Runyi​ Yao


Causal inferences in discourse

My DPhil project focuses the causal inference and expectations in discourse processing. I am investigating questions: What are discourse segments, and how are they linked to underlying event representations? How do readers establish causal relationships between discourse segments? What cues and factors guide causal inferences and expectations in discourse, and what factors constrain them?

Here are some findings of this project:

Publications

  • Yao, R. Altshuler, D., and Husband, E.M. (2026). Discourse expectations are sensitive to the Right Frontier Constraint. Discourse Processes. 1-13. [ link ]
  • Yao, R. Altshuler, D., and Husband, E.M. (submitted). An incremental QUD approach to causal discourse expectations.
  • Yao, R, Husband, E.M., and Altshuler, D. (2024). Topichood and temporal interpretation of DPs guide clause-internal, causal coherence. Proceedings of Sinn und Bedeutung 28, 1034–1048. [ link ]

Conference Presentations

  • Yao, R., Altshuler, D., and Husband, E.M. (2025). Discourse expectations are sensitive to the right frontier constraint. Poster presented at the 38th annual conference on Human Sentence Processing, College Park, MD. [ abstract ][ poster ]
  • Yao, R., Altshuler, D., and Husband, E.M. (2024). Implicit questions-under-discussion raise expectations only in at-issue main clause. Poster presented at the 37th annual conference on Human Sentence Processing, Ann Arbor, MI. [ abstract ] [ poster ]
  • Yao, R., Husband, E.M., and Altshuler, D. (2023). Topichood and temporal interpretation of DPs guide clause-internal, causal coherence. Poster presented at the 28th annual Sinn und Bedeutung Conference, Bochum, Germany and the 10th biannual XPrag Conference, Paris, France. [abstract][poster][ slides ]
  • Yao, R., Sasaki, K., Altshuler, D., and Husband, E.M. (2023). Asymmetric processing effects of intra-sentential explanation coherence. Poster presented at the 29th annual Architectures and Mechanisms for Language Processing Conference, San Sebastian, Spain. [abstract][ poster ]
  • Yao, R., Sasaki, K., Altshuler, D., and Husband, E.M. (2023). Explanation coherence inside sentences, but only offline. Poster given at the 36th Human Sentence Processing Conference, Pittsburgh, PA. [ abstract ] [ poster ]
 

Predictive mechanisms in language comprehension

Human brains are prediction machines. My work in this area investigates predictive mechanisms in language processing, with a particular focus on semantics prediction and discourse-level prediction.

My recent work has focused on the relationship between lexical probability and predictions in online processing. I investigate the role of lexical probability (e.g., surprisal) in prediction, and to what extent it can account for human processing behavior. In addition, I am also interested in how probabilistic measures derived from large language models relate to human behavioral and judgment data, and what this comparison reveals about the mechanisms underlying language comprehension.

This line of work started from my research on the role of Mandarin classifiers in prediction updating. Methodologically, I combine behavioral experiments with computational approaches, using Transformer-based language models to derive probabilistic measures that inform statistical models of comprehension.

For instance, see:

  • Yao, R., Fan. Y., and Husband, E.M (2025). Mandarin classifier informativity gates prediction updating in maze reading. Poster presented at the 38th annual conference on Human Sentence Processing, College Park, MD. [ abstract ] [ poster ]
  • Yao, R., Fan. Y., and Husband, E.M (2024). Rapid prediction updating driven by Mandarin classifiers in Maze Reading. Poster presented at the 3rd Architectures and Mechanisms for Language Processing Asia, Singapore. [ abstract ] [poster]

Question-under-Discussion

I started learning the Question-under-Discussion (QUD) framework when studying implicit causality verbs - a class of verbs that implicitly suggest a cause for an event. I've found the QUD framework to be a convenient theoretical framework for psycholinguistic investigation of discourse-level phenomena. Some of my research aims to explain certain discourse-level effects through the QUD framework, comparing it to alternative theories of discourse coherence like Segmented Discourse Representation Theory (SDRT). I'm also interested in further formalizing and refining the QUD framework itself. My ongoing work with Dr. Katja Jasinskaja is investigating the role of QUDs in terms of discourse attachment and anaphora resolution, comparing the predictions of the QUD framework to the Right Frontier Constraint.

For instance: 

  •  Jasinskaja, K., and Yao, R (2025). The Right Frontier in the QUD Framework. Talk given at The 25th Szklarska Poręba Workshop on the Roots of Pragmasemantics, Poland. [ abstract ] [ slides ]