Accessibility statement

Sequences, scenes and syntax: towards a neural model of working memory representations in language comprehension

Thursday 6 February 2020, 4.00PM to - 5.30pm

Speaker(s): Ellen Lau (University of Maryland)

Figuring out the neural implementation of short-term relational representations for language comprehension would be of great value not just for theories of cognitive neuroscience, but also for theories of linguistic and semantic knowledge. Alan Baddeley famously proposed the existence of multiple, independent working memory systems that support these kinds of representations. In this talk, I'll outline what I take to be three core working memory circuits for language comprehension, each with a data format and neural implementation informed by functional considerations: (1) a posterior temporal - frontal circuit that buffers phonological-articulatory information in sequence format, (2) an anterior temporal - inferior parietal circuit that buffers semantic information in 'scene' format, and (3) a posterior middle temporal region that supports both long-term memory for individual lexical-syntactic units and working memory for the syntactic relations between them in a given sentence. As (1) follows from a longstanding existing cognitive neuroscience literature on the phonological loop, my discussion will focus on the semantic and syntactic representations in (2) and (3). I will end by speculating about ways in which such neural implementation hypotheses could ultimately have implications for syntactic and semantic theory.

Location: B/M/023

Admission: Everyone