¹ Cf. Prichard, H. A., Moral Obligation and Duty and Interest: Essays and Lectures, Oxford, 1968, ch. 7, p. 198Google Scholar: ‘Once call some act a promise and all question of whether there is an obligation to do it seems to have vanished.’ Prichard here focuses on promises. It seems he might equally well have said: ‘Once say that people have made an agreement and all question of whether they are obligated to conform to it seems to have vanished.’

² Hart, H. L. A., ‘Are There Any Natural Rights?’, Philosophical Review, lxiv (1955), 179, n. 7Google Scholar. The following quotation in the text below is from the same passage.

³ Later Hart amplifies point (3), referring to ‘special transactions between individuals or … some special relationship in which they stand to each other’ (183), and to ‘previous transactions and relations between individuals’ (190) (my emphasis in both quotations).

⁴ See Rawls, John, A Theory of Justice, Cambridge, Mass., 1971Google Scholar; Simmons, A. John, Moral Principles and Political Obligations, Princeton, 1979Google Scholar; Klosko, George, The Principle of Fairness and Political Obligation, Lanham, Md., 1992Google Scholar.

⁵ Cf. Brandt, R. B., ‘The Concepts of Obligation and Duty’, Mind, lxxiii (1965)Google Scholar.

⁶ Hart, 183 ff. He focuses on the ‘special rights’ correlative with obligations.

⁷ Hart, 183.

⁸ He may, along with many others, assume that an agreement is simply an exchange of promises. I dispute this assumption in Gilbert, , ‘Is an Agreement an Exchange of Promises?’ Journal of Philosophy, xc (1993)Google Scholar, repr. with revisions as ch. 13 of Gilbert, Living Together: Rationality, Sociality, and Obligation, Lanham, Md., 1996. See also the text, below.

⁹ I first discussed joint commitment in Gilbert, , On Social Facts, London, 1989 (second printing Princeton, 1992)Google Scholar. See the introduction to Gilbert, Living Together, for a comparison of my exposition there with subsequent elaborations.

¹⁰ See Gilbert, On Social Facts. I there argue at length that the concept of a joint commitment is involved in our everyday concepts of group languages, social groups, social conventions, group beliefs and other central social phenomena.

¹¹ For related discussion see Gilbert, , ‘What is it for Us to Intend?’, Contemporary Action Theory, vol. II, ed. Holmstrom-Hintikka, G. and Tuomela, R., Dordrecht, 1997, 79 fGoogle Scholar.

¹² Many fine-grained accounts of common knowledge have been produced. The first published discussion is that in Lewis, David, Convention: A Philosophic Study, Cambridge, Mass., 1969Google Scholar. Some of these accounts have been criticized for lack of realism. For an attempt to produce an account such that common knowledge is not ‘too high for humanity’, and some further references, see Gilbert, On Social Facts.

¹³ The parties to a joint commitment can always initiate certain side-understandings. Suppose Anne says, ‘Shall we meet for lunch on Tuesday?’; Ben replies, ‘Yes …’, somewhat reluctantly. Anne may then offer: ‘Don't worry; you can always call me in advance and say you've changed your mind.’ One might describe the situation here as follows. Anne and Ben have a plan to meet, that is, they are jointly committed to accepting a certain plan as a body. They also have a special side-understanding to the effect that they will both act as if Anne is not in a position to hold Ben to the commitment – provided he calls her first.

¹⁴ I have discussed the case of belief in a number of places, beginning with Gilbert, , ‘Modelling Collective Belief, Synthese, lxxiii (1987)Google Scholar, repr. in Gilbert, Living Together. The Introduction to Living Together contains an overview of this material. On goals see Gilbert, On Social Facts, ch. 4, and Gilbert, , ‘Walking Together: A Paradigmatic Social Phenomenon’, Midwest Studies in Philosophy, xvGoogle Scholar, The Philosophy of The Human Sciences, ed. French, P. A., Uehling, T. E., and Wettstein, H. K., Notre Dame, 1990Google Scholar, repr. in Living Together.

¹⁵ I have previously discussed this issue most extensively in Gilbert, , ‘Agreement, Coercion, and Obligation’, Ethics, ciii (1993), sect. 3Google Scholar. The discussion here is significantly different. There I start by proposing a partial characterization of obligation in general, a characterization derived from reflections on the broad sweep of everyday usage: if I have an obligation to do A, I have a reason to do A that is not the ‘creature of my own will’ in the sense that an arbitrary act of my own will was sufficient to create it and is sufficient to destroy it. I note that, in contrast to personal decisions or other forms of personal commitment, joint commitments give reasons that satisfy the proposed necessary condition, reasons which it seems otherwise appropriate to speak of as ‘obligations’ (our term ‘obligation’ comes from the Latin word ‘ligare’, to bind; failure to act on joint commitment-based reasons involves breaking what is saliently a bond). Here I do not attempt any initial characterization of obligations in general. My main aim in both discussions is to make plausible the idea that in some important contexts our talk of ‘obligations’ (and ‘rights’) may be responsive to the understanding that, simply, a joint commitment is present.

