English prosodic marking of Information Structure by L1-Japanese second language learners

Abstract

This dissertation examines second language (L2) use of prosody to convey and comprehend Information Structure. In English, contrastive information and new information are typically higher in pitch and longer in duration, whereas given information is often deaccented. Contrastive information and new information are often differentiated by distinct types of pitch accents: the former marked with L+H*, the latter with H*. Tokyo Japanese, by contrast, uses pitch accent primarily to express lexical meaning rather than information status, but contrastive information can be indicated by expanded pitch range, similar to English L+H*, without a change in either duration or lexically-determined pitch accent type. In light of such differences between English and Japanese, this study investigates how L2 learners develop target-language prosodic marking of Information Structure, attending in particular to whether this is initially modulated by native language (L1) properties. L1-Japanese L2 learners of English were tested on three prosodic patterns–deaccentuation, a regular high pitch accent (H*), and a contrastive pitch accent (L+H*)–and their link to Information Structure via three tasks: a prosody-in-context naturalness rating task, an eye-tracking listening comprehension task, and a production task. It was found that native English speakers associate given information with deaccentuation, contrastive information with L+H*, and new information with accentuation (both H* and L+H*) in all three tasks. Advanced L2 learners could map given information with deaccentuation and contrastive information with L+H* in the rating and production tasks, but only the mapping between given information and deaccentuation was demonstrated in the eye-tracking listening task. Less proficient L2 learners were able to associate given information with deaccentuation and contrastive information with L+H* in the rating task, but no evidence for prosody-discourse association emerged in the production or eye-tracking listening tasks. The outcomes of the study suggest that L2 learners do not necessarily transfer their L1 prosody-discourse mappings (for all tasks), and that they can master target-language prosodic marking of Information Structure that is not instantiated in the L1. Furthermore, how successfully they map prosody to information status depends on several factors: learners’ language proficiency, processing difficulty of particular mappings, task difficulty, and whether learners’ attention is explicitly drawn to prosody.