What do you think?
Rate this book
Convention: A Philosophical Study
Convention was immediately recognized as a major contribution to the subject and its significance has remained undiminished since its first publication in 1969. Lewis analyzes social conventions as regularities in the resolution of recurring coordination problems-situations characterized by interdependent decision processes in which common interests are at stake. Conventions are contrasted with other kinds of regularity, and conventions governing systems of communication are given special attention.
228 pages, Paperback
First published January 28, 1969
22 people are currently reading
459 people want to read
About the author
David Lewis
375 books72 followersLibrarian Note: There is more than one author by this name in the Goodreads database.
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 10 of 10 reviews
August 21, 2016
Lewis attempts to rescue the notion of analyticity, which was put into question by his mentor, W.V.O. Quine. This notion, regarding propositions, or what are otherwise known as declarative sentences, has it that some propositions are analytic, that they are true regardless of their interpretation. It was supposed by philosophers, in a long-standing tradition, that the analyticity of a proposition was a result born purely of the rules of forming sentences out of terms and the meaning of those terms. So one could say that "All bachelors are unmarried" is an analytic proposition because the predicate "unmarried" is contained in its subject "bachelors."
Quine questioned how terms got their meanings with an increasing force of argument in his career, culminating in what would become the most famous philosophical paper written in the 20th century. Yet his colleagues could not bring themselves to giving up the notion of analyticity, and would say that meaning could be explained by the conventional use of language. Many would be happy with that. But what then is a convention? More importantly, what sense can one put to a convention when there could never have been a convening?
It is this last question which Lewis answers in his book. He does a remarkable job of it, using aids in areas of science such as game theory, linguistics, and the use of signals from information theory. His cunning use of these in the work of explicating the notion of convention will make the book worth reading even if you had no other reason. But you will have others reasons. Conventions are interesting in themselves, though you may not have stopped to think of them before. They invade every aspect of social life. It is perhaps surprising that this book is not more widely read by those who have a vocational interest in subjects which are invaded by conventions. I'm thinking of sociologists, psychologists, and economists in particular.
The book is also quite entertaining, Lewis being not only a very lucid writer but an amusing one as well.
Despite all the book's perks I do think its argument falls short of its thesis. When Lewis finally comes to the definition of analyticity, in terms of linguistic conventions, he sets it down also in modal terms, with reference to the so called "possible worlds." These are Lewis' favourite objects of contemplation, as one notes from another of his famous works. However, the modal notion of possible worlds is no less questionable than the one of analyticity which he sets out to rescue. Indeed, you can start with either one and derive the other.
In the last few pages of Conventions, Lewis confesses to this problem but tries to cure the reader of any worry with a few words. It is not enough. When one is going to rest the entirety of an argument on one point it deserves more than a few passing remarks in the conclusion. Do not let that deter you! It is a wonderful book, filled in its mere 200 pages with much that will feed your own thinking.
Quine questioned how terms got their meanings with an increasing force of argument in his career, culminating in what would become the most famous philosophical paper written in the 20th century. Yet his colleagues could not bring themselves to giving up the notion of analyticity, and would say that meaning could be explained by the conventional use of language. Many would be happy with that. But what then is a convention? More importantly, what sense can one put to a convention when there could never have been a convening?
It is this last question which Lewis answers in his book. He does a remarkable job of it, using aids in areas of science such as game theory, linguistics, and the use of signals from information theory. His cunning use of these in the work of explicating the notion of convention will make the book worth reading even if you had no other reason. But you will have others reasons. Conventions are interesting in themselves, though you may not have stopped to think of them before. They invade every aspect of social life. It is perhaps surprising that this book is not more widely read by those who have a vocational interest in subjects which are invaded by conventions. I'm thinking of sociologists, psychologists, and economists in particular.
The book is also quite entertaining, Lewis being not only a very lucid writer but an amusing one as well.
Despite all the book's perks I do think its argument falls short of its thesis. When Lewis finally comes to the definition of analyticity, in terms of linguistic conventions, he sets it down also in modal terms, with reference to the so called "possible worlds." These are Lewis' favourite objects of contemplation, as one notes from another of his famous works. However, the modal notion of possible worlds is no less questionable than the one of analyticity which he sets out to rescue. Indeed, you can start with either one and derive the other.
