Jonathan Dunn

LecturerJonathan Dunn

Elsie Locke Rm 206
Internal Phone: 90305
I specialize in computational models of both grammar learning and grammatical variation.


Research Interests


My research is centered around three ideas:

First, that meaning and usage are essential parts of language
Second, that computational models can encode and test linguistic theories
Third, that linguistics should be applied to practical problems

I have focused on the intersection of form, usage, and meaning: Construction Grammar, Dialectal Variation, and Metaphor. I describe this collection of work as computational cognitive linguistics because it uses computational modeling to formalize and test ideas from Cognitive Linguistics. On a practical level, my work provides models that can be used to annotate written corpora: Language Identification, Dialect Identification, and CxG Parsing.

I currently am working on global-scale dialectometry as the combination of grammar induction and geospatial text classification. The goal is to model regional syntactic variation so accurately that the model can be used to predict an individual’s region-of-origin.

Recent Publications

  • Dunn J. (2019) Frequency vs. Association for Constraint Selection in Usage-Based Construction Grammar. In Proceedings of the NAACL 2019 Workshop on Cognitive Modeling and Computational Linguistics.: 117-128.
  • Dunn J. (2019) Modeling Global Syntactic Variation in English Using Dialect Classification. In Proceedings of the NAACL 2019 Sixth Workshop on NLP for Similar Languages, Varieties and Dialects: 42-53.
  • Dunn J. (2018) Corpus-Based Dialectometry Using Construction Grammars. Salt Lake City, UT: Linguistic Society of America Annual Meeting, 1 Jan 2018.
  • Dunn J. (2018) Finding variants for construction-based dialectometry: A corpus-based approach to regional CxGs. Cognitive Linguistics 29(2): 275-311.
  • Dunn J. (2018) Modeling the Complexity and Descriptive Adequacy of Construction Grammars. In Proceedings of the Society for Computation in Linguistics 1(1) 10: 81-90.