Speech perception abilities in early and late bilinguals

Among the early L2 learners we can distinguish between sequential and simultaneous bilinguals. Early L2 learners can be contrasted to late L2 learners who learned their second language relatively late in life. How do early and late L2 learners compare to native speakers in terms of reaching proficiency in L2?  In quiet, L2 speakers show the same performance as native speakers in speech perception tasks (e.g. Nabelek & Donahue, 1984) while in background noise,  their speech perception in the second language is more affected than in the first language (Florentine, 1985a, b; Takata & Nabelek, 1990). This effect has been suggested to relate to listeners’ age (Bergman, 1980), the time-period of L2 study (Florentine, 1985a, b), and the environmental situation under which listening occurs (Takata & Nabelek, 1990).

Research by Florentine (1985b) showed that exposure to L2 from infancy onwards helped two L2 listeners to perform as well as L1 speakers on speech perception tasks in the presence of increasing noise. By contrast, L2 listeners who had been exposed to L2 only after puberty did not perform at the same level as L1 listeners of American English even after massive exposure. Moreover, L2 listeners did not make use of any contextual cues, which contrasts to the effects seen in L1 listeners. These data are interpreted as indicating a sensitive period after which learning a second language negatively affects L2 listeners’ perception of L2 in noise. Thus, in noise, L2 speakers’ performance on speech perception tasks has been shown to depend on the age at which they acquire L2 (Florentine, 1985b; Mayo et al., 1997).

It was shown that in speech perception tasks with noise, early learners of L2 performed better and benefitted more from sentence-level contextual information compared to late but very proficient L2 learners. However, early L2 learners’ ability to perceive L2 in noise has been suggested to be decreased and not be like that of native listeners’ due to intervention by L1 experience (Mayo et al., 1997). It has therefore been argued that early L2 learners’ better performance might be due to the age at which L2 was acquired and not the average time length of L2 exposure. Consequently, if L2 study is not started in early childhood, L2 listeners will have difficulty perceiving L2 in noise even with extensive exposure. This has been illustrated by early L2 learners showing higher levels of tolerating noise than late L2 learners. However, L1 English listeners had higher noise-tolerance levels than early L2 learners of English (Mayo et al., 1997).  L1 listeners have thus been claimed to recover quickly from noise-induced disturbance because of their linguistic knowledge of established L1 categories (Bradlow & Alexander, 2007). On the contrary, late L2 learners are not able to recover their speech perception that is disrupted by noise as quickly as L1 listeners because their lacking linguistic knowledge of L2 causes their recovery from noise to be too slow (Bradlow & Alexander, 2007).

Does this mean that late L2 learners will not be able to attain nativelike L2 proficiency? Luckily, findings that found nativelike attainment in L2 late learners, and non-native like attainment in early L2 learners suggest the answer is no (Birdsong, 1999, 2006; Bongaerts, 1999). For example, a study, in which native Vietnamese early and child learners of English were tested on a grammaticality judgement task, showed how they did not perform better or worse than native English speakers (MacDonald, 2000). Specifically, early Vietnamese learners were revealed to have lasting grammatical accents as well, suggesting that some early L2 learners do not show nativelike proficiency at all. One can therefore say that it is not certain that nativelike proficiency will result from an early exposure to a second language. In another study, native English speakers were asked to rate the English speech samples for accent that had been produced by L1 Dutch speakers who started learning English at about the age of 12 (Bongaerts, 1999). About half of these proficient learners of English were mistaken for native speakers of English suggesting that it is possible for late L2 learners to attain proficiency similar to native speakers with regard to pronunciation. This possibility of successful late learning of a second language was further supported by studies that replicated the previous study with learners of languages that are not closely related to L1 such as proficient Dutch L1 learners of French (Bongaerts, 1999), and late learners of Hebrew who had a different native language (Abu-Rabia & Kehat, 2004). Thus, an early start in learning a second language is not prerequisite in acquiring unaccented speech in L2.

Physiological evidence provided additional support in form of a study, in which it was shown how a group of adult participants who were trained on an artificial language demonstrated a similar pattern of brain activation during the processing of that language that is observed when native speakers process a natural language (Friederici et al., 2002). In particular, the observations of early negativity and late positivity (N400 and P600) as elicited by syntactic violations imply the processes of swift automatic parsing, and decelerated repair and reanalysis that are also observed during native speakers’ processing of syntactic violations. This finding seems to suggest that both early and late learners of a language make use of the same brain mechanism during the processing of that language. These results are supported by other ERP studies that have shown ERP patterns evoked in fluent L2 users that are largely observed in native speakers as well (Hahne & Friederici, 2001) while differences in ERP patterns were revealed when native and non-proficient speakers’ processing was compared (Ojima et al., 2005). Similarly, brain imaging studies found the same brain areas being activated when fluent L2 speakers and native speakers process a language (Perani et al., 1998) while brain activation in different regions were observed when native and non-proficient speakers’ processing was compared (Dehaene et al. 1997).

