slides - Linguistics

The Role of Canonicity in Shaping Perceptual 

Biases in Speech Processing 

Charles B. Chang 1 and Alan Mishler 

University of Maryland, College Park 

1 cbchang@umd.edu 

13th Conference on Laboratory Phonology 

July 27, 2012

Outline 

Background 

Experiment 1: 4AFC Identification 

Experiment 2: AX Discrimination 

Discussion 

1 Background 

Transfer in L2 Speech Perception 

Unreleased Stops: English vs. Korean 

Research Questions 

2 Experiment 1: 4AFC Identification 

Methods 

Results 

3 Experiment 2: AX Discrimination 

Methods 

Results 

4 Discussion 

Nature of Language Transfer 

Future Directions 

Charles B. Chang, Alan Mishler Canonicity in Speech Processing

Background 



Discussion 




Non-veridicality of non-native speech perception 

Non-native (L2) listeners tend to be worse than native (L1) 

listeners at perceiving phonemes in a language. 

[Nábělek and Donahue 1984, Bradlow and Pisoni 1999, Garcia Lecumberri and Cooke 2006, Cutler et al. 2008] 

cognitive constraints 

[Goh 2000] 

partial knowledge of the L2 vocabulary, grammar, and culture 

[Bloomfield et al. 2011] 

phonological interference (viz., transfer) from L1 

[Lado 1957, Best 1994, Best 1995] 


Background 



Discussion 

Native language biases as obstacle 




L1 transfer results in L2 performance that is either worse than L1 

performance (“negative” transfer) or not very different from L1 

performance (“positive”, i.e. neutral, transfer). [Odlin 1989] 

Negative transfer: L1 processing of similar sounds as the same 

phoneme becomes a hindrance in an L2 where the sounds are 

different phonemes. [Flege 1995, Cutler 2001, Best and Tyler 2007] 

L1 Japanese speakers have trouble with English /l/ and /r/, 

even after learning to pronounce the distinction reliably. 

[Goto 1971, Sheldon and Strange 1982, Yamada and Tohkura 1992, Yamada 1995] 

L1 English speakers have trouble with Japanese short and long 

consonants and vowels. [Han 1992, Tajima et al. 2008] 


Background 



Discussion 

Native language biases as innocuous 




Neutral transfer: parallelism between L1 and L2 phoneme 

inventories results in little trouble with perception of L2 sounds. 

Japanese: [l] and [r] are variants of the same phoneme > 

great difficulty distinguishing between English /l/ and /r/. 

German: /l/ and /r/ are different phonemes > little difficulty 

distinguishing between between English /l/ and /r/. 

[Iverson et al. 2003] 

Possibility of advantageous L1 transfer? 

Test case: perception of unreleased stops. 


Background 



Discussion 




Final stops in English: released or unreleased 

Frequency (Hz) 

Frequency (Hz) 

6000 

6000 

0 

0 p u p p u tt 

1 

Time (s) 

0 

0 p u p p u tt 

1 

Time (s) 

released: 

[p h 5p (h) ] vs. [p h 5t (h) ] 

unreleased: 

[p h 5p^] vs. [p h 5t^] 


Background 



Discussion 




Frequency vs. canonicity of unreleased stops in English 

Final stop variants in American English demonstrate a disparity 

between frequency (rate of occurrence) and canonicity (relative 

status). 

Final voiceless stops are frequently unreleased at all POAs. 

/p/: produced without release 40–60% of the time 

/t/: produced without release 43–72% of the time 

/k/: produced without release 15–75% of the time 

[Byrd 1993, Davidson 2011] 

Released stops are recognized as canonical. 

unreleased [t^] more frequent than released [t (h) ] word-finally 

released [t (h) ] more effective prime [Sumner and Samuel 2005] 


Background 



Discussion 




Canonicity of unreleased stops in Korean 

Final stops in Korean are obligatorily unreleased (i.e., frequent + 

canonical). [Sohn 1999] 

Final stop contrast has a heavy functional load (many minimal 

pairs/triplets). 

