Time-course of adaptation to regional and foreign accents Caroline Floccia, EA3188, Université de Franche-Comté, France; CNRS UMR6057, Aix-en-Provence, France & School of Psychology, University of Plymouth Jeremy Goslin, School of Psychology, University of Plymouth & Frédérique Girard, EA3188, Université de Franche-Comté, France [email protected] The study of accent perception and recognition offers not only a privileged insight of the phonological and prosodic representations used in speech perception, but also the methods used by our language processing device to cope with variability. Investigating the stages of adaptation to accents, Girard, Floccia and Goslin (2004) showed that the presentation of an unfamiliar regional accent initially induces a 30 ms cost in word identification, and a 150 ms cost in the case of a foreign accent. Clarke and Garrett (2004) showed that the costs associated with the processing of foreign accents only returned to baseline levels after the presentation of four to five two-second sentences, signifying that full adaptation had been achieved. In the present study we investigate whether this time-course also holds for regional accents, and also whether this adaptation process can transfer across speakers of the same accent. Thirty three participants originating from the south-west of the UK took part in a lexical decision task for targets placed at the end of carrier sentences, where they had to decide as quickly as possible if the last item was a word (like “puppet”) or a pseudo-word (like “gamlet”). All participants were first presented with a baseline block (B1) of sentences uttered by speaker with a ‘local’ (Plymouthian) accent. This block was then followed by two blocks of new sentences whose accent varied according to the experimental condition. In the foreign accent condition the second block (B2) was produced by a French speaker, followed by the final block (B3) produced by another French speaker. In the regional accent condition B2 was produced by an Irish speaker, with B3 by another Irish speaker. Finally, in the control condition B2 and B3 were produced by two new ‘local’ speakers (also Plymouthian). The order of presentation for the B2 and B3 blocks was counterbalanced across participants. We found that the very first presentations of the foreign and the regional accents after the baseline block produced increased reaction times of 161 ms and 95 ms respectively when compared to the control condition. After the presentation of three sentences, less than twelve seconds of signal, there was no longer any significant difference in reaction times between the conditions, indicating that adaptation had taken place. This result replicates the foreign accent findings of Clarke and Garrett (2004) and also extends our findings to regional accent adaptation. We also found that the moment of speaker change within a given accent did not elicit any significant delay, showing that the accent adaptation can transfer and generalise across speakers whatever the quality of the accent or its perceptual distance from the native speech style. These results suggest that the speech perception system is sensitive to variations due to unfamiliar accents, and that the rapid time-course of adaptation reveals an efficient learning mechanism used for different speech styles. Additionally, our findings indicate that the phonological and prosodic representations stored when processing an unfamiliar accent are sufficiently abstract to resist to within-accent speaker changes. These results are discussed in light of Norris, McQueen & Cutler’s (2003) recent accounts for perceptual learning.