II. Politics: The Hyper-Realist View

I want to begin by considering a position I will call “hyper-realism” which if true would explain why politics is not susceptible to normative philosophical investigation. It depicts politics as a struggle for dominance between human beings whose aims are conflicting, but who need to act collectively in order to achieve those aims. So its central component is the exercise of power — power in the sense of the ability to get other people to fall in line with your own aims rather than pursue theirs. Although in the course of acting politically people may proclaim certain beliefs or espouse certain principles, these are not to be taken at face value, but should be regarded simply as instruments of persuasion, or ways of disguising the real purpose of action. Insofar, therefore, as philosophy seeks to establish the truth, or justifiability, of beliefs, it entirely misses the point when applied to politics. Since beliefs or principles are not action-guiding in any case, assessing them normatively is a waste of time. Philosophy is either irrelevant, or if it does end up offering justifications for the creed of some group or another, is simply complicit in a struggle for power whose existence it fails to recognize.

I am not sure if anyone holds hyper-realism in exactly the stark form just identified, but Raymond Geuss in a recent book comes fairly close. ^{Footnote 5} The oddity, perhaps, is that Geuss takes himself to be defending “the realist approach to political philosophy,” whereas the realist account of politics that he lays out appears to preclude any form of genuine philosophical engagement with it. There is room, of course, for “political thought” in the sense of thinking prudentially about how best to advance your aims against opposition in a variety of circumstances. But this would be unlikely to issue in any general principles, let alone principles of the kind that political philosophers have regarded it as their business to defend (principles of justice, for example). As Geuss himself remarks, “politics is more like the exercise of a craft or art, than like traditional conceptions of what happens when a theory is applied.” ^{Footnote 6} The skill involved consists to a large extent in the ability to respond flexibly as the situation in which you are pursuing your aims changes and new obstacles must be overcome. It is possible to advance some maxims about how to succeed politically, as Machiavelli for example did in The Prince, but it seems to me obfuscating to call this political philosophy if we assume that the latter has to do with normative justification in some form that is not merely prudential; at this stage I am leaving it open what that form might be. ^{Footnote 7}

My aim in specifying “hyper-realism” is in part to distinguish it from the kind of realism that I will associate particularly with the work of Bernard Williams ^{Footnote 8} (and that does, I think, give an interesting answer to my title question), but also in part to suggest what politics must be like if (action-guiding) political philosophy is to be possible at all. If politics were only a matter of individuals driven by their interests and competing for power with one another, then the scope it offered for philosophical reflection would be very limited. It would still be possible to engage in what I have elsewhere called “political philosophy as lamentation,” where the real world of politics is contrasted, unfavorably, with an alternative world (or heaven) in which certain moral virtues such as justice are displayed. ^{Footnote 9} But this does not yield normative principles suitable to offer guidance when decisions have to be made about public policy or institutional change. However, the hyper-realist view of politics is narrow and one-sided. Struggles for power do of course take place, but they do so in an institutional context in the absence of which the struggle would not be political. Politics, I suggest, must be regarded as a means of resolving disagreements and solving collective action problems through procedures and by means of institutions that nearly all of those subject to them regard as authoritative. The decisions arrived at must, if necessary, be coercively enforced, but most of their addressees must take their authoritative character as a reason for compliance. Thus the question of legitimacy — what gives an institution or a procedure the appropriate kind of status such that its deliverances are treated as authoritative — is central to politics. What form these institutions and procedures take is quite variable, and this in turn will affect the part played by political activity in the narrow sense — as portrayed by hyper-realists — within politics in the broader sense I am now specifying. If the decision-rule is “whoever holds the conch shell shall decide for the group,” then it is predictable that there will be a struggle to get hold of the conch shell. If, at the other extreme, decision making takes the form of deliberation in which a near-consensus is required to render the decision authoritative, then there may still be argumentative maneuvering, but much less opportunity to exercise the cruder forms of power. In between, we find familiar practices, such as splitting the difference through compromise, taking turns to decide, and so forth, that allow all participants to feel that they have something to gain by accepting authority.

It may be useful at this point to assemble some reminders of the many and diverse activities that we would routinely consider to be “political.” ^{Footnote 10} Restricting the list just to domestic democratic politics, we might include: the members of a constituent assembly writing or revising a constitution; heads of spending departments negotiating with Treasury officials over the contents of a forthcoming budget; a parliamentary or congressional committee scrutinizing a proposed piece of legislation; party activists canvassing voters for support at an impending election; NGO members lobbying elected representatives over some policy issue; candidates for party leadership trying to win over rank and file members to their cause. Notice that only the last of these comes even close to exemplifying the hyper-realist view of politics as a struggle for dominance. All of them involve, to varying degrees, discussion, negotiation, persuasion, and therefore, in a broad sense, the giving of reasons that are in turn susceptible to normative assessment. If, as I have been assuming, normative assessment of political institutions, procedures, laws, and policies is the main task of political philosophy, there is no reason to think that the nature of politics precludes this.

