ACLC Seminar | Hielke Vriesendorp: Exemplars vs. allophones: the role of lexical frequency and listener experience in sociolinguistic processing

Organized by the Amsterdam Center for Language and Communication! (ACLC)

• Speaker: Hielke Vriesendorp
• Location: P.C. Hoofthuis
• Room: 1.15
• Title: Exemplars vs. allophones: the role of lexical frequency and listener experience in sociolinguistic processing
• Abstract: Do we recognise a Northern English speaker saying ‘way’ as Northern because we have memories of other times we’ve heard Northern speakers say ‘way’ before, or because we recognise the abstract concept of a monophthongal FACE vowel (/eɪ/ realised as [e:]) in their pronunciation of ‘way’? Work on exemplar theory suggests that the way listeners are able to recognise the social meaning of linguistic variation is because social information is a part of the detailed (episodic) memories of words – exemplars – we use to process speech. However, as new innovations within the theory open up the possibility that listeners may generalise across these exemplar memories to create differently sized or differently abstract representations (hybrid exemplar theory). This raises the question of whether it is detailed exemplars that carry social meaning or more abstract linguistic representations like allophones. I discuss two accent recognition tasks which provide evidence that both are used, but partly depending on listener experience: the findings suggest that all listeners are able to rely on sublexical representations for accent recognition, as accent recognition was far above chance in non-word stimuli. However, it was found that inexperienced listeners of an accent benefited significantly less from hearing real word stimuli rather than non-word stimuli compared to experienced listeners, this suggests that lexical representations are more dominant in sociolinguistic processing for experienced listeners, whilst sub-lexical representations are more dominant in inexperienced listeners who may not have access to similarly strong lexical representations