Slide 1

Cross-linguistic perception of Thai tones is shaped by the functional prominence of lexically-contrastive pitch in L1Vance Schaefer and Isabelle DarcyDepartment of Second Language StudiesIndiana University New Sounds 2013Montreal, Quebec, CanadaConcordia UniversityMay 17-19, 2013

0 Tone Tone languages use variations of voice height = pitch, or F0 to distinguish words.

Patterns : LEVEL or CONTOUR1 What is tone? Pitch variations that affect the meaning of a word Ladefoged (2006:248)Pitch variations include LEVEL and CONTOUR tones.The most known example is perhaps Mandarin Chinese.We have the following:CLICK TO SEE 1) WORD, MEANING, HIGH LEVELCLICK TO SEE 2) THE VISUAL ARROWCLICK ON THE SOUND FILE TO HEAR.

Every one of these means something different. But to most of us, they seem quite similar perhaps due to L1 influences.

*Aware of neutral tone = no pitch phonologically, but the pitch level of the previous syllable in a word affects the pitch phonetically

1

Thai tones4 n: face fallingn: thick rising

n: aunt high level

na: rice field mid leveln: custard apple low level

Source: Contour shapes of Thai tones in citation form. Representative examples from one speaker. From Zsiga & Nitisaroj, 2007, p. 3472 contour tones3 level tones

2 These are the target stimuli from Thai.

As you can see, there are THREE LEVEL TONES and TWO CONTOUR TONES.

They sound as follows.2Tone perception by native speakersNative speakers perceive tones as linguistic categories Van Lancker & Fromkin, 1973; Wang, Jongman & Sereno, 2001

Tonal information also constrains lexical access Lee, 20076 3 Speakers of non-tonal languages have less sensitivity to tonal contrasts than people with previous tonal experience (e.g., Wang et al., 2006). Yet the question remains whether all non-tonal language speakers experience similar difficulties with tonal contrasts (e.g., Burnham et al., 1996). 3Tone perception by non-native speakersSpeakers of a tonal language display high accuracy in non-native tone perception Wayland & Guion, 2004

Speakers of non-tonal languages have less sensitivity to tonal contrasts than people with previous tonal experience Hall, Chang & Best, 2004, for French listeners; Gandour & Harshman, 1978; Wang, Behne, Jongman & Sereno, 2004, among others7 4 Speakers of a tonal language display high accuracy in the perception of tones in another tone language (non-native tone perception)In contrast, speakers of languages that do not have tones show less sensitivity to tonal contrasts.4Do all non-tonal language speakers perform equally in non-native tone perception? There are differences AMONG non-tonal language speakers in non-native tone perception e.g., L1 pitch accent speakers perform at comparable accuracy levels to L1 tone language speakers Burnham et al., 1996; So, 2006

Languages differ in the extent and function to which they use F0 variations:

All languages use pitch for intonation at the level of phrases while only some use pitch for distinctions at the word level8 5 5Non-lexicalLexically-contrastive pitch usage Tone e.g., Mandarin Chinese, Thai, Vietnamese Pitch-accent languagesHigh pitch on the accented mora, determining the pitch level (H or L) of preceding/following moras (+ more rules)e.g., Japanese, Swedish e.g., A-me rain (HL) vs a-ME candy (LH) Word-stress languagesPitch variation as one correlate of lexically-contrastive word stress e.g., English, German, Spanish. e.g., REcord vs reCORD Intonation - only languagesThese languages do not use lexically-contrastive pitch, but like all languages we know of, they use intonation (phrase domain)e.g., Korean, French

10 Lexical

6 Phonological systems differ in several ways. Segmental. Syllabic structure. Phonological processes. Suprasegmentals. And most importantly for this talk, tone.

CLICK

YOU NEED TO CLICK AT THE COMPUTER TO GET TO THE NEXT SLIDE ABOUT WHAT IS TONE AND MANDARIN TONE.

