Greg Detre
Monday, December 02, 2002
Putnam is attacking the view that meaning can be equated with mental representations. Putnam picks out three assumptions as characterising this, the Aristotelian theory of meaning:
The three main points he makes to undermine this view are that:
o Sometimes our mental representations are too impoverished on their own, and we defer to experts. In the case of the words for �elm� and �beech� for example, the layman�s mental representations for these words contain little more than some phonetic and syntactic information, the knowledge that they are trees and that they are different from each other, and an opinion of where expert knowledge can be found to provide more information. That details of meaning can reside in expert corners of a linguistic community is labelled the �linguistic division of labour�.
o Sometimes the environment contributes to meaning. This is a large step. Putnam uses his Twin Earth thought experiment to help us imagine two language communities whose references for the term �water� are different, i.e. unbeknownst to them, the two substances have different ultimate constitutions and subtly different, unobserved behaviour. However, because both substances have identical observable properties, the words have the same phonetic and syntactic properties, and so lead to the formation of identical mental representations (i.e. concepts/intensions) among the speakers. As more advanced experimental and theoretical knowledge is accrued though, scientists would come to say that �water� on Earth does not mean the same as �water� on Twin Earth. Putnam argues from this that �water� on Earth and Twin Earth never meant the same thing, despite the equivalence of mental representations. In this way, the environment contributes to the meaning � if the referent of the word is different, though the mental representation remains the same, its meaning is different.
o There has to be a social component to meaning. For a start, who we as individual speakers consider to be an expert on a given subject may vary, potentially leading to more than one meaning for a given contentious term.
These can all be seen as indexical components � that is to say that there is something beyond my qualitative (non-indexical) mental representation, whether it�s the other speakers in my language community, the referents as they happen to be in my particular environment (even beyond my knowledge), or knowledge held by people that I defer to at a given moment � all of these factors relativise and fix meaning as being an interaction between my mental representation and things outside my head.
Putnam is very convincing. But he�s also quite fickle. I don�t know what his latest views are, but it wouldn�t surprise me to find that he has conceded some of the force of his argument regarding the contribution of the environment to meaning. In the case of the Twin Earthers, I think I do want to say that the meaning of �water� for them is the same as ours until the relevant discoveries of chemistry have been made, although it is of course different after that. One way to achieve this would be to say that the word �water� gets reinvented on both Earth and Twin Earth as its expert meaning becomes more and more detailed than its layman meaning, and that it happens to get reinvented in two different ways on the two planets. Unfortunately though, this commits us to saying that a word gets reinvented every time we learn something more about its referent, which seems rather far-fetched.
It almost seems that we can detach this referential aspect of meaning from his argument for the indexicality of meaning based on the linguistic division of labour and social components, and still have made progress from the Aristotelian theory. As far as I could tell, this is what Fodor has done, in differentiating between �narrow content� (the semantic representation in Mentalese) and �broad content� (the function which gives the referents in each possible world), and so arguing that the �ordinary notion of meaning is referentially ambiguous�.
This dislodges the problems regarding reference to discussions about how broad content interfaces with the wider world in one direction and the semantic representations in the other. In order to defend this position, one would have to show why the Twin Earth argument for the contribution of the environment to meaning doesn�t hold, and concede most of the the mentalist ground too. Picking up from the argument made above, we could try and do this, not in terms of laymen and experts knowing parts or the whole of meanings, but in terms of multiple meanings. We might then say that the meaning of �gold� for a layman is simple and involves an index to experts who know more. The experts themselves have two meanings for the word �gold� � the lay definition, and their more scientific one. It is then only this scientific one that changes as they find out about the compounds H2O and XYZ.
In truth though, Putnam�s account is explanatorily richer and less clumsy. Moreover, I�m inclined to feel that the LoT hypothesis is not worth saving for other reasons. Putnam�s account is also less expensive � it doesn�t make any requirements about each person�s mental representations being identical in order for them to be able to talk to each other.