Editing David Lewis (philosopher) (section)

== Early work on convention ==
Lewis's first monograph was <cite>''Convention: A Philosophical Study''</cite> (1969), which is based on his doctoral dissertation and uses concepts of [[game theory]] to analyze the nature of social conventions. It won the [[American Philosophical Association]]'s first Franklin Matchette Prize for the best book published in philosophy by a philosopher under 40. Lewis claimed that social conventions, such as the convention in most states that one drives on the right (not on the left), the convention that the original caller will re-call if a phone conversation is interrupted, etc., are solutions to so-called "'co-ordination problems'". Co-ordination problems were at the time of Lewis's book an under-discussed kind of game-theoretical problem; most game-theoretical discussion had centered on problems where the participants are in conflict, such as the [[prisoner's dilemma]].

Co-ordination problems are problematic, for, though the participants have common interests, there are several solutions. Sometimes one of the solutions is "salient", a concept invented by the game-theorist and economist [[Thomas Schelling]] (by whom Lewis was much inspired). For example, a co-ordination problem that has the form of a meeting may have a salient solution if there is only one possible spot to meet in town. But in most cases, we must rely on what Lewis calls "precedent" for a salient solution. If both participants know that a particular co-ordination problem, say "which side should we drive on?", has been solved in the same way numerous times before, both know that both know this, both know that both know that both know this, etc. (this particular state Lewis calls [[common knowledge (logic)|common knowledge]], and it has since been a frequent topic of discussion among philosophers and game theorists), then they will easily solve the problem. That they have solved the problem successfully will be seen by even more people, and thus the convention will spread in the society. A convention is thus a behavioral regularity that sustains itself because it serves the interests of everyone involved. Another important feature of a convention is that a convention could be entirely different: one could just as well drive on the left; it is more or less arbitrary that one drives on the right in the US, for example.

Lewis's main goal in the book, however, was not simply to provide an account of convention but rather to investigate the "platitude that language is ruled by convention" (<cite>Convention</cite>, p.&nbsp;1.) The book's last two chapters (<cite>Signalling Systems</cite> and <cite>Conventions of Language</cite>; cf. also "Languages and Language", 1975) make the case that a population's use of a language consists of conventions of truthfulness and trust among its members. Lewis recasts in this framework notions such as truth and analyticity, claiming that they are better understood as relations between sentences and a language rather than as properties of sentences.