Japanese F0 drop (Sugiyama, 2008, 2012)

6

Functional scale of pitch contrasts

11

Adapted from Van Lancker, 1980: 210

7 We could arrange these language types in a functional scale, such as one proposed by Van Lancker in 1980. In this figure, we see that the way a language uses pitch contrasts vary in domain (from small to large), and also in the function that the pitch will fulfill (from tonal contrasts to marking e.g. focus structure or affect). The five languages this study examines fall in different sections. READ THEM AND EXPLAIN THEM:Thai and Chinese are very similar here, as they both use tone as a phonological contrast and at the level of the segment/ syllable.Then, Japanese and English are also similar in that they both F0 contrasts lexically, at the level of words.Finally, Korean here would fall in the last section, as standard Korean does not use F0 contrasts lexically, but only through intonation at the sentence level. 7Pitch prominence typology and predictions for tone perception accuracyLanguageDomainProminenceTone (Mandarin)Lexical, syllableMaximalPitch-accent (Japanese)Lexical, wordHigh-intermediate (pitch is exclusive)Word stress (English)Lexical, wordLow-intermediate (pitch is non-exclusive)Intonation-only (Korean)Non lexicalLow 8 Here we combine the idea of functional use of pitch with the prominence idea. We organize the languages we use in this study according to the role of pitch for each. The three languages up here all use pitch to distinguish words (even though the domain at which they do so varies) whereas Korean uses pitch in a domain larger than a word and does not use pitch to contrast words.

Based on this, we established a scale of prominence going from maximal to low. We also have an additional distinction between English and Japanese based on whether or not pitch can be exclusively used to distinguish words.

CLICK and say:So, we predict that if the prominence of pitch to distinguish words in your first language is high, then your accuracy in tone perception will also be high. Low/low

8Pitch prominence typology and predictions for tone perception accuracyLanguageDomainProminenceTone (Mandarin)Lexical, syllableMaximalPitch-accent (Japanese)Lexical, wordHigh-intermediate (pitch is exclusive)Word stress (English)Lexical, wordLow-intermediate (pitch is non-exclusive)Intonation-only (Korean)Non lexicalLowPredicted Sensitivity/ Accuracy in tone perception 8 Here we combine the idea of functional use of pitch with the prominence idea. We organize the languages we use in this study according to the role of pitch for each. The three languages up here all use pitch to distinguish words (even though the domain at which they do so varies) whereas Korean uses pitch in a domain larger than a word and does not use pitch to contrast words.

Based on this, we established a scale of prominence going from maximal to low. We also have an additional distinction between English and Japanese based on whether or not pitch can be exclusively used to distinguish words.

CLICK and say:So, we predict that if the prominence of pitch to distinguish words in your first language is high, then your accuracy in tone perception will also be high. Low/low

9Pitch Prominence Hypothesis13 Similar predictions are found in previous studies

Feature Hypothesis McAllister, Flege, & Piske, 2002: L2 perception of Swedish vowel length contrasts by native speakers of Estonian, English, and Spanish

Linguistic relevance of a dimension in L1 shapes the brain response to L2 contrasts (with MMN data) Nenonen, Shestakova, Huotilainen, & Ntnen, 2003

We predict accuracy of cross-language tone perception based on prominence of pitch in the L1 9 Referencing the Feature Hypothesis (McAllister et al., 2002), we put forward a more narrowly defined Pitch Prominence Model to explain the perception of non-native lexically-contrastive pitch: we hypothesize that the degree to which linguistic pitch differentiates lexical items in the L1 (i.e., prominence) will shape the nave perception of a non-native tonal system.

We expect the following hierarchy of performance in the accuracy of non-native/Thai tone perception as shaped by the prominence of lexically-contrastive pitch usage in the L1.

X-axis is prominence in the languages in the study as seen in an earlier slide.10Prominence predicts accuracy11 Maximal --- Prominence of contrastive pitch at the word level --- NonePredicted AccuracyKoreanEnglishJapaneseMandarin14 10 We expect the following hierarchy of performance in the accuracy of non-native/Thai tone perception as shaped by the prominence of lexically-contrastive pitch usage in the L1.

X-axis is prominence in the languages in the study as seen in an earlier slide.11 Methodology

DONT mention excluding Thaiyet.

We also tested two Thai speakers as a control to validate the design of the task.12Participants

N = 2 Thai native speakersN = 10 Mandarin speakersN = 11 Japanese speakersN = 10 English speakersN = 10 Korean speakers- Graduate students- Generally involved in language studies/linguistics- Students in the US11 DONT mention excluding Thaiyet.