Can you become bilingual?

The observation that our ability to learn speech in different languages does remain functional over the course of life is good news for all of us who recently thought of taking up a second language. No matter what your age is, or which language you considered of learning, as long as you create yourself the ideal environment for you to learn a second language, with effort it is possible. For example, I learnt my second language after learning my first language; hence I am a sequential bilingual. The fact that both languages were rhythmically different languages and do not share the same sound system, helped me with learning both languages more efficiently, and to this day I am still learning bits and pieces in these two languages that help me understand how I can refine my communication skills. The learning therefore never stops! Thus, when you train yourself intensively in perceiving and producing the sounds of the second language, show the motivation and enthusiasm to sound nativelike, with massive L2 input, it will be possible for you to achieve your aim of becoming a fluent L2 speaker. 

References

 

Abu-Rabia, S., & Kehat, S. (2004). The critical period for second language pronunciation: Is there such a thing? Ten case studies of late starters who attained a native-like Hebrew accent. Educational Psychology, 24, 77-98.

Bergman, M. (1980). Aging and the perception of speech. Baltimore: University Park Press.

Birdsong, D. (1999). Introduction: Whys and why nots of the critical period hypothesis for second language acquisition. In D. Birdsong (Ed.), Second language acquisition and the critical period hypothesis (pp. 1-22). Mahwah, NJ: Erlbaum.

Birdsong, D. (2006). Age and second language acquisition and processing: A selective overview. Language Learning, 56, 9-48.

Bongaerts, T. (1999). Ultimate attainment in L2 pronunciation: The ease of very advanced late L2 learners. In D. Birdsong (Ed.), Second language acquisition and the critical period hypothesis (pp.133-149). Mahwah, NJ. Erlbaum.

Dehaene, S., Dupoux, E., Mehler, J., Cohen, L., Paulesu, E., Perani, D., et al. (1997). Anatomical variability in the cortical representation of first and second language. NeuroReport, 8, 3809-3815.

Florentine, M. (1985a). Non-native listeners’ perception of American-English in noise. Proceedings of Inter-Noise ’85, 1021-1024.

Florentine, M. (1985b). Speech perception in noise by fluent, non-native listeners. Proceedings of the Acoustical Society of Japan. H-85-16.

Hahne, A., & Friederici, A. D. (2001). Processing a second language: Late learners’ comprehension mechanisms as revealed by event-related brain potentials. Bilingualism: Language and Cognition, 4, 123-141.

 

Mayo, L. H., Florentine, M., & Buus, S. (1997). Age of second-language acquisition and perception of speech in noise Journal of Speech, Language & Hearing Research, 40, 686–93.

McDonald, J.L. (2000). Grammaticality judgements in a second language: Influences of age of acquisition and native language. Applied Psycholinguistics, 21, 395-423.

Nabelek, A.K., & Donahue, A.M. (1984). Perception of consonants in reverberation by native and non-native listeners. Journal of Acoustical Society of America, 75, 632-634.

Ojima, S., Nakata, H., & Kakigi, R. (2005). An ERP study of second language learning after childhood: Effects of proficiency. Journal of Cognitive Neuroscience, 17, 1212-1228.

Perani, D., Paulesu, E., Sebastian-Galles, N., Dupoux, E., Dehaene, S., Bettinardi, V., et al. (1998). The bilingual brain: Proficiency and age of acquisition of the second language. Brain, 121, 1841-1852.

Takata, Y., & Nabelek, A.K. (1990). English consonant recognition in noise and in reverberation by Japanese and American listeners. Journal of Acoustical Society of America, 88, 663-666.

  
 

Language learning in bilinguals’ early development

One fact about humankind is that a large portion of human beings on this earth speaks more than one language. As mentioned before, our speech perception abilities undergo a binary modification towards the end of the first year of life in that our ability to differentiate between non-native contrasts decreases.  The question therefore arises what impact a bilingual context would have on such a typical phonetic development? Would the impact vary depending on whether an infant learns the second language at the same time as the first language or shortly after the first language?

Bilinguals who were exposed to their first (L1) and second (L2) languages from birth, are known as simultaneous bilinguals (Meisel, 1989). Simultaneous bilinguals differ from bilinguals who learned their second language when their lexical and phonological knowledge of their first language had already partially established. Those bilinguals are known as sequential bilinguals. While sequential bilingualism is considered to cause a transfer from the first onto the second language, simultaneous bilingualism is marked by both languages evolving relatively autonomously from one another (de Houwer, 2005; Meisel, 1989, 2001).