[p ^ ap^] ‘rice’, [p ^ at^] ‘field’, [p ^ ak^] ‘outside’ 

[t ^ ap^] ‘answer’, [t ^ at^] ‘close’, [t ^ ak^] ‘chicken’ 

[k ^ ap^] ‘pay back’, [k ^ at^] ‘be similar’, [k ^ ak^] ‘each’ 


Background 



Discussion 




Native language biases as advantageous? 

Question: Are unreleased stops in American English better 

perceived by Americans (L1 English speakers) or by Koreans in the 

U.S. (L2 English speakers with L1 Korean)? 

Factors influencing listeners’ performance favor both groups. 

Americans: greater experience with American English 

efficient processing of English speech 

familiarity with language-specific properties of unreleased stops 

[Tsukada 2006, Tsukada et al. 2007, Tsukada and Roengpitya 2008] 

Koreans: greater experience with unreleased stops 

bias to attend to formant transition cues to final stops 

[Abramson and Tingsabadh 1999, Cho and McQueen 2006] 


Background 



Discussion 




A canonicity benefit in speech perception 

Hypothesis: Canonicity leads to a perceptual advantage. 

Prediction: Koreans will be better than Americans at perceiving 

unreleased stops in English. 

without lexical information (Experiment 1) 

with lexical information (Experiment 2) 


Participants 

Background 



Discussion 

Methods 

Results 

Talkers: 

2 native speakers of English from Maryland 

both m., age 19 and 25 yr 

no experience with obligatorily unreleased stops 

Listeners: 

25 native speakers of English born/raised in the U.S. 

11 m., mean age 21.1 yr (SD 5.4) 


25 native speakers of Korean born/raised primarily in S. Korea 

11 m., mean age 27.6 yr (SD 7.5) 

mean AOA to the U.S. = 21.1 yr (SD 9.0) 


Stimuli 

Background 



Discussion 

Methods 

Results 

56 bisyllabic English nonce words of the shape C1V1C2V2C3 

C1 = ô, V1 = {5, @}, C2 = z 

V2 = {i, u, A, eI, oU, AI, Aô}, C3 = {p^, t^, k^, ∅} 

stress = initial or final 

e.g., ô@"zAIp^ (rezipe) v. "ô5zAIp^ (ruzzipe) 

Recorded in a sound-attenuated booth. 

Zoom H4N audio recorder, Audix HT5 head-mounted mic 

44.1-kHz sampling, 24-bit resolution 

items presented via English spelling 

Releases removed and intensity normalized using Praat. 

[Boersma and Weenink 2011] 

ensured that oral closure was realized 

“dereleased” stops perceived similarly to unreleased stops 

[Lisker 1999] 


Procedure 

Background 



Discussion 

Methods 

Results 

Task: speeded 4AFC identification of finals in English nonce words 

(played in a carrier sentence: sample trial) 

stimuli presented through headphones in E-Prime 

responses entered on a serial response box 

Analysis: logistic mixed-effects modeling of identification accuracy. 

[Jaeger 2008, Dixon 2008] 

starting with random-effect terms for Participant and Item 

adding fixed-effect terms for Final (unreleased stop or 

non-stop), Group (Americans or Koreans), and a Final x 

Group interaction 


Background 



Discussion 

Methods 

Results 

Identification of unreleased stops in English 

Accuracy (%) 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

stop sonorant 

Final 

Group 

Americans 

Koreans 

Model improved by: 

Final 

[χ 2 (1) = 106.71, p < 0.0001] 

Group 

Charles B. Chang, Alan Mishler Canonicity in Speech Processing 

[χ 2 (1) = 4.633, p = 0.031] 

Final x Group 

[χ 2 (1) = 5.534, p = 0.019] 

Korean advantage for: 

final stops 

[β = 0.355, p = 0.017] 

No difference in RTs 

[F (1, 48) = 1.656, p = 0.204]

Background 



Discussion 

Methods 

Results 

Identification of unreleased stops in English 

Accuracy (%) 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

stop sonorant 

Final 

Group 

Americans 

Koreans 

Korean advantage for 

final stops applies to 

English syllable nuclei 

absent from Korean: 

/eI, oU, AI, Aô/. 

