ADR-009: philosophy-of-language joins the topics

Context

On 2026-09-26 the close reading of Grindrod, Porter & Hansen, Distributional Semantics, Holism, and the Instability of Meaning (LIT-208), found that its philosophy-of-science tag was not justified. The paper is philosophy of language: meaning holism, and how similar two speakers’ meanings must be for communication to succeed. Duhem–Quine appears only as the lineage of the view it defends. Its other tag, linguistics, is blurbed as language “as linguists pose [it], not as models encode [it]”, and this paper is about meaning as models encode it. So neither of its tags was right. The reader named the gap and recommended the word, which is the move the record asks for (ADR-006, ADR-007).

Four other held works are about meaning or assertion in the same sense:

• Marabini on Brandom’s inferentialism and the meaning–entitlement connection (LIT-106);
• Almagro & Isern-Mas, Blunting concepts, on how popularising psychotherapy language changes what its words mean (LIT-214);
• Mallory, Large Language Models Are Stochastic Measuring Devices (LIT-140), which calls error theory the most defensible view of what LLM text means;
• Bottou & Schölkopf, Borges and AI (LIT-178), which denies that a language model asserts anything.

Decision

Add philosophy-of-language to topics, with a label and a blurb, and add it to the philosophy tag group (ADR-003).

work	tags now
Grindrod et al.	philosophy-of-language alone: both old tags were wrong
Marabini	philosophy-of-language first, then logic, then linguistics
Almagro & Isern-Mas	philosophy-of-language first, then linguistics, then ethics
Mallory	philosophy-of-language added; linguistics removed, as in the cognition pass
Bottou & Schölkopf	philosophy-of-language added

logic keeps “inferentialist accounts of meaning” in its blurb. Marabini carries both words, because the paper is about both inference and meaning.

Alternatives

• Keep linguistics for these works. That is the nearest wrong word. Its own blurb excludes meaning as models encode it, and theories of meaning (Brandom, holism) are philosophy rather than linguistics.
• Stretch cognition to mental and linguistic content. cognition is the mind as information processing. What a sentence means, and whether a model asserts, are different questions from how a mind processes information.