In the last few pages of Conventions, Lewis confesses to this problem but tries to cure the reader of any worry with a few words. It is not enough. When one is going to rest the entirety of an argument on one point it deserves more than a few passing remarks in the conclusion. Do not let that deter you! It is a wonderful book, filled in its mere 200 pages with much that will feed your own thinking.
November 29, 2021
Let F be a conventional signaling system which is made up of a range of contingency plans for communicators (Fc) and a domain of contingency plans for audience members (Fa).
Let s be the state of affairs that “this book is incredible and makes me remember why I fell in love with my discipline”.
I desire that you act according to F & I expect that you will act according to Fa & Fa|Fc = F → I have reason to desire that I act according to Fc
I observe that s holds & I have reason to desire that I act according to Fc & Fc(s) = σ → I do σ
Where σ is giving this book a 5 star review with a whimsical reference to its fourth chapter
and
Fa(σ) = r
Where r = you reading this book yourself and discussing it with me.
Let s be the state of affairs that “this book is incredible and makes me remember why I fell in love with my discipline”.
I desire that you act according to F & I expect that you will act according to Fa & Fa|Fc = F → I have reason to desire that I act according to Fc
I observe that s holds & I have reason to desire that I act according to Fc & Fc(s) = σ → I do σ
Where σ is giving this book a 5 star review with a whimsical reference to its fourth chapter
and
Fa(σ) = r
Where r = you reading this book yourself and discussing it with me.
June 29, 2016
In this book, David Lewis sets out to provide a working definition of convention and then attempts to demonstrate that language is conventional in the more precise sense in which he uses the word. A convention, according to Lewis, is a regularity in the behavior of members of a group across similar situations, when and only when it is both true and known by the members of the group that across the similar situations 1) almost everyone will conform to the regularity, 2) members of the group will have mostly the same preferences regarding the possible actions, and 3) the members of the prefer would prefer themselves and the other members to conform to the regularity as to some other regularity (adapted from 78). Lewis demonstrates that a person could construct a simple language that operates according to this definition, and so be conventional in his sense. He also tries to show that a complex language could be conventional. However, even if his reasoning that a complex language could be conventional is sound (I'm not sure if it is; it was a difficult argument for me to understand), it does not follow that language in the real world is conventional, which is what Lewis would like to prove. Take the second necessary condition above, that members of the group have mostly the same preferences regarding possible actions. For language to be conventional, then, speakers of a language (that is, members of the said group) would have mostly the same preferences regarding language use (possible actions). As a matter of fact, however, speakers have an infinite variety of ways to generate sentences and an infinite variety of ways to initiate and respond in conversation, to take a couple of examples. If this is the case, then no speaker of a language could have most of the same preferences for how to use the language. Therefore, language is not conventional. A more complicated story of language is told in linguistics, namely in its conception of an ur-language as a natural endowment of human beings, which itself is capable of producing the complex variety of spoken, written, and signed languages on the basis of adjustments in cognitive 'parameters' regarding syntactic structures, the adjustments being the result of environmental 'triggers' vis-a-vis human beings. This hypothesis is the predominant view these days, and which seems to have the most support.
Even though Lewis might be wrong about language, it does not deter from the usefulness of his working definition of a convention and its application to other aspects of human behavior, especially in its potential explanatory power for how humans socially organize. Matters ranging from driving on the right side of the road in the United States to the Milgram experiment could fall under the category of human adherence to convention. Thus, just because Lewis did not adequately explain language does not mean that the concept he attempted to use to explain language should be dismissed.
