The articulatory settings of bilingual Canadian EnglishFrench speakers

1
The articulatory settings of bilingual Canadian
English-French speakers
Ian Wilson, Bryan Gick, Fiona Campbell, Eric
Vatikiotis-Bateson University of British Columbia

• Ultrafest III - Apr.14, 2005
• Tucson, AZ

2
Use of ultrasound for speech research

• Becoming much more common but still has
  unresolved methodological issues
• Choice of methodology depends on
• Experiment setting (lab vs. field)
• Type of subjects (adult, child, impaired, )
• Type of study (biofeedback, etc.)
• Bryan Gick will expand on this tomorrow

3
The present study

• Normal adults
• Speakers of English Québécois-French
• Data collected in a laboratory

4
Methodological issues

• What to measure?
• Track given point on tongue? (EMA x-ray
  microbeam)
• Shape of tongue surface in 2D, 3D
• Cross-sectional area under curve of tongue
• Relative timing of tongue gestures
• Distance from tongue surface to fixed point

5
Measuring tongue surface to fixed point

• What fixed point?
• Transducer surface / centre
• Opposing surface (i.e. palate, pharyngeal wall)

6
Tongue to fixed point more issues

• Need to make opposing surface visible
• Have subject swallow
• CT scan
• Measurement location based on
• Location of opposing surface point
• Location of tongue gesture
• Controlling and/or correcting for head movement
  relative to probe
• 1. Optotrak/video 2. Helmet 3. Restraint
• Data extraction
• By hand
• Software for edge detection / curve fitting

7
Back to the present study

• Articulatory Setting (AS)
• When speaking a foreign language, ones
  articulators seem to have a whole different
  underlying posture

8
Different languages sound different

• Why??
• Different phonemes
• Different phonologies
• Different articulatory settings
• gross oral posture and mechanics of a language
  - Honikman (1964)

9
Implications of language-specific AS (1)

• Second language acquisition
• Quantitative evidence to support newer L2
  teaching methodologies (Mompeán-González, 2003)
• Speech motor control
• If AS is weighted average of postures for all
  phonemes in a language (Laver, 2000), then
  supports view of AS being functionally
  determined, reducing travel cost of articulators
  (Rosenbaum et al., 1995)

10
Implications of language-specific AS (2)

• Models of speech production
• de Bots (1992) model of bilingual speech
  production contains articulator (Levelt, 1989)
  that uses set of non-language specific speech
  motor plans
• Predicts bilinguals have 1 AS shared between
  languages could be based on type frequency
• Hesitation pauses schwa
• AS may be reflected in these (e.g. AS for French
  has protruded lips schwa is rounded)

11
Implications of language-specific AS (3)

• Development of languages
• Esling (2000) possible that instances of
  language change are accomplished by slightly
  altering AS, which in turn produces minute,
  sub-phonemic changes in the phonetic quality of
  certain susceptible segments.

12
How to measure AS

• Gick, Wilson, Koch Cook (2004 Phonetica)
• Link AS to inter-speech posture (Barry, 1992
  Gick, 2002) reducing segmental interference

13
Gick et al. measurements taken

• 1 pharynx width
• 2 VPP width
• 3 TB to palate
• 4 TT to palate
• 5 upper central incisors to jaw
• 6 upper lip protrusion
• 7 lower lip protrusion

Significantly different across languages
14
AS in bilinguals

• Why bilinguals?
• Within a speaker, can compare across languages
  without worrying about physiological differences
• Bilinguals may economize in their tongue gestures
  (Wilson, 2003) and VOT (Watson, 1990, 1991)
• Is AS like a speech target in the sense that
  economy can play a part?
• If AS is simply functional, this predicts
  speaking mode (bilingual vs. monolingual) will
  affect AS (Grosjean, 1998)

15
Subjects

• 10 monolingual Canadian-English speakers
• 10 monolingual Quebecois-French speakers
• 10 bilingual English-French speakers

16
Stimuli

• Monolingual subject trials
• At least 6 blocks of 30 utterances (180 possible
  rest positions per subject)
• Bilingual subject trials
• 2 English blocks, 2 French blocks, 2 mixed
  language blocks
• Before mixed language blocks, subject is informed
  that language of the next sentence is randomly
  selected
• Phonetic context of first and last syllables
  controlled for as much as possible across
  languages

17
Method Experimental set-up
18
Method Optotrak marker set-up
19
Ultrasound data for MLD (bilingual)

• Bilingual trial B-mode M-mode ultrasound

20
Method Correction for head movement

• Palate trace done after swallow can then measure
  shortest distance from alveolar ridge to tongue

• Optotrak tracks head as rigid body probe is
  fixed position is known by Optotrak
• Rotate/translate palate trace about probe
  (MATLAB)

