the four groups.
It was only possible to collect./Ba/- /Da/' data on four infants^ two
of whomwere tested on,the Hindi,contrasts, and two of whom were only
tested on the English contrasts. Since the two who were only .tested on the
English contrasts were never tested on the Hindi.sounds, and since other-
infants who were tested on the Hindi sounds f a i l e d to reach c r i t e r i o n or
were not tested on the Ba/Da, no chi-square comparisons were made., These
data were analyzed i n terms of mean number of t r i a l s to shaping and to
c r i t e r i o n as summarized i n Table 7. Only i n d i v i d u a l s who reached c r i t e r i o n
were included i n computing these means. T-tests for correlated data were
then performed comparing the means l i s t e d i n Table 7. No s i g n i f i c a n t d i f
ferences were found among the three d i f f e r e n t sound p a i r s i n ei t h e r number
of t r i a l s to shaping, or number of t r i a l s to reaching c r i t e r i o n . Since the
N was so small, a 2-group, one-tailed t - t e s t was performed to compare
33
Table 7
Mean Number of Infant T r i a l s on Speech Contrasts
Infants .
Trial to shaping Trials; to passing
N = 4 Ba/Da 5,5 20.2
N = 11 Ret/Dent 10,9 26,5'
N = 10 Un/ASD 8,1 22,36
34
/Ta/ / t a / and /Ba/ /Da/. Although the d i f f e r e n c e between the means on
number of t r i a l s to c r i t e r i o n was c l e a r l y not s i g n i f i c a n t , the d i f f e r e n c e
on number of t r i a l s to shaping almost reached s i g n i f i c a n c e at p .05 l e v e l b h
(p = .06), A s i m i l a r analysis on / t / /d / vs. /Ba//Da/ did not approach
s i g n i f i c a n c e .
An analysis of variance was run on the number of t r i a l s to c r i t e r i o n
Cor s t o p p i n g ) f o r the four groups on the two p a i r s of Hindi sounds.
C e l l means are shown i n Table 8, and the r e s u l t s of the analysis of v a r i
ance are shown i n Table 9. As can be seen, the main e f f e c t s for both
Group (I-TV) and Sound P a i r were s t a t i s t i c a l l y s i g n i f i c a n t . Using
Tukey's method of planned comparisons, i t was found that the group main
e f f e c t could be accounted for iby a . s i g n i f i c a n t d i f f e r e n c e between the
Group 1 (Hindi adults) and a l l other groups on number of t r i a l s to c r i
t e r i o n . No s i g n i f i c a n t differences emerged between Groups T l , I I I and
IV. The other main e f f e c t was simply that o v e r a l l , the unvoiced, unaspir-
ated/unvoiced aspirated sound p a i r was easier than the r e t r o f l e x / d e n t a l . '
(Ba//Da/ could not be included i n the ANOVA as i t would not f i t into a
repeated measures design).
Tnfarits and adults were run as long as possible during a t e s t i n g session i n the hope that they would reach c r i t e r i o n , Whenever an infant became overly fussy, or sleepy during a t e s t i n g session, we would stop.' Conversely when an adult started s i g n a l l i n g he/she was fatigued we would stop.' A d d i t i o n a l l y , t e s t i n g of adults who f a i l e d to show any recognition of the sound change was terminated a f t e r 25 t r i a l s .
35
Table 8
C e l l Means (M) and Standard D e v i a t i o n s (SD) f o r Number of T r i a l s to C r i t e r i o n
Group I Group I I G r o u p . I l l Group IV
M SD M SD M SD M SD
Ta/ta 10.4 .548 27.58 9.65 25,6 10.38 27.7 3.498
T h / D h 10.2 .447 22.5 11.31 18.5 5.60 1 9 . 3 ' 5.056
Table 9
Source Table f o r Number o f T r i a l s to C r i t e r i o n ; A n a l y s i s of Var i ance
Sum of Squares Degrees of Freedom
MS F r a t i o
Between Group 2247.099 3 749.03 13,307*
E r r o r 1857.516 33 56.29
Sound P a i r 446.842 1 446.84 6.797*
Grp x S n d p a i r 160,814 3 53.61 0.815
W i t h i n Groups 2169.508 33 65.74
• • • \ • •
36
DISCUSSION
The r e s u l t s of t h i s study y i e l d support for the hypothesis:
^Humans are born,with the a b i l i t y to discriminate.the universal set of d i s t i n c t i v e l i n g u i s t i c features. Through development there i s a decline i n t h i s a b i l i t y as the perceptual space becomes organized to approximate the phonemic categories used i n the native language".
There was no d i f f e r e n c e between the infants and the Hindi adults, or be
tween the two groups of English speaking adults, but there were s i g n i f i
cant differences i n discriminatory a b i l i t i e s between a l l other p a i r s of
groups for the r e t r o f l e x / d e n t a l d i s t i n c t i o n , A higher proportion of the
infants and the Hindi adults could discriminate t h i s Hindi contrast than
could e i t h e r group of English speaking adults, English adults with l i m i
ted t r a i n i n g performed no differentafrom English adults without t r a i n i n g ,
showing that minimal t r a i n i n g with adults did not f a c i l i t a t e the d i s c r i
mination. I n t e r e s t i n g l y , the one adult English speaker who was able to
discriminate the contrast reported that seven years p r i o r to his being
tested on the Hindi contrast he had spoken T w i f o r a period of three
months, Retroflex/Dental i s used i n Twi to d i f f e r e n t i a t e phonemic cate- ,
gories (Fromkin, 1974), This subject, l i k e the Hindi speakers, reported
the contrast as being very obvious, and was surprised others could not
discriminate the difference,
Results from the comparison of the four groups on the other Hindi h h
sound p a i r , /d / and / t / were les s c l e a r , Although the analysis of
proportions did not quite reach s i g n i f i c a n c e planned comparisons were
performed to try to determine whether there was any pattern to the pos- 1
s i b l e difference between the four groups. ' The only s i g n i f i c a n t d i f f e r e n c e
37
was between the Hindi adults and the naive English adults. I t should be
pointed out that two infants out of-twelve f a i l e d to perform t h i s d i s
crimination. One.of'the i n f a n t s , however, was very c l e a r l y able to make
th i s d i s c r i m i n a t i o n , but simply f a i l e d to perform the required task. Dur-r
ing the t e s t i n g session, she would turn her eyes toward the speaker when
ever the sound changed, but she would, not perform a f u l l head turn. Eye
movements were not r e i n f o r c e d , nor could they be counted as correct r e
sponses. Had we been able to count them, however, t h i s infant would have
reached c r i t e r i o n , .arid the o v e r a l l Chi^square for the comparison would
have reached s i g n i f i c a n c e (p = .02). In a d d i t i o n , the i n f a n t vs. naive
adults comparison would have reached s i g n i f i c a n c e . (p = .03).; The compari
son;- between the E n g l i s h adults with t r a i n i n g , and both the Hindi adults
and English infants would s t i l l not have been s i g n i f i c a n t , however.
