David Gauthier, a contemporary American philosopher, argues in Morals by Agreement that all moral norms and principles must be acceptable to all rational persons. He supports the received view that rational persons attempt to maximize their self-interest, and so moral constraints, to be proper, must be in the self-interest of all people. Gauthier argues that it is rational to accept constraints when such constraints make all better off, as in the Prisoners' Dilemma . In the Prisoners’ Dilemma (which is well studied by game theorists), it is apparent that acting in one’s immediate self-interest sometimes leads to bad results. That is, if everyone acted without regard to the interests of others, all might lose, while if people consider the good of others as well as their own good, all might be better off. Assuming that many such situations exist, and that each can be solved with a moral norm that constrains action, Gauthier recommends a basic morality based on rational self-interest.   Top

Problem of the free-rider

    But the problem of the free-rider intrudes; a free rider takes advantage of the fact that others do the right thing. (Think of the person who goes to the head of a line; unless others followed the rule, that person could not gain such an easy advantage.) Why not fake compliance with a moral norm and take advantage of others? Gauthier insists that moral norms are rational, but that, to be effective, they must constrain everyont's self-interest. That is the problem: to make even a clever deceiver see that following moral constraints is in his or her self-interest. For norms to work, each person must make a prior commitment. Gauthier believes that such a commitment can be rational because a deceiver is often recognized, and punished, for what he or she is. When recognized, that person will not be able to participate in rational gains. So Gauthier's solution is that people are translucent -- that is, we can often see through their behavior, and understand that they are not being genuine.
    This solution is not adequate. A moral constraint is only a moral constraint when it is not in a person's interest. If we can see through people, then they are more or less forced to act morally. These actions are not based on a moral constraint but on a social constraint. A social constraint may be one that we can morally endorse; even the faker may want to endorse social constraints. They may be designed to detect cheats, or to prosecute them. This is not what Gauthier wishes; he wants an individuals to be committed to moral rule or norm, one that all rational persons may be able to endorse as morally proper. Gauthier wants to show that it is rational for each individual to be constrained by moral norms. This argument fails because some individuals may believe that it is more rational to attempt to fake compliance. When an individual truly believes that he or she will be detected, then the constraint is not a moral constraint, but a prudential constraint.    Top

Deriving what ought to be from what is

    Gauthier's account is an attempt to use a nonmoral value, acting rationally, as a way to defend moral values. Such attempts generally fall under the rubric of deriving an ought from what is. Gauthier attempts to derive moral values from nonmoral basic individual interest expressed as individual rationality. Such efforts run into a problem similar to that raised by G.E. Moore, a British philosopher writing after the turn of the twentieth century. Whenever a nonmoral value is offered as a defining feature of moral value, Moore tests the claim by examining whether it makes sense to ask whether the nonmoral value is really good. If the nonmoral value defines goodness, then the question should make little sense. But if our sense of goodness goes beyond the non-moral value, then the question makes good sense. (More about Moore's test.) We can indeed evaluate non-moral values in moral terms. When we find ourselves evaluating the proposed basic value, we then understand that the proposed definition does not provide the meaning of moral terms.
    Gauthier does not attempt to define moral value in terms of rationality; instead, he seeks to derive moral values from his conception of rationality. However, we may still loosely apply Moore's test. Is this value, rationality, really a basic value from which moral values may be derived? This question makes perfect sense, and calls for an answer. Once we give an answer, we tend to use other values, including non-moral values, to defend the status of the basic value. In this process, those other values gain a standing equal to that of the basic value, or perhaps even a higher standing.   Top

Preferences and moral values

    Basic nonmoral values like rationality and pleasure are controversial. We can easily think of pleasures that appear to be evil. Rationality is more complex because it depends on preference fulfillment. Suppose a person is brainwashed into accepting a preference ranking that is coercively imposed from outside. Gauthier understands that some preferences are not morally acceptable. So he builds into his conception of rationality some standards by which preferences may be evaluated. Preferences must be autonomously formed, consistent, stable, and carefully considered. This complicates the analysis by introducing, within the basic nonmoral value, unevaluated and undefended moral values, such as autonomy or controlling one's own destiny. Even after introducing these values, we may still question whether the individual rationality can adequately serve as the basis for moral value, and this question continues to make good sense. The answer will lead to further use of values, moral and otherwise.
Philosophers frequently point out that the distinction between moral values and nonmoral values is arbitrary. Gauthier's conception of preferences is a case in point. If a brainwashed person says, "I prefer to give all my money to a cult leader," we may say that the person doesn't really prefer that. This is not the person's genuine preference. A moral value, that preferences be independently formed, has been added to the supposedly non-moral status of a preference.
    Our analysis shows that norms or moral rules cannot be readily derived as morally binding from any basic non-moral values. Some norms are widely thought to be morally binding, and can be supported or rejected in terms of other moral values. The hope of finding some nonmoral touchstone from which to determine moral value has so far proved ineffective.     Top