21
Results for MLD - tongue tip

• Tongue tip to alveolar ridge distance
• English
• mean 21.13 pix
• (N35 std dev2.91)
• French
• mean 17.44 pix
• (N42 std dev3.81)
• Bilingual mode
• mean 23.06 pix
• (N43 std dev3.09)

E
F
Bil
Group means 95 confidence intervals)
22
Results for MLD - vertical lip aperture

• Vertical lip aperture
• English
• mean 23.13 mm
• (N20 std dev1.88)
• French
• mean 23.47 mm
• (N20 std dev1.12)
• Bilingual
• mean 23.82 mm
• (N21 std dev1.30)
• No significant differences

23
Results for MLD - horizontal lip aperture

• Horizontal lip aperture
• English
• mean 50.52 mm
• (N20 std dev0.76)
• French
• mean 51.60 mm
• (N20 std dev1.68)
• Bilingual
• mean 49.79 mm
• (N21 std dev0.40)

E
F
Bil
Group means 95 confidence intervals)
24
Significance of findings so far

• Support for existence of language-specific AS
  not contextually determined
• AS in bilingual mode does not fall between 2
  monolingual mode settings
• implies AS is not determined simply by motor
  control constraints such as articulator travel
  cost
• N.B. speech is a low-energy system
• when language target unknown, subjects
  inter-speech posture is closer to absolute rest
  position (i.e. a position out of speech mode)
• Seemingly uneconomic

25
Next Steps

• What is the relationship of type token
  frequency to AS?
• Is AS different for natural speech vs. read
  speech? Nonsense words vs. real words? (i.e. is
  it task dependent)
• What is the relationship between L2 pronunciation
  proficiency and AS?
• What is perceptually salient in AS (i.e. if
  learned, how is it learned?) Can it all be read
  in the face?!

26
Thank you!

• Thanks to
• Jason Chang
• Shaffiq Rahemtullah
• Doug Pulleyblank
• and all our cooperative subjects.

27
References

• Barry (1992) Comments on Chapter 2 (Browman and
  Goldstein) in Docherty Ladd, Papers in
  laboratory phonology II gesture, segment,
  prosody (pp. 65-67) Cambridge Cambridge
  University Press.
• de Bot (1992) A bilingual production model
  Levelts Speaking model adapted. Applied
  Linguistics 13 1-24.
• Esling (2000) Crosslinguistic aspects of voice
  quality. In R. D. Kent M. J. Ball, (Eds.),
  Voice quality measurement (pp. 25-35). San Diego
  Singular.
• Gick et al. (2004) Language-specific articulatory
  settings Evidence from inter-utterance rest
  position. Phonetica 61 220-233.
• Grosjean (1998) Studying bilinguals
  Methodological and conceptual issues.
  Bilingualism Language and Cognition 1 131-149.
• Heffner (1950) General phonetics. Madison, WI
  The University of Wisconsin Press.
• Honikman (1964) Articulatory settings. In D.
  Abercrombie, D. B. Fry, P. A. D. MacCarthy, N. C.
  Scott, J. L. M. Trim (Eds.) In Honour of Daniel
  Jones (pp. 73-84). London Longman.
• Laver (2000) Phonetic evaluation of voice
  quality. In R. D. Kent M. J. Ball, (Eds.),
  Voice quality measurement (pp. 37-48). San Diego
  Singular.
• Levelt (1989) Speaking. From intention to
  articulation. Cambridge, MA MIT Press.
• Mompeán-González (2003) Pedagogical tools for
  teaching articulatory setting. Poster presented
  at ICPhS 15, Barcelona.
• Munhall et al. (1994) X-ray film database for
  speech research, ATR technical report TR-H-116.
  ATR Human Information Processing Research
  Laboratories, Kyoto.
• OConnor (1973) Phonetics. Harmondsworth
  Penguin.
• Öhman (1967) Peripheral Motor Commands in Labial
  Articulation. STL-QPSR 4/1967 RIT Stockholm.
• Perkell (1969) Physiology of speech production
  Results and implications of a quantitative
  cineradiographic study. Cambridge MA MIT Press.
• Rosenbaum et al. (1995) Planning reaches by
  evaluating stored postures. Psych. Rev. 102
  28-67.
• Sweet (1877) A handbook of phonetics. Reprinted
  in 1970 by McGrath Publishing Co., College Park,
  Maryland.
• Wallis (1653) Grammatica linguae anglicanae.
  Edited/translated by J. A. Kemp, 1972. London,
  U.K. Longman.
• Watson (1990) Acquiring the voicing contrast in
  French A comparative study of monolingual and
  bilingual children. In J. N. Green W.
  Ayres-Bennett (Eds.), Variation and change in
  French Essays presented to Rebecca Posner on the
  occasion of her sixtieth birthday (pp. 37-60).
  London Routledge.
• Watson (1991) Phonological processing in two
  languages. In E. Bialystok (Ed.), Language
  processing in bilingual children (pp. 25-48).
  Cambridge Cambridge University Press.