Thus, at best, the i n c l u s i o n of t h i s infant i n the analysis would have
lent only l i m i t e d support to the main hypothesis;
In an attempt to determine whether the Hindi. contrasts required a
greater number of t r i a l s to c r i t e r i o n than did the E n g l i s h contrast /Ba/
/Da/, t - t e s t s were performed on the. i n f a n t data. Although the means were
greater f o r the Hindi contrasts (see Table 7), t h i s d i f f e r e n c e did not
reach s i g n i f i c a n c e i n a correlated one-tailed t - t e s t . This f i n d i n g was.
s u r p r i s i n g since i t was predicted that the English contrast would,be
somewhat easi e r , at l e a s t i n shaping t r i a l s , sirice many of the i n f a n t s
were already r e p e a t i n g /ba/'s and /da/'s i n both t h e i r i m i t a t i v e and
spontaneous babbling. This lack of s i g n i f i c a n c e suggests that even though
some language appropriate sounds have entered the i n f a n t s ' productive
38
r e p e r t o i r e , f l e x i b i l i t y may s t i l l be ma in ta ined ,a t the p e r c e p t u a l l e v e l .
An a n a l y s i s of v a r i a n c e conducted on number o f t r i a l s to c r i t e r i o n
y i e l d e d two s i g n i f i c a n t e f f e c t s . "Groups" and "Sound P a i r s " . Planned
comparisons showed o n l y the d i f f e r e n c e between H i n d i a d u l t s and a l l o ther
groups t o . b e ' s i g n i f i c a n t . ' This i s not s u r p r i s i n g g iven that H i n d i
adu l t s performed s i m i l a r l y to E n g l i s h speakers i n a p i l o t experiment con
ducted u s i n g E n g l i s h con t r a s t s as s t i m u l i , That i s , they reached p e r f e c t
scores almost immediately on n a t i v e c o n t r a s t s . Of more i n t e r e s t , how
eve r , i s the l a c k of s i g n i f i c a n c e between p a i r s of the o the r three groups.
The i n f a n t s may have faced a more d i f f i c u l t t ask than the a d u l t s f o r
three reasons.'
(1) The i n f a n t s had to c o n t i n u a l l y d i v i d e t h e i r a t t e n t i o n between
the toys be ing manipulated by the a s s i s t a n t and the speech sounds. The
adu l t s on the o ther hand, always had t h e i r a t t e n t i o n d i r e c t e d a t . t h e
speech sounds,
(2) Al though the v i s u a l r e i n f o r c e r may have been more•rewarding to
the i n f a n t s than to the a d u l t s , presumably the fear o f f a i l u r e was not as
s t r o n g . (Green and Swets, 1966) have suggested tha t fear o f f a i l u r e may
be equa l t o monetary re inforcements as a m o t i v a t i n g f o r c e fo r a d u l t s .
C e r t a i n l y , the a p o l o g i e s , r a t i o n a l e s , and compla in ts o f f e r e d by each
adu l t sub jec t who d i d not reach c r i t e r i o n onthese c o n t r a s t s y i e l d e d
s u b j e c t i v e support f o r t h i s n o t i o n !
(3) Infants are desc r ibed as be ing more a f f ec t ed by s t a t e f l u c t u a
t i o n s (E i senberg , 1976) than adu l t s and should thus be expected to make
more e r r o r s on an e q u a l l y easy task due to these s t a t e f l u c t u a t i o n s . That
i s , be ing t i r e d , hungry, e t c , . , i s desc r ibed as i n t e r f e r i n g much more
39
with an infant's attention to a task than would s i m i l a r f l u c t u a t i o n s
i n an adult. A d d i t i o n a l l y , day to day differences i n q u a l i t i e s such
as s o c i a b i l i t y , excitement, fear of strange places, etc, a f f e c t an i n
fant ' s perf ormance i n a c o n t r o l l e d experimental s e t t i n g . ' The e f f e c t s
of such factors on thel performance of adults i s assumed to be minimal.
Since there was no s i g n i f i c a n t difference between the number of
t r i a l s to c r i t e r i o n (or stopping) for adults and i n f a n t s , . i t cannot be
argued that.the infants were simply given a greater number of t r i a l s to
" l e a r n " the contrasts. Thus the hypothesis'that infants can discriminate
non-native sounds with l i t t l e or no learning although adults may not be
able t o d o so was substantiated,
There i s some question as to whether VRTSD should continue to be
considered a conditioning paradigm or whether,it should be reconceived
as a d i s c r i m i n a t i o n paradigm with feedback. Since the monkey iri a box
can hardly.be seen as a potent adult r e i n f o r c e r , one could/argue.. ^that
fear offfailure.wastthe -motivating!:foreeDbehirid^adult^pefformahce^.
A d d i t i o n a l l y , infant behavior suggested some kirid of "competence" motiva
t i o n rather than a.direct stimulus r e i n f o r c e r as being primary. Once
an infant had "learned" to perform a head turn response upon a change
i n the. background stimulus the r e i n f o r c i n g value of the toy seemed to'
be diminished, A f t e r several t r i a l s , i n fants would t y p i c a l l y concentrate
on the assistant manipulatirig toys, would then swing t h e i r head around
toward the speaker and back again when the speech sound changed and
continue watching the assistant rather than the r e i n f o r c e r , ' I t seemed
as i f infants only wanted to know they could acti v a t e the toy animal,
40
but did not p a r t i c u l a r l y care to watch i t . In addition, i t would be
hard to argue that subjects could " l e a r n " which acoustic signals they
should heed given the multiple natural tokens and the l i m i t e d number
of t r i a l s .