III. Bernard Williams: Political Moralism and Political Realism

Our next question, however, is about the form that such assessment takes, and in particular the sense in which, allegedly, it should be “political” rather than “moral.” Here we turn to Williams’s critique of “political moralism,” which for him is represented paradigmatically by utilitarianism on the one hand and Rawls’s theory of justice on the other. ^{Footnote 11} The charge is that in these theories political arrangements are being assessed by standards developed by moral reflection that is prior to and independent of politics. What makes this assessment particularly inappropriate, according to Williams, is that ethical disagreement is one of the circumstances that politics exists to respond to — not by resolving it, but by coming down on one side of the argument or the other. To assume that we can apply our favored theory of justice to political questions is to assume, contrary to fact, that such a theory will command widespread assent once laid out. Were that indeed so — were there an underlying consensus on moral principles relevant to public affairs — there would hardly be a call for politics itself.

Williams’s alternative — his version of political realism — makes legitimacy the central problem of political philosophy. What he calls “the first political question” is how the state can secure “order, protection, safety, trust and the conditions of cooperation” while at the same time satisfying the “basic legitimation demand” that its manner of doing so should be acceptable to the people it governs. ^{Footnote 12} Williams denies that this “basic legitimation demand” is itself a moral principle, in the usual sense, but is instead “a claim that is inherent in there being such a thing as politics” as opposed to a society in which one group of people simply exercises coercive power over the remainder. ^{Footnote 13} Whatever else it does, then, political philosophy must supply an answer to this question, and Williams asserts that utilitarianism, Rawls’s theory, and other versions of “political moralism” all fail to do so.

This position, although set out by Williams in somewhat fragmentary form, deserves close examination. Three issues can be distinguished. The first is the status of the goals whose fulfillment motivates “the first political question.” Why couldn’t these goals be regarded by someone unfriendly to Williams’s form of realism as moral objectives or constraints that any political arrangement must meet? The second is the force of his appeal to politics as something presumed to be preferable to its alternative, characterized negatively by Williams as “one lot of people terrorizing another lot of people.” The third is the “basic legitimation demand” itself. How is it to be satisfied? Who needs to be convinced that the state that governs them is legitimate? What kind of reasons might be necessary or sufficient for this purpose? ^{Footnote 14}

Williams interprets “the first political question” in Hobbesian terms, and his thought seems to be that the principles and values that make up the standard fare of political philosophy — freedom, distributive justice, democracy, and so forth — cannot be pursued until people have been safeguarded against “continuall feare, and danger of violent death,” ^{Footnote 15} so that any political institution must be judged in the first place by its capacity over time to provide basic security against these evils in the face of changing circumstances. But what makes the basic goods that Williams lists distinctive? Why aren’t they just important moral values, potentially subsumable under wider principles such as utility? (Bentham, for example, described security as the most important of the subordinate ends that a utilitarian legislator should pursue — the others being subsistence, abundance and equality. ^{Footnote 16} ) Williams’s thought here is perhaps that they are undeniable in the way that other principles and values are not. When discussing human rights, for example, he argues that there is a small list of rights whose purpose is to protect people against evils that are universally recognized, such as torture and the denial of religious expression. ^{Footnote 17} But, first, even if like Williams we accept that ethical disagreement is pervasive, it is surely not definitive of a moral principle that it should in fact be contested. So somebody who wants to argue that political philosophy is a branch of moral philosophy could reply to Williams that all he has shown us is that politics must involve the pursuit of a subclass of moral goods, one of whose features is their uncontestable status. But then, second, are they really uncontestable? Perhaps at some very basic level they are: nobody, presented with Hobbes’s famous description of the state of nature, would deny that it was better to escape from it. But consider the only slightly less famous speech by Harry Lime in The Third Man about the comparative effects, for the creation of great art, of thirty years of warfare, terror, murder, and bloodshed under the Borgias, and five centuries of peace, democracy, and brotherly love in Switzerland. Italy in this period had solved the Hobbesian problem far less successfully than the Helvetian Confederation, but meanwhile had produced “Michelangelo, Leonardo da Vinci and the Renaissance” rather than the cuckoo clock. Is Harry’s assessment incomprehensible to us? Can’t the basic goods that Williams highlights be traded off against other values when we are asked to make choices between different political arrangements? Isn’t this in a meaningful sense an ethical question?

Consider next Williams’s account of politics as a human practice. It has two main features. First, it is activity whose purpose is to reach a decision, not to establish truth. Williams tells us to regard those on the other side as opponents, not as “arguers who are simply mistaken, or as fellow seekers after truth.” ^{Footnote 18} Second, although the decisions that are reached must of course be enforced against dissenters, this is not simply a case of coercion, since the institutions and procedures by which these decisions are reached are such as to satisfy the Basic Legitimation Demand (BLD). This means, at least in the case of most of those who have to experience their effects, that the decisions are regarded as authoritative, even by those who disagree with their substance (I return later to the position of people for whom the BLD is not satisfied).

