We frequently understand utterances more deeply than their explicit content suggests. For example, "Mary had cheesecake for dessert" often implies she had only cheesecake, and "I ate some of your cookies" suggests I didn't eat all of them. We make these inferences swiftly during real-time language processing. In this project, with
David Barner, we investigate how this feat is accomplished: Do listeners initially adopt a non-logical, pragmatic interpretation or begin with a logical meaning and then enrich the conditional statement via inference?
Using web-based eye tracking, with
Eva Wittenberg, we focus on how negation and the availability of alternatives shape our mental representations during this process. A negated counterfactual, like 'If there had been no zebras, there would have been no lions in the zoo', suggests a world without these animals, while implicitly confirming their existence. These negated counterfactuals help answer the implicit Question under Discussion (QuD) by reducing uncertainty and creating specific mental images. This interaction indicates that the cognitive effort in comprehending counterfactuals primarily arises from uncertainty over alternative states and QuD accommodation.
In natural communication settings, ambiguous referring expressions like 'she', 'that', 'there', 'then' are rarely encountered without context. They usually appear after a meaningful discourse, enabling the construction of plausible hypotheses about their reference. Even within coherent discourse, identifying the correct reference among multiple alternatives is complex, involving understanding who or what the utterance is about, the event, the participants and their roles, and the relationship of the utterance to other discourse elements. Focusing on Turkish, with
Duygu Özge, we examine how null and overt pronouns are interpreted across different verb types, coherence relations, valency, and word order.