Current Lab Research Overview

• Sentence processing and Ambiguity resolution
• Semantic memory
• Morphology

Sentence processing and ambiguity resolution
Back

Current work in the lab focuses on the role of event knowledge in comprehension. As one example, many verbs have multiple senses, reflecting variations in the types of event the verb refers to, and the different senses tend to occur in different syntactic structures. In earlier work my colleagues and I showed that verb sense influenced the interpretation of temporary structural ambiguities. Currently we are examining whether such meaning-structure correlations influence comprehension of verbs like burn, which have both a change-of-state and a causative sense (the popcorn burned; the popcorn burned his fingers). Our results to this point suggest that comprehenders rely on information beyond anything encoded in a verb's lexical representation. For example, it has recently been proposed that reduced relative clauses (as in The horse raced past the barn fell) are difficult to understand because the lexical semantics of the verb clashes with the requirements of the reduced relative construction (McKoon & Ratcliff, Psychological Review, 2003, 2005). In a set of experiments and corpus analyses, my colleagues and I found that the lexical semantic account makes the wrong predictions about ease of comprehension. Moreover, using a regression model, we found that the lexical semantic account did not predict the occurrence of reduced relatives for verbs in large corpora. Instead, ease of comprehension was determined by factors that influence ambiguity resolution, as constraint-based accounts predict.

Semantic memory
Back

A related line of work investigates the types of semantic relationships encoded in semantic memory. In a series of priming studies, we have found that verbs prime their typical thematic role fillers, and, conversely, salient role fillers prime the verb. Again, however, the results are not strictly verb-related: nouns referring to events also prime typical event participants, and salient aspects of common events, like locations or instruments, prime other aspects of the event as well. Overall, these results studies suggest that event-based knowledge is a key component of semantic memory, and it is rapidly activated during normal comprehension.

In a separate line of research, I have been investigating the rule-like nature of many aspects of grammar. Much of the recent work on this topic has focused on English past tense morphology, because that has a regular, rule-like pattern (e.g., bake-baked) and a set of irregular forms (e.g., take-took). Priming studies have tended to show differences in the two types of verbs; the question of interest is whether or not these behavioral differences are due to a categorical difference in how the verbs are represented and processed. Recently my colleagues and I compared priming for regular and prototypically irregular verbs with priming for verbs like sleep-slept, which are irregular but take a version of the regular suffix. The results of these studies suggest a continuity in how regulars and irregulars are processed, rather than a categorical divide.