Matilde Marcolli
California Institute of Technology, CA·Mathematics & Statistics
Machines can now generate fluent text, but imitating language is not the same as explaining it. A precise theory of language as a computational system is still missing.
- Fields
Mathematics & Statistics
Computer Science
Social Science
- Topics
mathematical linguistics
generative linguistics
Minimalist Program
- Grants
- 1 known grant, through Jul 2028
What they work on
- Extend the Hopf algebra Markov chain formulation of Merge, the operation that builds syntactic trees in the Minimalist Model, so it captures all of syntax including the interfaces with morphology and semantics.
- Build rigorous mathematical models of externalization, the conversion of internal syntactic structure into spoken or written form, and of the syntactic parameters that vary across languages.
- Develop computable quantitative invariants that allow direct mathematical comparison between human language formation and the behavior of large language models.
- Quantify how far large language models reconstruct the computational process of syntax from data encoding only semantic proximity between words.
More about the research
Human language is a highly structured natural phenomenon that, like a physical system, can be modeled mathematically with real predictive power. This project builds out such a model and uses it to compare how humans and machines form language.
Why it matters
Language technologies are spreading faster than the theory that would explain them. Exact mathematical models of syntax, paired with computable measures comparing people and machines, could give linguists and AI researchers rigorous tests of whether language models capture genuine linguistic structure, and clearer evidence about what those systems actually learn.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Elena Koulidobrova
Central Connecticut State University, CT·Social Science
Build treebanks, collections of sentences annotated for grammatical structure in the Universal Dependencies format, that capture how evidential markers (grammatical signals of information source) and focus features appear in sentences.
Eszter Ronai
Northwestern University, IL·Psychology
Identify the real-time processing signals of focus, the marking of new or emphasized information, in languages with fixed versus flexible word order, using eye-gaze data and statistical modeling.
Kyle Mahowald
University of Texas at Austin, TX·Social Science
Measure how speakers of a variety of languages choose deictic words such as "this," "that," "here," and "there" to refer to objects and locations, using cross-linguistic behavioral experiments.
The school
California Institute of Technology: 18 labs on PI Pages, $445M a year on research, 200 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.