We study the cognitive processes that underlie language knowledge and use, using both experimental and computational approaches. Our research addresses linguistic meaning, form, and the relation between them.


Why is language so ambiguous? How do we resolve ambiguity when we encounter it? The lab conducts work on lexical, syntactic, and pragmatic ambiguity using behavioral and statistical techniques.

Trott, S., & Bergen, B. (2023). Word meaning is both categorical and continuousPsychological Review.

Trott, S., & Bergen, B. (2022). Languages Are Efficient, But For Whom? Cognition 225.

Trott, S. & Bergen, B. (2021). RAW-C: Relatedness of Ambiguous Words in Context (A New Lexical Resource for English). In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics.

Jones, C. & Bergen, B. (2021). The Role of Physical Inference in Pronoun Resolution. In Proceedings of the 43rd Annual Conference of the Cognitive Science Society. Austin, TX: Cognitive Science Society.

Trott, S. & Bergen, B. (2020). Why do human languages have homophones? Cognition 205(104449).

Trott, S. & Bergen, B. (2018). Individual Differences in Mentalizing Capacity Predict Indirect Request ComprehensionDiscourse Processes, 1-33.

Embodied simulation for language comprehension

How do comprehenders interpret and encode the meaning of the language they’re exposed to? Part of the answer might be that they use sensorimotor systems to activate and enact the sights, sounds, and actions that the language describes and implies. This is the embodied simulation hypothesis, explored in the work below, among many other papers from the lab.

Bergen, B. (2012) Louder than Words: The New Science of How the Mind Makes Meaning. New York: Basic Books.

Schütt, E., Dudschig, C., Bergen, B. & Kaup, B. (2022). Sentence-based mental simulations: Evidence from behavioral experiments using garden-path sentences. Memory and Cognition.

Sato, M. and B. Bergen. (2013). The case of the missing pronouns: Does mentally simulated perspective play a functional role in the comprehension of person?Cognition 127(3):361-74.

Bergen, B, S. Lindsay, T. Matlock, and S. Narayanan. (2007). Spatial and linguistic aspects of visual imagery in sentence comprehensionCognitive Science 31: 733-764.

Large Language Models

Large computational language models are beginning to approach human competence in a variety of tasks. What do their successes and failures teach us about how humans learn and use language?

Trott, S., Jones, C., Chang, T., Michaelov, J., & Bergen, B. (2023). Do Large Language Models know what humans know? Cognitive Science 47(7): e13309.

Chang, T. & B. Bergen (2023). Language Model Behavior: A Comprehensive Survey. Computational Linguistics.

Michaelov, J., S. Coulson, & B. Bergen (2023). Can Peanuts Fall in Love with Distributional Semantics? Proceedings of the 45th Annual Meeting of the Cognitive Science Society. Austin, TX: Cognitive Science Society.

Michaelov, J. & Bergen, B. (2022). Do language models make human-like predictions about the coreferents of Italian anaphoric zero pronouns? Proceedings of the 29th International Conference on Computational Linguistics.

Michaelov, J., Coulson, S., & Bergen, B. (2022). So Cloze yet so Far: N400 Amplitude is Better Predicted by Distributional Information than Human Predictability JudgementsIEEE Transactions on Cognitive and Developmental Systems.

Jones, C., Chang, T., Coulson, S., Michaelov, J., Trott, T., & Bergen, B. (2022). Distributional Semantics Still Can’t Account for Affordances. Proceedings of the 44th Annual Conference of the Cognitive Science Society. Austin, TX: Cognitive Science Society.

Chang, T. A. & Bergen, B. K. (2021). Word acquisition in neural language modelsTransactions of the Association for Computational Linguistics.


In many languages, words with similar meanings tend to also have similar forms. In English, words relating to light and vision tend to start with gl-, like glitter, gleam, and glisten. Our works asks whether these associations are psychologically real and whether can we discover them using statistical methods.

Bergen, Benjamin. (2004). The psychological reality of phonaesthemesLanguage, 80(2).

Gutierrez, E.D., Roger Levy, & Benjamin Bergen. (2016). Finding Non-Arbitrary Form-Meaning Systematicity Using String-Metric Learning for Kernel Regression. Proceedings of the Association for Computational Linguistics.

The Science of Swearing

Profanity works differently from other kinds of language–it manifests differently in the brain, it affects the body differently, and it changes differently over time. Our work addresses the aspects of cognition that can be uniquely revealed through taboo language.

Bergen, B. 2016. What the F: What Swearing Reveals About our Language, Our Brains, and Ourselves. New York: Basic Books.

Bergen, B. (2019). Do gestures retain mental associations with their iconic origins, even after they become emblematic? An analysis of the middle-finger gesture among American English speakersPLOS ONE 14(4): e0215633.