We also tested two Thai speakers as a control to validate the design of the task.13AXB categorization

500 msAccuracy rates and reaction times

12 What happens in an AXB task is that you hear three stimuli in a row and you have to decide whether the middle one is more like the first one or second one. To tap a more phonological level of processing, we use two different voices. Voice one is female and voice two is male.

What is usually looked at in an AXB task are accuracy rates and reaction times.14Experimental conditions- Monosyllabic words & nonwords presented in triplets (48 test, 48 control)- All test words were open syllables- 3 test conditions:

Test ConditionsControl ConditionDirection (n=12)Height (n=12)Mixed (n=24)Control (n=48)rising-fallinglow-midlow-risinglow-fallingconsonantvowelrising-fallinglow-highmid-risingmid-fallingrising-fallingmid-highhigh-risinghigh-falling13 48 triplets: monosyllables differing in segments but not in tone48 triplets: monosyllables differing in tone but not in segments

15 Results

16Accuracy rates in each group*ns* = significant effect of group Significant interaction between group and condition: F (3, 37) = 11.3, p < .001 Effect of group is significant for test condition only : F(3, 67.3) = 11.3, p < .001 Predicted hierarchy of accuracy: Mandarin (M = 87% correct), Japanese (M = 77% correct), English and Korean (M = 67 % correct for both).14 This is the accuracy rate for each group. On the left side, it is the accuracy on the test condition and on the right side it is the accuracy on the control condition. The black bar represents the Thai listeners and so, we excluded them from the analysis because first of all there are only two and because we had words in the stimuli, there were able to approach the task using lexical knowledge.

So, what we see here is that there is a significant effect of group/language only on the test condition. We can clearly see that globally the data on the text conditions conforms to the accuracy scale we predicted.17Reaction times in each group Interaction was not significant: F (3, 37) = 2.4, p = 0.08 15 We obtained a similar pattern for reaction times, which seems to mirror the accuracy scores, but it was statistically not significant, suggesting that all groups were similar in that respect. We clearly dont have a strategic effect.

18Conclusions Influence of the L1 phonological system

The functional prominence of lexically- contrastive pitch in L1 shapes cross- linguistic perception of Thai tones

Globally, our findings confirm previous results obtained across studies and add strength by allowing a direct comparison with the same methodology16

What we learned from the study is thatREAD19Discussion: Overall performance

Equal accuracy between English and Korean in tone discrimination was not predicted. Why?

Are English less accurate than expected?F0 is rarely used alone to distinguish words in English, perhaps creating the same performance as if F0 was not used at all to signal lexical contrast (English = Korean)Stress constrains lexical access only to a limited extent in English (Cooper, Cutler & Wales, 2002)By contrast, when F0 can be used alone to distinguish words, as in Japanese, performance is higher

Are Koreans more accurate than expected?Influence of L2 English on Koreans? Exposure to a pitch-accent Kyungsang dialect?

17

1-3 one argumentHowever, perhaps the Koreans are points 4 and 520Individual Korean Dialectal Differences

dialectal exposure defines prominence, as exemplified by data from three Korean-speakers of the Kyungsang pitch-accent dialect, who outperformed non-pitch-accent dialect speakers (i.e., Seoul). 21Kyungsang Korean

Kyungsang = GyeongsangCholla = Jeolla Dialectal boundaries Lee & Ramsey, 2000

18 ()gyeongsang dialect22Lexical pitch in KoreanKyungsang listeners show categorical perception of Pitch accent patterns Kim & de Jong, 2007; Kim, 2011

Limited advantage in the nave perception of Japanese pitch accent Sukegawa, Choi, Maekawa & Sato, 1995

Emergence of lexical pitch in standard Korean among younger speakers Silva, 2006

19 There is a study in Japanese concerning the perception of Japanese pitch accent by Korean speakers of the Kyungsang dialect in an AXB task of nonce words, showing one possible advantage in nonnative pitch perception over speakers of standard Korean (i.e., for CVH syllables where H indicates a long vowel, as opposed to CV-CV and CV-N syllables) (Sukegawa, Choi, Maekawa & Sato, 1995). Sukekawa, Y., Choi, H., Maekawa, K., & Sato, S. (1995). Perception of pitch accent by Korean learners of Japanese and its implications. Denshi Joho Tsushin Gakkai gijutsu kenkyu hokoku: shingaku giho, 95(41)/19950518, 61-66.

