stuebinm quoted Word and Object by Willard Van Orman Quine
A rabbit scurries by, the native says ‘Gavagai’, and the linguist notes down the sentence ‘Rabbit’ (or ‘Lo, a rabbit’ ) as tentative translation
— Word and Object by Willard Van Orman Quine, Patricia S. Churchland, Dagfinn Føllesdal (Page 56)
So i left this book lying around for some while but picked it up again now, and by chance had stopped at the end of chapter 1, so now read §7 (the first part of chapter 2), which happens to contains the first mention of "Gavagai", which is basically the reason i'm reading this book.
Chapter 1 was basically setup for this: it talks mostly about how language depends on previous experience, and can't really be separated from it (but "speakers of the same language have perforce come to resemble one another").
But then gavagai is about what to do if that fails, which he calls "radical translation": what to do if there is no shared history, no interpreter, nothing? The linguist in the quote might guess it means 'rabbit', but how can they be sure? if they ask a native "gavagai?" while pointing at a rabbit, how could they even distinguish agreement from disagreement? (can they even be sure the native speaker understands what the "pointing" gesture means?)
The whole story has (afaict) become a bit of a classic for talking about this, but honestly i find it uncomfortable: why does Quine have the linguist ask the native questions, then guess from the reactions, gradually trying to improve? What about the native, can't they decide to maybe teach this confused person walking around pointing at random things and asking "gavagai"? Even if they speak entirely unrelated languages, surely the linguist's perception of the 'natives' will be informed by preconceptions about how 'natives' behave? (Quine's is, at least; why else are all examples about rabbits and a giraffe?)
I dunno, I'm not sure this actually makes any difference to the point he's trying to make — could the linguist even tell if the native's trying to teach them? —, except, well, Quine works pretty hard to construct a situation in which the language learner has all the agency and at the same time acts as an observer mostly separated from the situation (i wonder if there are surviving accounts from the few cases in recorded history where truly "radical" translation had to be done? how did they do it? i bet it wasn't this one-sided a process)