The.second main e f f e c t "Sound P a i r s " was a t t r i b u t a b l e to / t ^ / / d ^ /
taking f e w e r . t r i a l s to discriminate for a l l groups,, Such a r e s u l t i s
consistent with the r e s u l t s from both the Analysis of Porportions and
from the t - t e s t s i n that both these s t a t i s t i c a l analyses gave some siip^
port f o r the notion that /t*V/d*V was-intermediate i n d i f f i c u l t y between
the common English /Ba//Da/ and the, more rare Hindi /Ta//ta/ contrasts.
Two explanations could be offered for t h i s d i fference,
(1) A "language experience" confound could have been possible for
two reasons.' F i r s t , v o i c e l e s s , aspirated vs. breathy voiced stops are
used phonemically across a wider range of natural languages'than i s the
r e t r o f l e x vs. dental place of a r t i c u l a t i o n d i s t i n c t i o n . Thus adults,
and even i n f a n t s , would be more l i k e l y to have had exposure to t h i s
sound contrast.
Second, there i s controversy as to the correct d e s c r i p t i o n of the
four categories of voice onset time used iri Hindi, Although many l i n : r •
guists(e.g., Ladefoged, 1975) assert that breathy v o i c i n g must be desr-
cribed as a unique category of voice onset time, some l i n g u i s t s have
described breathy v o i c i n g as being a voiced, aspirated stop, ' I f t h i s
l a t t e r d e s c r i p t i o n i s correct, one could argue that the.voiceless vs.
voiced d i s t i n c t i o n could provide English l i s t e n e r s with a p a r t i a l cue
to the d i s c r i m i n a t i o n . (English d i f f e r e n t i a t e s voice vs. v o i c e l e s s un^
aspirated stops).
41
C2) An explanation based on the notion of "perceptual distance" i s
also p o s s i b l e . Tt may be that although the human ear can nec e s s a r i l y
discriminate a l l phonemic d i f f e r e n c e s , the perceptual d i s s i m i l a r i t y of
some'of these differences i s greater f o r some feature d i s t i n c t i o n s than
i t i s for others. In this regard a s t r i c t l y phonological d e s c r i p t i o n
(based on meaningful articulat'ory differences) may not always be complet
The r e l i a n c e on such a d e s c r i p t i o n i n t r a d i t i o n a l l i n g u i s t i c s rather
than oh a phonetic d e s c r i p t i o n (based more on acoustic differences) may
have been misleading. That i s , a s i n g l e feature d i f f e r e n c e i s a useful
t o o l f o r explaining phenomena that y i e l d a perceptual (and meaningful)
invar i a n t but not an acoustic one. This does not, however, imply pere
ceptual (or acoustic) equality among a l l feature differences,'
A perceptual distance explanation i n i t i a l l y appears to involve an
acoustic rather than a l i n g u i s t i c , speech perception mechanism. On
clo s e r examination however, a more complex picture emerges, suggesting
p o s s i b l e p a r a l l e l l e v e l s of processing with d i f f e r e n t i a l access to an
acoustic or a l i n g u i s t i c l e v e l depending upon the most'effective strateg
for the task.
In summary, the infants were found to discriminate a l l sound pa i r s
according to phonemic category with comparable ease (with a s l i g h t sugr
gestion that the Hindi /Ta//ta/ was more d i f f i c u l t to shape), Tt was,
also shown that Hindi adults could discriminate, a l l sound p a i r s , whereas
English adults could not discriminate /Ta//ta/ at a l l , and only some
h h h h English adults could discriminate / t //d /', O v e r a l l , the ft //d / d i s crimination appeared to be easier than the /Ta//ta/ d i s c r i m i n a t i o n . An important a d d i t i o n a l piece of information concerns the categorizing
phenomenon.observed, That i s , i n d i v i d u a l s whocould perform to. c r i
t e r i o n for a sound p a i r would continue to perform at that l e v e l through
out a t e s t i n g session without an increase i n f a l s e p o s i t i v e rate, (In
fants would t i r e a f t e r an a d d i t i o n a l 10 t r i a l s or so). Once a c r i t e r i o n
was established, the i n d i v i d u a l s did not return to p r e - c r i t e r i o n l e v e l s
of performance. ' Because natural tokens were used, there were obvious
differences between each token within a speech sound category, Those d i f
ferences were only responded to by i n d i v i d u a l s who could not reach c r i
t e r i o n on a sound pair i . Individuals who categorized the sounds (reached
c r i t e r i o n at the phoneme l e v e l ) did not respond to within category
v a r i a t i o n , although they could hear differences when asked.
If i t i s accepted that neither Hindi sound p a i r involved tokens that
are used phonemically i n English ( i . e . , i f the language experience con
found explanation i s discovered) recourse to a categorizing explanation
based on experience with a language i s not po s s i b l e . The explanation
1 would l i k e to o f f e r i s thus based on the notion of twor-level of proces
sing, an acoustic and a l i n g u i s t i c l e v e l , much as suggested by Wood
(1974). When possibl e , people seem to employ a l i n g u i s t i c l e v e l of pro
cessing, and respond to the speech sounds i n a c a t e g o r i c a l manner d i s r e
garding within category differences, When t h i s strategy f a i l s , they
employ an acoustic l e v e l of processing,
An acoustic phonetic analysis of the speech sounds chosen for t h i s
studysupports such an explanation. The phonological and phonetic des
c r i p t i o n s of r e t r o f l e x vs. dental sounds are i d e n t i c a l . The phonetic
d e s c r i p t i o n of v o i c e l e s s aspirated vs., breathy voiced i s d i f ferent, howr,
ever, from the phonological d e s c r i p t i o n . That i s , although these sounds
43
d i f f e r i n only a s i n g l e phonemic feature, the voice onset time d i f f e r
ence may be at l e a s t two steps apart on a continuum. This i s because
the Hindi murmured stops are neither voiced nor aspirated, Tt could be
assumed' that they a c t u a l l y d i f f e r i n at le a s t two ways from vo i c e l e s s
aspirated stops (Ladefoged, 1975), Tt could then be argued that the peril h
ceptual distance between ft / and/d / i s greater than.that f o r /Ta//ta/.
