Looks like an interesting read, thank you @v9v.
Just when my night was going through a meditative sleep about basing ontological models using change as fundamental block. Identity is such a brittle choice as foundation, even if it's a great tool in many situations otherwise.
Many ancient cultures use behavior as identity. It certainly has a charm.
Was it the Navajos whose language doesn't have nouns, only verbs? A noun is a kind of illusion of eternal identity. A chair is only chair-ing for the moment as a configuration of matter that was doing something else before, and will fall apart and transform into doing something else in the future.
IIRC Navajo has a pretty robust noun-verb distinction. However there definitely are other languages where nouns and verbs behave very similarly, e.g. most famously Salishan languages. That said, there don’t seem to be any natural languages in which nouns and verbs are completely indistinguishable — there’s always some minor difference in how they behave.
I don't think noun as a grammatical class is an issue, all the more if we take for granted that grammar themselves are mere inferences modeling what's happening on average when producing some utterance, or at least a very simplified representation of a leaned version of the utterance.
It might become more problematic when using a term such as substantive which can connotate some ontological beliefs about the nature of the word or what it refers to, or their relationship.
English is already very generous with conversion of word type without morphological impact in general. I heard Mandarin don't have even that kind of string lexical typology bound to every item in the vocabulary, but I didn't check the details to be transparent.
Thanks anyway for the hints on the other languages