“If it talks likes a duck, walks like a duck, it is functionally a duck.”
I believe behaviorist/functionalist definitions of mental capacity (incl. language) are important because they represent how language is used in practise. It seems like the linguists you reference have elevated meaning into the test of linguistic competence maybe because they are trying to preserve human exceptionalism.
However, it could also be that linguists are unimpressed with recent ML-driven models of language not only because the models fail on certain probes but also because we’ve all become habituated to twitter bots and automated phishing scams which both rely on crude (by today’s standards) language automation. I’ve worked at a publisher before which received dozens of manuscripts which were linguistically well formed but void of meaning.
Maybe we’re all overly obsessed with essence and authenticity in the same way that a forged painting has zero value even if it is physically (99%) indistinguishable to the original.
“If it talks likes a duck, walks like a duck, it is functionally a duck.”
I believe behaviorist/functionalist definitions of mental capacity (incl. language) are important because they represent how language is used in practise. It seems like the linguists you reference have elevated meaning into the test of linguistic competence maybe because they are trying to preserve human exceptionalism.
However, it could also be that linguists are unimpressed with recent ML-driven models of language not only because the models fail on certain probes but also because we’ve all become habituated to twitter bots and automated phishing scams which both rely on crude (by today’s standards) language automation. I’ve worked at a publisher before which received dozens of manuscripts which were linguistically well formed but void of meaning.
Maybe we’re all overly obsessed with essence and authenticity in the same way that a forged painting has zero value even if it is physically (99%) indistinguishable to the original.