You Can't Fight in Here! This is BBS!
2026-04-10 • Computation and Language
Computation and Language
AI summaryⓘ
The authors use a fictional conversation between two experts to explore debates about what modern language models (LMs) can teach us about human language. They point out two common misunderstandings: that LMs are only simple statistical tools like older models, and that current LM research is the final word on language science. They argue for broader research that addresses these concerns to better understand both human language and language models. Their goal is to improve the science around how language works and how AI models process it.
Language modelsLinguistic competenceMarkov modelsString statisticsComputational linguisticsNeuroscienceCognitive scienceAI researchHuman languageLanguage sciences
Authors
Richard Futrell, Kyle Mahowald
Abstract
Norm, the formal theoretical linguist, and Claudette, the computational language scientist, have a lovely time discussing whether modern language models can inform important questions in the language sciences. Just as they are about to part ways until they meet again, 25 of their closest friends show up -- from linguistics, neuroscience, cognitive science, psychology, philosophy, and computer science. We use this discussion to highlight what we see as some common underlying issues: the String Statistics Strawman (the mistaken idea that LMs can't be linguistically competent or interesting because they, like their Markov model predecessors, are statistical models that learn from strings) and the As Good As it Gets Assumption (the idea that LM research as it stands in 2026 is the limit of what it can tell us about linguistics). We clarify the role of LM-based work for scientific insights into human language and advocate for a more expansive research program for the language sciences in the AI age, one that takes on the commentators' concerns in order to produce a better and more robust science of both human language and of LMs.