Thus employing an acoustic l e v e l of processing would be more successful
for t h i s sound p a i r , and could lead to more accurate d i s c r i m i n a t i o n by
non-native speakers. Since-the data suggested that phonemic discrimina
t i o n of /t*V/d*V was easier (required fewer t r i a l s to reach c r i t e r i o n )
for a l l groups than /Ta//ta/, a l i n g u i s t i c l e v e l of processing can also
be invoked enabling l i s t e n e r s to ignore acoustic differences that' are
i r r e l e v a n t to natural languages.
44
CONCLUSIONS
This study yielded d e f i n i t e support for the:idea that infants
possess the propensity to discriminate the universal set of l i n g u i s t i c ,
features. The checks b u i l t into the study make i t d i f f i c u l t to argue
that infants were simply learning these contrasts i n a very few t r i a l s .
The r e s u l t s therefore support the view that i n f a n t s actively* impose'a
structure (but a structure constrained by b i o l o g i c a l perceptual capar
b i l i t i e s ) on continuous'auditory input, ;, Such an innate.constraint would
allow the infant to segment the continuous input of speech into the units
used to convey,meaningful differences within the language •. environment to
which the infant i s exposed.
Some support was given for the idea that there may be a decline i n
speech perceptual a b i l i t i e s with "either age/or l i n g u i s t i c experience.
The f i n d i n g that t h i s decline was more evident with one than the other
Hindi sound p a i r makes i n t e r p r e t a t i o n more d i f f i c u l t , . Two explanations
were offered f o r these r e s u l t s . It. was suggested that the Non-Hindi h h
population may not have been naive with respect, to the / t //d / c o n t r a s t .
A l t e r n a t i v e l y , it: was suggested that perceptual distance may have
accounted for-the d i f f e r e n c e s , and two l e v e l s of processing were involved
i n t h i s explanation,
REFERENCES
Abbs, M . S . , & M i n i f i e , F . D , E f f e c t of a c o u s t i c cues i n f r i c a t i v e s on
pe rcep tua l confus ion i n p r e schoo l c h i l d r e n . J o u r n a l o f the
A c o u s t i c S o c i e t y o f - A m e r i c a , .1969, 46, ' 1535-1542.
Abramson, A . S . , & L i s k e r , L . ' D i s c r i m i n a b i l i t y a long the v o i c i n g c o n
t inuum; Cross language t e s t s . I n Proceedings o f the S i x t h I n t e r n a
t i o n a l Congress o f Phone t i c Sc i ences . Prague: Academia, .1970.
Asher , J , J . ' , & G a r c i a , R, The o p t i m a l age to l e a r n a f o r e i g n language.
Modern Language J o u r n a l , 1969, 53, 334-341,
Bower, T . G . R . • A pr imer o f i n f a n t development, San F r a n c i s c o : W.H.
Freeman arid Company, 1977.
Brown, R . W . , & Lenneberg, E , H , S tud ies i n l i n g u i s t i c r e l a t i v i t y . In
E . E . Maccoby, T . M . Newcombe and E . L . H a r t l e y , Readings i n S o c i a l .
Psycho logy . New Y o r k ; Henry H o l t & Company, 1958,
Bruner , J , S . • On pe rcep tua l r e a d i n e s s , P s y c h o l o g i c a l Review, 1957, 64,
123-152.
Carramazza, A . , Yeni -Komshian , G . N . , ' Z u r i f , . E . B , , & Carbone; E . The
a c q u i s i t i o n of a new p h o n o l o g i c a l c o n t r a s t : The case of stop c o n -
sonants i n F r e n c h - E n g l i s h b i l i r i g u a l s , J o u r n a l of ^the A c o u s t i c So
c i e t y of Amer ica , 1973, 54, 421-428.
Chomsky , N . , & H u l l e , M. The sound p a t t e r n o f E n g l i s h . New Y o r k :
Harper & Row, 1968,
Condon, W . S . , & Ogston, W,D, Speech and body motion synchrony, o f the
speaker -hearer , Iri D . L , Horton arid J . J . - J enk ins (Eds',) , The p e r
c e p t i o n of language, Columbus, Oh io : Char les E . M e r r i l l P u b l i s h i n g
Company, 1971. 1
46
D i x , M . R . , & H a l l p i k e , C . S . The peep show: New technique f o r pure tone
audiometers i n young c h i l d r e n . B r i t i s h M e d i c a l J o u r n a l , 1947, 719.
E i l e r s , R. , & M i n i f i e , F , . F r i c a t i v e d i s c r i m i n a t i o n i n e a r l y i n f a n c y .
J o u r n a l of Speech and Hear ing Research, .1975, 18 , 158r-167.
E i l e r s , R. , W i l s o n , W . R , , Moore, ' J , M ; Speech p e r c e p t i o n i n the language
innocent arid the language wises A study i n the p e r c e p t i o n of v o i c e
onset t ime . I n P r e s s : J o u r n a l o f C h i l d Language,
Eimas, P . D . , S i g u e l a n d , E , R , , Ju sczyk , P . , : & V i g o r i t o , 3] Speech percep t
t i o n i n i n f a n t s . S c i e n c e , 1971, 171, 303r-306.
Eimas, P . D , A u d i t o r y and l i n g u i s t i c processes of cues f o r p l a c e o f a r t i
c u l a t i o n by i n f a n t s . P e r c e p t i o n , and Psychophys ics , ' 1974, 16_,, 513-
521,
Eimas, P . D . Speech p e r c e p t i o n i n e a r l y i n f a n c y , In L . B , Cohen and P .
Salapatek ( E d s . ) , In fan t p e r c e p t i o n , ' New Y o r k : Academic P r e s s , 1975.
Eimas, P . D . A u d i t o r y and phone t i c coding o f the cues f o r speech: D i s
c r i m i n a t i o n of the ( r -1 ) d i s t i n c t i o n by young i n f a n t s , P e r c e p t i o n
and Psychophys i c s , 1975, 18, 341-347.
