Natural Semantic Metalanguage sets out number of semantic primes. A semantic prime is a meaning which cannot be reduced to a combination of simpler meanings. The Wikipedia article has a table of them, and they come and go as NSM research progresses.

My question is:

(1) How do people discussing semantic primes be sure that they're thinking of the same meanings?

Take sᴀʏ, for example. Is sᴀʏ the meaning of a person using their speech apparatus to make non-musical, meaningful sounds intended to be heard and interpreted by others? Or is it unnecessary that the sounds be intended for others, as when a person speaks aloud their thoughts in private to themselves? Or is it unnecessary that what is interpreted by others be produced by a person's speech apparatus and heard, as when I communicate with you by written word? Or any of whatever possibilities exist beyond the three I just gave?

Perhaps I could reword my question:

(2) How may I acquire the meanings of semantic primes and have confidence that I have acquired them correctly?

  • "They come and go as NSM research progresses". Wiezbicka 2021 claims the list has settled around 65 items and discusses one cpntroversial item at length. The difference you are trying to make seems to be one along the lines of inside-voice, to think outloud, and declaring. That's indeed a very good question at the heart of linguistics. Potentially covered by "implicature" and Gricean maximes, I'd say the question is too imprecise and it's no surprise the accepted answer ignores that part. The example I have in mind is per say, its relation to per se and quotatives such as German so.
    – vectory
    Jan 1 at 22:43

1 Answer 1


NSM primes do have fixed meanings that don't include all of the polysemy present in a language. For example, the YOU prime is strictly singular, so you might need to be careful not to just use YOU for the English word "you" if your meaning was plural, you'd also have to use one of the quantifier primes.

Primes can be identified in a particular language through a set of "canonical sentences", which is available to download. There isn't a particular canonical sentence that covers speaking to yourself alone in a room, but I don't think it would be considered a different sense than the prime sense of SAY. (Consider the difference between "say" and "tell" in English, where "tell" really requires another person.)

The primes don't cover metaphorical extension. For example, it's very natural to use the word "say" to cover what we type or write to each other. But in NSM you'd have to describe that using the LIKE prime. Or perhaps a lot of the time the actual medium of communication is irrelevant, and you could just use the SAY prime.

The most detailed description of each prime that I'm aware of is in Anna Wierzbicka's 1996 book "Semantics: Primes and Universals". I'm not sure if there's a more recent book that goes into as much detail for every prime, or if the more recent amendments to the list of primes are just discussed case by case. But this book's chapter called "A Survey of Semantic Primitives" is very helpful, giving a detailed explanation of each prime.

The section on SAY is not very long, so I can just quote it all here (page 50):

The universal concept of SAY can be illustrated with the following canonical sentences:

I said something to you.
People say something bad about you.
I want to say something now.

Like the indefinability of mental predicates (e.g. THINK), the indefinability of SAY can best be appreciated by looking at contortions and vicious circles in the attempted dictionary definitions of this word.

The concept of SAY plays an important role in speech as a basis of different illocutionary forces (e.g. in questions which imply: 'I want you you say something'), in the thematic organization of utterances ('I want to say something about this'), and in the basic "subject-predicate" structure of sentences ('I'm thinking about X; I say Y'). In the lexicon, its most important function lies in the categorization of discourse, since the distinctions between different "speech acts" and "speech genres" shape, to a considerable extent, our interpretation of human interaction. (See Wierzbicka 1987a.)

  • 1
    This is exactly the information I was looking for, thank you! The canonical sentences in particular are most useful. Dec 31, 2023 at 4:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.