Andrei Marmor - On Convention
Andrei Marmor - On Convention
Andrei Marmor - On Convention
ON CONVENTION 1
having a given convention in the first place, and what are the reasons for
complying with a convention which is already practised. I will try to show
that the answers to these two questions are not always the same. First,
however, some other issues ought to be addressed. As it is the arbitrariness
of conventions which renders them problematic, I will begin by offering a
definition of the arbitrariness of conventions, exploring some of the impli-
cations that follow from it. Next I consider Margaret Gilbert's alternative
analysis, arguing that it fails to account for the normativity of conventions
in some crucial ways. As between these two opposed attitudes, I will argue,
Lewis seems to have got it right, but his analysis only applies to one type
of conventions. There is, I will argue, another type of conventions not
explicable on the grounds offered by Lewis.
1. ARBITRARINESS DEFINED
more important and fundamental than the reason for uniform conformity.
People who believe that one is morally required to show respect for all
human beings as such, for example, would believe that they have a sufficient
reason to follow this rule even in a society where no such respect is actually
practised.
Gilbert raised some doubts about this account, relying on the following
counter example: suppose that it has been a convention in a given commu-
nity that people should send thank-you notes after being invited to dinner
parties. As it happens, conformity with this convention has dwindled and
presently most people in this community no longer abide by the conven-
tion. Gilbert suggested that this does not mean that there is no longer a
convention in this community to send thank-you notes after being invited
to dinner. The lack of general conformity, she claims, does not necessarily
affect the appropriate characterization of this rule as a convention or not. t9
Now, I suspect that one would tend to reply that what we face here is pre-
cisely what is described: a convention which has dwindled. But the point
is more substantial that this, and for the following reason: sending thank-
you notes after dinner invitations is one alternative way, amongst many
others, of expressing gratitude for the invitation, whether a convention to
that effect exists or not. In other words, expressing gratitude for the invi-
tation is in itself valuable, regardless of conventions, and any reasonable
means of doing that would normally serve the purpose. The existence of
the convention can render it the appropriate way of expressing gratitude
in a certain community, only in one special manner. Namely, in that this is
the way people actually follow and hence expect others to follow as well.
Their expectation makes sense only as conditional on the actual conformi-
ty of others. There would be no point whatsoever in expecting people to
comply with one particular alternative which is, after all, arbitrary, unless
this alternative is actually followed by the community at large.
Once again, such a structure of reasoning is markedly opposed to the
way we think about morality: one can be criticized for not complying
with a moral rule, even if most others in his community fail to comply
as well. One cannot be criticized, however, for failing to comply with a
convention which is not actually practised. People can believe, of course,
that a specific uniformity of action is desirable, even if it is not practised.
Belief in the reasons for the desirability of having a convention in certain
circumstances, however, does not constitute a reason for complying with
it. In other words, 'It ought to be that it ought to be that P ' normally entails
that 'It ought to be that P ' . Whereas, 'It ought to be that it is a convention
that P ' does not entail that 'It is a convention that P ' .
35 8 ANDREIMARMOR
One may wonder, however, whether it is not the case even with respect
to some fundamental moral rules, that their efficacy affects their validity.
The obligation to keep one's promises is an example that immediately
comes to mind here. Suppose we could imagine a certain community where
keeping promises is not practised: expressions of a promise don't engender
an expectation that it may be kept, and promises are not held to impose
obligations on the promisor. Would we say that if promises are not sustained
by practice, the obligation to keep one's promises is not a valid rule? And
if so, would it show that efficacy affects the validity of fundamental moral
rules as well? Or would we rather say that the obligation derives from a
moral rule which is valid, and that in our imaginary community there is
simply no opportunity for complying with it?
The answer would seem to depend on one's views on the nature of
promissory obligations. According to one line of thought, which is some-
times called The Practice Theory, the validity of promissory obligations
derives from a combination of the social practice of promising and a gen-
eral moral duty to uphold valuable practices. On this view, the analysis of
promissory obligations is a two-stage affair. First, there must be a social
practice, which consists in the rules people actually follow, and the expec-
tations, intentions, etc., they associate with those rules. Second, there is a
moral judgement that the practice is valuable, and hence, that there is a
general moral duty (or, perhaps a more specific obligation of Fair Play, as
Rawls maintained,) to sustain and uphold it. According to a different line of
thought, promissory obligations are instances of a more general moral duty
which we owe to other people when we have led them to form expectations
about our future conduct in certain circumstances.2 On this Expectation
Theory of promises, the role of the conventional practice of promising is
much more limited; it consists in providing us with certain conventional
means of expressing a certain type of intentions and creating expectations
about our future conduct. Hence according to the Expectation Theory, in
our imaginary community we would simply lack many opportunities to
undertake promissory obligations in conventionally established ways.