[β = 0.492, p = 0.002] 


Participants 

Background 



Discussion 

Methods 

Results 

Talkers: 

2 native speakers of English from Maryland 

both m., age 19 and 25 yr 


Listeners: 

25 native speakers of English born/raised in the U.S. 

11 m., mean age 21.1 yr (SD 5.4) 


25 native speakers of Korean born/raised primarily in S. Korea 

11 m., mean age 27.6 yr (SD 7.5) 

mean AOA to the U.S. = 21.1 yr (SD 9.0) 


Stimuli 

Background 



Discussion 

Methods 

Results 

48 minimal pairs of monosyllabic English words 

stop/stop (e.g., lip–lick), stop/zero (e.g., peek–pee) 

balanced for spoken frequency using COCA [Davies 2008-] 

Recorded in a sound-attenuated booth. 

Zoom H4N audio recorder, Audix HT5 head-mounted mic 

44.1-kHz sampling, 24-bit resolution 

items presented via English spelling 

Releases removed and intensity normalized using Praat. 

[Boersma and Weenink 2011] 

ensured that oral closure was realized 

“dereleased” stops perceived similarly to unreleased stops 

[Lisker 1999] 


Procedure 

Background 



Discussion 

Methods 

Results 

Task: speeded AX discrimination of English minimal pairs 

(contrasting in terms of final: sample trial 1, 2) 

stimuli presented through headphones in E-Prime 

responses entered on a serial response box 

Analysis: linear mixed-effects modeling of perceptual sensitivity, as 

indexed by d’. [Macmillan and Creelman 2005] 

starting with random-effect term for Participant 

adding fixed-effect terms for Contrast (stop/stop or 

stop/non-stop), Group (Americans or Koreans), and a 

Contrast x Group interaction 


Background 



Discussion 

Methods 

Results 

Discrimination of unreleased stops in English 

d' 

2.0 

1.5 

1.0 

0.5 

0.0 

stop/stop stop/zero 

Contrast 

Group 

Americans 

Koreans 

Charles B. Chang, Alan Mishler Canonicity in Speech Processing 

Model improved by: 

Contrast 

[χ 2 (1) = 63.862, p < 0.0001] 

Contrast x Group 

[χ 2 (1) = 8.274, p = 0.004] 

Korean advantage for: 

stop/zero contrasts 

[β = 0.321, p = 0.041] 

No difference in RTs 

[F (1, 48) = 1.675, p = 0.202]

Background 



Discussion 



Positive effects of phonological transfer 

Koreans were better than Americans at perceiving unreleased stops 

in English. 

Experiment 1: better at identifying unreleased stops in English 

nonce words. 

Experiment 2: better at discriminating between the presence 

and absence of an unreleased stop in real English words. 

Advantageous L1 transfer: 

Canonicity of unreleased stops in Korean ≫ high frequency of 

unreleased stops in American English + L1 processing 

advantage + L1-specific phonetic knowledge. 

Canonicity of released stops in English seems to result in a 

reliance on burst cues to final stops. [Wang 1959] 


Background 



Discussion 

Generality of advantageous transfer 



Koreans’ perceptual advantage was not a coincidence. 

Experiment 1: better at identifying unreleased stops following 

English vowels dissimilar from any Korean vowel. 

Experiment 2: better at discriminating stop/zero pairs of 

English words that are meaningful. 

General benefit of L1 transfer: 

Koreans abstract from their L1 experience a general ability to 

extract information about a final stop from formant transition 

cues (which can then be applied to new L2 contexts). 