Also, Lewis provides a working definition in his book of a social contract to contrast it with convention. A social contract, according to Lewis, is a regularity in the behavior of members of a group across similar situations when and only when it is both true and known that 1) members of the group in some relevant situation will conform to the regularity, and 2) the members of the group prefer to conform to the regularity than not to conform to the regularity, where not conforming to the regularity would lead no members of a group to anticipate a regularity across situations (paraphrased from 88-89). If this definition seems too close to a convention, Lewis further clarifies in the following way. "For convention, we require that each agent prefer general conformity to conformity by all but himself [sic:], ignoring his preferences regarding states of general nonconformity. For social contract, we require that each agent prefer general conformity to a certain state of general nonconformity, ignoring his preferences regarding conformity by all but himself [sic:]" (90). That is, for a convention, a person would just prefer that she and the others conform to this regularity and not to some others, whereas in a social contract, the alternative is no regularity at all, and so of utmost importance to the person is that everyone conforms to the regularity, without which there would be no regularity to conform to at all. For example, Hobbes argues that without conforming to the rule of law (the regularity), human beings resign themselves to a life that is nasty, brutish, and short (no regularity at all).
Overall, it was a good read. You should read it if you like philosophy. If not, you probably wouldn't like it.
Even though Lewis might be wrong about language, it does not deter from the usefulness of his working definition of a convention and its application to other aspects of human behavior, especially in its potential explanatory power for how humans socially organize. Matters ranging from driving on the right side of the road in the United States to the Milgram experiment could fall under the category of human adherence to convention. Thus, just because Lewis did not adequately explain language does not mean that the concept he attempted to use to explain language should be dismissed.
Also, Lewis provides a working definition in his book of a social contract to contrast it with convention. A social contract, according to Lewis, is a regularity in the behavior of members of a group across similar situations when and only when it is both true and known that 1) members of the group in some relevant situation will conform to the regularity, and 2) the members of the group prefer to conform to the regularity than not to conform to the regularity, where not conforming to the regularity would lead no members of a group to anticipate a regularity across situations (paraphrased from 88-89). If this definition seems too close to a convention, Lewis further clarifies in the following way. "For convention, we require that each agent prefer general conformity to conformity by all but himself [sic:], ignoring his preferences regarding states of general nonconformity. For social contract, we require that each agent prefer general conformity to a certain state of general nonconformity, ignoring his preferences regarding conformity by all but himself [sic:]" (90). That is, for a convention, a person would just prefer that she and the others conform to this regularity and not to some others, whereas in a social contract, the alternative is no regularity at all, and so of utmost importance to the person is that everyone conforms to the regularity, without which there would be no regularity to conform to at all. For example, Hobbes argues that without conforming to the rule of law (the regularity), human beings resign themselves to a life that is nasty, brutish, and short (no regularity at all).
Overall, it was a good read. You should read it if you like philosophy. If not, you probably wouldn't like it.
November 3, 2014
A lot of writers throw around "convention" as a sort of cop-out explanation for a confusing phenomenon -- why do we called a dog "dog" in the US and not some other word? why do they use rupees in India and not some other currency? Lewis pinpoints the qualities that these scenarios, especially language (the mother of all conventions), have in common, and places the word "convention" back on solid ground.
This is essential reading for social psychologists and linguists trying to use the word in a meaningful, consistent way. Certain parts of the argument have not aged terribly well, especially the parts drawing fine distinctions between Lewis' theory and other contemporary proposals that haven't been as long-lived. This is understandable, given its status as a PhD thesis (and woe to all current PhD students who have to compare their own doctoral work to this masterpiece), but makes it a bit of a slog to pick out the important parts, sometimes.
This is essential reading for social psychologists and linguists trying to use the word in a meaningful, consistent way. Certain parts of the argument have not aged terribly well, especially the parts drawing fine distinctions between Lewis' theory and other contemporary proposals that haven't been as long-lived. This is understandable, given its status as a PhD thesis (and woe to all current PhD students who have to compare their own doctoral work to this masterpiece), but makes it a bit of a slog to pick out the important parts, sometimes.
May 27, 2021
A classic text in the history of philosophy that is short but, perhaps, not for the faint of heart. It adds some richness to have read Hume's Treatise already, which Lewis refers to on occasion. But anyone interested in convention or topics where one often sees the phrase "This is a matter of convention...." should probably read at least the first half of this book at least once.
For those familiar with game theory, I imagine this will be a pleasant read. Lewis has a nice writing style, especially for a philosopher: he is very clear, generous with examples, and even sometimes funny. The book is also well-organized.
