Asymmetry of onsets and codas in language acquisition: Implications for phonological theories

Despite the range of syllable structures in the languages of the world, typological investigations have demonstrated that CV is the only syllable type that is always licit across the world languages. To what extent can the cross-linguistic syllable typology be accounted for by the learning biases in children’s phonological representation? How may these biases be modulated by language experience and come to shape learners’ phonological systems? In this dissertation, evidence concerning these questions is garnered from studies of infant and adult speech perception in laboratory settings, as well as from a corpus investigation of children’s production. Chapter 2 examines 8-, 12-, and 15-month-olds’ learning of an experimental phonotactic pattern in which a category of speech sounds is restricted to either word-medial onset or coda position. Eight-month-olds fail to learn either positional restriction, 12-month-olds learn the pattern only in the Onset condition, while 15-month-olds are able to learn the pattern in the Coda condition. Chapter 3 explores 24-month-olds’ sensitivity to word-medial onsets and codas in a more challenging word-learning task. Here, toddlers show better recognition of the target object when the label is mispronounced in coda than when it is mispronounced in onset position. These results suggest that children are biased towards full phonological representation of onset over coda consonants. Chapter 4 seeks out to explore the role of the input on children’s production of fricatives. Results demonstrated that the role of the input is stronger in coda position than in onset position, again suggesting that onset is learned earlier than coda. In order to explore the developmental pattern as well as the influence of language experience on phonotactic learning related to syllable position, Mandarin and English adult speakers are tested on a subjective judgment task in Chapter 5, and both groups show equal sensitivity to onsets and codas. In sum, these studies suggest that 1) infants bring with them learning biases in attending to subsyllabic constituents; and 2) these biases may be modulated by language experience. In order to account for these developmental patterns, based on PRIMIR (Curtin et al., 2011; Werker & Curtin, 2005), I propose a modified and broader model in Chapter 6 which takes into account these subsyllabic attentional patterns. In addition, the underlying mechanisms of the learning biases as well as their implications are discussed