Pr

1
Perception of native and non-native fricative
contrasts by Ukrainian-Australian English
bilinguals and Australian English monolinguals.
Tamara A. Kencalo, Catherine T. Best, Michael
D. Tyler Louis M. Goldstein MARCS Auditory
Laboratories, University of Western Sydney,
Australia Haskins Laboratories, New Haven CT,
U.S.A. Yale University, New Haven CT, U.S.A.

• Articulatory Organ Hypothesis (AOH)
• A recent AP proposal is that infants attend more
  to the articulatory organ that is active in a
  gesture than to the specific gestures a given
  organ makes. The AOH predicts infants will
  continue to discriminate between-organ contrasts
  but will show a decline for within-organ ones
  (Goldstein Fowler, 2003).
• Infant perceptual findings support the AOH
  (Best McRoberts, 2003)
• Within-organ contrast shows developmental
  decline by 10 months
• Between-organ contrast shows no decline
• We extended PAM and AOH to adults differing in
  language experience
• Monolinguals vs. Bilinguals whose languages
  differ in the target contrasts

• Procedure
• Discrimination
• AXB procedure
• separate tests for each language (order
  counterbalanced)
• separate subtests per contrast within language
  (counterbalanced)
• 8 blocks of 16 trials per subtest
• tokens presented equally in AAB, ABB, BBA, BAA
  trials per subtest
• 8 practice trials
• Interstimulus interval (ISI) 1 s ITI (trials)
  3.5 s IBI (blocks) 5 s
• Categorisation
• Randomized presentation of individual tokens
• Separate tests for each language (order
  counterbalanced)
• Participants wrote what each consonant sounded
  like to them
• Monolinguals ENG orthography
• Bilinguals UKR or ENG orthography
• They then gave a goodness-of-fit rating (1 very
  poor, 7 excellent)

Précis Adults discriminate many non-native
consonant contrasts poorly, but exceptions offer
key insights about listeners knowledge of their
native phonological systems. The Perceptual
Assimilation Model (PAM Best, 1995) predicts
monolinguals perception, but has not been
extended to bilinguals. We tested PAM predictions
for bilinguals vs. monolinguals, as compared to
those for an adult extension of the Articulatory
Organ (AOH) hypothesis (Goldstein Fowler,
2003), which posits that infants find contrasts
within a single articulatory organ less
discriminable than contrasts between two
different organs. Early bilingual
(Ukrainian-Australian English) and monolingual
(Australian-English) adults were examined for
perception of between-organ versus within-organ
non-native and native fricative contrasts.
Results were consistent with PAM predictions,
going beyond basic AOH predictions. Thus,
bilingual vs. monolingual language experience
affects adults non-native speech perception,
requiring modifications of the AOH as it applies
to adults.
Discussion and Conclusions Language experience
played a crucial role in perception of the NCN
contrasts, according to the significant Stimulus
Language x Organ x Language Background
interaction and the group differences in
assimilation of the between-organ NCN contrast.
This suggests that perception of the two
non-native contrasts was strongly affected by L1
exposure, ENG for the monolingual group and UKR
for the bilinguals. The groups NCN assimilation
patterns and discrimination levels strongly
support PAM predictions, more than they support
the basic AOH predictions originally developed
for infants. Specifially, bilinguals displayed
SC assimilation of the within-organ NCN contrast
(tongue body ?-x) to their UKR velar x.
They also showed poorer discrimination of this
contrast than of the between-organ NCN contrast
(tongue body vs. root ?-h), which the
majority assimilated as TC. By comparison,
monolingual ENG listeners assimilated all NCN
fricatives to h, displaying predominantly SC
assimilation of both NCN contrasts. They showed
equivalent discrimination of the ?-x and
?-h contrasts, consistent with PAM
predictions, but failing to support a simple AOH
between gt within organ advantage for these
contrasts. A possibility worth considering in
future is that for adult speakers of a language
like ENG, which does not make contrastive use of
tongue body versus root, these are not treated as
separate organs (single organ tongue
dorsum). Interestingly, the majority of
bilinguals categorised the non-native fricatives
to their L1 (UKR), even though all have spent the
majority of their lives in Sydney, an English
dominant city, all are English-dominant
themselves, and all were participating in a study
conducted entirely in ENG. Comparing the present
findings in a single-language dominant city to
studies in bilingual cities such as Barcelona
(Pallier et al., 1997 Bosch et al., 2000
Navarra et al, 2005), it is clear that in both
situations, early proficient bilinguals are
likely to assimilate non-native consonants to
their L1, perhaps especially when it is
advantageous for non-native perception.
The Present Study Early proficient bilingual
(Ukrainian UKR - Australian English AusE) and
monolingual (AusE) adults completed perceptual
tests on two non-native voiceless fricative
contrasts from Nuu Chah Nulth (NCN), a native
Canadian language, and two native (L1) or early
second language (L2) English (ENG) voiceless
fricative contrasts. One contrast from each
language was a between-organ distinction, the
second a within-organ distinction (see Figure
2) NCN ?a-ha tongue body vs. root uvular
vs. pharyngeal ?a-xa tongue body uvular
vs. velar ENG ?a-fa lips vs. tongue
tip interdental vs. labiodental ?a-sa
tongue tip interdental vs. alveolar ENG has no
posterior (guttural) fricatives except voiceless
h (larynx glottal). Ukrainian (UKR) has xa
(tongue body velar) and voiced ? (larynx
glottal).
Background Adults have difficulty discriminating
some, but not all, non-native consonant contrasts
(Best et al., 1988, 2001). The Perceptual
Assimilation Model (PAM) (Best, 1995) posits that
the ease or difficulty of a given non-native
contrast depends on the perceived similarity of
its consonants to those of the native phonology.
Results Discrimination 3-way ANOVA on correct
responses The Stimulus Language x Organ
interaction was significant, F(1, 26) 13.41, p
.001, ?p2 .340, which was qualified by a
3-way interaction with Language Background, F(1,
26) 4.33, p 0.05, ?p2 .14 (Figure 3).
Both groups discriminated ENG gt NCN, and
discriminated ENG within-organ ?-f better
than between-organ ?-s. But the bilinguals
alone discriminated NCN between-organ ?-h
better than within-organ ?-x.