23Pitch accent in Korean Kyungsang dialectMinimal pairs of 3 lexical accent patterns

a. [moi]: HL vs. LH feed, conspiracyb. [moe]: HL vs. HH sand, the day after tomorrow c. [yamo]: LH vs. HH wool, adoptive mother From Kim, 2011; Kim & de Jong, 2007

20 Kim & de Jong, 2007Kim, J.-S., & de Jong, K.J. (2007). Perception and Production in the Pitch Accent System of Korean. In J. Trouvain and W. J. Barry (Eds.), ICPhS XVI (pp. 1273 1277).24PredictionsIf the L1 phonological system determines accuracy, Kyungsang Korean dialect speakers should outperform non-Kyungsang speakers

We examine individual performance for the Korean group

21 Kim & de Jong, 2007Kim, J.-S., & de Jong, K.J. (2007). Perception and Production in the Pitch Accent System of Korean. In J. Trouvain and W. J. Barry (Eds.), ICPhS XVI (pp. 1273 1277).25Korean performance on combined test items

22 Here we see that all the Pitch accent speakers are on the higher end of the accuracy spectrum. 26Korean performance on control items

23 But on the control condition, we see that the two subgroups behave comparably.27Accuracy rates for each Korean subgroupWe conclude that the Korean group most likely performed More accurately than expected because of the dialect differences within that group24 This is the same graph as we presented before except that we split here the two Korean subgroups, to compare their performance to the other groups. For the test condition, we see that theres a clear difference between the two subegrups, and we also observe that the blue bar (the non-puitch accent) is in fact less accurate than the English, which conforms to our original prediction. The pitch accent subgroup, by contrast OUTperforms the English, and is more comparable to the Japanese. 28Take home message Influence of the L1 phonological system - in a narrow sense, i.e. L1 dialect

The functional prominence of lexically-contrastive pitch in L1 shapes cross- linguistic perception

Further support for the Feature Hypothesis (McAllister et al., 2002): Accuracy of perception of non-native phonological dimensions is shaped by the prominence of that dimension in the L1 phonological system

For pitch: Exclusivity and domain size matter to determine prominence

25 What we learned from the study is thatREAD29Kathleen Bardovi-Harlig Laurent DekydtspotterKen De JongStephanie Dickinson Mariko KondoKeiko KuriyamaPhilip LeSourd Charles Linner zelikRex SprouseDavid StringerSecond Language Psycholinguistics Lab membersSLRF audienceLabPhon audienceSLS seminar classmates30

Acknowledgements30ReferencesBurnham, D., Francis, E., Webster, D., Luksaneeyanawin, S., Attapaiboon, C., Lacerda, F., & Keller, P. (1996). Perception of lexical tone across languages: Evidence for a linguistic mode of processing. In H. T. Bunnell & W. Idsardi (Eds.), Proceedings of the Fourth International Conference on Spoken Language Processing (Vol. 1, pp. 25142517). Wilmington, DE: Applied Science and Engineering Laboratories.Cooper, N., Cutler, A., & Wales, R. (2002). Constraints of lexical stress on lexical access in English: Evidence from native and non-native listeners. Language and Speech, 45(3), 207-228.Gandour, J., & Harshman, R. (1978). Crosslanguage differences in tone perception: a multidimensional scaling investigation. Language and Speech, 21, 133.Hall, P. A., Chang, Y-C. & Best, C.T. (2004). Identification and discrimination of Mandarin Chinese tones by Mandarin Chinese vs French listeners. Journal of Phonetics, 32, 395-421.Kim, J.-S. (2011). Perception of Lexical Pitch Accent by Kyungsang and Cholla Korean Listeners. In W.-S. Lee, & E. Zee (Eds.), Proceedings of the 17th International Congress of Phonetic Sciences 2011 [ICPhS XVII] (pp. 1070-1073). Hong Kong: Department of Chinese, Translation and Linguistics, City University of Hong Kong.

