van Gompel, Roger

My research is aimed at understanding how people comprehend and produce language. An important part of my research involves the investigation of how people process the syntactic structure of sentences (e.g., Van Gompel, 2013). Perhaps surprisingly, our research (Van Gompel, Pickering, & Traxler, 2001; Van Gompel, Pickering, Pearson, & Liversedge, 2005) suggests that sentences that are syntactically ambiguous are often easier to read than sentences that are disambiguated. This provides evidence against the commonly held view that different syntactic analyses of an ambiguous structure compete. Instead, the results suggest that difficulty is due to structural reanalysis.

In other research, we have investigated how the processing of sentences is affected by recent exposure to sentences with a very similar structure, a phenomenon often referred to as structural priming (Arai, Van Gompel, & Scheepers, 2007; Carminati, Van Gompel, Scheepers, & Arai, 2008). It appears that structural priming in language comprehension occurs with the same structures as in language production, but interestingly, priming effects in comprehension are much more lexically driven. In research with Leila Kantola (Kantola & Van Gompel, 2011), we have also used structural priming to investigate the syntactic representations of bilinguals: Our results show that structures that exist in two different languages have one single mental representation.

Another important part of my research is concerned with how people comprehend anaphoric expressions (e.g., Järvikivi, Van Gompel, Hyönä, & Bertram, 2005; Van Gompel & Majid, 2004) and produce them. In collaboration with Kumiko Fukumura, we have recently investigated when people produce either a pronoun or a repeated name or noun phrase to refer back to an earlier introduced discourse entity. Our results show that the choice for a particular anaphoric expression is driven by, amongst others, the similarity of a discourse entity to other entities (Fukumura & Van Gompel, in press; Fukumura, Van Gompel, Harley, & Pickering, in press) and syntactic factors, but interestingly, not by the likelihood that people refer to an entity (Fukumura & Van Gompel, 2010). I am currently also collaborating with computational linguists to test the psychological plausibility of computational models of reference, and we have edited a special issue of the journal TopiCS in Cognitive Science on this issue (Van Deemter, Gatt, Van Gompel, & Krahmer, 2012).

I currently teach Psychology of Language (Levels 3, 4 and 5) and Statistics and Methodology (Levels 2 and 5).