arXiv Open Access 2025

Studies with impossible languages falsify LMs as models of human language

Jeffrey S. Bowers Jeff Mitchell

Lihat Sumber

Abstrak

According to Futrell and Mahowald [arXiv:2501.17047], both infants and language models (LMs) find attested languages easier to learn than impossible languages that have unnatural structures. We review the literature and show that LMs often learn attested and many impossible languages equally well. Difficult to learn impossible languages are simply more complex (or random). LMs are missing human inductive biases that support language acquisition.

Topik & Kata Kunci

cs.CL

Penulis (2)

Jeffrey S. Bowers

Jeff Mitchell

Format Sitasi

APA MLA BibTeX

Bowers, J.S., Mitchell, J. (2025). Studies with impossible languages falsify LMs as models of human language. https://arxiv.org/abs/2511.11389

Akses Cepat

Lihat di Sumber

Informasi Jurnal

Tahun Terbit: 2025
Bahasa: en
Sumber Database: arXiv
Akses: Open Access ✓