Since children will have lost their sensitivity to non-native speech sounds by the end of the first year of life, in sequential bilingualism, when children are exposed to the second language, they will need to regain their sensitivity to the non-native phonetic sounds including sounds that are distinct in the second language. For example, a child with Japanese as L1, who is now learning English as L2, would need to restore the perceptual difference between speech sounds /r/ and /l/ that in contrast to English language, belong to the same phonetic category in the Japanese language (Goto, 1971; Miyawaki et al., 1975). This way, they will be able to differentiate between /l/ and /r/, and understand that, for example, ‘lead’ and ‘read’ are words in English that differ in meaning.

What does the situation look like in simultaneous bilinguals who grew up in a bilingual language environment since birth? Previous research showed that infants are able to create and learn phonetic categories and contrasts because they are sensitive to how the phonetic values of the phonetic elements are statistically distributed in a language system (Maye, Werker and Gerken, 2002). Then, what is the impact of this sensitivity to how speech sounds are statistically distributed in the language input on infants who grew up in a bilingual context? If the role of continued exposure to contrasts in both languages is important, this would mean that by the age of 8 months, infants as simultaneous bilinguals would have created two phonetic categories. However, a distributional overlap between two contrasts in one language and one speech sound in the other language that represents an acoustically intermediate speech sound may give rise to a single extended phonetic category in the simultaneous bilingual infants that includes all three speech sounds (Bosch & Sebastian-Galles, 2003 a, b). The simultaneous bilingual infant would then have difficulties discriminating between these sounds.

This would imply that compared to monolingual infants’ perceptual abilities, simultaneous bilingual infants’ ability to create contrastive categories that are particular to a language is postponed by this cross-language distributional overlap of speech sounds. Researchers predicted that the extent to which speech sounds occur in both languages, i.e. the frequency occurrence of speech sounds, may counteract the impact that this delay has on bilingual infants’ discrimination ability (Sundara et al., 2008). They compared monolingual English infants’, monolingual French infants’ and bilingual French-English infants’ discrimination ability of English and French instances of /d/ (Sundara et al., 2008). They found that due to the high frequency of specific speech sounds tested, similar to monolingual English infants, bilingual 10-12 month olds were able to differentiate between different exemplars of French and English /d/ in spite of overlapping distributions of French and English /d/ (Sundara et al., 2008). Thus, apart from confirming previous evidence that statistical distributional learning assists infants in language learning (Saffran, 2003), the results specifically suggest that for a particular phonetic contrast, the question whether bilinguals follow a different developmental trajectory from matched monolingual controls depends on how frequent speech sounds from phonetic categories occur in actual speech, and on their cross-language distributional overlap.

References

Bosch, L., Sebastian-Galles, N. (2003a). Language experience and the perception of a voicing contrast in fricatives: Infant and adult data. In M.J. Sole, D. Recasens, & J. Romero (Eds.), Proceedings of the 15th International Congress of Phonetic Sciences (pp. 1987-1990). Barcelona: Causal Productions.

 

De Houwer, A. (2005).  Early bilingual acquisition: Focus on morphosyntax and the Separate Development Hypothesis. In J.F. Kroll  & A. M. B. de Groot (Eds.), Handbook of bilingualism: Psycholinguistic approaches (pp.30-48). New York: Oxford University Press.

Goto, H. (1971). Auditory perception by normal Japanese adults of the sounds L and R. Neuropsychologia, 9, 317-323.

Maye, J. Werker, J.F., & Gerken, L. (2002). Infant sensitivity to distributional information can affect phonetic discrimination. Cognitive Psychology, 82, B101-B111.

 

Meisel, J. (1989). Early differentiation of languages in bilingual children. In K. Hyltenstam & L. Obler (Eds.), Bilingualism across the lifespan. Aspects of acquisition, maturity and loss (pp. 13-40). Cambridge, UK: Cambridge University Press.

Meisel, J. M. (2001). The simultaneous acquisition of two first languages: Early differentiation and subsequent development of grammars. In J. Cenoz& F. Genese (Eds.), Trends in bilingual acquisition (pp.11-41). Amsterdam/Philadelphia: John Benjamins.

Miyawaki, K., Strange, W., Verbrugge, R., & Liberman, A. M. (1975). An effect of linguistic experience: The discrimination of [r] and [l] by native speakers of Japanese and English. Perception & Psychophysics, 18, 331-340.

Saffran, J.R. (2003). Statistical language learning: Mechanisms and constraints. Current Directions in Psychological Science, 12, 110-114.

Sundara, M., Polka, L., & Molnar, M. (2008). Development of coronal step perception: Bilingual infants keep pace with their monolingual peers. Cognition, 108, 232-242.

