Our perceptions of the semantic relations that hold between objects, events and situations have far-reaching consequences for our ability to learn and reason about our environment. In particular, recognizing that two events or objects are similar to each other, or that one is a cause or a consequence of the other, is central for understanding how the world works. Our ability to distinguish different semantic relations is also fundamental for language comprehension, and a growing body of research suggests that our ability to infer what relations hold between different events is necessary for successful pronoun interpretation. Pronouns (e.g. they, she, it) play a central role in the information management process that underlies successful communication, but, due to their semantically impoverished nature, appear to pose a challenge for the language comprehension system. Recent research suggests that a successful account of pronoun interpretation needs to take into account the semantic coherence relations that hold between sentences (e.g. X is a result of Y, X is similar to Y). However, the linguistic and cognitive properties of these coherence representations are not yet well- understood: It is not known whether these representations are specific to language (and possibly specific to the pronoun interpretation process), or whether they are more domain-general and shared between language and non-linguistic cognitive representations. In addition, the question of how coherence effects interact with other factors that have been found to influence pronoun interpretation is still open. Our long-term goal is learn more about how different linguistic and visual cues contribute to the semantic relations that people perceive or infer to hold between events, how these relations are represented and how they influence language processing. The proposed research is the first step towards this goal, and focuses on investigating the nature of the coherence- related representations that underlie pronoun interpretation. To do this, we develop a novel experimental paradigm that combines visual-world eye-tracking with priming. The proposed research has 3 specific aims: (1) to test for direct evidence for coherence-related representations being activated during pronoun interpretation, (2) to determine how domain-general these coherence representations are, and (3) to investigate whether and to what extent pronoun interpretation is guided by linguistic representations, in particular anaphoric dependency representations, that cannot be derived from coherence. The significance of this research extends beyond pronoun interpretation. Although these experiments investigate the representations underlying pronoun interpretation, they constitute the first steps in a broader research continuum that aims to further our understanding of how people recognize and represent semantic relations between events. The new eye-tracking+priming paradigm potentially offers a unified way of investigating causality and similarity in two different domains, language and vision, allowing us to investigate the perception of causality and similarity in a new light by benefiting from cross-fertilization between linguistic and visual research on coherence relations.
Research on the mental representations involved in the interpretation of pronouns (she, it, they) - in particular the extent of their linguistic and cognitive generality - can potentially contribute to our understanding of the discourse-level challenges faced by second-language learners, children, and linguistically- and/or memory- impaired individuals (e.g., individuals with aphasia or Alzheimer's disease). In addition, the proposed research has the potential to bring together insights from linguistic and visuo-spatial research, thereby furthering our understanding of how humans recognize and represent semantic relations between events - an aspect of cognition that is crucial for our ability to learn and reason about our environment.