E i s e n b e r g , R . B . A u d i t o r y competence : i n e a r l y l i f e : The roo t s of communi
c a t i v e b e h a v i o r . B a l t i m o r e : U n i v e r s i t y Park P r e s s , 1976,
F romkih , V . , & Rodman, R, An i n t r o d u c t i o n to language, H o l t , .Rinehart &
Wins ton , , I n c . , 1974,
G ibson , E . J . ' P r i n c i p l e s , o f pe r cep tua l l ea r r i i r ig .arid development. • New
Y o r k : Apple tonr -Century-Cfof t s , Mered i th C o r p o r a t i o n , 1969.
G i l b e r t , J , H , V . "Speech p e r c e p t i o n i n c h i l d r e n 1 ' . In A . Cohen & S . E .
Nooteboom ( E d s . ) , S t r u c t u r e arid process i i i speech p e r c e p t i o n .
H e i d e l b u r g : Spr inge r V e r l a g , 1975, pp. 312-338.
47
G l a n v i l l e , B.B., Levenson, R. , & Best, C, A cardiac measure.of cerebral
asymmetries i n infant auditory perception., Developmental Psychology,
1977, 1^.(1), 54-59.
Green, D,M,, & Swets, J.A. Signal detection theory and psychophysics.'
New York: John Wiley, .1966,
Hillenbrand, J.-, •& M i n i f i e , P.D., & Edwards, T.J. Tempo of.formant f r e
quency change as a cue i n infant speech discrimination,'! Paper pre
sented at the b i e n n i a l meeting for the Society f o r Research i n
Child Development, New,Orleans, Louisiana, 1977.
Holmberg, T.L,, Morgan, K,A,, & Kuhl, P,A. Speech perception i n early
infancy: Discrimination of f r i c a t i v e consonants. Journal of the
Acoustical Society of America, 1977, 62, Supplement 1, S99.
Jakobson, R,, & H a l l e , M, Fundamentals of language. The Hague: Mouton
Publishers, 1956.
Jusczyk, P.W. Perception of s y l l a b l e - f i n a l stop consonants by two-month
old i n f a n t s . Perception and Psychophysics, 1977, ,21, 450-454.
Jusczyk, P.W., Copari,.H.C., & Thompson, E,J. Perception of glides i n
m u l t i s y l l a b i c utterances by i n f a n t s . Paper presented at: 94th
meeting, Ac o u s t i c a l Society of America, Miami, F l o r i d a , December
16, 1977,
Jusczyk, P.W.,- Rosner, B., Cutting, J, , Fourd, C , & Smith, L. Categori
c a l perception of non-speech sounds by two-month-old i n f a n t s .
Perception & Psychophysics, 1977, 21, 50-54.
Kimura, D. ' Functional asymmetry of the b r a i n i n d i c h o t i c l i s t e n i n g .
Cortex, 1967, 3, 163-178.
48
K u h l , P . K . Speech p e r c e p t i o n in , e a r l y i n f a n c y : P e r c e p t u a l constancy f o r
vowel c a t e g o r i e s . J o u r n a l of the A c o u s t i c S o c i e t y o f Amer ica , 1976,
60, Supplement S90,
K u h l , P . K . , & M i l l e r j J . D , Speech p e r c e p t i o n by the c h i n c h i l l a : V o i c e d -
v o i c e l e s s d i s t i n c t i o n i n a l v e o l a r p l o s i v e consonants . Sc i ence ,
1975, 190, 69-72.
Ladefoged, P . A course i n p h o n e t i c s . ISfew Y o r k : Harcour t Brace
J o v a h o v i c h , I n c . , 1975,
L a i n e , M, (Ed.) I n t r o d u c t i o n to s t r u c t u r a l i s m . . New Y o r k : B a s i c Books;
I n c . , P u b l i s h e r s , 1970.
Lasky^ R . E . , S y r d a l - L a s k y , A , , ,& K l e i n , R . E . VOT d i s c r i m i n a t i o n by four-
and s i x .and a , h a l f month o l d i n f a n t s from Spanish environments .
J o u r n a l of Exper imenta l C h i l d - P s y c h o l o g y , 1975, 20_, 215T-225.
Lenneberg, E . H . B i o l o g i c a l foundat ions o f language, New Y o r k : W i l e y ,
1967.
Liberman, A . M . , Cooper, F . , Shankwei le r , D, , & Studdert-Kennedy, M.
P e r c e p t i o n of the speech code. P s y c h o l o g i c a l Review, 1967, 74,
. 431-459.
L i s k e r , L . , & Abramson, A . S , . The v o i c i n g dimensionst Some experiments i n
comparative p h o n e t i c s . In Proceedings of the S i x t h I n t e r n a t i o n a l
Congress o f Phone t ic Sc i ences . Prague: Academia, 1970,
M a r a s c u i l o , L . A . Large s c a l e m u l t i p l e comparisons. P s y c h o l o g i c a l Bu-ller-
t i n , 1966, 69.(5), 280-290.
M c L a u g h l i n , B , Secondr-language l e a r n i n g i n c h i l d r e n . P s y c h o l o g i c a l
B u l l e t i n , 1977, 8 4 , 4 3 8 - 4 3 9 ,
49
M i l l e r , N . E . Theory and experiment r e l a t i n g p s y c h o a n a l y t i c displacement
to s t imulus - response g e n e r a l i z a t i o n . J o u r n a l o f Abnormal S o c i a l
Psycho logy , 1948, 4 3 , ' 1 5 5 - 1 7 8 ,
M i l l e r , G .A . , & N i c e l y , P . An a n a l y s i s of pe r cep tua l confus ions among
some E n g l i s h consonants, J o u r n a l of .the A c o u s t i c a l S o c i e t y of.
Amer ica , 1955., 27, 338-352.
M i n i f i e , F, 'D, ( P r i n c i p a l i n v e s t i g a t o r ) , N , T , H . P rogress Repor t , November,
1976.
