Computational-Linguistics on Marginaliahttps://sguzman.github.io/marginalia/tags/computational-linguistics/Recent content in Computational-Linguistics on MarginaliaHugoen-usSat, 22 Aug 2026 00:00:00 +0000The Learner in the Corpus: What LLMs Actually Show About Chomskyhttps://sguzman.github.io/marginalia/research/how-llms-challenge-chomskyan-assumptions-analytical-report/Tue, 03 Mar 2026 00:00:00 +0000https://sguzman.github.io/marginalia/research/how-llms-challenge-chomskyan-assumptions-analytical-report/A source-audited assessment of what neural language models actually establish about Chomskyan learnability claims. Specific computational results weaken some poverty-of-the-stimulus and impossible-language arguments, while child-scale data, model inductive bias, harder syntactic phenomena, and cognitive mismatch block a blanket <code>Chomsky refuted</code> verdict.