Much recent second language (L2) research has investigated sentence processing and the extent to which this is similar to native language (L1) sentence processing. These studies have largely concentrated on the importance of syntactic, lexical, thematic and contextual information in L2 interpretation. Although prosody is also important in L1 processing, very few L2 studies have examined this. What research has been done in this area has shown that some learners demonstrate an interpretive response to prosody some of the time, but it is unclear what sets these cases apart from those in which learners showed no interpretive response to prosody. The study presented in this dissertation examined L1 English-L2 French learners and was designed to explore (1) whether all learners can respond to prosody for interpretation and (2) if so, what factors obscure the effects of prosody on L2 interpretation. Specifically considered were syntactically-based preferences and computational complexity due to length and structure.;Experiments involved four ambiguity types of differing complexity: (1) lexical, involving word boundaries, (2) adjective modification of one or both conjuncts in a coordinate NP, (3) PP attachment to the verb or to the adjacent noun, and (4) relative clause attachment to the first or second noun in a complex genitive NP. Experimental items were recorded with two distinct prosodies, each with a specific phrasing intended to favor one of the possible interpretations. Participants performed judgment tasks, which indicated whether they accepted the sentence they heard as congruous with a given interpretation, and identification tasks, which indicated their preferred interpretation of the sentence.;Results clearly showed that all learners responded to prosody for lexical disambiguation. Among the other three ambiguities, learners demonstrated the greatest response to prosody on the ambiguity involving adjectives, and the least response to prosody on the relative clause ambiguity, a pattern which is in keeping with a hypothesis that computational complexity can influence the extent to which available information is fully incorporated into the parse. In addition, the specific patterns of interpretive preferences provide strong evidence that syntax is integrated into the parse in early stages of L2 processing and that syntactically-based interpretive preferences are not always overcome by overt prosody.
展开▼