M i y a w a k i , K . , Strange , .W, , Verberge y Llberma.n, , A . M , , & Fu j imura , 0 . >
An e f f e c t of l i n g u i s t i c expe r i ence i .The d i s c r i m i n a t i o n of ( r ) and
(1) by n a t i v e speakers of Japanese and E n g l i s h , P e r c e p t i o n and
Psychophys i c s , 1975, 18 ( 5 ) , 331-340, -
M o f f i t , A . R . Consonant cue p e r c e p t i o n by twenty^ t o ' t w e n t y - f o u r week
o l d i n f a n t s . ; C h i l d Development, 1971, 42, 717-731.
M o f f i t , A . R , C a t e g o r i c a l d i s c r i m i n a t i o n among the -acoust ic cues fo r
b i l a b i a l -stop consonants by young i n f a n t s . (Unpublished m a n u s c r i p t ) .
Moore, J . M , , W i l s o n , W , R . , & Thompson, G. V i s u a l re inforcement o f head-
t u r n . responses i n i n f a n t s under twelve months of a g e . J o u r n a l o f
Speech arid Hear ing D i s o r d e r s , 1976.
Morse, P . A . Infant speech p e r c e p t i o n : A p r e l i m i n a r y model and review o f
the l i t e r a t u r e . I n R . L , S c h i e f e l b u s h and L . L . L l o y d (Eds , )
Language p e r s p e c t i v e s - a c q u i s i t i o n , r e t a r d a t i o n , and i n t e r v e n t i o n .
B a l t i m o r e : U n i v e r s i t y Park, P r e s s , 1974.
P e n f i e l d , W. , & Robe r t s , L . Speech arid b r a i n mechanisms.. P r i n c e t o n :
P r i n c e t o n U n i v e r s i t y P r e s s , 1959.
50
Piaget,. J." • The o r i g i n s of i n t e l l i g e n c e i n c h i l d r e n . New York: Inter
national U n i v e r s i t i e s Press, 1952.
P i s o n i , D.P. Speech perception. In Status Report, Haskihs Laboratories, .
1977.
P r a t t ; K.C. The neonate. In L. Carmichael (Ed.), Manual of c h i l d psy
chology. New.York: John Wiley & Sons, Inc., 1960, Pp. 215-291. :
Samuel, A.G. The e f f e c t of d i s c r i m i n a t i o n t r a i n i n g on' speech perception: .
Noncategorical perception. Perception and Psychophysics, 1977, 22.,
(4), 321-330.
Singh,,S., & Black, J.W. Study of twenty-six i n t e r v o c a l consonants as
spoken and recognized by four language groups, Jourrial' of the
Ac o u s t i c a l Society of America, 1966, 39,; 371r-387.
Stern, H.H. Optimal age: Myth or r e a l i t y ? The Canadian Modern Language
Review, 1976, 32(3), 283-294.
Stevens, K.N., & K l a t t , D.H. Role of formant t r a n s i t i o n s i n the voiced-
vo i c e l e s s d i s t i n c t i o n f o r stops. Journal Of the A c o u s t i c a l Society
of America, 1974, 55_, 653-659.
Stevens, K.N., & Blumstein, S.E. Quantal aspects of consonant productiori
and perception; A study of r e t r o f l e x stop consonants. Journal.of
Phonetics, 1975, 3, 215-233.
Stevens, S.S. Psychology and the science of science, Psychological
B u l l e t i n , 1939, 36_, 221-263.
Streeter, L.A. Language perception of two-month-old i n f a n t s shows
e f f e c t s of both innate mechanisms and experience. Nature, 1976,
25,9, 39-41.
51
Suzuki, T. & Ogiba, Y, A technique of pure tone audiometry f o r . c h i l d r e n
under 3 years of age? Conditioned o r i e n t a t i o n r e f l e x (C.O.R.)
audiometry. Review of Laryngology, 1960, 81, 33,
Suzuki, T., & Ogiba, Y, Conditioned o r i e n t a t i o n r e f l e x audiometry.
Archives of Otolaryngology, .1961, 74, 192-198,
Tees; R.C. ' Perceptual development i n mammals, ' In C, G o t t l i e b (Ed,),'
Studies on the development of behavior and the nervous system:
Neural and behavioral s p e c i f i c i t y . New York: Academic Press, 1976j
*:PpS 1281-326.
Trehub, S.E, A u d i t o r y - l i n g u i s t i c s e n s i t i v i t y i n i n f a n t s . Doctoral d i s s e r
t a t i o n . M c G i l l U n i v e r s i t y , Montreal, 1973, .
Trehub, S.E, The d i s c r i m i n a t i o n of foreign speech contrasts by infants
and adults. C h i l d Development, 1976, 47(2), 466-472.
Williams, L,. Speech perception and production as a function of exposure
to a second language. Unpublished Doctoral D i s s e r t a t i o n , Harvard
Uni v e r s i t y , 1974.
Williams, L. Infant d i s c r i m i n a t i o n of synthetic speech s t i m u l i contrast
ing i n place of a r t i c u l a t i o n with and without a release burst.
Manuscript i n preparation.
Wilson, W.R. Assessment of auditory a b i l i t i e s i n i n f a n t s . Paper pre
sented as a.short course at American speech and Hearing Association,
Western Regional Convention, May, 1976. .
Wilson, W.R., Lee, K.H., Owen, G., & Moore, J.M. Instrumentation for
operant infant auditory assessment. Ch i l d Development and ..Mental
Retardation Center, U n i v e r s i t y of Washington, Undated manuscript.
52
Witelson, S.F., & P a l l i e , W, L e f t hemisphere s p e c i a l i z a t i o n f o r language
i n the Newborn: Neuroanatomical evidence of asymmetry. Brain, 1973,
96̂ , 641-646.
Wood, .CO.. P a r a l l e l processing at auditory- and phonetic information i n
speech and di s c r i m i n a t i o n . Perception and Psychophysics, 1975, 15,-
501-508.
