Wordlikeness Judgements, Phonotactic Probability, Accent and Repetition in Auditory Recognition of Novel Words in a Foreign Language

Abstract

This research considers the hypothesis that the acquisition of vocabulary in a second language is (inter alia) dependent on the acquisition of a knowledge of the phonotactic and phonetic structure of the second language. The observation that children acquire considerable knowledge of the phonotactic and phonetic structure of their native language before they begin to speak, coupled with the finding that phonological memory in both children and adults is correlated with native language “wordlikeness”, suggests that implicit knowledge of the phonotactic and phonetic structure of the native language is implicated in vocabulary development - in particular the ability to rapidly acquire ("fast-mapping") the form of novel, native (but typically not foreign) words. This thesis considers the argument that the lack of such experientially-derived, implicit phonotactic knowledge can explain many of the difficulties experienced by second language learners in the acquisition of vocabulary in the early stages and that judgements of wordlikeness represent an index of the learning burden presented by novel foreign vocabulary. The thesis critically reviews the hypothesis that the "Phonological loop”, a short-term memory component of the working memory model developed by Baddeley & Hitch (1974), is implicated in the acquisition of both native and second language vocabulary. It is noted that the evolution of the research into the relationship between short-term memory and vocabulary acquisition has been dominated by the emergence of evidence that suggests that long-term rather than short-term processes may play a more critical role. The proposition is considered that short-term memory and associated rehearsal processes may not fully account for the sort of phonological learning associated with language acquisition and that learning the phonological form of new words is dependent on the gradual acquisition of probabilistic phonotactic and phonetic knowledge. Such knowledge is assumed to be acquired automatically and pre-attentively from simple exposure to the spoken language. A simple model of phonological acquisition is developed which is described in terms of the principles of Hebbian learning in the context of speech perception. "Acquiring the phonetic encoding system of a language involves acquiring probability distributions over the phonetic space”, (Pierrehumbert, 2003). Acquiring the word encoding system of a language further involves the acquisition of its probabilistic phonotactics. It is proposed that the ability to "fast-map" novel words of a second language develops with the acquisition of this probabilistic knowledge - knowledge which can only be acquired as a consequence of significant auditory exposure to the spoken language. The thesis describes the results of an experiment which tests the major assumptions of the model by manipulating subjective wordlikeness judgements, phonotactic probability, accent and repetition in a spoken word recognition design using real words of a foreign language (Russian). The results of the research presented in this thesis suggest that second language vocabulary learning in a pedagogical setting may be enhanced by extensive auditory exposure to the target language in the early stages of acquisition. The thesis concludes that the value of "incomprehensible input" at the earliest stages of second language acquisition merits further research.