Page 132 - Encyclopedia of Philosophy of Language

Truth and Meaning
and by what means they are enforced. The most sys- tematic, and widely debated, attempt to explain the concept of a convention comes from the philosopher David Lewis and it is his account that will provide the principal focus for this article. But the grounds on which some philosophers reject a convention-based account of language altogether will also be inves- tigated. (For a fuller discussion of the related notion of a 'rule' applied to language see Rules.)
1. David Lewis's Account of the Concept of a Convention
The idea that languages are conventional, that the meaning of a word is a matter of convention, is fre- quently entertained, though rarely with any accompanying conception of what a convention is. One cannot just adopt the familiar view of a con- vention as an explicit verbal agreement, since then the account becomes circular: conventions are needed for language and language is needed for conventions. It was not until David Lewis's work on the concept (Lewis 1969) that a fully developed alternative was available. Lewis built upon earlier work by T. C. Schelling on game theory (Schelling 1960) and by David Shwayder (1965) on the concept of a rule.
The basic ideas are as follows. Conventions are regularities in action. Conventions are social, they concern interaction between members of a community, or population. Conventions coordinate actions of different members in particular types of situation. Conventions tend to perpetuate themselves since the existence of a convention in a population gives the members reasons of self-interest to conform to it. Conventions are arbitrary; where there is a con- vention to act in a certain way there are also other ways of acting that would achieve coordination equ- ally well. A regularity in action is conventional only if there is, in a sense, knowledge in the population that the regularity is conventional; more precisely, members must know that other members conform for the same reason as they do themselves.
One of Lewis's prime examples is the convention to drive on the right hand side of the road. This achieves coordination since there is a common interest of avoiding head-on collision. It is arbitrary, since keep- ing to the left would serve this interest equally well. Drivers prefer to keep to the right only insofar as they expect other drivers to keep to the right and they expect other drivers to prefer keeping to the right for the same reason.
The final definition, (in Lewis 1969), runs as follows:
A regularity R in the behavior of members of a popu- lation P when they are agents in a recurrent situation S is a convention if and only if it is true that, and it is common knowledge in P that, in almost any instance of S among members of P:
(a) almost everyone conforms to R 110
(b) almosteveryoneexpectsalmosteveryoneelsetocon- form to R
(c) almost everyone has approximately the same pref- erences regarding all possible combinations of actions
(d) almost everyone prefers that any one more conform to /?, on condition that almost everyone conforms to R
(e) almost everyone would prefer that any one more conform to R', on condition that almost everyone conform to R'
where R' is some possible regularity in the behavior of members of P in 5, such that almost noone in almost any instance of S among members of P could conform both lo R'and to/?.
The notion of common knowledge, although defined differently by Lewis (1969), is roughly this: common knowledge in P that p means that almost everyone in P knows that p, and almost everyone in P knows that almost everyone in P knows that p, and almost everyone in P knows that almost everyone in P knows that almost everyone in P knows that p, and so on. The common knowledge condition in the definition is to ensure that members know that other members have the same reasons for conforming as they have themselves. Keeping to the right would not, according to Lewis, be a convention if everyone believed that everyone else kept to the right by sheer habit.
2. Criticism of Lewis's Account
Lewis's account has been widely appreciated and it undoubtedly captures important aspects of many social phenomena. It has also, however, been heavily criticized, mostly for being too strict in a number of ways, that is, for imposing conditions that are not necessary for something to be a convention.
One such objection is that a convention need not be a regularity in action. It is perfectly possible, by means of explicit agreement, to create a convention for one particular occasion, e.g., a particular sign. Moreover, there may be social conventions which are not generally conformed to, not even in the majority of cases.
There may well be conventions which create con- formity even in the absence of any preference for conformity per se. Fashion in mode of dressing may simply influence people's tastes, but should be regarded as conventional nonetheless.
A language is conventional even if speakers of the language are not aware of the possibility of alternative languages. It is, moreover, conventional even if the speakers deny that the meanings of its words are arbi- trary. So the common knowledge requirement is too strong.
These objections suggest that Lewis has over- rationalized conventions. On the other hand, however, Lewis has also been criticized for not suc-