53
APPENDIX A
Background i n Speech Decoding and i n D i s t i n c t i v e Features
It i s now well established that the cochlea,,is completely formed
by 26 weeks i n utero, and that by t h i s time both the middle and inner ear
structures have reached f u l l adult s i z e . (Eisenberg, 1976). Auditory
nerve f i b e r s begin to myelinate during the. s i x t h month i n utero, so that
by b i r t h even.the auditory cortex i s myelinated i n the- normal f u l l - t e r m
infant (Eisenberg, 1976). Although i t was once believed that the i n f a n t ;
could not make any d i f f e r e n t i a l r e s P o n s e s associated with the complex
c h a r a c t e r i s t i c s of auditory s t i m u l i (Pratt, I960), i t i s now 'well docun
mented that most newborns, even premature babies and those with abnormali
t i e s of the c e n t r a l nervous system can discriminate sounds according to
various acoustic variables (Eisenberg, 1976).
Several acoustic parameters r e l a t e d to the decoding of speech have
been explored. Each of the parameters which may be p a r t i c u l a r l y impor
tant i n speech has been shown to be within the auditory c a p a b i l i t i e s of
the newborn: (1) duration, or the t o t a l amount of time consumed by a s t i r -
mulus; (Eisenberg, 1976)5 (2) frequency, i n Hertz (hz), or the number of
sine wave r e p e t i t i o n s per second i n a pure tone s i g n a l (Trehub, 1973);
(3) sound pressure l e v e l , or the ph y s i c a l i n t e n s i t y of the s i g n a l . i n
decibels with reference to acoustic zero (Eisenberg, 1976), (4) dimen
s i o n a l i t y , or the kind and amount of variance within a complex auditory
stimulus (Eisenberg, 1976).
Unfortunately, speech sounds cannot be e a s i l y analyzed i n terms of
t h e i r constituent parts for three main reasons.
54
(1) Speech i s continuous, thus acoustic cues are highly overlapping,
as shown by the diagram below. | ^ j | g
TIME Figure l i Schematic diagram showing how the overlap of
a r t l c u l a t o r y features produced encoding i n the conversion to sound, From Liberman, 1967),
(2) Phonetic segments do not have inva r i a n t properties. For example,
the /d/'s i n the continuum from / d i / to /du/ would each be perceived as
a /d/ sound, but the acoustic dimensions of these /d/ portions would be
e n t i r e l y d i f f e r e n t as i l l u s t r a t e d i n the abbreviated spectograms below.
(From Liberman, 1967).
3000 1800
0 Idi de d9 da d) do du
Figure 2. Abbreviated "d M spectograms,
In addition, accents, voice q u a l i t y , whispering, etc, a l l create v a r i a t i o n
i n the acoustic s i g n a l , but do n o t a l t e r a l i s t e n e r ' s i n v a r i a n t perception.
(3) Phonetic segments do not stand i n a one-to-one r e l a t i o n to the
acoustic s i g n a l , i n most non-speech and "unencoded" speech sounds, i t i s
possible to judge the manner i n which the energy of the sound i s d i s t r i
buted, whereas i n encoded speech sounds, i t i s impossible to judge the
55
a c o u s t i c cues u n d e r l y i n g the s i g n a l , '
Al though so f a r i t has proved imposs ib l e to demonstrate i n v a r i a n t
a c o u s t i c cues u n d e r l y i n g segments o f speech, there are (as i n d i c a t e d above)
p e r c e p t u a l i n v a r i a n t s . That i s , many speech sounds (except vowels and
f r i c a t i v e s ) are pe rce ived i n a c a t e g o r i c a l f a sh ion when presented i n a
l i n g u i s t i c con t ex t . Thus a l though a speech s i g n a l may ..vary . a long an
a c o u s t i c continuum, i t w i l l not be p e r c e i v e d as a new ca tegory u n t i l i t
c rosses a phoneme boundary. The meaning o f ' t h i s p e r c e p t u a l invar iance . i s
a mat ter of cons ide rab l e debate. Some i n v e s t i g a t o r s p o s i t a s p e c i f i c
l i n g u i s t i c p r o c e s s e s ' ( E i m a s , 1976) , whereas o thers c l a i m i t to be a s imple
a c o u s t i c p rocessor more s e n s i t i v e t o ' t h e c o m p l e x i t i e s of m u l t i d i m e n s i o n a l
s i g n a l s (Stevens- & K l a t t , 1974), ' A d d i t i o n a l l y , the f i n d i n g o f w i t h i n -
category d i s c r i m i n a t i o n under s p e c i a l c i rcumstances has l ead some i n v e s
t i g a t o r s to p o s i t two p a r a l l e l modes o f p r o c e s s i n g , a c o u s t i c and phone t i c
(Wood, 1974) .
In an attempt to understand t h i s pe rcep tua l i n v a r i a n c e , t r a d i t i o n a l
l i n g u i s t i c s desc r ibes a s t r u c t u r a l a n a l y s i s of the r e g u l a r i t y between
phonemes and a r t i c u l a t i o n , and p o s i t s a u n i v e r s a l set of phonemes i n
terms of a r t i c u l a t o r y f e a t u r e s . S e v e r a l c l a s s i f i c a t i o n schemes have been
developed, some i n terms o f o n l y phonemic d e s c r i p t i o n (Jakobson & H a l l e ,
1956), and o thers i n c l u d i n g r u l e s to account fo r recombina t ion i n t o mor
phemes (Chomsky & H a l l e , 1968). An example of fea tu re d e s c r i p t i o n i s
g iven i n . T a b l e 1.
As can be seen i n Table 1, each phoneme can be d e s c r i b e d i n terms
o f a r t i c u l a t o r y fea tures arid phonemic c o n t r a s t s can be de sc r i bed i n terms
*A s t imulus dimension i s cons idered to be p e r c e i v e d c a t e g o r i c a l l y i f the spac ing of s i g n a l s a long that dimension i s found to be the same i n d i s c r i m i n a t i o n experiments as i n i d e n t i f i c a t i o n exper iments .
T a b l e 1 Tho prime features.
