Complement Coercion as the Processing of Aspectual Verbs: Evidence from Self-Paced Reading and fMRI

Verbs like begin and enjoy have been assumed to select for events and would coerce an entity-denoting complement to an event in semantic representation; such type-shifting operation engenders additional processing cost. However, recent studies show that this effect is observable only for aspectual verbs (e.g. begin), but not psychological verbs (e.g. enjoy), suggesting the set of “coercion verbs” is heterogeneous. We hypothesize that aspectual verbs select for structured individuals instead of events. They lexically encode a set of functions, each applying on a dimension (e.g. eventive, informational, spatial) denoted by the complement. The parser must determine the verb’s specific function and the dimension from the complement. The processing cost results from (1) exhaustive activation of the verb’s lexical functions and (2) resolution of ambiguity created by dimension extraction from the complement. We show that processing aspectual verbs was more costly than psychological verbs and that it recruited Wernicke’s area for cost 1 and LIFG for cost 2.