Those who are daunted by the game theory can either get themselves a quick intro (as I did) or can muddle through. I think a lot of content, especially in the latter half of the book, can still be absorbed even without the formal backing. But I do think it better to wrestle with the game theory. There are some good youtube videos I found (Game Theory 101, which has a textbook as well) that give enough introduction to make Lewis's discussion comprehensible. Without that, it will be less clear *why* convention should be thought of in the way Lewis does-- for Lewis, a convention is essentially a Nash equilibrium, a stable strategy in a game, with additional details.
For those who enjoy that, and like (or are at least not turned off by) the mathematical side, Skyrm's "Signals" is a nice follow-up book that discusses how signalling systems evolve.
For those familiar with game theory, I imagine this will be a pleasant read. Lewis has a nice writing style, especially for a philosopher: he is very clear, generous with examples, and even sometimes funny. The book is also well-organized.
Those who are daunted by the game theory can either get themselves a quick intro (as I did) or can muddle through. I think a lot of content, especially in the latter half of the book, can still be absorbed even without the formal backing. But I do think it better to wrestle with the game theory. There are some good youtube videos I found (Game Theory 101, which has a textbook as well) that give enough introduction to make Lewis's discussion comprehensible. Without that, it will be less clear *why* convention should be thought of in the way Lewis does-- for Lewis, a convention is essentially a Nash equilibrium, a stable strategy in a game, with additional details.
For those who enjoy that, and like (or are at least not turned off by) the mathematical side, Skyrm's "Signals" is a nice follow-up book that discusses how signalling systems evolve.
November 15, 2025
Mein Kaufbeleg vom 24. September 1997 ist der Beweis für ein intellektuelles Schnäppchen: Nur 12 DM für David Lewis’ „Convention“. Das ist Wühltisch-Preis für ein Buch, das im Grunde die gesamte Gesellschaft als spieltheoretisches Koordinationsproblem erklärt – etwa warum wir alle auf derselben Straßenseite fahren, obwohl es nirgendwo ausdrücklich steht. Aber der eigentliche Ritterschlag – oder die ultimative Einschüchterung, je nach Tagesform – lauert im Vorwort.
Dieses stammt von niemand Geringerem als W. V. Quine. Ein Vorwort von Quine in einem sprachphilosophischen Werk zu haben, ist, als würde der Papst persönlich die Einleitung zu deinem Gebetbuch schreiben. Man öffnet das Buch mit einer gewissen Ehrfurcht, denn man weiß: Wenn Quine, diese einschüchternde Autorität, die über das Sein oder Nichtsein ganzer Entitäten per „ontological commitment“ entschied, ein Buch für gut befand – wer war man dann, mit läppischen 12 DM in der Hand, um noch Widerspruch zu wagen? Das war keine Einleitung. Das war eine akademische Absolution.
Dieses stammt von niemand Geringerem als W. V. Quine. Ein Vorwort von Quine in einem sprachphilosophischen Werk zu haben, ist, als würde der Papst persönlich die Einleitung zu deinem Gebetbuch schreiben. Man öffnet das Buch mit einer gewissen Ehrfurcht, denn man weiß: Wenn Quine, diese einschüchternde Autorität, die über das Sein oder Nichtsein ganzer Entitäten per „ontological commitment“ entschied, ein Buch für gut befand – wer war man dann, mit läppischen 12 DM in der Hand, um noch Widerspruch zu wagen? Das war keine Einleitung. Das war eine akademische Absolution.
November 10, 2022
Very interesting treatment of conventions. The book uses game theory to define conventions as solutions to coordination games which include expectations and higher order expectations as part of the equilibrium. I somehow skimmed the sections on applications of conventions to language (which I suppose was the main purpose of the book, but I am not so much interested in linguistics...)
December 9, 2020
在这本书中，路易斯用博弈论考察了我们通常认为是约定俗成的种种情况，并用形式化很强，准确度很高的语言详细的讨论了约定俗成的定义。最后又讨论语言的本性，指出语言确实是约定俗成的。
February 14, 2013
The title is Convention: A Philosophical Study. (It hasn't got anything directly to do with art and performance). The book discusses in detail and at length many areas of philosophy, ultimately coming to a well-argued conclusion about knowledge formation and convention. I loved this book. Great ideas and beautifully written.
Displaying 1 - 10 of 10 reviews