• PAM Predictions
• Two Category (TC) When two non-native
  consonants are assimilated to two native
  categories, discrimination is excellent.
• Single Category (SC) When two non-native
  consonants are assimilated to one native category
  equally, discrimination is poor.
• Category Goodness (CG) When two non-native
  consonants are assimilated to the same native
  category but with varying goodness of fit,
  discrimination is good to very good.
• Uncategorised-Categorized (UC) When one
  non-native consonant is unclearly categorized
  while the other is clearly assimilated to a
  native category, discrimination is good to
  excellent.
• Assimilations depend on similarity of native and
  non-native articulatory gestures, per
  Articulatory Phonology (Browman Goldstein,
  1992).

Figure 2. Schematic representation of the
within-organ versus between-organ contrasts in
native English fricatives (left) and non-native
NCN fricatives (right), according to the AOH
Hypothesis.
Future research Future research should
investigate the PAM and AOH hypotheses and
include alternate non-native within and
between-organ contrasts for ENG monolinguals and
UKR-ENG bilinguals to investigate whether adult
discrimination of within and between-organ
consonant contrasts differs for other target
consonants or languages and is not language
specific. Other types of bilinguals should also
be tested, such as Arabic-English bilinguals for
whom both the pharyngeal ha and the uvular ?a
fricatives exist in their L1 to extend the
present findings to additional L1 systems.
Figure 3. Mean percent correct discrimination,
Stimulus Language x Organ x Language Background
interaction. Error bars represent standard errors
of the mean.

• Categorisation ?2 on group differences in
  assimilation types
• Categorization of each NCN consonant determined
  per participant
• Categorised gt 50 of labels as the same native
  consonant
• Uncategorized lt50 of labels in any single
  native category
• Assimilation types determined for each NCN
  contrast per participant
• SC Categorized to same native consonant, ns
  diff. in ratings
• CG Categorized to same native consonant,
  significant rating diff.
• TC Categorized to two different native
  consonants
• UC One NCN consonant Categorized, the other
  Uncategorized
• Monolinguals categorised all NCN consonants to
  ENG h. Bilinguals used mostly UKR responses,
  especially x for NCN x and ?. Bilinguals
  assimilated NCN between-organ ?-h as TC
  significantly more often than monolinguals did,
  ?2(3, 28) 12.79, p lt .005. Both groups
  assimilated the NCN within-organ ?-x contrast
  primarily as SC (see Figure 4)

