In Word and Object, Quine wrote about how we can never be sure as to what a word actually means in and translates into our own language if we were a linguist studying an un-contacted native tribal language.

The example he gives is the word "gavagi" which is supposed to translate into English as "rabbit" but this could refer to "rabbithood", "undetached rabbit parts", "temporal slices of rabbits", and so on, within the native tribal community and we'd never really know with much certainty.

Wouldn't this also apply to a young child learning the native language for the first time as well?

Wouldn't this young child start to form their own concept of "gavagi" independent of what his or her teachers think that "gavagi" means or refers to?

  • See Quine on Indeterminacy of Translation: "Some philosophers hold that the idea of indeterminacy is absurd, or that it amounts to an extreme form of scepticism about whether we ever understand one another, or whether correct translation is possible at all." Commented Sep 11, 2017 at 6:54
  • I think you may have confused indeterminacy of translation with inscrutability of reference. The former is about sentences whose meaning we cannot determine, in that example 'Gavagi' is not a word as such, but the one word sentence 'Gavagi!', spoken to express the particular neural stimulus caused by a rabbit. The failure to determine the meaning of the word 'gavagi' is Quine's inscrutability of reference theory, in 'Word and Object'. Perhaps you could clarify what it is that you were reading when this question arose?
    – user22791
    Commented Sep 11, 2017 at 7:52
  • But your example is sound: the issue is about language learning. We (humans) do not learn language through "words defintiions". Commented Sep 11, 2017 at 9:34
  • I was watching a lecture on this. But it seems, after I read a bit more he says that we can have multiple translations that are equally "right", not that we can't ever have any successful translation at all between two languages. If you are a young child born into this tribe, trying to learn this language, you will also run into the same problems that are present in any language. You can never have a perfect understanding of another person's intended meaning due to differences in brain states.
    – user28485
    Commented Sep 11, 2017 at 14:18

1 Answer 1


It certainly would, and you can observe it doing so. Our understanding of one another cannot be complete. There can be no true communal standard as to what 'rabbit' actually means, and as we move away from the central anchor of 'rabbit-hood' we will disagree on what features things need to retain in order to remain 'rabbits'.

This variation is sterner and broader in children than adults. Notoriously, way too many things are doggies for a while, and kids disagree a whole lot more on what doggy-hood is depending on whether they have contrasting notions of kitty-hood or horsey-hood mixed into their experience. They make those guesses and fight with one another about them, or test them by how adults respond to their usage. Language is a game more like 'tea time' than like 'Monopoly', we learn what the rules are by taking part, not by being told.

Humans don't learn definitions or grammar as children, because we know that is impossible, basically from logic like Quine's (more specifically, but less clearly, Wittgenstein's). Instead, we converge on vague complexes of meaning and usage until we have enough structure to conceive of the notion of a definition and its surrounding grammar. Then we change the way we talk about words, because that structure is a lot more efficient, even if it is less than honest.

This is part of the motivation for Chomsky presuming children have a native ability to vet guesses about meaning against pre-established patterns of grammar. It seemed insurmountable to him that children could both deal with vagueness and learn grammar at the same time, without some built-in instinct for what parts of language were which sort of thing. But apparently we do.

You must log in to answer this question.