The roles of familiarity, intelligibility and attitude in the processing of L1 accents

Abstract

This thesis presents an experimental investigation into the roles of familiarity, intelligibility and attitude in the lexical processing and short-term narrative recall of first language (L1) speech. Lev-Ari and Keysar (2012) suggest that the lexical and semantic representations of second language (L2) speech by L1 listeners are less detailed, potentially due to assumptions of lower linguistic proficiency in L2 speakers. These assumptions would not typically apply to L1 speakers. However, even for L1 speakers, listeners vary in terms of their familiarity with the speakers’ accent, how intelligible it is to them and their attitudes towards it, all of which can affect speech processing. Three online experiments were conducted to research the effects of familiarity, intelligibility and attitude. A lexical decision task was used to measure lexical processing. Recall was measured with a change detection task that included short stories and variations of the semantic proximity of the change. The experiments included manipulations of, first, the participants’ familiarity with the speakers’ accent; second, listening conditions (through added background noise) and; third, the participants’ attitudes towards speakers. Regarding lexical processing, the results showed a familiarity benefit and a window of opportunity for adaptation, which opened when lexical access was demanding but not too challenging. Negative attitudes were associated with poorer performance. For recall, a familiarity benefit only surfaced in noise. Recall performance further increased when the changed words were semantically related and when participants held negative attitudes towards speakers. These findings can be accounted for within a framework of semantic priming and item-specific versus relational processing. Recall is supported by semantic priming and item-specific processing. The latter seems to be induced by unfamiliar accents (in quiet), adverse listening conditions and negative attitudes. Taken together, this suggests that the less-detailed representations suggested for L2 accents do not generally apply to L1 accents.