Background 



Discussion 



Variable nature of phonological transfer 

The conceptualization of L1 transfer in terms of negative or 

neutral effects on L2 speech perception is inadequate. 

occurrence of positive L1 transfer (boosting L2 perception 

beyond L1 levels) 

L1 transfer influences L2 perception in a manner dependent upon 

the alignment of properties in L1 and L2. 

absence of examples of advantageous L1 transfer due to the 

kinds of L1-L2 pairings examined in the literature 

favorable L1-L2 alignment > advantageous L1 transfer 


Background 



Discussion 



Advantageous transfer in other domains? 

L2 speakers are expected to display a perceptual advantage over 

L1 speakers in other cases of non-coincidence between frequency 

and canonicity. 

e.g., nasalized vowels in English vs. nasalized vowels in French 

Is English vowel nasalization better perceived by L1 French or 

L1 English speakers? 

The full range of transfer effects appear varied, suggesting fruitful 

avenues of further research in cross-linguistic speech perception. 


Acknowledgements 

Appendix 

Acknowledgments 

References 

Center for Advanced Study of Language, University of Maryland 

Department of Hearing and Speech Sciences, University of Maryland 


References I 

Appendix 

Abramson, A. S., Tingsabadh, K. (1999). 

Thai final stops: Cross-language perception. 

Phonetica 56, 111–122. 


References 

Best, C. T. (1994). 

The emergence of native-language phonological influences in infants: A perceptual assimilation model. 

In The Development of Speech Perception: The Transition from Speech Sounds to Spoken Words, ed. J. C. 

Goodman and H. C. Nusbaum, pp. 167–224. Cambridge, MA: MIT Press. 

Best, C. T. (1995). 

A direct realist view of cross-language speech perception. 

In Speech Perception and Linguistic Experience: Issues in Cross-Language Research, ed. W. Strange, pp. 

171–204. Baltimore, MD: York Press. 

Best, C. T., Tyler, M. D. (2007). 

Nonnative and second-language speech perception: Commonalities and complementarities. 

In Language Experience in Second Language Speech Learning: In Honor of James Emil Flege, ed. O.-S. 

Bohn and M. J. Munro, pp. 13–34. Amsterdam, The Netherlands: John Benjamins Publishing. 

Bloomfield, A., Wayland, S. C., Rhoades, E., Blodgett, A., Linck, J., Ross, S. (2011). 

What makes listening difficult? Factors affecting second language listening comprehension. 

Technical Report TTO 81434 E.3.1, University of Maryland Center for Advanced Study of Language, 

College Park, MD. 

Boersma, P., Weenink, D. (2011). 

Praat: Doing phonetics by computer, Version 5.3. http://www.praat.org. 


References II 

Appendix 


References 

Bradlow, A. R., Pisoni, D. B. (1999). 

Recognition of spoken words by native and non-native listeners: Talker-, listener-, and item-related factors. 

Journal of the Acoustical Society of America 106, 2074–2085. 

Byrd, D. (1993). 

54,000 American stops. 

UCLA Working Papers in Phonetics 83, 97–116. 

Cho, T., McQueen, J. M. (2006). 

Phonological versus phonetic cues in native and nonnative listening: Korean and Dutch listeners’ perception 

of Dutch and English consonants. 


Cutler, A. (2001). 

Listening to a second language through the ears of a first. 

Interpreting 5, 1–23. 

Cutler, A., Garcia Lecumberri, M. L., Cooke, M. (2008). 

Consonant identification in noise by native and non-native listeners: Effects of local context. 


Davidson, L. (2011). 

Characteristics of stop releases in American English spontaneous speech. 

Speech Communication 53, 1042–1058. 


References III 

Appendix 


References 

Davies, M. (2008-). 

The corpus of contemporary American English: 425 million words, 1990–present. 

Available online at http://corpus.byu.edu/coca/. 

Dixon, P. (2008). 

Models of accuracy in repeated-measures designs. 

Journal of Memory and Language 59, 447–456. 

Flege, J. E. (1995). 

Second language speech learning: Theory, findings, and problems. 



Garcia Lecumberri, M. L., Cooke, M. (2006). 

Effect of masker type on native and non-native consonant perception in noise. 

