Seidl, A. 1 , Danielson, K. 2 , Onishi, K. H. 3 , Alamian, G. 3 & Cristia, A. 4, 5
1 Purdue University
2 University of British Columbia
3 McGill University
4 Centre Nationale de la Recherche Scientifique
5 Laboratoire de Sciences Cognitives et Psycholinguistique
There is a substantial literature describing how infants become more sensitive to differences between native phonemes (sounds that are both present and meaningful in the input) and less sensitive to differences between non-native phonemes (sounds that are neither present nor meaningful in the input) over the course of development. A more difficult problem is what to do with allophonic dimensions, where variation in the input is not recruited for lexical encoding. For example, the sound "I" in "I'm" should map to the same underlying sound category as "I" in "I'd", despite the fact that the first vowel is nasal and the second oral. Adult listeners process such pairs of sounds differently than sounds that map onto different phonemes, so learners must come to treat them differently as well. Recent evidence suggests that infants' sensitivity to the contrast between a pair of sounds that are allophones of the same phoneme in the ambient language declines as early as 11 months, while that between allophones of different phonemes is maintained. We investigated the hypothesis that phonological status may be apparent in the acoustic cue distributions representing tenseness and nasality in the speech of English and French, monolingual and bilingual mothers. While tenseness is phonemic in American English, it is allophonic in Quebec French, and the opposite is true for nasality. Results corroborated our hypothesis, suggesting that, to some extent, infants' perception of allophonic pairs could degrade as a function of exposure to less clear acoustic cue distributions. We integrate these results with previous behavioral and computational research. Collectively, this work suggests that the calculation of phonetic similarity, the computation of complementary distributions, and possibly protolexical knowledge operate in concert to guide infants towards a functional interpretation of sound variants that are present in the input, yet not lexically contrastive.