¹⁶ ‘Betrayal’ may seem too strong a term for one's sense of things in some such cases. Is there not a difference between my failing to accord with a decision when this involves my acting against deeply held moral beliefs, say, and a case where nothing like this is at stake? Certainly there is a difference. Perhaps, though, there can be small betrayals. One is unlikely to react grievously to them, but one may understand them to have occurred. The suggestion in the text assumes this.

¹⁷ See, for instance, Kenny, Anthony, Action, Emotion, and Will, London, 1963, p. 218Google Scholar: ‘on our theory, to have the intention to bring it about that p, is to say-in-one's-heart “Let it be the case that p”’; and p. 220: ‘… a command uttered to oneself’.

¹⁸ This way of representing the order involved in a personal commitment, explicitly referring to the addressee, allows one to differentiate between the orders involved in personal commitments and those involved in joint commitments. In this it contrasts with Kenny's representation in Kenny, p. 218 (see previous note).

¹⁹ Cf. Kenny, n. 17 above.

²⁰ See n. 21 below on what may be a necessary condition for orders that are not self-addressed.

²¹ Self-addressed orders are clearly special in so far as the orders by reference to which people learn the term ‘order’ are addressed by one person to another (often within some background institution). It can be argued that if such an ‘other-addressed’ order is to be more than a merely purported order, there must be a joint commitment that binds the issuer and the addressee, to the effect that the addressee is to do what the issuer says.

²² It may seem that nothing could count as disobedience to a self-addressed order. Would one's apparently failing to conform to one's order not show that one has in effect rescinded it? The examples in the text above of how one might fail to conform to one's decision suggest not. For instance, one might attempt to conform to one's order but fail. Thus I might order myself to leave the house at 6 p.m. but, without so intending, not leave myself enough time to make the necessary preparations. As I scurry around trying to meet my deadline, I have not yet rescinded the order. Eventually I have to admit I have failed in my attempt; I have disobeyed my order. This note was prompted by a comment from Roger Crisp.

²³ Appropriate abstentions may be included under ‘actions’ here. In so far as the parties are always jointly committed to do something as a body, it appears that something will always be owing from both.

²⁴ So also might have concluded the betrayal argument and the argument from the jointness of the commitment as these have so far been represented here. If Anne would indeed betray Ben by violating the joint commitment, she must owe him conformity to it. Similarly, if Anne's violation of the commitment that is theirs does indeed offend against Ben, she owes him her conformity. The joint order argument may make the protasis of these conditionals more persuasive. In any case all of these arguments point to the one conclusion.

²⁵ One might also speak of the ‘joint rescission’ feature, ‘rescission’ being the corresponding term in contract law.

²⁶ ‘… if he has not concurred.’ The case is not quite so simple, but for present purposes the point can stand. An important question is the outcome of one party's default. For some remarks on this see Gilbert, , Living Together, pp. 13–15CrossRefGoogle Scholar. See also the text below, on the interdependence of the obligations of joint commitment.

²⁷ Rights and obligations of joint commitment are of course in some sense waiveable, as in the example where Anne says to Ben, ‘You can always call me in advance and say you've changed your mind’. See n. 13 above. Such waivers occur, I take it, only in the context of special procedures and understandings.

²⁸ Often people say that they are sorry, but they can't do such-and-such after all. This, too, will not constitute the unilateral abrogation of a joint commitment, however precisely we should describe its significance. Quite often, of course, what is in fact at issue in such cases is preference rather than inability. It is understood that a contrary preference cannot ensure that one is let off the hook. (‘I really don't want to go.’ ‘But you said you would!’, etc.)

²⁹ For related discussion see Gilbert, ‘Rationality, Coordination, and Convention’, Synthese, lxxxiv (1990)Google Scholar, repr. in Gilbert, Living Together; and Gilbert ‘What is it for Us to Intend?’.

³⁰ In The Metaphysics of Morals, sect. II (‘On Contract Right’), subsect. 18, Kant writes of ‘an act of the united choice of two persons’. I am indebted to Christine Korsgaard for referring me to Kant's discussion in this section after hearing me speak about joint commitment and obligation. Much of what I say about joint commitment in this paper and elsewhere seems to accord quite well with views Kant expresses. His key term is ‘common will’ (also ‘united will’) rather than ‘joint commitment’. He argues compactly that the existence of a common will gives the parties rights of a special sort to each other's actions. He does not allude to the necessity of joint abrogation.

³¹ Why have the term ‘agreement’ if one has the term ‘decision’ and can tack on the pronominal adjective ‘our’? On this see Gilbert, ‘Agreements, Coercion, and Obligation’, sect. 3.

³² See Gilbert, , ‘Is an Agreement an Exchange of Promises?’, Journal of Philosophy, xc (1993)Google Scholar, repr. with some revisions in Gilbert, Living Together. See also Gilbert, ‘Agreements, Coercion, and Obligation’.

