theory in the philosophy of language that a certain part of language (e.g. a term, a complete sentence) can only be understood through its relations to a (previously understood) larger segment of language
Discovered by embedding cosine similarity (sentence-transformers MiniLM, 384-dim).