Now, it is true that according to the Practice Theory of promises, the
efficacy of the conventions which constitute the practice would be taken
to affect the validity of promissory obligations. As long as the practice is
largely followed, every one has a reason to comply as well. And if there is
no practice, then there is nothing One would have reason to sustain. But this
only shows that according to the Practice Theory, keeping one's promise
is basically an instance of complying with a convention. Proponents of
the Practice Theory, faced with the imaginary community where promises
are not sustained by practice, would not wish to say that their basic moral
0NCONVENTION 359
2. GILBERT ON CONVENTION
feature of conventions is clearly not supported by the facts. Why does she
repeatedly affirm the contrary? It is, after all, quite unlikely that Gilbert
has made the simple mistake of generalizing from some cases to all. It is
more likely that her thesis is related to a more fundamental aspect of her
analysis, and it is not difficult to see what it is.
As I have already mentioned, Gilbert has failed to realize that arbitrari-
ness is a conceptually defining feature of conventions, due to her miscon-
strual of what arbitrariness consists in. Having presumed arbitrariness to
mean something like indifference, she has been content to prove that con-
ventions are not arbitrary in this sense. But in this she has already depleted
the concept of a convention of its most important philosophical aspects. 23
From a normative point of view, it is the arbitrary feature of conventions
(properly defined, and not in terms of indifference), that renders social
conventions a peculiar type of rules, calling for an explanation of the kinds
of reason people could have for following a convention. Whether these
reasons consist only in the solution of recurrent co-ordination problems,
as Lewis maintains, or in other possible types of reason as well, remains
to be seen. But unless this normative puzzle is appreciated as such, there is
no hope of accounting for the normativity of conventions. Gilbert's failure
to realize this has led her to an impoverished account of the normativity
of conventions - resting on such notions as 'group fiat' and 'acceptance'
- and hence to her over-emphasis of the collectivity aspect, mentioned
earlier. But these sociological concepts contribute very little, if anything
at all, to an explication of the normativity of conventions, and for several
reasons.
First, it is important to note that Gilbert's account of conventions does
not make any allowance for a distinction between following a convention
and following a generally accepted reason. But it is essential to keep
these two apart. In other words, "a jointly accepted principle of action" is
not necessarily a convention, whether it is, or isn't, essentially collectivity
involving. Suppose it is generally accepted in my college that we should do
certain things to promote women's opportunities in pursuing an academic
career. Suppose that this is a principle we jointly accept; we all accept,
for example, that the college should provide child care facilities, that it
should provide more flexible working hours for women, etc. It is our
college fiat, as it were, to follow these principles of action, and we believe
that this principle entails certain obligations with respect to our actions
in certain circumstances. On Gilbert's account, we would be described as
following a convention here. But this is surely wrong. Our jointly accepted
commitment to enhancing women's opportunities in college would be ill
described as a convention. It is simply a principle of action we have come
362 ANDP,EI MARMOR
Why, and under what circumstances, does it make sense to follow a rule
which is arbitrary? David Lewis has provided an illuminating answer:
the rationale of following a convention, he claimed, consists in the fact
that conventions are solutions to recurrent co-ordination problems. A co-
ordination problem arises, according to Lewis, 29 when several agents try
to achieve uniformity of action by each doing whatever the others will do.
Thus within a given set of options, each and every agent has a stronger
preference for acting as the other agents will, than his preference of any one
364 ANDREIMARMOR
constitutive of values and reasons for action, but this is a fact we have to
account for and it should not be obscured by philosophers' preoccupation
with co-ordination conventions.
4. SUMMARY
NOTES
I I am indebted to Timothy Williamson and Joseph Raz with whom I have discussedthese
matters at length, and I am grateful for their invaluable comments on a draft of this paper.
I am also indebted to Brian Bix, Ruth Gavison, Alon Harel, Edna Ullmann-Margalit, and
the editors of Synthese for their helpful comments and suggestions.
2 Conventions ought to be distinguished here from rules which are designed to solve
'prisoner's dilemma' conflicts. In the latter case too, but for different reasons, efficacyis
ONCONVENTION 369
essential in rendering the norm valid. Throughout this article, however, I shall assume that
conventions and PD norms are clearly distinguishable, and I shall not be concerned with the
latter. For a detailed account of the distinction see Ullmann-Margalit(1977, pp. 114-121).
3 See Lewis (1969, p. 78). A revised version &his definition was later published in Lewis
(1975).
4 See, for example, Grandy (1974), Postema (1982), Schiffer (1972, p. 152), and Ullmann-
Margalit (1977, p. 74ff).
5 Similar suspicions, and for very similar reasons, I share with respect to the preoccupation of
writers on the topic with game-theoretical analyses. The predominance of game-theoretical
analysis of conventions has not only engendered these idle definitional projects, as it tended
to over-emphasize some aspects of the topic, on the expense of many others. In any case, I
shall not be concerned here with game theories at all.