Journal of the Acoustical Society of America 119, pp. 2445–2454. 

Goh, C. C. M. (2000). 

A cognitive perspective on language learners’ listening comprehension problems. 

System 28, 55–75. 

Goto, H. (1971). 

Auditory perception by normal Japanese adults of the sounds “L” and “R”. 

Neuropsychologia 9, 317–323. 


References IV 

Appendix 


References 

Han, M. (1992). 

The timing control of geminate and single stop consonants in Japanese: A challenge for nonnative speakers. 

Phonetica 49, 102–127. 

Iverson, P., Kuhl, P. K., Akahane-Yamada, R., Diesch, E., Tohkura, Y., Kettermann, A., Siebert, C. (2003). 

A perceptual interference account of acquisition difficulties for non-native phonemes. 

Cognition 87, B47–B57. 

Jaeger, T. F. (2008). 

Categorical data analysis: Away from ANOVAs (transformation or not) and towards logit mixed models. 


Lado, R. (1957). 

Linguistics Across Cultures: Applied Linguistics for Language Teachers. Ann Arbor, MI: University of 

Michigan Press. 

Lisker, L. (1999). 

Perceiving final voiceless stops without release: Effects of preceding monophthongs versus 

nonmonophthongs. 

Phonetica 56, 44–55. 

Macmillan, N. A., Creelman, C. D. (2005). 

Detection Theory: A User’s Guide, 2nd edition. Mahwah, NJ: Lawrence Erlbaum Associates. 


References V 

Appendix 


References 

Nábělek, A. K., Donahue, A. M. (1984). 

Perception of consonants in reverberation by native and non-native listeners. 


Odlin, T. (1989). 

Language Transfer: Cross-Linguistic Influence in Language Learning. Cambridge, UK: Cambridge University 

Press. 

Sheldon, A., Strange, W. (1982). 

The acquisition of /r/ and /l/ by Japanese learners of English: Evidence that speech production can 

precede speech perception. 

Applied Psycholinguistics 3, 243–261. 

Sohn, H.-M. (1999). 

The Korean Language. Cambridge, UK: Cambridge University Press. 

Sumner, M., Samuel, A. G. (2005). 

Perception and representation of regular variation: The case of final /t/. 


Tajima, K., Kato, H., Rothwell, A., Akahane-Yamada, R., Munhall, K. G. (2008). 

Training English listeners to perceive phonemic length contrasts in Japanese. 



References VI 

Appendix 


References 

Tsukada, K. (2006). 

Cross-language perception of word-final stops in Thai and English. 

Bilingualism: Language and Cognition 9, 309–318. 

Tsukada, K., Nguyen, T. T. A., Roengpitya, R., Ishihara, S. (2007). 

Cross-language perception of word-final stops: Comparison of Cantonese, Japanese, Korean and Vietnamese 

listeners. 

In Proceedings of the 16 th International Congress of Phonetic Sciences, ed. J. Trouvain and W. J. Barry, 

pp. 1781–1784. Dudweiler, Germany: Pirrot. 

Tsukada, K., Roengpitya, R. (2008). 

Discrimination of English and Thai words ending with voiceless stops by native Thai listeners differing in 

English experience. 

Journal of the International Phonetic Association 38, 325–347. 

Wang, W. S. Y. (1959). 

Transition and release and perceptual cues for final plosives. 

Journal of Speech and Hearing Research 2, 66–73. 

Yamada, R. A. (1995). 

Age and acquisition of second language speech sounds: Perception of American English /r/ and /l/ by 

native speakers of Japanese. 




References VII 

Yamada, R. A., Tohkura, Y. (1992). 

Appendix 


References 

Perception of American English /r/ and /l/ by native speakers of Japanese. 

In Speech Perception, Production and Linguistic Structure, ed. Y. Tohkura, E. Vatikiotis-Bateson, and 

Y. Sagisaka, pp. 155–174. Tokyo, Japan: Ohmsha.

slides - Linguistics

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?