
For the final LEL seminar this semester, we will have our very own Colin Bannard speaking on “Accounting for multiword frequency effects in a compositional model of language production”. The talk will be on Tuesday 17th May at 4pm in Samuel Alexander A201 and on Zoom. All are welcome!
Accounting for multiword frequency effects
in a compositional model of language production
Colin Bannard
University of Manchester
Multiword frequency effects on language production, where the frequency of a sequence of words has an impact on the timing or fluency of speech, have been reported for a diverse range of experimental tasks. It has often been stated (including by me) that these effects present challenges to existing models of speech production, but there has to-date been no full articulation of what those challenges are, and no fully-worked-out alternative models proposed. In this talk I will start from Levelt’s classic “blueprint for a speaker”, explore the implications of multiword frequency effects for this and related models, and present some initial thoughts as to how multiword frequency effects could be accounted for in a full and similarly compositional account of language production. I will finish by describing a couple of recent observational studies in which the relative frequency of different multiword sequences results in errors of commission in children’s speech and considering the more challenging task of accounting for these kinds of effects.
The featured photo has been cropped. The version here includes Frankie the lab.
Posted on May 16, 2022 by manling