• Predictions
• AOH predicts that within-organ contrasts are less
  discriminable than between-organ contrasts even
  for adults, and regardless of language experience
• PAM predicts that
• Monolinguals should assimilate all NCN fricatives
  to the only ENG guttural fricative, h. If two
  NCN consonants differ in perceived goodness of
  fit to h (CG assimilation), discrimination will
  be good if no goodness difference is perceived
  (SC assimilation) discrimination will be poor.
• Bilinguals should assimilate NCN ? and x to
  UKR x (SC assimilation) and discriminate it
  poorly. They may assimilate NCN pharyngeal h to
  UKR ? or ENG h (glottal), showing TC
  assimilation of ?-h and excellent
  discrimination or they may assimilate ha to
  UKR x, with CG or SC assimilation and less-good
  to poor discrimination of ?-h.

• Articulatory Phonology (AP)
• Phonological elements in speech are defined by
• Articulatory organs of the vocal tract
• Constrictions made by the active articulators
  (see Figure 1)
• at specific constriction locations
• using specific constriction degrees

Acknowledgements This research was supported by
NIH grant DC00403 (Principal Investigator C.
Best). Many thanks to Elizabeth Beach who
assisted in the stimulus development of the
English tokens. We are most grateful to all the
participants who agreed to take part in this
research, and to our Nuu Chah Nulth speaker and
to colleagues Bryan Gick and Ian Wilson, who
helped us record the NCN stimulus materials at
the University of British Columbia.

• Method
• Participants
• Monolinguals Australian-English (AusE)
• n 16 (10 female, Mage 22.69, range 18-35
  years)
• Bilinguals Ukrainian (UKR)-AusE
• n 12 (7 female, Mage 33.42, range 20-49
  years)
• Stimulus Materials
• 4 tokens each
• 3 target ENG /Ca/ syllables by native female
  AusE speaker)
• 3 target NCN /Ca/ syllables by native female NCN
  speaker)
• Tokens were selected for best match on duration,
  amplitude and frequency characteristics, based on
  acoustic analyses (Praat).

References Best, C. T. (1995). A direct realist
perspective on cross-language speech perception.
In W. Strange (Ed.), Speech perception and
linguistic experience Issued in cross-language
research (pp. 171-204). Timonium, MD York
Press. Best, C., McRoberts, G. W. (2003).
Infant perception of non-native consonant
contrasts that adults assimilate in different
ways. Language and Speech, 46, 183-216. Best, C.
T., McRoberts, G. W., Goodell, E. (2001).
American listeners perception of nonnative
consonant contrasts varying in perceptual
assimilation to English phonology. Journal of the
Acoustical Society of America, 109,
775-794. Best, C. T., McRoberts, G. W.,
Sithole, N. M. (1988). Examination of perceptual
reorganization for nonnative speech contrasts
Zulu click discrimination by English-speaking
adults and infants. Journal of Experimental
Psychology Human Perception and Performance, 14,
345-360. Bosch, L., Costa, A.,
Sebastian-Galles, N. (2000). First and second
language perception in early bilinguals. European
Journal of Cognitive Psychology, 12,
189-221. Browman, C. P., Goldstein, L. (1992).
Articulatory phonology An overview. Phonetica,
49, 155-180. Goldstein, L., Fowler, C. A.
(2003). Articulatory phonology a phonology for
public language use. In N. Schiller and A. Meyer
(Eds.) Phonetics and Phonology in Language
Comprehension and Production (pp. 159-208).
Berlin Mouton de Gruyer. Navarra, J.,
Sebastian-Galles, N., Soto-Faraco, S. (2005).
The perception of second language sounds in early
bilinguals New evidence from an implicit
measure. Journal of Experimental Psychology
Human Perception and Performance, 31,
912-918. Pallier, C., Bosch L.,
Sebastian-Galles, N. (1997). A limit on
behavioral plasticity in vowel acquisition.
Cognition, 64, B9-B17.
Figure 4. Total mean percentages of each
Assimilation Type for NCN Pharyngeal-Uvular and
Uvular-Velar contrasts, according to
categorization and ratings by monolinguals and
bilinguals.
Figure 1. Schematic representation of the active
articulators and their geometrical organisation
in the vocal tract, according to AP theory.