6 As Wittgenstein famously has shown, not all the concepts we use have a defining feature.
'Family resemblance' concepts "have no one thing in common which makes us use the
same word for all" its applications (Wittgenstein 1953, section 65).
7 Wittgenstein (1967, section 144).
s The relevant part of Lewis' definition reads as follows: "There is at least one alternative
R' such that the belief that the others conformed to R' would give everyone a good and
decisive practical or epistemic reason to conform to R' likewise; such that there is a general
preference for general conformity to R' rather than slightly-less-than-general conformity
to R'; and such that there is normally no way of conforming to R and R' both'. And, the
condition above, among others, is a 'matter of common (or mutual) knowledge..." Lewis
(1975, pp. 5-6).
9 I do not wish to deny that conventions can provide reasons for belief. On the contrary, I
believe that with suitable modifications, the account which follows should be applicable to
conventional belief systems as well.
10 Lewis (1969, pp. 104-105).
l J Lewis' failure to realize this point has been the cause of some confusion in the critical
literature that ensued. For example, his repeated emphasis on regularities of behaviour in
his definition of convention, gave rise to the criticism that conventions don't necessarily
involve situations which occur with frequency. See Gilbert (1989, pp. 344-345). But upon
noticing the fact that conventions are necessarily rules, the criticism is rendered spurious.
The idea of following a rule does not involve any requirement of frequency. For a further
account of the distinction between rules and regularities of behaviour, see Marmor (1992,
pp. 41-42).
12 See Burge (1975). Notably, Lewis himself was not unaware of the problem raised by
Burge, but seems to have ignored its implications. See Lewis (1969, p. 75).
13 Burge (1975, p. 250).
14 The opposite, of course, is not the case; it is not the case that each and every anti-realist
stance with respect to a given class of statements entails conventionalism with respect
to that class of statements. One can deny the possibility of realism with respect to the
characterization of morality, for example, without being committed to the view that moral
prescriptions are conventions. In other words, it makes perfect sense to hold the following
two theses concomitantly: that fundamental moral rules are not conventions; and that there
is no such a thing as an 'objective' moral reality. Those who maintain that moral truth
is always discoverable by reason, need not necessarily choose between endorsing either
realism or conventionalism; they can reject them both.
is For a discussion of some of the complexities involved here, see Marmor (1992, pp.
86-97).
370 ANDREI MARMOR
pp. 152ff); Jamieson (1975); and Burge (1975), (whose criticism, though, seems to be more
limited).
32 Schiffer (1972, p. 152).
33 In fact, he did claim something like this in a similar context, see Lewis (1975, pp. 26-27).
34 On the absurdities of Platonism in such a context, see Raz (1991, p. 91).
REFERENCES
Burge, T.: 1975, 'On Knowledge and Convention', The Philosophical Review 84, 249--55.
Gilbert, M.: 1989, On Social Facts, Routledge, London.
Gilbert, M.: 1990, 'Rationality, Coordination, and Convention', Synthese 84, 1-21.
Grandy, R.: 1974, 'Convention, a Philosophical Study - Book Review', Journal of Philos-
ophy 74, 129-39.
Jamieson, D.: 1975, 'David Lewis On Convention', Canadian Journal of Philosophy 5,
73-81.
Lewis, D.: 1969, Convention, A Philosophical Study, Blackwell, Oxford.
Lewis, D.: 1975, 'Languages and Language', Minnesota Studies in the Philosophy of
Science 7, 3-35.
Marmot, A.: 1992, Interpretation andLegal Theory, Oxford University Press, Oxford.
Miller, S.: 1990, 'Rationalising Conventions', Synthese 84, 23-41.
Postema, G.: 1982, 'Coordination and Convention at the Foundations of Law', The Journal
of Legal Studies 9, 165-203.
Rawls, J.: 1971, A Theory of Justice, Oxford University Press, Oxford.
Raz, J.: 1975, Practical Reasons and Norms, Hutchinson, London.
Raz, J.: 1977, 'Promises and Obligations', in P. Hacker and J. Raz (eds.), Law Morality
and Society, Essays in the Honour of H.L.A. Hart, Oxford University Press, Oxford, pp.
210-228.
Raz, J.: 1991, 'Mixing Values', Aristotelian Society sup. 83-100.
Scanlon, T.: 1990, 'Promises and Practices', Philosophy and Public Affairs 19, 199-226.
Schiffer, S.: 1972, Meaning, Oxford University Press, Oxford.
Ullmann-Margalit, E.: 1977, The Emergence of Norms, Oxford University Press, Oxford.
Wittgenstein, L.: t 953, Philosophicallnvestigations, (G.E.M. Anscombe trans.), Blackwell,
Oxford.
Wittgenstein, L.: 1967, Zettle, (Anscombe & Von Wright trans.), Blackwell, Oxford.
Faculty of Law
Tel-Aviv University
Remat Aviv
Tel Aviv 69978
Israel