UC Santa Barbara

Children acquire their L1 entirely through interactions with other speakers; in the same way, L2 learners benefit from participating in conversations with native speakers. I take a discourse-functional view of native speaker competence, assuming that positive evidence (language in use), comprising both frequency and contexts of usage, plays an important role in native speakers' mental representations and acquisition of grammar. If we assume that any comfortably proficient L2 speaker cannot have acquired that language ability solely from textbooks or classroom instruction, then the question arises: in what ways does conversation help language learners acquire a discourse-based grammar mirroring that of native speakers? I address this question across three case studies, using data from twelve conversations, each between one native speaker and one non-native speaker of Japanese.

The first case study (Chapter 3) investigates the types of explicit and implicit interactional feedback (comprising negative as well as positive evidence) that native speakers provide in conversation with non-native speakers. I conduct a qualitative analysis of points in the conversation related to the non-native speaker's language ability, including instances of explicit metalinguistic discussion and of implicit feedback, such as recasts of non-native speaker utterances, and I suggest the ways in which these function as potential L2 learning mechanisms.

In the other two case studies, I examine the question: To what extent do the conversational grammars of non-native speakers exhibit the same relationships between grammatical form and discourse function as the conversational grammars of native speakers? In Chapter 4, I conduct a qualitative and quantitative analysis of Noun-Modifying Constructions (NMCs) and compare the usage of NMCs among native vs. non-native speakers, based on discourse factors, finding that the non-native speakers, especially those with more experience living in Japan, produced NMCs with frequencies and distributions similar to those of the native speakers.

In Chapter 5, a quantitative analysis of subject argument realization, I use a mixed-effects model to show that both native and non-native speakers' patterns of subject realization are influenced by discourse-pragmatic factors such as givenness and contrast, and that the patterns observed among non-native speakers mirror those in native speakers' speech. I propose that non-native speakers can only demonstrate such similar sensitivities to discourse-pragmatic factors if exposed to native-speaker-like frequencies of use in conversation. The findings of this case study allow for a glimpse into the impact of positive evidence from natural input on L2 learners' acquisition of discourse-based grammar.