Guest Speaker: Prof. Brian Dillon

Four observations about dependency formation, and a proposal

Brian Dillon (UMass Amherst)

How do we leverage working memory resources to form syntactic dependencies in comprehension? Despite an enormous amount of work on this topic, a firm grasp on the cognitive mechanisms that allow this feat remains elusive. In this talk, I will give evidence from recent work that supports four observations about the cognitive mechanisms that subserve dependency formation: One, the context of encoding linguistic items into memory matters; Two, relational hierarchical constraints matter; Three, interference is not always cue-driven; and Four, interference often results in illusory item-feature conjunctions. I think that none of these four observations are well captured by current models of how syntactic information is encoded and manipulated in working memory during real-time dependency formation. Time permitting, I will describe a new proposal that aims to capture these observations, and discuss successes and challenges for this approach.