     
 

Modifications in speech to enhance intelligibility in the English language

What is the one tool that we use to solve problems and advance culture through the retention and transmission of knowledge? Exactly, it is language. What happens when we first learn a language? Early in life we appear to be able to differentiate between virtually all phonetic units in the world languages (Kuhl et al., 2008). And around the age of nine months infants adjust to the information of their native language. For example, cross-sectional and longitudinal studies showed that while 6-8 months old English infants distinguished between non-native consonant contrasts easily such as in the Hindi language, 10-12 months old infants had difficulty with this task (Werker & Tees, 1984a; Werker & Lalonde, 1988). Similarly, infants’ ability to distinguish between non-native vowel categories was reported to decrease as well (Werker & Tees, 1984a; Werker & Lalonde, 1988).

Thus, we can establish that our speech perception abilities undergo a binary modification towards the end of the first year of life (Kuhl et al., 2008). At around the same time in the English language, mothers talk to their infants in a way that has come to be known as motherese or infant-directed speech (IDS). Infant-directed speech involves acoustic changes such as increased mean pitch, increased emotional affect, and slow speech rate. Mothers also exaggerate the differences in the sound of vowels: Thus, if you have the words ‘sheep’, ‘shoe’ and ‘shark’, you will make the vowels ‘ee’, ‘oo’ and ‘aa’ sound more different from one another. This is known as increased vowel space (Kuhl, Andruski, Chistovich, Chistovich, & Kozhevnikova, 1997; Uther, Knoll, & Burnham, 2007).

All these acoustic changes are considered to facilitate language acquisition in infants. Prior research that compared IDS to clear read speech to test if those changes in IDS are uniquely didactic did indeed demonstrate increased vowel space in IDS compared to clear read speech (Kangatharan, 2016). Exaggerated vowel space has previously been correlated to speech intelligibility in IDS (Bradlow, Torretta, & Pisoni, 1996). It is exactly this acoustic aspect that was also observed in speech to adult foreigners (Uther et al., 2007).

The observation of exaggerated vowel space in IDS and in foreigner-directed speech (FDS) can be understood in the context of the Hyper-HypoSpeech theory by Lindblom (1990) according to which any changes in your speech signal are considerably influenced by who the listener is and what the speaking environment is. It is considered that speech production is a listener-oriented modification of speech. In this regard, the observation of exaggerated vowel space in IDS and FDS can be explained in the English language by the fact that both groups have the same linguistic needs (Uther et al., 2007). Thus, one can categorize these two groups as language learners as they are in clear need of an atypically clearer speech input, and speech with exaggerated vowel space is supposed to maximize speech intelligibility.

Recent research specifically found exaggerated vowel space to occur in speech to foreign-sounding rather than foreign-looking listeners (Kangatharan, 2016). This confirms in line with the H&H theory that speech is modified with regard to the listener’s linguistic needs rather than visual appearance. It was also demonstrated that speech with expanded vowel space resulted in increased speech intelligibility in both native English speakers and in second language learners of English in contrast to speech without expanded vowel space. The findings indicate that both first and second language learners of English can perceptually benefit from verbal adaptions in the English language if they have a didactic function (Kangatharan, 2016).

References

Bradlow, A. R., Torretta, G. M., & Pisoni, D. B. (1996). Intelligibility of normal speech. I. global and fine-rained acoustic-phonetic talker characteristics. Speech Communication, 20, 255–272.

 

Kangatharan, J. (2016). “The role of vowel hyperarticulation in clear speech to foreigners and infants: PhD Thesis publication.” European Acoustics Association Young Acoustics Network Newsletter, 47, Issue 2, 4-5.

Kuhl, P.K., Andruski, J.E., Chistovich, I.A., Chistovich, L.A., Kozhevnikova, E.V., Ryskina, V.L., Stolyarova, E.I., Sundberg, U., & Lacerda, F. (1997). Cross-language analysis of phonetic units in language addressed to infants. Science, 277, 684–686.

Kuhl, P. K., Conboy, B. T., Coffey-Corina, S., Padden, D., Rivera-Gaxiola, M., & Nelson, T. (2008). Phonetic learning as a pathway to language: New data and native language magnet theory expanded (NLM-e). Philosophical Transactions of the Royal Society B, 363, 979-1000.

Lindblom, B. (1990). ‘Explaining phonetic variation: A sketch of the H&H theory’, 403-439 in Hardcastle W J & Marchal A (eds): Speech Production and Speech Modeling, Dordrecht:Kluwer.

Uther, M., Knoll, M. A., & Burnham, D. (2007). Do you speak E-NG-L-I-SH? A comparison of foreigner- and infant-directed speech. Speech Communication, 49, 2-7.

Werker, J. F., & Tees, R. C. (. (1984a). Cross-language speech perception: Evidence for perceptual reorganization during the first year of life. Infant Behavior and Development, 7, 49-63. 

Werker, J. F., & Lalonde, C. E. (1988). Developmental Psychology, 24, 672-683.