Jon Rawski
Assistant Professor of Linguistics
About
I teach and study the mathematics of human language and learning. My work places mathematical boundary conditions on the grammars underlying human language, and on how they can be learned.
Interests
- Mathematical Linguistics
- Cognitive Science
- Theoretical Computer Science
Recent Publications
I try to make my work accessible following the FAIR Principles, on this website or some other repository.
We show some divergences in expressive power when transformers are used as probabilistic and autoregressive sequence models.
We characterize transformer sequence models as transducers and show some expressivity bounds for several variants.
We give a pragmatic problem-centered account of theories’ epistemic virtues, like coherence, depth, and parsimony.
The Computational Power of Harmonic Forms
We overview vowel harmony computationally, describing necessary and sufficient conditions on phonotactics, processes, and learning.
Regular Reduplication Across Modalities
We show that reduplication in signed languages respects the upper bound of regular functions seen in spoken languages
