Healthcare and Medicine Reference
In-Depth Information
A
A
VOWELS
English /i/
Swedish /y/
Handshapes produced by deaf infants
second formant
2000
1000
first formant
0
B
Time
80
B
TEST STIMULI
Deaf
60
1800
prototype /i/
40
varient of /i/
1600
20
Hearing
prototype /y/
0
1400
0
10
11
12
Age (months)
13
14
15
100 300 500
First Formant (unit of pitch)
FIGURE 10.28 Manual communication by infants. A. The hand
signals produced by deaf infants were studied and codified in order
to detect hand shapes that correspond to American Sign Language
(ASL). B. Deaf infants produced hand shapes corresponding to ASL
(manual babbling) more often than hearing infants of the same age.
(Adapted from Petitto and Marentette, 1991)
C
PROCEDURE
present
prototype
present
variant
Baby turns head
if variant sounds
different
important when learning to “speak” with one's hands,
in general agreement with studies of bird song
development.
Given the complexity of language, it is not surpris-
ing that we are only beginning to understand the
neural mechanisms that support human communica-
tion and how it develops. The ability of infants from
two countries to recognize their native vowel sounds
was studied to find out whether early experience
effects perception. Six-month-old infants from Sweden
and the United States were asked to judge two vowel
sounds, one from their own country and one from the
other country (Kuhl et al., 1992). The English vowel
was a /i/ sound, as in the word “fee”. The Swedish
vowel was a front rounded vowel /y/ sound, as in the
Swedish word “fy.” Vowel sounds are composed of a
unique set of frequencies, called formants , and the /i/
sound has slightly higher formants than the /y/ sound
(Figure 10.29A). Most adult English-speaking listeners
can categorize a sound as being like a /i/ sound if the
first and second formants are reasonably close to the
ideal. This ability to generalize is thought to prevent
confusion since individual voice quality varies a good
deal, particularly between children, adult females, and
adult males. To see whether infants are able to catego-
rize vowel sounds, ideal /i/ and /y/ vowels, called
prototypes , were generated, and slight variations were
80
American infants
70
/i/
60
50
/y/
40
Greater distance between variant & prototype
FIGURE 10.29 Recognition of language-specific phonemes. A.
The frequency spectrum of an American vowel (/i/) and a Swedish
vowel (/y/) are shown. Each vowel is composed of two major
frequency bands, called formants. B. The ability of American and
Swedish infants to recognize their native vowel sounds was exam-
ined with a range of computer-generated stimuli. An 'ideal' version
of each vowel, called a prototype , was produced, along with vowels
with small changes to one of the formant frequencies (called vari-
ants ). C. Infants were trained to turn their head if the second of two
vowels sounded different than the first. The data for American
infants show that their native /i/ sound can be recognized even
when formant frequency changes a good deal. However, the same
frequency changes for the Swedish /y/ led to decreased recognition.
(Adapted from Kuhl et al., 1992)
Search Pocayo ::




Custom Search