20th WCP: Partial Contractarianism and Moral Motivation

I begin by making what I take to be a crucial distinction. This distinction separates two levels at which a contract theory may operate. At the first level the contractarian theory is directed at the question of moral motivation. That is, it takes the idea of agreement to be the source of motivation to be or become moral. The agreement thus serves to bring into the moral domain agents who, prior to the agreement, were not moral agents. At the second level the contractarian theory is directed at the question of the content and justification of our most general normative principles and values. That is, it takes the idea of agreement to be the source of both content and justification. For convenience I will describe a theory which is contractarian at both levels as complete, and a theory which is contractarian at only one level as partial.

The problem of moral motivation, when understood as a problem of enticing non-moral agents into the moral domain, is a specific problem only for a contractarian theory which is complete or which is partial at level one. A contractarianism which is partial at level two has no special obligations, qua contractarian theory, to answer the Why be moral? question. In other words, such a theory does not offer, and does not aim at offering, a contractarian answer to the Why be moral? question since it is not concerned with moral-non-moral distinction. The early Rawls (1971) and Gauthier (1975,1986) both offer complete theories, while the later Rawls (1980) and Thomas Scanlon (1982) offer theories which are partial at level two (I will drop the ‘at level two': this can be assumed unless I indicate otherwise). They set aside the question of original moral motivation and proceed with a contractarian account of the normative and justificatory features of morality.

Of course, partial theories do need to say something about moral motivation, but the question they address is quite different from the one the early Rawls and Gauthier address. Partial theories are not exempt from giving an account of moral agency and this account must, in the first place, include an explanation of how agency is structured such that individuals are motivated to act in accord with morality. More importantly, in the second place, they have to show how this account is compatible with a contractarian approach to the rest of morality. What partial theories do not do is offer a contractarian account of ‘moral baptism', an account of how pre-social, non-moral agents come to be motivated, through a process of agreement, to be, or be disposed to be, moral.

The most forceful considerations favouring a complete contractarian theory are offered by David Gauthier who argues that the problem of moral motivation is resolved by showing that it is rational to be moral. My aim in this presentation is to show that this argument fails.

For Gauthier, moral outcomes are utility maximising in conditions of scarcity, competition and where cooperation is likely to yield a surplus, and, importantly, where there exists a general disposition to cooperate. Cooperation to achieve moral outcomes because such cooperation is likely to be utility maximising, is one leg of Gauthier's notion of constrained maximisation. The second leg consists of an affective commitment to impartial principles of cooperation (moral principles) which guide the actions and decisions of constrained maximisers. Both legs are rationally vindicated by their utility maximising possibilities. Constrained maximisers do better for themselves than straightforward maximisers, and constrained maximisers who internalise the principles of cooperation, developing an affective commitment to them, do better than wavering constrained maximisers who are likely to jump at the first chance to take advantage of their fellow cooperators.

The credence this picture enjoys, from the perspective of rationality, lies in the instrumental value of constrained maximisation for a creature set on maximising long term personal benefit. From the perspective of morality though, it is precisely the instrumental character of the agent's commitment which threatens to defeat its claims to be an account of moral motivation. On Gauthier's view this threat is met by internalisation and sentimental commitment to the impartial principles of cooperation. This commitment is more than a mere inclination. It enjoys a place at the very centre of the agent's motivational structure.

It would seem that the task set by the Why be moral? question has been accomplished. We have a reason to be moral, that is to say, we can offer a justification which has a universal character, compelling anyone similarly situated. And, moreover, we are motivated to act in accord with morality by a deeply seated affective concern for the well-being of others, as expressed in the principles of cooperation for mutual advantage. It would seem that reason and morality are here blended into a simple, coherent picture in which reason serves morality and morality serves reason. But, I will argue, on closer examination an instability becomes apparent and, once identified, will prise apart the rational and the moral, and simultaneously separate justification and motivation. One of the merits of this argument, if it succeeds, is that it does not rely on the usual complaint against constrained maximisation, namely, that it is likely to be more rational to masquerade as a constrained maximiser than to become one. My argument concedes the possibility of constrained maximisation in all the senses Gauthier intends.

There are two aspects of commonsense morality which seem to escape Gauthier's reduction of the moral to the rational. These are: the apparent importance of the moral character of motive, and the apparent importance of an other-regarding concern. Gauthier attempts to accommodate this residue by attributing to his agents an ‘affective capacity' for morality, thereby discarding the ‘economic man' of the prisoners' dilemma and embracing the ‘liberal individual' (1986:330ff). Gauthier recognises, and indeed argues, that economic man, who in machine-like fashion chooses strictly in accord with the demands of utility maximisation, fails the test of moral agency.

‘Economic man lacks the capacity to be truly the just man. He understands the arguments for moral constraint, but he regards such constraint as an evil from which he would be free.' (1986:328)

There are three features of the economic man's existence which make the constraints a necessary ‘evil'—a lack of self-sufficiency, a lack of a dominating power, and the rewards of cooperation. The economic man is a reluctant participant in the game of moral cooperation, and if possible would escape the constraints morality imposes. Weakness and dependency in the face of competing others are obstacles to straightforward maximisation. If these obstacles could be overcome, economic man would have unlimited licence to exploit and benefit from others. Place the ring of Gyges on the finger of the most just person and she would return to her natural state—a maximiser of individual utility, blind to the constraints of justice which would instantly lose their claim on her reason.

Economic man is therefore the egoist forced to labour beneath the weight of an unwelcome moral order. The motives of economic man remain the motives of an exploiter, and the scope of his concerns does not stretch beyond the narrow limits of his present and future self. Any ‘moral' deeds performed are a sham, as Gauthier clearly recognises (1986:316):

Were the conception of economic man our last word about ourselves, morality would have a precarious future. For moral constraints can have no hold on those who see in them only instruments of domination.

It is here that the idea of an ‘affective capacity' for morality enters Gauthier's picture. It is used to fill out the account of human nature, so that it is a nature capable of both genuine moral motive and genuine other-regarding concerns. I will argue below that this move fails. However, before presenting this argument we should be clear about the place an affective capacity has in Gauthier's theory. To start with, the rationality of moral constraint depends entirely on considerations addressed to, and understandable by, economic man. An affective capacity for morality is not part of the why of moral constraint.

An affective capacity for morality ... introduces no constraints, but disposes one emotionally and motivationally to adhere to constraints previously and independently accepted. (1986:328)

The affective capacity thus has to do with how independently accepted rational constraints become fully moral actions. The justification of rational constraint is thus the intellectual predecessor of the motivation to act in a genuinely moral fashion. In this way economic man evolves into the liberal individual. But this is not to say that the economic man disappears from view. Gauthier writes (1986:329):

We may distinguish the pure non-tuist, whose affections cannot be engaged except by his own concerns, from the person who demands a non-tuistic rationale for constraint, but whose affections may then be engaged by the constraints so justified.

The rationale for morality is no different for the economic man than it is for the liberal individual. What distinguishes the two is the latter's feeling for morality. And, of course, this feeling is itself rationally justified.

It will be useful to distinguish constrained maximisation as practised by economic man, and constrained maximisation as practised by the liberal individual by signifying the former by CM^emand the latter by CM^li. First note two arguments which do not succeed against either type of constrained maximisation. It might be claimed that an agent would have done better by remaining a straightforward maximiser. While this may be true, even if it is, this would not affect the rationality of choosing constrained maximisation. It is the expected outcome which matters when judging the rationality of choices, and if the world contrives to upset rationally held expectations, then the agent may be disappointed but her choice no less rational. In these circumstances the rationality of choosing constrained maximisation remains intact.

The second unsuccessful argument claims that an agent might do better as a straightforward maximiser whatever others do. This is to say that the future is opaque and opportunities may arise which a straightforward maximiser can exploit but which a constrained maximiser would be precluded from exploiting. This must be conceded but, again, the rationality of the choice of constrained maximisation rests entirely on expected outcomes and, if we accept the lessons of the prisoners' dilemma, in an interdependent situation in which most others are constrained maximisers, it would be rational to choose constrained maximisation for oneself.

We have rehearsed the reasoning leading Gauthier from CM^em to CM^li. The claim being assessed here is that CM^li is a resolution of the Why be moral? problem. Two questions arise. Firstly, can CM^em cultivate genuinely moral motives and real other-regarding sentiments? And secondly, what assurance is there that these motives and sentiments are actually genuine? We can set aside the first question without too much discussion. Gauthier need not claim that the separation between CM^em and CM^li is a temporal one, so that the former precedes the latter in time. He could suppose that such motives and sentiments are present, at least in the form of a capacity, from birth or soon after. CM^em is an abstraction from a person rather than a stage in a person's life.

The second question I claim leads to difficulties for Gauthier's view. Moral motivation is genuine only if the moral character of the action is what moves the agent to act. It is the absence of this character as a possibility for CM^em that disqualifies him from membership of the moral community. Recall here that CM behaviour, including having and exhibiting moral motives and sentiments, is rational and justified for CM^em, but awareness of this justification (ie. awareness of the gain in expected utility) must be motivationally invisible for CM^li. Otherwise the moral character of his actions would be compromised. The ladder of rationality, once mounted, must be kicked away. Once CM^li has properly embraced his moral personality he would reject reasoning which leads CM^em to CM^li as irrelevant as a justification to be morally constrained. Thus telling a genuinely moral person, moved by moral reasons, that he benefits from his moral behaviour and that this is a reason and a justification for so acting, must be pointless from the perspective of the question of motivation. Thus the original justification for CM^li's constraint must be beyond his reach. It is not that he could not understand such reasoning, but that he could not be moved by it, and therefore could not take it as a basis for his actions.

Let me put this point in a slightly different way before bringing out the implications. CM^em has utility maximisation as the basis of motivation and CM^li has other-regarding sentiment as a basis for motivation. CM^li cannot opt to view a proposed act as a utility maximiser except in an abstract non-motivating sense. But if this is not possible then CM^li is deaf to any justification which is based on maximising grounds. CM^li can understand, of course, how utility maximisation could be a justification for CM^em, but this is the kind of understanding we can have, say, of a murderer's justification for killing. Understanding such a justification neither moves us to kill nor does it entail taking this justification as our own. Therefore, once the ladder has been kicked away the justification falls with it.

Nevertheless, it could still be argued that CM^li does better in utility maximising terms than either CM^ems or straightforward maximisers. Surely then the rationality of being a CM^li is demonstrated, and, from the perspective of reason, fully justified? This is true, but what it misses is crucial. If we take a sufficiently distant view of CM^li's co-operative strategies, the rationality of his actions becomes evident in just the ways Gauthier so forcefully shows. However, from the inner view of a CM^li the rationality of his strategies can play neither the role of justification nor the role of motivation. From the inner view, motivation and justification have come apart.

We said earlier that the Why be moral? question is answered when we bring together motivation and justification and I have argued that they come apart when we push hard enough against Gauthier's reply. I have shown that motivational invisibility is a requirement for genuinely moral action for CM^li but that motivational invisibility defeats the requirement for genuine justification. In short, CM^li cannot both take herself to be justified (except in a self-deceptive way) and act from a properly moral motive. Of course, it is open to an advocate of Gauthier's argument to reply that from where we stand CM^li's actions are justified in an appropriate way and his motives are morally pure. But this reply is unsatisfactory on at least two counts.

Firstly, we can divide the Why be moral? question in two: Why should we be moral? and Why should I be moral?. The first question is easily answered. Collectively we do better by complying with agreed constraints than we would by exercising unrestrained liberty, as Hobbes so vividly showed. It is the second question which is so difficult to answer. Why should any individual comply with agreed constraints when afforded opportunities to free-ride? Gauthier partially succeeds in answering this question if we allow that CM^em may act to the letter of the moral law, though not in its spirit. CM^li is supposed to bridge the gap between letter and spirit, but in making this journey the rational justification which initiated it is completely lost from view. Put bluntly, what is required here is a rational justification which is present to, and capable of motivating an agent to entertain, and where appropriate, to perform genuinely moral actions.

Secondly, if the inner view does not matter as far as justification is concerned, then there is no difference between the process Gauthier advocates and brainwashing people to be morally good. We have already ruled out such strategies. Thus the Why be moral? question demands that we take up the position of the inner view, and if from this position justification and motivation do not meet, then the question has not been answered.

I conclude that Gauthier's argument fails to answer the ‘Why be moral?' question. I have shown that the move from constrained maximisation as agreed to by straightforward maximisers of utility, to constrained maximisation as employed by the liberal individual does not succeed in melding justification and motivation. However, this does not discredit the broader project of moral contractarianism provided this project is understood in the partial rather than the complete sense outlined at the beginning of this presentation.