Convention: A Philosophical Study
David Lewis's 1969 PhD-thesis-turned-monograph — a game-theoretic analysis of conventions as solutions to coordination problems and the foundation of language
Tradition: Analytic philosophy of language / formal pragmatics
A convention is a stable solution to a coordination problem — and the conventions of language are not arbitrary but rationally selected for mutual benefit
Convention is the published version of Lewis's 1967 Harvard PhD dissertation, written under W. V. O. Quine. Its central thesis: a convention is a regularity of behaviour in a population such that members of the population would prefer to conform to the regularity conditional on others doing so — a stable solution to a coordination problem, in the sense of game theory developed by Thomas Schelling. The book applies this analysis to language: the conventions of natural language (that "cat" means cat in English) are not arbitrary but are rationally selected stable solutions to the coordination problem of communication. Lewis's account answered Quine's sceptical challenge to language conventions ("Truth by Convention," 1936) and shaped much subsequent work in formal pragmatics (Schiffer, Stalnaker, Grice). It also introduced what became standard apparatus in analytic philosophy: the rigorous use of game theory and possible-worlds semantics in philosophical argument, and the analysis of "common knowledge" as a precondition of conventional behaviour.
Author
Editions cited
- Convention: A Philosophical Study (Harvard UP, 1969; reprinted Blackwell, 2002, with new introduction by the author shortly before his death)
School Embodiments
The book is paradigm analytic philosophy: a formally rigorous treatment of a philosophical concept (convention) using the apparatus of game theory and modal logic.
"A regularity R is a convention iff in any instance of S among members of P, (1) everyone conforms to R, (2) everyone expects everyone to conform to R, (3) everyone prefers to conform to R conditional on others doing so." (Convention, ch. 2)
The functional analysis of conventions — they exist because they solve coordination problems — is pragmatist in shape; Quine's anti-foundationalism is the methodological background.
"Conventions are not contracts; no one ever agrees explicitly to them. They emerge because they are useful and persist because of their utility." (Convention, ch. 3)
Lewis treats language and other conventional phenomena naturalistically, without postulating any normative authority beyond the rational interests of agents in coordination.
"There is nothing mysterious about conventions; they are exactly what game theory shows them to be — stable solutions to coordination problems." (Convention, ch. 1)
Conventions are not discovered but constructed by populations through repeated interaction; the truth-conditions of conventional sentences are population-dependent.
"What it is for L to be the language of population P is determined by the mutual expectations and conformity-behaviour of P's members, not by anything intrinsic to L itself." (Convention, ch. 5)
The treatment of language as a system whose elements have meaning through their relational coordination-roles within a community is structuralist in spirit, though Lewis comes from a different tradition.
"Meaning is a feature of the system of conventions, not of the words taken individually." (Convention, ch. 4)
Lewis's confidence that rational agents in coordination situations will converge on Nash-equilibrium-style solutions, and that this delivers the structure of language, is rationalist in spirit.
"Agents who are rational, knowledgeable about each other's rationality, and aware of the coordination problem will arrive at one of the equilibria. The convention is the equilibrium that has been selected." (Convention, ch. 2)
The book answers Quine's sceptical challenge to language conventions and so descends, by way of opposition, from the Carnap-Quine debate within logical positivism.
"Quine has shown that language conventions cannot be founded on explicit agreement. I shall show that they can be founded on tacit coordination." (Convention, Preface)
Analytic-philosophical tradition.
Modal-realist tradition.
Internal Tensions
The "common knowledge" condition (every agent knows R, every agent knows that every agent knows R, etc., to arbitrary depth) has been challenged as too strong by Schiffer and others, who developed weaker "mutual belief" alternatives. Whether the analysis really avoids Quine's scepticism (or merely pushes it back to questions about how populations come to share common knowledge in the first place) remains contested. The game-theoretic framework was generative — it shaped Grice, Stalnaker, Brandom — but later work by Searle and Brandom argued conventions involve normative-deontic elements that the rational-choice analysis underweights.
I. Time
The temporal stability of conventions — they emerge through repeated interactions and persist because deviation is locally costly.
Attributes
II. Space
The population (P) as the social space within which the convention holds — Lewis's analysis is explicit that conventions are population-relative.
Attributes
III. Matter
The embodied agents whose coordination problems give rise to conventions.
Attributes
IV. Observer
The rational agent whose first-order conformity and higher-order expectations together constitute the convention.
Attributes
V. Energy
The strategic energies of rational coordination — the costs of deviation, the benefits of conformity, the convergence on equilibrium.
Attributes
VI. Information
Common knowledge as the information-theoretic precondition of convention — every agent knows the regularity, knows that others know it, knows that others know that they know it, ad infinitum.
Attributes
Personas with the nearest attribute fingerprint
Historical figures whose own classification on the same six-dimensional grid lands closest to this work's. Computed by attribute-agreement on coordinates both address.
Computed school proximity
The work's attribute fingerprint scored against all schools using the same quiz scorer. Useful as a sanity check on the hand-curated embodiments above.
How Convention: A Philosophical Study resolves each dilemma
51 resolved positions across 4 dimensions, including 6 distinctive where the majority of schools go the other way · 6 unaligned.
Each dimension is sorted so minority positions come first. Mainstream positions are folded into an expandable list.
Time · 9 dilemmas · 3 distinctive
Persistence, the future, and the direction of becoming.
6 mainstream positions
Matter · 7 dilemmas, all mainstream
Observer · 37 dilemmas · 3 distinctive
Mind, agency, and the knower's relation to the known.