Projects in the Lab
Dialects: When bilinguals speak, both fluent language systems become activated in parallel and exert an influence on speech production. As a consequence of maintaining separation between the two linguistic systems, bilinguals are purported to develop enhanced executive control functioning. Like bilinguals, individuals who speak two dialects must also maintain separation between two linguistic systems, albeit to a lesser degree. This line of research is dedicated to investigating parallels and differences between bilingual and bidialectal cognitive and linguistic processing.
Swinging Lexical Network: The investigation of semantic context effects has served as a valuable tool in investigating mechanisms of language production. Classic semantic interference effects from the picture-word interference, semantic blocking, and speech error elicitation paradigms have provided influential support for a competitive lexical selection mechanism. However, recent interest in semantic facilitation effects from non-categorical semantic relations has stimulated a discussion on whether or not context effects reflect competition during lexical selection. Together with my colleague Rasha Abdel Rahman, we have proposed a framework of lexical selection by competition that is sensitive to the activation of lexical cohorts. Support for our swinging lexical network hypothesis is drawn from a variety of 'unexpected' interference effects, for example from semantic associates. We suggest that by adopting the assumptions of our proposal we can parsimoniously account for a majority of the discussed semantic facilitation and interference.
Sentence Production: When formulating sentences, speakers often reuse the constituent structure of recently encountered sentences, the so-called structural persistence effect. Results from this paradigm have been taken as evidence against strong lexical involvement in sentence production. Rather, some recent models of sentence production view the formulation process as an interaction between event structure and general syntactic constraints, rather than more subtle verb-specific constraints. My research aims at providing evidence for lexically-driven syntactic processes in production as well as uncovering the scope of the persistence phenomenon. Some recent findings show units smaller than the phrase can be primed, suggesting that priming does not strictly involve traditional subcategorization frames.
Gesture: When people speak, they often produce gestures that are temporally and semantically coordinated with the concurrent speech. However, rates of gesture production vary depending on the situation and the individual. Several factors have been shown to influence within-speaker variation including the degree to which the elicitation task taxes the working memory system. This suggests that the processing capacity of individuals might also influence gesture rates, underlying between-speaker differences. In my research I've looked both at the sources of between- and within-speaker gesture rate variation.
A quantitative analysis of semantic associations: Semantic associates, in contrast to category coordinates, have not captured the attention of production researchers. This has led some scholars to view semantic association effects, both in comprehension and production, an indexing lexical co-occurrence rather than semantic processing. To evaluate this assumption, I use corpora and statistical grammars developed by computational linguists to evaluate the co-occurrence claim. Together with Sabine Schulte im Walde (Stuttgart University), we have evaluated a) the types of semantic relations reflected by semantic associations, as supported by lexical ontologies, and b) the correlation between association strength and target-associate co-occurrence frequencies in corpora.