Professor Lapin and his assistant were very excited at the prospect of building a lexicon for a previously unknown language. They had only recently discovered the lost tribe of Leporidae and today they were to begin recording the meanings of the words in their language.
The first word to be defined was ‘gavagai’. They had heard this word being used whenever a rabbit was present, so Lapin was about to write ‘gavagai = rabbit’. But then his assistant interjected. For all they knew, couldn’t ‘gavagai’ mean something else, such as ‘undetached rabbit part’ or ‘Look! Rabbit!’? Perhaps the Leporidae thought of animals as existing in four dimensions, over time and space, and ‘gavagai’ referred only to the part of the rabbit present at the moment of observation? Or perhaps ‘gavagai’ were only observed rabbits and unseen rabbits had a different name?
The possibilities seemed fanciful, but Lapin had to admit that they were all consistent with what they had observed so far. But how could they know which one was correct? They could make more observations, but in order to rule out all the possibilities they would have to know more or less everything about the tribe, how they lived and the other words they used. But then how could they even begin their dictionary?
Source: Word and Object by W.V.O. Quine (MIT Press, 1960)
Anyone who speaks more than one language will be well aware of certain words that cannot be easily translated from one to another. The Spanish, for example, talk about the ‘marcha’ of a city or party. This is similar but not identical to the Irish word ‘craic’, which is also hard to translate exactly into English. The closest equivalent might be ‘buzz’ or ‘good time feel’ but to know what ‘marcha’ or ‘craic’ means you really have to get under the skin of the language and culture to which they belong.
Similarly, there is not one translation of the verb ‘to be’ in Spanish. Rather, there are two, ‘ser’ and ‘estar’, and which one you need to use depends on differences in the meaning of ‘be’ which the English lexicon does not reflect. And it is not enough to know that ‘esposas’ means ‘wives’ in Spanish to have a full command of the word. You also need to know that it means ‘handcuffs’, and have an awareness of the traditional Spanish machismo.
What the story of the ‘gavagai’ suggests is that all words are like ‘craic’, ‘marcha’, ‘ser’ and ‘esposas’ in that their meanings are tied intimately to the practices of a culture and the other words in the language. Whenever we translate a word into another language we lose these crucial contexts. Most of the time, we can get away with this, since the meanings are similar enough for us to be able to use the word and function in the community of speakers that use it. Hence if Lapin thinks of ‘gavagai’ as ‘rabbit’, he’ll probably get on fine, even if there are subtle differences in meaning between the two. But to understand the true meaning of ‘gavagai’ he must focus on the language and community in which it is embedded, not his English concepts and practices.
Why does this matter? We are apt to think of words as functioning as a kind of label for ideas or objects, which enables people who speak different first languages to talk about the same things and have the same ideas. It’s just that they use different words to do so. On this model, words have a one-to-one relation to their meanings or the things they refer to.
But if we take the ‘gavagai’ story seriously, we need to change this picture radically. Words do not stand in a one-to-one relation with things and ideas. Rather, words are interconnected with each other and the practices of their speakers. Meaning is ‘holistic’, in that you can never truly understand one word in isolation.
If we accept this, all sorts of strange consequences follow. For example, what does it mean for any statement to be true? We tend to think that ‘the rabbit sat on the mat’ is true just if there is a rabbit which sat on the mat. Truth is about a correspondence between a sentence and a state of affairs. But this simple relation is not possible if the meaning of a sentence depends on the language and culture in which it is embedded. Instead of a simple correspondence between sentence and facts, there is a complex web of relations between facts, sentence, the wider language and culture.
Does that mean truth is relative to language and culture? It would be too quick to jump to that conclusion, but from the starting point of meaning holism, it might well be possible to walk slowly to it.
See also
| 19. | Bursting the soap bubble |
| 23. | The beetle in the box |
| 74. | Water, water, everywhere |
| 85. | The nowhere man |