• Syntax Without Rules: The Distributional Skeleton of Language

    The most consequential finding in modern linguistics did not come from a linguistics department. It came from scaling next-token prediction: large language models acquire grammatical competence — subject-verb agreement across long dependencies, recursive embedding, tense consistency, productive morphology, pragmatic sensitivity — without being taught any grammar at all.

  • Memory as Trajectory Regeneration

    The storage-retrieval model of memory has been the default framework in cognitive science since Ebbinghaus. Experiences are encoded, consolidated into stable representations, and later retrieved. The metaphor shifts with each generation — wax tablets, filing cabinets, hard drives — but the computational claim remains the same: memory is a lookup...

  • World Properties Without World Models: What Word Embeddings Know About Geography

    A central claim of the autoregressive brain framework is that distributional structure learned over sequential experience is sufficient to support contextually appropriate generation about the world — without the system needing to maintain explicit world models, stored representations of physical reality, or any mechanism that “looks at” the external environment....