³³ For some references see Gilbert, ‘Is An Agreement an Exchange of Promises?’. Some philosophers refer simply to ‘mutual promises’. In discussion of the theory that agreements are ‘exchanges’ of promises I construe ‘exchange’ in a broad sense: Anne and Ben exchange promises if Anne makes a promise to Ben and Ben responds with a promise to Anne. Ben's promise may be explicitly responsive to Anne's as in ‘Since you have promised to …, I promise to …‘.

³⁴ This is not intended to imply that no other obligations are produced as well. But these are the obligations explicitly specified in the agreement.

³⁵ For a discussion of other candidate promise-exchanges see Gilbert, ‘Is An Agreement an Exchange of Promises?’.

³⁶ I shall not attempt a full discussion of this idea here. I hope to pursue it in a longer study. For some further discussion see Gilbert, ‘Is an Agreement an Exchange of Promises?’, sect. 9.

³⁷ Kant, The Metaphysics of Morals, sect. II, subsect. 20, makes a clear distinction between rights that stem from (or inhere in) contracts, and another kind of right – a moral right? – which, he says, has to do with ‘the choice of all united a priori’. Playing on this Kantian phrase, contract right (and rights and obligations of joint commitment) could be referred to in contrast as stemming from ‘the choice of some united a posteriori’. I discuss what may amount to the same contrast in Gilbert, ‘Agreements, Coercion, and Obligation’, without presupposing a Kantian perspective on the moral domain. See also the text below.

³⁸ Whether or not knowledge that one has an obligation of joint commitment is best characterized as moral knowledge, it is surely going to be knowledge of something morally significant. This may or may not be of some comfort to the philosophers in question. See Gilbert, , Living Together, ch. 12, p. 297Google Scholar.

³⁹ Hart, 185.

⁴⁰ Hart contrasts ‘special rights’, which have correlative obligations, with ‘general rights’, ‘which are asserted defensively, when some unjustified interference is anticipated’ (187). ‘“You have no right to stop him reading that book” refers to the reader's general right. “You have no right to stop him reading that book” denies that the person addressed has a special right to interfere though others may have’ (ibid, n. 13).

₄₁ Ibid.

⁴² This interpretation is born out at Hart, 186, ‘The social-contract theorists …’.

⁴³ For extended discussions relating to the notion of a joint enterprise see Gilbert, On Social Facts, ch. 4, ‘Walking Together’, and, more recently, ‘What is it for Us to Intend?’. For an account of social rules as joint commitment phenomena see Gilbert, , ‘Social Rules: Some Problems for Hart's Account, and An Alternative Proposal’, Law and Philosophy, forthcoming, 1999CrossRefGoogle Scholar.

⁴⁴ I shall not attempt here to discuss the claim that membership in a political society involves one or more joint commitment phenomena, but see Gilbert, On Social Facts, on social groups in general, and (in relation to political obligation), Gilbert, , ‘Group Membership and Political Obligation’, Monist, lxxvi (1993)Google Scholar, and ‘Reconsidering the “Actual Contract” Theory of Political Obligation’, Ethics, cix (1999)Google Scholar. One relevant issue is the question of large group joint commitments. In everyday life we envisage joint enterprises involving from two to millions of people (as when we speak of a country at war). Can joint commitments be formed in large populations in which, among other things, no given member is directly acquainted with more than a relatively few other members? On this see Gilbert, , ‘Reconsidering’, 243 fGoogle Scholar; also 249–53.

⁴⁵ Hart, 186.

⁴⁶ Hart, 185.

⁴⁷ Hart, 190 f.

⁴⁸ Simmons, p. 102.

⁴⁹ Hart, 191.

⁵⁰ Ibid.

⁵¹ Simmons, pp. 178 f.

⁵² Cf. Card, Claudia, ‘Gratitude and Obligation’, American Philosophical Quarterly, xxv (1988), 120Google Scholar: ‘The benefactor does not have a right to one's acting in accord with [responsibilities of gratitude] but only deserves it (or doesn't).’

⁵³ Hart, 186 f.

⁵⁴ Discussions of familial obligations need to be clear on whether they concern a biological matter or more of a social one. In speaking of a ‘natural’ relationship Hart appears to have the former in mind. Many joint commitments are likely to arise in the context of contacts between the members of social as opposed to biological families.

⁵⁵ For comparison of this type of (commonly) so-called ‘obligation’ with obligations of joint commitment, see Gilbert, , Living Together, ch. 12, esp. pp. 296–301Google Scholar.

⁵⁶ It is possible that, historically, talk of ‘obligations’ in English originally latched on to cases ofjoint commitment. See Brandt's discussion of the way in which use of the term ‘obligation’ has expanded.

⁵⁷ Versions of this paper have been read at the University of Connecticut, and at the inaugural meeting of the British Society for Ethical Theory, held at Lancaster University, July 1997. I am grateful for the comments I received on these occasions. I thank Donald Baxter, Roger Crisp, Onora O'Neill and John Troyer for reading and commenting on draft material, and Christine Korsgaard for relevant discussion. Responsibility for the ideas expressed here is mine alone.