Perception of coarticulatory nasalization by speakers of English and Thai: Evidence for partial compensation

Abstract

The conditions under which listeners do and do not compensate for coarticulatory vowel nasalization were examined through a series of experiments of listeners' perception of naturally produced American English oral and nasal vowels spliced into three contexts: oral (C_C), nasal (N_N), and isolation. Two perceptual paradigms, a rating task in which listeners judged the relative nasality of stimulus pairs and a 4IAX discrimination task in which listeners judged vowel similarity, were used with two listener groups, native English speakers and native Thai speakers. Thai and English speakers were chosen because their languages differ in the temporal extent of anticipatory vowel nasalization. Listeners' responses were highly context dependent. For both perceptual paradigms and both language groups, listeners were less accurate at judging vowels in nasal than in non-nasal (oral or isolation) contexts; nasal vowels in nasal contexts were the most difficult to judge. Response patterns were generally consistent with the hypothesis that, given an appropriate and detectable nasal consonant context, listeners compensate for contextual vowel nasalization and attribute the acoustic effects of the nasal context to their coarticulatory source. However, the results also indicated that listeners do not hear nasal vowels in nasal contexts as oral; listeners retained some sensitivity to vowel nasalization in all contexts, indicating partial compensation for coarticulatory vowel nasalization. Moreover, there were small but systematic differences between the native Thai- and native English-speaking groups. These differences are as expected if perceptual compensation is partial and the extent of compensation is linked to patterns of coarticulatory nasalization in the listeners' native language.

How we hear what is hardly there: Mechanisms underlying compensation for /t/-reduction in speech comprehension

“…In Experiments 2 and 3, we have undertaken another attempt to make the context effect disappear. We used a discrimination task here, following the example set by several other researchers (Beddor & Krakow, 1999;Fitch, Halwes, Erickson, & Liberman, 1980;Kingston & Macmillan, 1995;Macmillan, Kingston, Thorburn, Dickey, & Bartels, 1999), arguing that the discrimination task reveals processing at an earlier level than the identification task. If the context effect occurs at early auditory levels, it should be observed in a discrimination task as well.…”

Phonetic Diversity, Statistical Learning, and Acquisition of Phonology

“…However, the extent of these effects depends on the system of lexical contrasts. Beddor and Krakow (1999) show that English and Thai listeners differ in perceptual compensation for coarticulatory nasalization. In production, Thai has less anticipatory coarticulation in VN sequences than English does.…”

On the causes of compensation for coarticulation: Evidence for phonological mediation

“…There is some evidence that learning is involved in compensation for assimilation. Beddor and colleagues (Beddor, Harnsberger, & Lindemann, 2002;Beddor & Krakow, 1999;Darcy, Peperkamp, & Dupoux, in press) showed that languages differ in their fine-grained detail of coarticulatory pattern and that listeners' compensatory patterns are adjusted to that. If listeners from different language backgrounds are presented with identical items, they show more compensation for native-like patterns of coarticulation.…”

“…How are such coarticulated phonemes then recognized? A number of studies have shown that coarticulation is compensated for in perception (e.g., Beddor & Krakow, 1999;Liberman, Delattre, & Cooper, 1952;Mann, 1980;Mann & Repp, 1981). Mann and Repp (1980), for instance, showed that listeners will still accept a fricative with some cues for lip rounding as / / if it is followed by a rounded vowel, whereas the same fricative is interpreted as / / if followed by an unrounded vowel.…”