Laughter as Same-Turn Self-Repair Initiation in L2 Oral Proficiency Interview

Abstract

Repair is the effort that speaker makes to smooth utterances or conversations in order to achieve the ongoing progressivity of the turn. Among the types of repair, self-repair is mostly preferred (Schegloff et al., 1977) in talk-in-interaction, and in self-repair, self-initiated self-repair is mostly preferred, especially when it happens in the same turn. Therefore it has received much attention. In the gate-keeping interviews for Chinese candidates to get access to the target college in Britain, there occurs in L2 a large number of same-turn self-initiated self-repair cases, in which the initiation is no longer the single lexical words, or the non-lexical pauses like cutoff , or sound stretches as has been studied, but the paralinguistic device, laughter. Despite the numerous studies on roles of laughter, there is hardly a mention of laughter employed as a self-repair initiation. Consequently, this study determines on the focus on laughter as a self-repair initiator in L2, and aims to identify its exact shape and positions in self-repair and relevant repairing segment, therefore concluding on its specific functions and its relations with delicacy. This study wishes to contribute to the researches on self-repair initiation in Conversation Analysis and on laughter in particular L2 testing context.

Proceedings of the 17th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2022)

“…Language learners use laughter as a 'trouble management device' during uncertainty (Looney and He, 2021), when pre-empting a problematic action (Petitjean and González-Martínez, 2015) and after making an error (Gao and Wu, 2018). In an analysis of UK university English proficiency interviews of 23 Chinese students, Gao (2020) found that laughter co-occurs with disfluencies in three ways: (i), on its own between the reparandum and repair phrase, (ii), alongside indicators of an interruption point such as pauses and word cutoffs, and (iii), simultaneously as laughed speech either during the repair phrase or the whole disfluency. Laughter has been shown to improve performance of models for other dialogue processing tasks such as dialogue act classification (Maraev et al, 2021) but as of yet, has not been applied as a feature to detect disfluencies in learner speech.…”