Event and Entity Coreference Across Five Languages: Effects of Context and Referring Expression

Current work on coreference focuses primarily on entities, often leaving unanalysed the use of anaphors to corefer with antecedents such as events and textual segments. Moreover, the anaphoric forms that speakers use for entity and non-entity coreference are not mutually exclusive. This ambiguity has been the subject of recent work in English, with evidence of a split between comprehenders' preferential interpretation of personal versus demonstrative pronouns. In addition, comprehenders are shown to be sensitive to antecedent complexity and aspectual status, two verb-driven cues that signal how an event is being portrayed. Here we extend this work via a comparison across five languages (English, French, German, Italian, and Spanish). With a story-continuation experiment, we test how different referring expressions corefer with entity and event antecedents and whether verbal features such as argument structure and aspect influence this choice. Our results show widely consistent, not categorical biases across languages: entity coreference is favoured for personal pronouns and event coreference for demonstratives. Antecedent complexity increases the rate at which anaphors are taken to corefer with an event antecedent, but portraying an event as completed does not reach statistical significance (though showing quite uniform patterns). Lastly, we report a comparison of the same referring expressions to refer to entity and event antecedents in a trilingual parallel corpus annotated with coreference.Together, the results provide a first crosslingual picture of coreference preferences beyond the restricted entity-only patterns targeted by most existing work on coreference. The five languages are all shown to allow gradable use of pronouns for entity and event coreference, with biases that align with existing generalizations about the link between prominence and the use of reduced referring expressions. The studies also show the feasibility of manipulating targeted verb-driven cues across multiple languages to support crosslingual comparisons.

Taking the perspective of narrative characters

An ongoing debate in the interpretation of referring expressions concerns the degree to which listeners make use of perspective information during referential processing. We aim to contribute to this debate by considering perspective shifting in narrative discourse. In a web-based mouse-tracking experiment in Dutch, we investigated whether listeners automatically shift to a narrative character’s perspective when resolving ambiguous referring expressions, and whether different linguistic perspective-shifting devices affect how and when listeners switch to another perspective. We compared perspective-neutral, direct, and free indirect discourse, manipulating which objects are visible to the character. Our results do not show a clear effect of the perspective shifting devices on participants’ eventual choice of referent, but our online mouse-tracking data reveal processing differences that suggest that listeners are indeed sensitive to the conventional markers of perspective shift associated with direct and (to a lesser degree) free indirect discourse.

Initial person reference in Providence Island Sign Language

When referring to non-present entities, speakers and signers can select from a range of different strategies to create expressions that range from extremely concise to highly elaborate. This design of referring expressions is based partly on the availability of contextual information that can aid addressee understanding. In the small signing community of Providence Island, signers’ heavy reliance on extra-linguistic information has led to their language being labelled as context-dependent (Washabaugh, de Santis & Woodward 1978). This study investigates the semiotic strategies that deaf signers in Providence Island use to introduce non-present third person referents, and examines how signers optimise specificity and minimise ambiguity by drawing on shared context. We examined first introductions to non-present people in spontaneous dyadic conversations between deaf signers and analysed the semiotic strategies used. We found that signers built referring expressions using the same strategies found in other sign languages, yet designed expressions that made use of contextual knowledge shared through community membership, such as geography, local spoken languages and traits of fellow islanders. Our signers also used strategies described as unusual or unattested in other sign languages, such as unframed constructed action sequences and stand-alone mouthings. This study deepens our understanding of context dependence by providing examples of how context is drawn upon by communities with high degrees of shared knowledge. Our results call into question the classification of sign languages as context-dependent or context-independent and highlights the differences in data collection across communities and the resulting limitations of cross-linguistic comparisons.

Pragmatic Prediction in the Processing of Referring Expressions Containing Scalar Quantifiers

Previous research in cognitive science and psycholinguistics has shown that language users are able to predict upcoming linguistic input probabilistically, pre-activating material on the basis of cues emerging from different levels of linguistic abstraction, from phonology to semantics. Current evidence suggests that linguistic prediction also operates at the level of pragmatics, where processing is strongly constrained by context. To test a specific theory of contextually-constrained processing, termed pragmatic surprisal theory here, we used a self-paced reading task where participants were asked to view visual scenes and then read descriptions of those same scenes. Crucially, we manipulated whether the visual context biased readers into specific pragmatic expectations about how the description might unfold word by word. Contrary to the predictions of pragmatic surprisal theory, we found that participants took longer reading the main critical term in scenarios where they were biased by context and pragmatic constraints to expect a given word, as opposed to scenarios where there was no pragmatic expectation for any particular referent.