Score: 1

Language models as tools for investigating the distinction between possible and impossible natural languages

Published: December 10, 2025 | arXiv ID: 2512.09394v1

By: Julie Kallini, Christopher Potts

BigTech Affiliations: Stanford University

Potential Business Impact:

Teaches computers to tell real languages from fake ones.

Business Areas:
Natural Language Processing Artificial Intelligence, Data and Analytics, Software

We argue that language models (LMs) have strong potential as investigative tools for probing the distinction between possible and impossible natural languages and thus uncovering the inductive biases that support human language learning. We outline a phased research program in which LM architectures are iteratively refined to better discriminate between possible and impossible languages, supporting linking hypotheses to human cognition.

Country of Origin
🇺🇸 United States

Page Count
3 pages

Category
Computer Science:
Computation and Language