56
Feature Abbreviated definition Phonological Exemplification
name of physical scale terms symbols language word gloss % value 1 Glottalic Rate of upward movement [cjectivel t' Uduk f t lick 100
of the glottis (pulmonic) t Uduk ter collect 50 [implosive] d" Uduk dek lift 0
2 Velaric Degree of suction of air in [ + click] 1 Zulu lata climb 100 mouth (-c l ick] t Zulu tatii third 0
3 Voice Degree of approximation of the (glottal stop] ? Javanese buka> open 100 arytenoid cartilages [laryngealized] b Hausa be:rd: rat 80
(voice) b llausa be:ra girl 60 [murmur] b f t Hindi br.al forehead 20 (voiceless) P Hindi pal take care of 0
4 Aspiration Time of onset of voicing with [aspirated] P" Thai p»a: split 100 respect to release of the [unaspirated] P Thai pa: forest 50 articulation [voiced] b Thai ba: shoulder 0
5 Place Distance from the glottis to the [bilabial) P Ewe e(3e Ewe 100 first constriction of the [labiodental] v Ewe tvt two 95 vocal tract [dental) I Malayalam .kutti stabbed 90
[alveolar] t Malayalam kutti peg 85 [retroflex] t Malayalam kutti child 80 [palato-alveolar] 1 English Jip sheep 75 (palatal) c Quechua caka bridge 70 [velar] k Qucchua kara expensive 60 [uvularl q Quechua qara skin 50 [pharyngeal] h Arabic had someone 30 [glottal] > Arabic 'alia God 0
6 Labial Degree of approximation of ( + labial) Igbo akpa bag 100 the centers of the lips (-labial] k Igbo aka hand 0
7 Stop Degree of approximation of (stop] t English tat tie 100 the articulators [fricative] s English sai sigh 90
[approximant] h English hai high 0 8 Nasal Degree of lowering of the ( +nasal) n English noo know 100
soft palate [-nasal] d English doo dough 0 9 Lateral Amount of airstream flowing ( +lateral) 1 English loo low 100
over the side of the tongue [-lateral] d English doo doe 0 10 Trill Degree of vibration of an [ +trill] r Spanish pero dog 100
articulator [- tr i l l ] r Spanish pero but 0 (11 Tap) Rate of articulator^ [ + tap] r Tamil srsm saw 100
movement? [-tap] j Tamil a:j3m depth 0 11 Sonorant Amount of acoustic (•fsonorantl, i English ' s A n i a r sunnier 80
energy [ — sonorant] j English Vnjan onion 70 13 Sibilant Amount of high frequency [ +sibilant] s English stn sin 100
(over 3000 Hz.) energy [-sibilant] e English Oin thin 20 14 Grave Ratio of low to h i g h [ +grave) f English pin, kin pin, kin 90
frequency energy (-grave) 0 English tin tin 60 15 Height Inverse of frequency of the [4 height] i Danish vi:6a know 95
first formant 13 h e i g h t ] e Danish ve:6a wheat 65 [2 height] c Danish VE:ds wet 50
•t [1 height] X Danish wade 10 16 Back Difference between frequency of I + back) u English hu who 80
formant two and formant one [-back] i English hi he 5 17 Round Inverse of distance between (4-round] y French ly read 90
corners of the lips ( — round) i French Ii bed 0 18 Wide Degree of advancement of [ + wide] i Igbo obi heart 100
tongue root I-wide] i Igbo ybj poverty 20 19 Rhotacized Lowering of the frequency of [ +rhotacized] 31 English bsjd bird 100
the third formant I — rhotacized) I English bid bid 0 20 Syllabic (No agreed physical scale) [ +syllabic] n English SAdn sudden 100
( - syllabic] n English SAn sun 0
of feature d i f f e r e n c e s . Support for such a system was presented by
M i l l e r and Nicely (1955) i n a now c l a s s i c study which demonstrated that
under masking conditions, errors i n perceptual responses increased as
the number of feature differences decreased. Similar .pairs (phonemes
d i f f e r i n g i n only one a r t i c u l a t o r y feature) were judged to be the hardest
to discriminate,; An example of a h i g h l y s i m i l a r p a i r of sounds would be
/b/ and /p/; the only a r t i c u l a t o r y feature difference iri v o i c i n g • ( v o i c
ing i s then c a l l e d the " d i s t i n c t i v e feature"), Most current speech per-,
ception research involves phoneme p a i r s which d i f f e r i n only one d i s - .
t i n c t i v e feature.
58
APPENDIX B : C o n s e n t Forms
A d u l t , ' ^ C o n s e n t Form THE UNIVERSITY OF BRITISH COLUMBIA
2075 Wesbrook Mall Vancouver, B.C., Canada
V6T 1W5 Department of Psychology
CONSENT FORM This experimental procedure has been requested by
I have been informed of the procedures and understand them. I also understand that the procedures may be terminated at any time at my request. PROCEDURE:
This is a study in speech perception. You will be sitting on a chair facing a loud speaker in the
testing chamber. A series of one syllable speech sounds will be played over the speaker. Your task will be to press a button whenever you detect a change in the speech sounds. Every correct discrimination response will be signalled by flashing lights. A record of all your responses will be kept.
If at any time you desire, there will be a break in, or termination of, the testing session.
My signature below certifies that I consent to the experimental procedure which has been described and which is to be conducted on the following d?-te: in the following place: • and designated in the following manner:
Date: . Name: Signature:
59
I n f a n t C o n s e n t Form
THE UNIVERSITY OF BRITISH COLUMBIA 2075 Uesbrook Mall
Vancouver, B.C., Canada V6T 1W5
Department of Psychology
CONSENT FORM
This experimental procedure has been requested by
I have been informed of the procedures and understand them. I also understand that the procedures may be terminated at any time at my request. PROCEDURE:
The infant will be held on his/her guardian's lap in the testing chamber. He or she will see a series of toys, and will hear a series of one syllable speech sounds. The infant will be watched by the experimenter in the adjoining room, and by the assistant in the testing chamber. Whenever the infant turns toward the loud-speaker during a test trial, the experimenter and the assistant will press a button. If this head turn has occurred when there is a change in the speech stimuli, a toy animal inside the dark plexiglass box will begin performing. A record of all the baby's responses to the speech sounds will be kept. The baby will be held by, and will be under the control of, the guardian at all times. Whenever the guardian desires, there will be a break in or termination of the testing session.
My signature below certifies that I consent to the experimental procedure which has been described and which is to be conducted on the following date: • In the following place: and designated in the following manner: .
Date: Name: SignaTUre: