Bertrand Russell was once asked if his extensive political activities were worth the neglect of his philosophical work. He replied that it depended on the practical results of these activities. If successful, he said, they were certainly worthwhile, but if not, they were of no avail - and should, we may take it, not have gotten in the way of philosophy. Let us look closer at the suggested contrast between politics and philosophy.

Philosophy consists in analysis and argumentation. A philosophical argument if it is clear, logically sound and rationally convincing. Its practical impact is neither here nor there. It does not matter, for example, if people are actually convinced by the argument as long as they have reason to be so.

Politics, of course, consists in more than argumentation. Yet analysis and arguments play an important role even in political life, and it is this aspect that will be considered here. What, then, distinguishes political and philosophical arguments? As Russell suggests, there are different standards of success. According to one view, practical impact counts for everything in politics. If so, the success of a political argument depends entirely on its causal force, and not at all on logical and analytical qualities. Hence, indoctrination is as good as persuasion if it works well. This characterization

* I thank Morten Egeberg, Jörgen Hermansson, Jon Hovi, Knut Midgaard and Stein
Tønnesson for helpful comments on earlier versions of this article. Two anonymous referees
also provided useful advice.

Resumé

The article discusses methodological issues in normative political theory. The basic assumption is that normative theory has a dual purpose in both establishing valid principles of differentiating from wrong and influencing actions and institutions in the right direction. The article starts by distinguishing between two approaches to normative political theory: one stresses the interpretation of existing ideas and conventions; the other takes on the constructivist of finding out what is really right. Then the relationship between theory and practice is explored. The question is how philosophical arguments can instigate practical reform. It is argued that practical considerations should be incorporated as an explicit element of normative political theory. The recommendation is, in particular, that philosophical theories enter into dialogue with the moral conventions of everyday life, whose normative force people already acknowledge.

is too stark, however. Many political arguments have a dual purpose: telling right from wrong, for example, and also affecting the course of political events in the right direction. Such arguments succeed only if they are both rationally convincing and effective when it comes to influencing beliefs and actions. We may conclude, then, that philosophical arguments are judged only by their analytical qualities, while political arguments are sometimes judged by effectiveness, and sometimes held to a dual standard of success, both practical and theoretical.

There is no denying that the practical aim is predominant in most political debates. While few participants get away with outright disregard of logical and analytical canons, many duck such requirements if they can score a rhetorical advantage with impunity. This is not, however, the way of political philosophers. They are rather apt to reverse the order of priority between the two purposes of political argument, setting greater store by logic and reason than practical accomplishments. Still, most political philosophers see their work as a species of political activity, and not as a purely philosophical trade. The questions of how political institutions should be arranged or what citizens may do to influence the exercise of authority, are rarely raised for just contemplative reasons. Political philosophy is an effort to affect social life by probing, bolstering and berating the intellectual foundations of practices and institutions.

The aim of this article is to establish that political philosophy has a dual purpose in both telling right from wrong and influencing actions and institutions in the right direction. Practical considerations, in short, should be incorporated as an explicit element of normative political theory. In particular, the recommendation is that philosophical theories enter into dialogue with the moral conventions of everyday life. Political philosophers should deal with ideas and principles that find wide resonance in society, in the sense that most people actually live by them, or at least acknowledge their normative force.

The article begins by distinguishing between two approaches to normative political theory: one stresses the interpretation of moral conventions, the other focuses on finding out what is really right and wrong. Following this, the relationship between theory and practice is taken up, as is the question of how ideas can instigate moral reform. The discussion touches on different tasks and divergent priorities for normative theory. These may seem to be in conflict, but they are really complementary, and all will be incorporated in the methodological framework which is suggested in concluding this article.

The Limits of Interpretation

The effort to develop a methodology of moral and political thinking dates

from antiquity. It goes back to the dispute between Socrates and his most notable opponent among the Sophists, Protagoras. According to Martha Nussbaum (1986, ch. 4), both Socrates and Protagoras held that everyday thinking about right and wrong is mostly confused and unsystematic. Hence, they agreed on the need for a techné - a science or method - which would lend order and system to moral and political deliberation. For the rest, however, they disagreed, recommending widely different ways out of moral confusion. Their contending positions bring out a lasting contrast between two tasks of political theory.

Some theorists concentrate on laying bare prevalent moral ideas. Their aim is to ascertain what people actually think about right and wrong, by charting expressed opinions and bringing implicit assumptions to light. Other theorists try to find correct answers to normative questions. They are preoccupied with the construction of theories and principles rather than the interpretation of existing beliefs. One may, accordingly, distinguish between interpretative and constructivist analysis.¹

Protagoras was engaged in interpretation. His modest aim was to inculcate people a habit of moral self-reflection - to make people 'more aware of the nature and interrelationships of [their] ethical commitments' (Nussbaum 1986, 104). He hoped that, as a result, whims, confusion and inconsistencies would give way to clear and coherent thought. Protagoras's techné of practical choice was a method of mental clearance work.

Nussbaum portrays Protagoras as a complacent conservative. While he finds that people are frequently bewildered in moral matters, he does not think their bewilderment runs very deeply. At root, Greek society is cemented by basically consistent and concordant beliefs about right and wrong. Fellow citizens do not just share social and political institutions, also have a moral outlook in common. They take part in the same moral convention, although their awareness of shared values is deficient and incomplete. Protagoras's aim as an expert on practical deliberation is to raise this awareness. His task is interpretation: to explicate and elucidate ideas that make up a real, but largely dormant, moral consensus.

Socrates had a much more inventive techné in mind, which reflected his quite different conception of what moral bewilderment implies. Socrates was not concerned about clarity and coherence as such; rather he wanted to root out confusion because he feared its adverse implications for social order. It leads, he suggested in the Euthypro, to disagreement about 'the just and unjust, the fine and shameful, the good and bad', and such disagreement will 'give rise to enmity and anger' (quoted in Nussbaum 1986, 106). The assumption is that moral bewilderment fosters social discord. Socrates distrusted people's ability to sort out their differences short of outright conflict. In view of this, he was not content to explicate

and interpret traditional morality, but saw the need to arbitrate conclusively between contending opinions and resolve moral problems with certainty. His was the constructivist task of finding out what is really right. Socrates raised critical questions about the validity of moral ideas in an attempt to distinguish between true and false beliefs. His techné would, as Nussbaum says, allow people to reach precise and decisive resolutions of practical controversies, in much the manner arithmetic and measurement permit when people disagree about number and size (Nussbaum 1986, 108-110).

The issue that divided Protagoras and Socrates is still central to methodological in normative theory. Michael Walzer is the foremost contemporary exponent of the interpretative approach. He holds that the proper 'way of doing philosophy is to interpret to one's fellow citizens the world of meaning that we share' (Walzer 1983, xiv). This idea has guided Walzer's works on different political issues. He has, among other things, done a thorough analysis of the particular 'world of meaning' that pertains to problems of justice in the distribution of social and economic goods. His assumption is that 'the members of a historical community' have in common certain 'sensibilities and institutions' about the justice and injustice of various patterns of distribution (Walzer 1983, 28). The task of the political theorist is to lay bare these sensibilities and intuitions. This is basically empirical work, though a training in philosophy is certainly useful for piecing together coherent arguments from the many loose ideas which float around in current debates.

Walzer stresses, however, that interpretation is not only done for its own sake. It also prepares the ground for a particular form of moral and political criticism - i.e. the exposure of hypocrisy and other divergences between thought and action. When we get to know what people believe about right and wrong, we can next ask if their actions and institutions measure up to professed principles. Walzer argues in particular (1983, 85) that the established system of welfare provisions in the United States is inadequate in the respect that 'the common understandings of the citizens [about security and welfare] point towards a more elaborate pattern". Such exposure of hypocrisy is 'the most ordinary, and it may also be the most important form of moral criticism' (Walzer 1977/1980, xv).

Walzer has aptly spoken of his approach to normative political theory as 'critical conventionalism'.² He uses social conventions as a basis of social criticism, but undertakes no critical examination of the conventions themselves. approach involves, as he says, no 'direct engagement' with 'the most profound questions of moral philosophy' (Walzer 1977/1980, xv).

We are rarely called upon to invent new ethical principles; if we did that, our criticism would not be comprehensible to the people whose behavior we wanted to condemn. Rather, we hold such people to their own principles, though we may draw these out and arrange them in ways they had not thought of before (1977/1980, xv).

Walzer, moreover, does not invent moral principles de novo, but reconstructs morality that, in his view, already exists in the morass of disorganized that lie behind everyday judgments of right and wrong. His aim is twofold: first, to clarify the structure of ordinary morality; second, to give a clear view of its 'critical force, . . . without the intervening confusion of prejudice or self-interest' (Walzer 1987, 16 and passim).

There are obvious problems with this approach to political theory. In the first place, conventions - 'common understandings' or a shared 'world of meaning' - do not exist in every society or each sphere of social life. Fellow citizens disagree, for example, about the most important considerations of distributive policy, and this raises an objection to Walzer's discussion of social justice: 'Our political arguments almost never begin in some shared understanding of the pertinent principles of distribution. Every important issue is a contest between competing models' (Dworkin 1983, 4). If so, critical conventionalism comes to a standstill, as it furnishes no definite standard for the assessment of distributive decisions.

This, Walzer replies, is as it should be. If 'radically different cultural traditions' offer competing standards, 'then it might be (morally) necessary to work out a political accommodation. Politics must sometimes substitute for justice, providing a neutral framework within which a common life slowly develops' (Walzer 1983a, 44). But the contrast between 'politics' and 'justice' is unclear. What kind of 'neutral framework' does Walzer have in mind? He can hardly envisage a political process that is devoid of ideas and arguments. It cannot be a question of substituting mute tests of numerical or physical strength for ideological and moral debate. It is hard even to imagine a political contest that does not basically consist in confrontation between more or less articulate opinions about right and wrong. The contraposition of 'politics' and 'justice' is not only confusing, but misleading. Conceptions of justice and other moral ideas are at the core of political debate. Although they often rationalize rather than inspire preferences and positions, they never disappear from view.

All this is to say that politics naturally (if not invariably) leads to a 'direct engagement' with 'the most profound questions of moral philosophy'. Political opinions shade off into philosophical ideas in so far as rationalization way to serious argument about right and wrong. Hence, the effort of philosophers to 'invent new ethical principles' is part of what politics is about. Although philosophical arguments may be subtler, more abstract and sometimes more pretentious than political opinions in general, they are not on that account essentially different from the rest.

This points towards a more general problem with Walzer's approach. It leaves the impression that conventional morality is an entrenched, fixed and immutable stock of ideas and practices - the mature product of a cultural process that has somehow come to an end and should not be

critically examined by standards that are not already enshrined in the existing convention. One is reminded of Edmund Burke's traditionalist defence of ideas and practices that take shape without purposive cultivation. Burke (1790/1981, 199) spoke, for example, of the English constitution as 'the happy effect of following nature, which is wisdom without reflection, and above it', and he warned against constitutional innovation on the ground that '[a] spirit of innovation is generally the result of a selfish temper and confined views'. As against Burke's conception, we may set the image of moral and political tradition as an open-ended process of examination and re-evaluation of more or less ingrained ideas and ways of thinking. This image may be both nearer to the truth about cultural life and more attractive as a model of how people should live.

[A]ll reasoning takes place within the context of a traditional mode of thought, transcending through criticism and invention the limitations of what had hitherto been reasoned in that tradition . . . Moreover, when a tradition is in good order it is always partially constituted by an argument about the goods the pursuit of which gives to that tradition its particular point and purpose. . . . when a tradition becomes Burkean, it is always dying or dead (Maclntyre 1981,206).

This is not to depreciate the role of conventional ideas in moral and political life. They certainly influence the way people think - their concepts and categories and the manner in which problems present themselves to them. But a tradition neither should nor will escape critical appraisal. Criticism and innovation are of a piece with cultural continuity. If so, the methodological approach of conventionalism is too narrow. It leaves out an integral part of moral thinking. Thus, normative theory cannot stop at interpretation, but is bound up with constructivist endeavours.

It deserves notice that even Walzer acknowledges the limitations of the interpretative approach when he encounters it in an extreme form. He has occasion to do so in Benjamin Barber's book The Conquest of Politics. Barber deplores the tendency of philosophers to bring participatory politics into disrepute by claiming to have discovered, once and for all, the right solutions to political issues. He sees an inherent hostility to democratic processes in such arguments. But Walzer disagrees. True, he says, a right to rule is occasionally, as in Plato's Republic, derived from the claim to know, but the authority philosophers pretend is also open to less sinister interpretations: 'Just as the citizens ought to stop and listen to a scientist who knows something about pollution when they are discussing environmental issues, so they might well stop and listen to a philosopher when they are discussing issues of distributive justice. We shouldn't give such people more power than the rest of us, only, sometimes, greater attention' (Walzer 1989,44). This is to say that philosophical constructivism is a fully legitimate endeavour which may contribute important insights to political debate.

Constructivism

Constructivism takes the critical examination of moral ideas, not their interpretation, to be the primary task of normative political theory. It stresses the philosophical aspect normative thinking: the systematic effort of separating valid and invalid ideas, justified and spurious beliefs, true and false opinions. But constructivist endeavours can go in many directions. Philosophers have proposed different ways of resolving moral disagreement by testing the validity of ideas about right and wrong.

One may, in the first place, examine specific beliefs about the moral character of particular actions and arrangements. The idea is that careful examination of such beliefs will allow us to distinguish between justified and unjustified points of view. This is the 'intuitionist' method. A notable exponent of intuitionism is W. D. Ross. He first suggested that we ascertain whether we actually grasp 'the facts about Tightness and wrongness' by comparing our opinions to 'the moral convictions of thoughtful and welleducated (Ross 1930,40-41). His assumption was that '[t]he existing body of moral convictions of the best people is the cumulative product of the moral reflection of many generations, which has developed an extremely delicate power of appreciation of moral distinctions' (Ross 1930, 41). But in a less aristocratic vein, Ross allowed that everyone is able to partake of this 'power of appreciation'. The fact, he said, that certain courses of action (such as telling the truth or promoting the good of other people) are prima facie good, will be 'evident without any need of proof to everyone who has 'reached sufficient mental maturity and have given sufficient attention to the proposition' (Ross 1930, 29).

If we put the allusion to self-evidence to one side, Ross's point is simply that moral (just like factual) beliefs derive from more or less careful consideration of the questions at hand. An opinion which is formed on sudden impulse or in a state of emotional agitation, is less reliable than one that is 'preceded and informed by the fullest reflection we can bestow on the act in all its bearings' (Ross 1930, 42). The test of validity is that moral beliefs are considered judgments, which command our confidence when we think carefully about them with adequate information and without the distorting influence of self-interest.

But this method does not conclusively resolve moral disagreement. Considered judgments diverge, and careful reflection is no guarantee of enlightened belief; it may rather give prejudice a boost.³ The intuitionist method is therefore increasingly supplemented with (or replaced by) a second test of the validity of moral beliefs. This is intended to filter out the distorting influence that factors like upbringing, class, and culture exert even on reflective opinions. One has to go beyond specific beliefs about the moral character of particular actions and institutions, to discover or

in vent general principles of right and wrong. These principles must be justified without reference to their implications for particular choices, or at least not only on such intuitionist grounds. What comes instead, or more often in addition, is a philosophical explanation of why the principles are correct.

Such an explanation can assume many forms. As an example, we may take John Rawls's (1971) theory of justice. It revolves around the notion of an 'original position' - a hypothetical contract - which is a device for systematically filtering out morally irrelevant considerations from normative We are to undertake a thought experiment in which we imagine ourselves choosing between contending principles of social justice behind a 'veil of ignorance'. The veil conceals knowledge of our personal histories, interests and outlooks. Thus, it ensures that judgments of justice cannot be attuned to individual preferences or personal conceptions of the good. If we were to choose a principle of justice under such circumstances, we would, according to Rawls, opt for one which gives priority to the protection of basic liberties and distributes wealth with a view to maximizing the material prospects of those who are worst off. The normative force of this principle is explained precisely by the fact that it would be chosen in the original position. It is valid, Rawls argues, because '[e]ach aspect of the contractual situation can be given independent support' (Rawls 1971, 21). Hence, by taking our cue from the hypothetical contract, we 'collect together into one conception a number of conditions on principles that we are ready upon due consideration to recognize as reasonable' (Rawls 1971, 21).

The idea of a hypothetical contract is but one of numerous devices th?t have been invoked in contemporary philosophy to justify general principles of right and wrong. Rawls's argument, moreover, is highly controversial, as is every other attempt that has been made to explain why one set of normative principles constitutes a philosophically favoured conception of right and wrong. This occasions the sceptical comment that we look in vain for the correct method of philosophical justification. Different theorists follow widely different guidelines in the endeavour to distance themselves from prejudices, personal inclinations and self-interest. Many directions lie open when it comes to escaping subjectivity, which suggests that a common method and concordant conclusions are out of reach. It seems that philosophical just carries moral disagreement to a higher level of abstraction - that it simply reproduces everyday disagreement in highsounding

Such qualms cannot be laid to rest, nor will they put an end to philosophicalpreoccupation morality. It should be noted, furthermore, that philosophy can make for progress in normative thinking despite the current state of widely divergent methodological recommendations. The

reason is two-fold. First, philosophical analysis may sustain progress in the sense of achieving greater transparency of normative thinking. If we move beyond specific judgments about particular cases to general principles of right and wrong, we will be better positioned to understand why people disagree about particulars. Such disagreement often reflects divergence with respect to basic moral outlooks, but people do not usually trace their judgments back to the ultimate premises. Philosophical explanation may, then, bring out the rationales of everyday beliefs and make clear why people disagree. True, it is unlikely to reduce disagreement right away, but an understanding of why moral opinions diverge is a prerequisite of eventually distinguishing between justified and unjustified beliefs.

Secondly, the current state of discordant philosophical theories is no proof that disagreement on methods and conclusions cannot be reduced. It may rather reflect the undeveloped state of moral and political philosophy. Derek Parfit (1984) suggests that the history of these disciplines may be in its infancy, and that it is far too early to judge their accomplishments.

Some people believe that there cannot be progress in Ethics, since everything has already
been said. ... I believe the opposite. How many people have made Non-Religious Ethics
their life's work? Before the recent past, very few. . . . Non-Religious Ethics has been
systematically studied, by many people, only since about 1960. Compared with the other
sciences, Non-Religious Ethics is the youngest and the least advanced (Parfit 1984, 453).

In view of this, Parfit concludes that 'it is not irrational to have high hopes' (1984, 454). The constructivist approach may hold promises that remain unfulfilled simply because rationalism has not yet been tried out with sufficient care.

Theory and Practice

The next question is how political philosophy can enter into current political debate, where practical considerations loom large. This question naturally accompanies both Walzer's critical conventionalism and most constructivist theories. It is more urgent the more important it is to ensure that false beliefs are rooted out. Socrates was, to recall, worried about the socially disruptive effects of moral disagreement. His techné of practical choice was designed for more than just straightening up distorted opinions. He rather assumed that rational inquiry into right and wrong contributes to - or, indeed, is a prerequisite of- social order.⁴

If philosophical arguments are to have impact on how people act, they must take the limits of human motivation into account. Where, then, do these limits go? It is commonly assumed that self-interest is the major reason why arguments about right and wrong may fall on deaf ears. People normally do what they believe to be best for themselves, and if the sacrifice

of acting otherwise is great, moral scruples are unlikely to stop them. This is a recurrent theme in normative theory. It preoccupies Protagoras, who, as A. E. Taylor (1926/1977) says, is 'wholly without any belief in the moral goodness of the unspoiled "savage" and . . . looks on morality as a product of civilization, a matter of imbibing a sound social tradition.' And it is no less central to modern moral philosophy. Thus, G. J. Warnock (1971, 23) observes that "limited sympathies", which foster indifference and even malignancy, are one of the 'wholly indisputable facts about people and the circumstances in which they exist'. Other-regarding attitudes do not belong to our original mental and emotional constitution; they have to be grafted on to it. Here lies, Warnock says, the "objective of morality": to curb selfinterestand limited sympathies so as to allay the 'human predicament' - the prospect that vital interests and urgent needs are neglected when they fairly easily could be met.

The acknowledgment of limited sympathies may have an air of triviality to it, but it still needs elaboration on several scores. In the first place, there is no necessary conflict between self-interest and morality. I am sometimes morally permitted to do things that are advantageous to myself and disadvantageous others because the benign effects on me count for more than the malign effects on others. While morality bars anyone from tendering interest without balancing it against the interests and claims of others, it does not demand self-abnegation, which is to subordinate personal concerns to other people's interests for no good reason. The problem is not self-interest as such, but selfishness.

In the second place, the natural tendency to do what is best for oneself is stronger the more one loses by acting otherwise. This gives no reason for surprise, but the question is whether we can differentiate more precisely between strong and weak motives of self-interest. Consider the "bank teller [who], faced with a credible gun threat, cannot, for all practical purposes, refuse to hand over the money' (Oppenheim 1981, 372). While not unfeasible, is practically impossible for almost anyone who finds himself in this kind of situation. Self-interest becomes an over-powering motive (and, incidentally, a motive which moral considerations will probably condone). Hence, demands to the effect that people sacrifice themselves in such situations are totally unrealistic. However, the notion of 'practical impossibility' permits different interpretations. 'It is', Felix Oppenheim says, 'not always possible to decide at what point the opportunity costs become so high or the risks so great as to render a contemplated course of action practically impossible or necessary'. There is, in theory, a borderline, but it 'is fluid and depends on the concrete situation' (Oppenheim 1981, 372).

Self-interest is, of course, a strong motive in many cases that do not
come near to practical impossibility. In some of these cases, too, it will be

unrealistic to demand that a person do things which are detrimental to himself. But there is no obvious way of differentiating in general terms between reasonable and unreasonable personal sacrifices. Borderlines are not only fluid, but even defy theoretical definition when we move below the level of practical impossibility.

In the third place, self-interest is not the only source of limited sympathy. Even the zealous pursuit of moral and political doctrines can foster indifference other people's welfare. Medieval crusaders and the Red Guardsmen China's 'cultural revolution' are cases in point. Moreover, Francis Hutcheson held that what appears to be selfish neglect of other people's interests and needs, is often done 'from a sense of virtue upon false opinions, and mistaken benevolence; upon wrong or partial views of public good, and the means to promote it'. He was convinced that '[i]t is not a delight in the misery of others, or malice, which occasions the horrid crimes which fill our histories; but generally an injudicious unreasonable enthusiasm for some kind of limited virtue' (quoted from Mackie 1980, 27-28).

In the fourth place, limited sympathy is not the only obstacle to moral reform by way of rational argumentation. Hutcheson's contemporary, David Hume, believes that the limits of imagination are more of a constraint than the limits of motivation when it comes to implementing moral principles. is particularly concerned about the question of how material goods ought to be distributed, which is important because 'the principal disturbance in society arises from those goods, which we call external, and from their looseness and easy transition from one person, to another' (Hume 1740/1985, 489). To stave off this disturbance, we need '[s]ome method ... by which we may distinguish what particular goods are to be assign'd to each particular person' (Hume 1740/1985, 502), and it is crucial that we find an effective method which works here and now. This is where the limits of imagination enter the picture. Hume assumes that some distributive arrangements have a certain naturalness about them; they bear the impress of Tightness, not because philosophical arguments can be mustered in their defence, but because they immediately suggest themselves and appeal, as it were by instinct, to anyone who thinks about the separation of possessions. Thus, 'the first possession always engages the attention most', and this suggests that resources should be the property of those who had them first (Hume 1740/1985, 505). Principles of right and wrong must, in general, be attuned to laws of cognitive psychology.⁵

If normative theory starts from philosophical premises rather than such empirical observations about the human mind, it will not, in Hume's view, come up with effective principles of distribution. He mocks the 'sublime theorists' who are 'possessed of reason, but unacquainted with human nature', and whose 'most obvious thought would be, to assign the largest

possessions to the most extensive virtue'. They do not realize 'that a rule, which, in speculation, may seem the most advantageous to society, may yet be found, in practice, totally pernicious and destructive' (Hume 1740/ 1985, 505).

The fifth and final comment is that self-interest and other sources of limited sympathy are not the only components of human motivation. There also exist motives and attitudes which make people responsive to moral considerations. Two observations are pertinent on this score; one has to do with emotional restraints on selfishness, the other with intellectual pillars of moral behaviour. The first point is that few people take no interest at all in how other people fare. Altruist sentiments make themselves felt in various ways with differing strength. The important question, which is notoriously difficult to answer, is whether such sentiments are normally strong or play only a marginal role. On this score, one is easily torn between contrasting images of human nature. For an apt illustration, we may turn to Hume once more. For all his scepticism about the prospect of moral reform, he still asserts that descriptions of human selfishness are often 'as wide of nature as any accounts of monsters, which we meet with in fables and romances'.

So far from thinking, that men have no affection for any thing beyond themselves, I am of opinion, that tho' it be rare to meet with one, who loves any single person better than himself; yet 'tis as rare to meet with one, in whom all the kind affections, taken together, do not over-balance all the selfish (Hume 1740/1985, 486-487).

The assertion is ambiguous. It may imply either that the total weight of 'kind affections' usually over-balances selfish inclinations, or that these affections will be decisive when they are all present (which may rarely be the case in concrete situations of choice). The second reading, which is least optimistic, probably captures Hume's intentions best. In any case, Hume proceeds to observe that the most powerful sentiments a person feels towards other people are directed at his family and friends: 'in his love to others [he] bears the greatest affection to his relations and acquaintance' (Hume 1740/1985, 487). Hence, selfishness is most likely to be kept in check by kind affections of rather restricted range.

Hume's double-vision is presumably typical of how many people see the relative importance of self-interest and altruist sentiments. Rejecting both the bogey of human beings as selfish monsters and an inflated image of 'benevolence to strangers' (Hume 1740/1985, 492), they end up with the idea that none is a pure egoist, but most take strong interest in the welfare of just some few others. This implies that impartial moral considerations will find scant support in human emotions.⁶

But altruist sentiments are not the only motivational basis of moral
reform. Henry Sidgwick (1907/1981, 5) held that 'when a man seriously

asks "why he should do" anything, he commonly assumes in himself a determination to pursue whatever conduct may be shown by argument to be reasonable, even though it be very different from that to which his nonrationalinclinations prompt'. This is to say that emotional motives which undermine morality are counteracted by a desire to be able to justify one's intentions and actions by rational argumentation. The existence of such a desire is, of course, subject to doubt. Hume would probably disregard it altogether,⁷ but on this score he may not be on the side of common sense.

If we juxtapose some of the assumptions just set out, an image of bifurcate motivation emerges. On the one hand, each person wants to do what is best for himself and those who are close to him; on the other hand, he does not want to be blamed for discriminating arbitrarily between interests of equal strength, whether they belong to himself, his kin or complete strangers. In the same vein, Thomas Nagel (1991, 3-4) says that there is 'a division in each individual between two standpoints, the personal and the impersonal': the latter 'produces ... a powerful demand for universal partiality and equality, while the personal standpoint gives rise to individualistic motives and requirements which present obstacles to the pursuit and realization of such ideals'. Impersonal considerations may not make up very powerful motives in many people, but there is no need here to speculate about the relative importance of the two standpoints. What matters is that rational arguments can appeal to an attitude which is no less real, though often much weaker, than self-interest and emotional ties to family and friends.

When a philosopher becomes concerned about the uncertain prospect of actually influencing actions through rational argumentation, one of three things usually happens. In the first place, the problem may be relegated to works of applied philosophy, which can get started only after basic questions of justification have been sorted out. This is, in particular, Rawls's strategy. His programme for 'ideal theory' implies that we work out 'the principles that characterize a well-ordered society under favorable circumstances' (Rawls 1971,245). The 'favourable' nature of the circumstances lies, among other things, in the absence of certain human weaknesses. It is supposed, for example, that people actually take their bearings from well-grounded moral principles, accepting and supporting whatever arrangements these principles vindicate. In general, ideal theory - which Rawls also calls 'strict compliance theory' - takes only the 'fixed constraints of human life' (Rawls 1971, 245) into account (like the fact that most material resources exist in limited supply), but it counts neither limited sympathy nor the limitations of the imagination among these constraints. According to Rawls, ideal theory is the proper place to begin because 'it provides . . . the only basis for the systematic grasp of [the] more pressing problems' of everyday life

(Rawls 1971, 9). His point is presumably that piecemeal solutions to particular problems may work at cross purposes, or add up to an undesirable whole, if there are no comprehensive principles to guide them, i.e. no notion of the 'nature and aims of the perfectly just society' (Rawls 1971, 9). Thus, we should first work out a blueprint of ideal arrangements and then worry about 'adjustments to natural limitations and historical contingencies' (Rawls 1971, 246).⁸

In the second place, a philosopher may believe that the best way of overcoming obstacles to moral reform is to refine his arguments. The force of reason may be trusted to (eventually) countervail selfishness and partiality if only moral truths are established beyond doubt. This implies that the philosopher can concentrate on the task of finding out what is really right. Compelling arguments have arguably influenced social events, and can do so again. Thus, Amartya Sen (1970, 122-123) believes that 'Rousseau's analysis of "injustice" and Marx's theory of "exploitation" . . . have had a bigger impact on the shape of the world than would have been predicted by the "hard-headed realist"'. In the same vein, Brian Barry (1981, 44) writes:

General ideas can give rise to movements that are liberating rather than destructive. The
best example is the anti-slavery movement in Britain and America. ... In both countries,
the opposition to slavery was, I think, almost entirely based on the moral sentiment that it
is wrong.

Yet, there are always 'hard-headed realists' who will maintain that self-interest is the decisive factor even when it seems under the sway of moral ideas. One may, for example, counter Barry's analysis of the American War with reference to President Lincoln's statement (of August 1862), that 'my paramount object in this struggle is to save the Union, and it is not either to save or destroy slavery. If I could save the Union without freeing any slave, I would do it; and if I could do it by freeing all the slaves, I would do it; and if I could do it by freeing some and leaving others alone, I would also do that' (quoted in Acheson 1967, 14).

In the third place, the concern about selfishness and other obstacles to moral reform may divest the attention from philosophical justification to psychological and sociological issues. This shift brings problems of moral education into focus, as well as many pragmatic questions that bear on the implementation of any given system of normative principles:

[W]ill [it] win our support? Are its principles teachable? Would it be reasonable for an individual or society to endorse it? How severe will be the psychological strain of adhering to it? Will people be able to comply with it successfully? Will they wish to do so? Will it satisfy our various purposes in having a shared morality in the first place? Could an alternative system do better? And so on (Shaw 1980, 133).

The methodological recommendations for normative theory which follow

steer a course between the last two responses to the question of how moral ideas can influence political practice. Both a blind trust in the force of reason and an exclusive preoccupation with psychological and institutional mechanisms are out of place. Instead, an approach is proposed that integratesthe tasks and concerns which have been brought up so far in this article.

Methodological Lessons

It has been argued above that normative political theory must go beyond the interpretation of moral conventions to engage itself with the constructivist of ascertaining what is really right. This is, indeed, part of what it means to take conventional morality seriously. It is kept in good order by the constant efforts of philosophers and others to elaborate, refine and change it. The constructivist approach has, in other words, a central role to play in normative theory. This is the first lesson to be drawn from the previous discussion.

Yet, arguments about right and wrong, however forceful in philosophical terms, may prove irrelevant to moral and political life. They risk falling to the earth because people are not motivated to abide by them. The problem relates to the practical aspect of normative theory. As previously noted, self-interest and emotional ties to family and friends are strong motives which easily cancel out impersonal and impartial considerations that appeal to our sense of moral decency. What can be done to avoid this outcome and, more generally, provide some bulwark against the practical irrelevance of normative theory?

Now, morality is not necessarily at odds with the personal standpoint. One's own interest sometimes comes before the interests of other people simply because it is weightier from a moral point of view. Moreover, close personal relationships may give rise to special obligations towards relatives and friends, and in so far as such obligations bear on the question of right action, there will be moral reasons to behave partially. They are not necessarily conclusive reasons, but must anyway be taken into account. Finally, there may be other grounds for favouring oneself or one's relatives and friends in a conflict of interest. There exist, accordingly, several possible bases of legitimate partiality. One way of providing against the practical irrelevance of normative theory, is to take these possibilities seriously. Normative theory is most likely to get through to people if it starts where practical deliberation normally begins: with the urge to do what is best for oneself or one's close associates. Thus, in addressing problems of right action, the first thing to ask is how far people may rightfully favour themselves or otherwise act as their limited sympathies incline them to. A

person will hardly give heed to impersonal and impartial considerations if
his personal standpoint does not get a proper hearing. This is the second
methodological lesson to be drawn.

Another thing one may do to provide against practical irrelevance, is to take moral conventions seriously. This third lesson is drawn from one of the arguments behind the conventionalist approach. Walzer's point was, to recall, that the most important form of moral criticism is the exposure of hypocrisy: holding people to their own principles. He argues, in particular, we should not 'invent new ethical principles', because 'if we did . . . our criticism would not be comprehensible to the people whose behavior we wanted to condemn' (Walzer 1977/1980, xv). The point is plain: the more political theorists reason in terms that are unfamiliar to their audience, the less likely their arguments are to be widely comprehended and heeded. People are more easily accessed if they are addressed in the language of their moral conventions.

This is a good reason to take conventional morality seriously, but no reason to tie normative theory up with prevailing opinions in a given society. The point is rather that moral criticism should start from an examination of the principles people already recognize as reasonable constraints self-interest and other personal motives. Then it will be possible to raise critical questions about justification and at the same time retain a secure foothold in conventional morality. Thus, a political theorist should not imagine an ideal moral system in place of a convention that does not measure up to critical standards. His task is instead, as J. L. Mackie (1977, 148) puts it, to bring pressures to bear on 'fragments of the [existing] system, so that they come gradually to be more favourable to what he sees as valuable and worthwhile'.

Still, whatever steps we take to allay the risk of irrelevance, considerable uncertainty surrounds the question of how normative theory enters into political life and affects political decisions. We can only speculate on how much impact it has on attitudes and action, and whether different modes of argument will be more or less effective in influencing events and institutions. to the latter question, it has been suggested, for example, that radical ideas of a Utopian bent may aggravate rather than ameliorate imperfect social conditions. The assumption is that they hold up impossible ideals which sap the motivational basis of established norms without putting anything in its place. This is, however, a controversial assumption - an axiom to conservatives and a self-serving prejudice to political radicals. The final lesson of this article is that normative theory should take this kind of pragmatic uncertainty into account when practical proposals for political reform are put forth.

---

NOTES

1. John Rawls and Brian Barry speak of 'constructivist' analysis in a related, but more specific, sense. See Barry (1989, 264-271).

2. In conversation (June 1985).

3. An example of this can be found in Dr. Johnson's account of his dispute with Mrs. Macaulay, 'a great republican': 'One day when I was at her house, I put on a very grave countenance, and said to her, "Madam, f am now become a convert to your way of thinking. I am convinced that all mankind are upon an equal footing; and to give you an unquestionable proof, Madam, that I am in earnest, here is a very sensible, civil, well-behaved fellow-citizen, your footman; I desire that he may be allowed to sit down and dine with us". I thus, Sir, showed here the absurdity of the levelling doctrine. She has never liked me since' (Boswell 1791/1952, 122).

4. Socrates' view of the role philosophy can play in social life has had a continual attraction on political theorists, some of whom share his somber assumption that philosophical confusion and discord disrupt society. Thus, R. M. Hare (1979, i) says that his work is done from 'a sense of urgency - a feeling that . . . philosophers might do more to help resolve important practical issues', which 'are issues over which people are willing to fight and kill one another; and it may be that unless some way is found of talking about them rationally and with hope of agreement, violence will finally engulf the world'. But Hare once took a more sober view of how much philosophers can hope to accomplish. To the charge that they 'ought to be using their philosophy to prove political conclusions', he replied: "it is not the function of philosophy to make up people's minds for them. It aims only at understanding; and its initial move is often to show that we do not understand what we think we understand. That is why it is so unpopular' (Hare 1972, 23).

5. Hume explains that 'the mind has a natural propensity to join relations, especially resembling ones, and finds a kind of fitness and university in such a union. From this propensity are deriv'd these laws of nature, that upon the first formation of society, property always follows the present possession; and afterwards, that it arises from first or from long possession. . . . [T]here is first a natural union betwixt the idea of the person and that of the object, and afterwards a new and moral union produc'd by that right or property, which we ascribe to the person' (Hume 1740/1985, 509-510).

6. I am impartial if the way I judge people and deal with them is not influenced by who they are or what special relationship they have to me. This implies that I do not discriminate between people on grounds of personal identity or favour those who are connected to me through, e.g., friendship or kinship. The only legitimate basis for discrimination is difference with respect to general characteristics that, as John Cottingham 86) says, 'would appeal to a neutral observer".

7. This view apparently follows from Hume's denial that any preference is in itself reasonable or unreasonable: "Tis not contrary to reason to prefer the destruction of the whole world to the scratching of my finger. 'Tis not contrary to reason for me to chuse my total ruin, to prevent the least uneasiness of an Indian or a person wholly unknown to me. 'Tis as little contrary to reason to prefer even my own acknowledg'd lesser good to my greater, and have a more ardent affection for the former than the latter' (Hume 1740/1985, 416).

8. As it turns out, however, Rawls does not keep within the framework of ideal-theory in justifying his principles of justice. He assumes that the parties in the original position 'will not enter into agreements they know they cannot accept, or can do so only with great difficulty' (Rawls 1971, 145, cf. ch. 111, sect. 29). Whether he still - as Robert Nozick (1974, 192-195) indicates - takes this practical problem too lightly, is another matter. Incidentally, it has been suggested that Rawls's principles of justice owe their attractiveness for him to the kind of instinctual mechanisms of imagination that Hume believes to underlie moral thinking (Harrison 1981, 91, n. 2).

REFERENCES

Acheson, D. 1967. 'Ethics in International Relations Today', in Raskin, M. G. & Fall, B. 8.,
eds., The Vietnam Reader. New York: Vintage Books.

Barry, B. 1981. 'Do Countries have Moral Obligations? The Case of World Poverty', in The
Tanner Lectures on Human Values. 1981. 11. Salt Lake City: University of Utah Press.

Barry, B. 1989. Theories of Justice. London: Harvester.

Boswell, J. 1791/1952. The Life of Samuel Johnson. New York: Random House.

Burke, E. 1790/1981. Reflections on the Revolution in France. Harmondsworth: Penguin
Books.

Cottingham, J. 1983. 'Ethics and Impartiality', Philosophical Studies 43, 83-99.

Hume, D. 1777/1975. Enquiries Concerning Human Understanding and Concerning the
Principles of Morals. Oxford: Clarendon Press.

Maclntyre, A. 1981. After Virtue. A Study in Moral Theory. London: Duckworth.

Nussbaum, M. 1986. The Fragility of Goodness. Luck and Ethics in Greek Tragedy and
Philosophy. Cambridge: Cambridge University Press.

Oppenheim, F. 1981. Political Concepts. A Reconstruction. Oxford: Basil Blackwell.

Shaw, W. H. 1980. 'Intuition and Moral Philosophy', American Philosophical Quarterly 17,
127-134.

Sidgwick, H. 1907/1981. The Methods of Ethics. Indianapolis: Hackett.

Walzer, M. 1983a. ' "Spheres of Justice": an Exchange', New York Review of Books 21 July.