Large Language Models and Impossible Language Acquisition: "False Promise" or an Overturn of our Current Perspective towards AI

In three linesExperimental study testing Chomsky's critique of LLMs: GPT-2 small and LSTM trained on syntactically impossible languages (reversed sentences, parity-based negations). GPT-2 shows lower perplexity on natural language (loss ratios up to 2.25× on reversed conditions), LSTM minimal differences. Authors propose functionalist paradigm against Chomsky's rationalist perspective.

Summary generated by Claude — human-verified