Vowel discrimination in early bilinguals: Evidence of phonetic reorganization

1
Vowel discrimination in early bilinguals
Evidence of phonetic reorganization
3aSC11 148th ASA meeting in San Diego 2004
2
Authors
Anders Højen James E. Flege Speech and Hearing
Sciences Division University of Alabama at
Birmingham
3
Abstract The issue of whether new categories may
be established for vowels found in a second
language (L2) but not the native language (L1)
remains controversial. The present study compared
the discrimination of English vowels by native
English speakers, early Spanish-English
bilinguals, and Spanish monolinguals (n20 each).
Vowels in the easy contrast (i-u) were
likely to be heard as two distinct Spanish vowels
by Spanish speakers, whereas the vowels in three
difficult contrasts (eI-I, A-Ã,
oU-U) were likely to be heard as a single
Spanish vowel. To avoid the ceiling effects often
seen in cross-language vowel discrimination
research, within-trial F0 variation was
introduced into the categorial AXB test used
here. The 64 trials testing each contrast were
presented in two randomized blocks differing in
inter-stimulus interval (0 vs. 1000 ms). Spanish
monolinguals had great difficulty discriminating
vowels in the 3 difficult contrasts, probably
because they classified the contrastive English
vowels as a single Spanish vowel. However, early
bilinguals discriminated these contrasts much
better, indicating English vowel category
formation. Age at onset of English-language
exposure, years of U.S. education, and amount of
L2 use (especially during the first years of
exposure) seemed to impact discrimination ability.
4
background

• Background
• It is generally believed a second language (L2)
  can be learned without problems in childhood
• However, research in Barcelona revealed important
  differences in the perception of Catalan vowels
  by native Catalan speakers vs. native Spanish
  speakers who learned Catalan in early childhood
• (Bosch et al. 2000 Pallier et al. 1997 Pallier
  et al. 2001 Sebastian-Galles Soto-Faraco,
  1999)

5

• What explains the difference between early
  bilinguals and L2 native speakers?
• A lack of neural plasticity?
• A lack of cognitive plasticity after L1
  acquisition?
• Insufficient input from L2 native speakers?
• Foreign-accented L2 input?
• An inability to establish new phonetic
  categories?
• Do all early bilinguals differ from L2 native
  speakers?

6

• This study
• Examined the discrimination of English vowels by
  three groups
• English monolinguals
• Spanish monolinguals
• Early native Spanish learners of English
• English contrasts examined
• The vowels in 3 difficult contrasts likely to
  be heard as a single Spanish vowel
• The vowels in one easy contrast likely to be
  heard as two distinct Spanish vowels

7
Aims

• Aims
• Develop a phonetically sensitive method to test
  for differences between early bilinguals and L2
  native speakers
• Requirement yield large differences
  between English Spanish monolinguals for the 3
  difficult contrasts, but little (or no)
  difference for the easy contrast
• Test for differences between the early
  Spanish-English bilinguals and both groups of
  monolinguals
• Explore individual differences in L2 vowel
  discrimination

8
Participants
Note Educ., total years of education AOE, age
at onset of L2 exposure L2-years, years of L2
exposure.
9

• Stimuli
• One native English speaker produced multiple
  tokens of 8 English vowels in non-words
• Half of stimuli produced in declarative sentence
  frame, with relatively low F0 half produced in
  interrogative frame, with high F0
• 4 consonant contexts (see Table)

10
contrasts
Classification by Spanish monolinguals
11
Easy vs. Difficult contrasts

• Acoustic analysis (see figure below) and
  classification of the English vowels by Spanish
  monolinguals suggested that
• The 2 English vowels in the easy contrast,
  would be heard as 2 distinct Spanish vowels (/i/
  and /u/)
• The 2 vowels making up each of the 3 difficult
  contrasts would be heard as a single Spanish
  vowel (/e/, /a/, and /o/)

