Found attachment of size 1035061.
Mon, 26 Oct 2026 07:18:35 -0700 Andy from private IP, post #18644366 /all Chomsky was wrong-- language is actually linear chunks with contextual next-phrase prediction For my fellow language nerds, Chomsky has been proved wrong. Language is a lexicon (words) that form linear chunks with contextual next-phrase prediction-- not a deep mathematical tree-like structure as advocated by Chomsky. This important article has big implications for Large Language Models, Natural Language Processing, Computer Science, Law, and many other fields. https://www.nature.com/articles/s41562-025-02387-z #Technology Mon, 26 Oct 2026 07:43:45 -0700 Andy from private IP Reply #16778964 I believe that because the typical person's short-term memory can contain perhaps seven words/entities/chunks, their brain looks back at the last seven entities for connections between them in order to enhance contextual next-phrase prediction. In addition, there is probably a graphical/spatial element to this, in which semantic connections appear in a look-back on the several most recent entities, which results in a stream of potential next-phrases that are refined as additional context becomes available.Replies require login.