26

Best, C. (1995). A direct realist view of cross-language speech perception. In Winifred Strange (ed.), Speech perception and linguistic experience: Issues in cross-language research (pp. 171-204). Baltimore, MD: York Press. Best, C. T., & Tyler, M. D. (2007). Nonnative and second-language speech perception: Commonalities and complementarities. In M. J. Munro & O.-S. Bohn (Eds.), Second language speech learning: The role of language experience in speech perception and production (pp. 13-34). Amsterdam: John Benjamins.Bohn, O.-S. (1995). Cross-language speech perception in adults: First language transfer doesn't tell it all. In W. Strange (Ed.), Speech perception and linguistic experience. Issues in cross-language research (pp. 273-304). Baltimore: York Press.Braun, B., & Johnson, E. K. (2011). Question or tone 2? How language experience and linguistic function guide pitch processing. Journal of Phonetics, 39, 585-594. Gandour, J. (1983). Tone perception in Far Eastern languages. Journal of Phonetics, 11, 49-175.

31ReferencesKim, J.-S., & de Jong, K.J. (2007). Perception and Production in the Pitch Accent System of Korean. In J. Trouvain and W. J. Barry (Eds.), Proceedings of the 16th International Congress of Phonetic Sciences 2007 [ICPhS XVI] (pp. 1273 1277). Dudweiler: Pirrot.Lee, I., & Ramsey, S. R. (2000). The Korean Language. Albany, New York: State University of New York Press.Lee, C-Y. (2007). Does Horse Activate Mother? Processing Lexical Tone in Form Priming. Language and Speech, 50(1), 101-123.McAllister, R., Flege, J. E., & Piske, T. (2002). The influence of L1 on the acquisition of Swedish quantity by native speakers of Spanish, English and Estonian. Journal of Phonetics, 30, 229-258.Nenonen, S., Shestakova, A., Huotilainen, M., & Naatanen, R. (2003). Linguistic relevance of duration within the native language determines the accuracy of speech-sound duration processing. Cognitive Brain Research, 16(3), 492-495.Silva, D. J. (2006). Acoustic evidence for the emergence of tonal contrast in contemporary Korean. Phonology, 23, 287-308.So, C. K. (2006). Perception of non-native tonal contrasts: Effects of native phonological and phonetic influences. In P. Warren, & C. I. Watson (Eds.), Proceedings of the 11th Australian International Conference on Speech Science & Technology. Auckland, New Zealand: University of Auckland.

27

Guion, S. G., & Pederson, E. (2007). Investigating the role of attention in phonetic learning. In O.-S. Bohn & M. Munro (Eds.), Language experience in second language speech learning (pp. 57-77). Amsterdam: John Benjamins.Lai, Y. & Zhang, J. (2008). Mandarin lexical tone recognition: the gating paradigm. In E. Tummons & S. Lux (Eds.), Proceedings of the 2007 Mid-America Linguistics Conference, Kansas Working Papers in Linguistics 30. 183-194. Lee, L., & Nusbaum, H. C. (1993). Processing interactions between segmental and suprasegmental information in native speakers of English and Mandarin Chinese. Perception & Psychophysics, 53, 157-165. Repp, B. H., & Lin, H. B. (1990). Integration of segmental and tonal information in speech perception: A cross-linguistic study. Journal of Phonetics, 18(4), 481-495.

32ReferencesSukegawa, Y., Choi, H., Maekawa, K., & Sato, S. (1995). Perception of pitch accent by Korean learners of Japanese and its implications. The Institute of Electronics, Information and Communication Engineers (IEICE) technical report: Speech 95(41), 61-66.Van Lancker, D. (1980). Cerebral lateralization of pitch cues in the linguistic signal. Papers in Linguistics: International Journal of Human Communication, 13, 201277.Van Lancker, D., & Fromkin, V. A. (1973). Hemispheric specialization for pitch and tone: Evidence from Thai. Journal of Phonetics, 1, 101109.Wang, Y., Behne, D. M., Jongman, A. & Sereno, J. A. (2004). The role of linguistic experience in the hemispheric processing of lexical tone. Applied Linguistics, 25, 449-466.Wang, Y., Jongman, A., & Sereno, J. A. (2001). Dichotic perception of Mandarin tones by Chinese and American listeners. Brain and Language, 78, 332348.Wayland, R. P., & Guion, S. G. (2004). Training English and Chinese listeners to perceive Thai tones: A preliminary report. Language Learning, 54, 681-712.Zsiga, E., & Nitisaroj, R. (2007). Tone features, tone perception, and peak alignment in Thai. Language and Speech, 50(3), 343-383.

28

Wood, C. C. (1974). Parallel processing of auditory and phonetic information in speech discrimination. Perception & Psychophysics, 15 (3), 501-508.

33