Implicit Causality And Consequentiality In Native And Non-Native Coreference Processing

This dissertation is composed of two studies that examined the role of implicit causality and consequentiality in coreference processing. Implicit causality (IC) refers to the phenomenon that certain interpersonal verbs bias the causation of the events described by the verbs towards either its subject (the first noun phrase NP1) or its object (the second noun phrase NP2). Implicit consequentiality (IR) refers to the phenomenon that certain verbs bias the consequence towards either NP1 or NP2. These IC and IR biases have been found to influence language comprehenders’ establishment of coreference. The first study examined whether intentionality of an event affects native English speakers’ re-mention biases of IC and IR. In two sentence-completion experiments, the strength of event intentionality was manipulated via intentionality-strengthening adverbs such as deliberately and intentionality-weakening adverbs such as accidentally. Results show that reinforcing intentionality changed IC and IR biases with participants showing increased references to NP1 in the IC context and NP2 in the IR context. The present study thus adds to a growing body of literature showing that IC and IR re-mention biases are not just determined by verb semantics but rather reflect a discourse phenomenon resulting from comprehenders’ causal inferences about the explanation for or the consequence of an event. The second study investigated advanced Chinese-speaking English learners’ use of IC and IR biases in establishing coreference. In two sentence-completion experiments that focused on IC and IR, respectively, participants wrote continuations to sentence fragments containing either NP1-biasing verbs or NP2-biasing verbs and ending with either a free prompt (e.g., NP1 verb-ed NP2 because…) or a pronoun prompt (e.g., NP1 verb-ed NP2 because he…). In both the IC and IR contexts, non-native speakers showed native-like re-mention biases in the free prompt condition. Moreover, like native speakers, non-native speakers produced more NP1 references in the pronoun prompt condition than in the free prompt condition. However, unlike native speakers, non-native speakers exhibited a “subject bias” in pronoun resolution by producing more NP1 references after NP2-biasing verbs. Overall, the study reveals that non-native speakers are able to generate predictions about the next-mentioned referent based on discourse-level information. The “subject bias” shown by non-native speakers in their resolution of pronouns indicates that when processing multiple sources of information, non-native speakers tend to resort to the cues that are easy to process, such as the subjecthood cue associated with the presence of pronouns