12
Figure with stimuli
13
procedure

• Procedure
• Each contrast tested by 64 AXB trials
• The 4 contrasts were tested in a single
  randomized block presented at 2 different ISIs
  (512 trials in all)
• Previous research (Cowan Morse, 1986 Pisoni,
  1973) suggested that lower percent correct scores
  would be obtained at 0-ms than 1000-ms ISI
• In half of trials (called incongruent), the F0
  variation suggested the wrong answer
• (example A-high B-high B-low)

14
Results
15
(No Transcript)
16

• The AXB test was phonetically sensitive
• The Spanish and English monolinguals differed
  little for the easy contrast (M
  88 vs. 98 correct)
• However, the Spanish monolinguals obtained much
  lower scores than the English monolinguals for
  the 3 difficult contrasts (M 58 vs. 96)

17
Results
2) The early bilinguals scores for the 3
difficult contrasts were much higher than the
Spanish monolinguals (M 89 vs 58) but
differed little from the English monolinguals (M
96)

• At 1000-ms ISI, no significant differences
  between early bilinguals and English monolinguals
  for any contrast
• At 0-ms ISI, early bilinguals scored lower than
  English monolinguals for 2 contrasts (I-eI,
  U-oU plt0.01)

18
Results
3) Most but not all early bilinguals performed
like the English monolinguals

• At 0-ms ISI, 4/20 early bilinguals obtained
  scores that fell outside the native English range
  (viz. 81-99 )
• The 4 highest- and the 4 lowest-scoring
  bilinguals differed in several respects

19
Table individual differences
Note Ranges in parentheses
20
Details of English use patterns
21

• Discussion
• Spanish monolinguals had great difficulty
  discriminating vowels in the 3 difficult
  contrasts, probably because they classified the
  contrastive English vowels as a single Spanish
  vowel
• Early bilinguals discriminated these contrasts
  quite well, perhaps indicating the formation of
  new English vowel categories

22
Discussion, 2

• As in previous research (Flege MacKay, 2004),
  amount of L2 use seemed to affect early
  bilinguals discrimination of L2 vowels
• Language use patterns during the first few
  years of L2 use may be especially important
• A frequent L2 use and possibly a relatively
  infrequent L1 use may be needed for early
  bilinguals to discriminate L2 vowels at
  native-like levels

23
Discussion, 3

• The results are consistent with the view that the
  phonetic system remains plastic in early
  childhood, and that establishment of L1 vowel
  system does not seriously compromise phonetic
  learning
• The results do not support the view, however,
  that all early bilinguals develop perceptual
  representations for L2 vowels that are exactly
  like those of L2 native speakers

24
References

• Bosch, L., Costa, A. and Sebastian-Galles, N.
  (2000). "First and second language vowel
  perception in early bilinguals," Eur. J. Cogn.
  Psychol. 12, 189-221.
• Cowan, N. and Morse, P.A. (1986). "The use of
  auditory and phonetic memory in vowel
  discrimination," Journal of the Acoustical
  Society of America 79, 500-507.
• Flege, J.E. and MacKay, I.R.A. (2004).
  "Perceiving vowels in a second language," SSLA
  26, 1-34.
• Pallier, C., Bosch, L. and Sebastian-Galles, N.
  (1997). "A limit on behavioral plasticity in
  speech perception," Cognition 64, B9-B17.
• Pallier, C., Colomé, A. and Sebastian-Galles, N.
  (2001). "The influence of native-language
  phonology on lexical access Exemplar-based vs.
  abstract lexical entries," Psychol. Sci. 12,
  445-449.
• Pisoni, D.B. (1973). "Auditory and phonetic
  memory codes in the discrimination of consonants
  and vowels," Perception and Psychophysics 13,
  253-260.
• Sebastian-Galles, N. and Soto-Faraco, S. (1999).
  "Online processing of native and non-native
  phonemic contrasts in early bilinguals,"
  Cognition 72, 111-123.