combining elements is not enough. A fundamental aspect of compositionality in human language is that it is productive. We do not just reuse a fixed set of combinations; we generate new ones, effortlessly.
I think this is a great take. And it has a nice implication against language purism:
If compositionality demands the gen of new elements, Language* demands compositionality, and any language* requires Language, then any language requires the gen of new elements. And yet purism is all about not using new elements - no neologisms, no borrowings, just take the language vocab “as is” and deal with it.
In other words, applying purism to a language means to not use said language. Language purists are thus fighting against the very thing they claim to defend.
*capital ⟨L⟩ for the human faculty; minuscule ⟨l⟩ for specific usages of it (like Arabic, Breton, Cherokee, etc.)
Back on non-human primates: I mentioned this in another thread, but IMO “we” (people in general) should stop seeing “is this language?” as a binary matter, and more like a gradient: “how close is this to language?”. What they’re doing is still not on the same level as we do, but it’s already beyond non-linguistic communication.
Hmm, purism can take many shapes, it’s not a strictly formulated stance (even though it might act like it is “scientific” because it minds etymology). It doesn’t have to be negative towards neologisms, in fact it can be very positive towards them if they’re based on native material and are meant to replace loanwords.
Fair point on the attitude changing towards neologisms.
Well, it was just a cheeky thought anyway.
Sure but what I am struggling to understand here is that these sets of communication elements must have been generated in the first place by the animals that use them, so this resolves nothing really. Is there something I am missing here?
Yeah, they didn’t propose how those sets of calls appear; only that they’re there, in other primates. So no, you aren’t missing it.
What I’m going to say is just a guess from my part. Those sets would appear like this:
- Simple call, conveying some simple info (for example: ook = “threat”)
- Call gets repeated to ensure others got the message. (for example: ook ook ook = “threat, threat, threat”)
- The number of repetitions gets associated with some additional info. (for example: ook ook = small threat, ook ook ook = big threat).
- The repetitions get some rhythm structure, to ensure others got the whole thing.
On #4 you already got a set. But all steps are on their own advantageous for the survival of the group.
However, once you got through all those steps, a problem appears: since the set itself is conveying info, how to ensure the info is not missed? Then you go back to #2, repeating the whole set to ensure others got the message.
