Gauthier's title is potentially misleading. The phrase “morals by agreement” suggests a social contract theory of morality according to which basic moral principles arise out of an actual or hypothetical agreement. John Rawls defends a hypothetical agreement version, arguing that the basic principles of justice are those that would be agreed to in an initial position of fair equality. I myself defend an actual agreement version, arguing that the moral principles that apply to a person derive from implicit conventions the person has accepted in dealing with other people. Gauthier's view is different from either of these sorts of contract theory. Instead, he holds that certain basic principles of impartiality are prior to actual agreements.

Gauthier's magnificent book erects a conception of morality, “morals by agreement,” on the foundation of his own theory of practical rationality. This is as it should be if, as he claims, following Hobbes and others, there is an initial “presumption against morality” (MA, p. 13) and no theory of morals “can ever serve any useful purpose, unless it can show that all the duties it recommends are also truly endorsed in each individual's reason” (MA, p.1), indeed, that it is a requirement of rationality that one always satisfy the requirements of morality (MA, p.5). This means, however, that the initial assumption against morality is inherited by his theory of practical rationality. His theory of morals therefore can serve a useful purpose only if his theory of rationality is sound. In this paper, I want to explore some of the more dubious aspects of that theory to see whether it can bear the heavy load of justification that “morals by agreement” places on it.

Part I of this essay will be devoted to Gauthier's principle of minimax relative concession. Part II will focus, more generally, on the variety of possible strategies available to liberal theory. In Part I, I will argue that the principle of minimax relative concession does not define “essential justice” as Gauthier claims. In Part II, I will argue that the difficulties facing Gauthier's strategy are common to other strategies of die same general kind. I will close by suggesting what I think may prove to be a more promising approach.

David Gauthier's Morals by Agreement presents a partial theory of distributive justice. It is partial because it applies only to the distribution of gains from joint endeavors, or what we may call the ‘social surplus’ from cooperation. This surplus is the benefit we receive from cooperation insofar as this is greater than what we might have produced through individual efforts without interaction with others. The central core of Gauthier's theory of distributive justice is his bargaining theory of ‘minimax relative concession’ or MRC. Whether his theory is compelling turns essentially on whether MRC is workable and compelling. It is this issue that I wish to address.

I take it as my assignment to criticize the Gauthier enterprise. At the outset, however, I should express my general agreement with David Gauthier's normative vision of a liberal social order, including the place that individual principles of morality hold in such an order. Whether the enterprise is, ultimately, judged to have succeeded or to have failed depends on the standards applied. Considered as a coherent grounding of such a social order in the rational choice behavior of persons, the enterprise fails. Considered as an extended argument implying that persons should (and possibly must) adopt the moral stance embodied in the Gauthier structure, the enterprise is, I dunk, largely successful. Considered as a set of empirically falsifiable propositions suggesting that persons do, indeed, choose as the Gauthier precepts dictate, the enterprise offers Humean hope rather than Hobbesian despair.

In Morals By Agreement, David Gauthier concludes that under certain conditions it is rational for an agent to be disposed to choose in accordance with a fair cooperative scheme rather than to choose the course of action that maximizes his utility. This is only one of a number of important claims advanced in that book. In particular, he also propounds a distinctive view concerning what counts as a fair cooperative arrangement. The thesis concerning the rationality of adopting a cooperative disposition is, however, logically independent of his substantive view of a fair cooperative scheme and is itself central to the project as a whole. Gauthier's concern is to establish that certain moral principles are those that fully rational, self-interested persons would agree to take as regulative of their dealings with one another – that a contractarian approach, in this sense, can provide an adequate basis for a theory of morality.

One of the most noteworthy features of David Gauthier's rational choice, contractarian theory of morality is its appeal to self-interested rationality. This appeal, however, will undoubtedly be the source of much controversy and criticism. For while self-interestedness is characteristic of much human behavior, it is not characteristic of all such behavior, much less of that which is most admirable. Yet contractarian ethics appears to assume that humans are entirely self-interested. It is not usually thought a virtue of a theory that its assumptions are literally false. What may be said on behalf of the contractarian?

There is a way of doing moral philosophy which goes something like this: If it can be shown that it is rational for perfectly selfish people to accept the constraints of morality, then it will follow, a fortiori, that it is rational for people capable of affective bonds, and thus less selfish, to do so. On this way of proceeding the real argument – that is, the argument for the actual constraints (theory or principles) to be adopted – proceeds with only fully rational individuals who have no other concern than to maximize their nontuistic (selfish) preferences. Then it is noted that the affective capacities of human beings actually make quite palatable the constraints that the fully rational persons with wholly nontuistic preferences have agreed upon.

(1) In his recent paper, “Justice as Fairness: Political not Metaphysical,” John Rawls makes use of a footnote to disown what to many readers must have seemed one of the most striking and original underlying ideas of his theory of justice, that it “is a part, perhaps the most significant part, of the theory of rational choice.” That Rawls should issue this disclaimer indicates, at least in my view, that he has a much clearer understanding of his theory, and its relationship to rational choice than he did at the time that he wrote A Theory of Justice. As I note in Morals by Agreement (pp.4–5), Rawls does not show that principles of justice are principles of rational choice. Hence, in appropriating the idea, I can claim diat I am undertaking a pioneering enterprise. No doubt Thomas Hobbes would have undertaken it had the resources of the theory of rational choice been at his disposal, but I do not intend to pursue counterfactuals in a search for historical antecedents. Moral theory as rational choice theory is, I claim, a new venture.