Leaving aside for later discussion what legitimation means here, I find nothing to quarrel with analytically in this account of politics. But someone might still ask: Why have politics rather than something else? There must be some element of evaluation hidden here, not made explicit by Williams. He must assume, first, the need for enforceable collective decisions over at least some areas of human life. Anarchists, presumably, will disagree, although it is unclear whether this is just a case of empirical disagreement about what might happen in the absence of politics in Williams’s sense, or whether there is evaluative disagreement as well — that is, whether the anarchist might find politics (understood as including enforcement of decisions) sufficiently repugnant that she would want to do without it even if the cost was, for instance, that many significant collective action problems went unresolved. More interesting, perhaps, is the question why it is better to be coerced by institutions that you regard as authoritative than by those that you do not, assuming that the substance of the coercion is the same in both cases. (Recall that at this point we are asking for an evaluation of politics as such, not democratic politics, or constitutionally limited politics, and so on.) The usual answer is that you are being given a reason (of a certain sort) in the first case, and no reason (other than “something bad will happen to you if you don’t do this”) in the second. But why should that be important? The usual answer again is that we show respect for our fellow human beings when we offer them reasons for compliance (albeit not reasons of the kind that the anarchists among them will accept) which we do not show when we simply coerce them. There is nothing wrong with this answer, but does it appeal to a distinctively political principle or value? The point about showing respect is going to hold in all cases in which authority is applied, which might be the authority of a parent over a child, or the authority of a football referee over one of the players. In each case we can claim that when authority is exercised legitimately, the relationship between commander and commanded is better than when compliance is simply enforced, by virtue of the fact that a reason for compliance can be given to those commanded. But “better” here seems to mean “ethically better.” Respect for persons is an ethical principle, not one that is narrowly political. ^{Footnote 19} This is not to deny that politics is the arena in which authority matters most, in which it is most important that the decisions that are made should be ones that are taken by legitimate procedures or institutions. Since the stakes are usually high, and those who fall on the losing side may be much worse off than they would have been had the decision gone the other way, they must recognize the authoritative standing of the process by which the decision was reached. In that sense one could say that legitimacy is a distinctively political value — it is a value that comes into play most prominently in that arena. But in the same sense, we could say that courage is a distinctively military value, because war is the arena in which it has its greatest scope and importance. Yet we also say that courage is a moral quality, that when we assess people as either courageous or cowardly, we are engaging in ethical evaluation. So it is not clear, so far, that by recognizing the question of legitimacy as the central political question, following Williams, we must deny that our defense of political authority (as contrasted with brute coercion) is itself an ethical defense.

Williams raises the question himself when he remarks “It may be asked whether the BLD [Basic Legitimation Demand] is itself a moral principle. If it is, it does not represent a morality which is prior to politics.” ^{Footnote 20} The expression “prior to” here is ambiguous, though. In one sense Williams is right, because if there were no such thing as politics, the issue of legitimation would not arise; the need for a justification arises precisely because politics involves institutions like states obliging people to comply with the decisions they reach. But in another sense it remains an open question whether what is doing the work when the legitimation demand is satisfied is a principle that is somehow internal to politics itself, or a moral principle that might also apply in domains other than politics, as I suggested above in relation to authority and its connection to the idea of respect for persons.

So we need to examine the BLD itself, and how Williams envisages it being fulfilled. This is one of the places at which his thinking is most difficult to pin down. He says, first of all, that there can be no universally applicable account of legitimation: what counts as legitimation will vary according to time and place, and will depend on the beliefs already held by the people over whom political rule is being exercised. In particular, although liberal principles are the only ones that will work for us — denizens of modern Western societies — there is no cause to assign them any kind of privileged status in general as solutions to the legitimacy problem. ^{Footnote 21} On the other hand, Williams is reluctant to say that just anything that “works” in the sense of inducing compliance will count, since he is alive to the problem of people being coercively drilled into accepting beliefs. He therefore applies a “critical theory principle” that disallows legitimation by means of beliefs that have only come to be accepted through the exercise of the coercion that they purport to justify. ^{Footnote 22} He also hedges on whether the form of legitimation that is being provided should merely be acceptable to the people in question, in the sense that it appeals to ideas that they ought to accept, given all of the other things that they believe, or should actually be accepted. He recognizes that no political order will be recognized as authoritative by everyone under its dominion — some people may, as he puts it, be “anarchists, or utterly unreasonable, or bandits, or merely enemies” ^{Footnote 23} — and this raises the question of what proportion of the population must accept a proposed legitimation for the BLD to be satisfied. These difficulties seem to me to reflect uncertainty on Williams’s part about how far he wants to offer a wholly “realist” or sociological account of legitimacy, where all that matters is that sufficiently large numbers of people should accept the justifying story that their state tells them (this, after all, is enough to solve the Hobbesian problem itself), or whether his ambition is to be more normative, such that we would be able to say to people in a certain context that they ought to accept the state’s authority because its claim to legitimacy is sound (albeit not necessarily liberal in form).

IV. Political Legitimacy: Williams versus Rawls

The next question to ask is about what, substantively, can count as legitimation. It is common in discussions of this topic to distinguish between “input” legitimation and “output” legitimation. That is, a political institution may claim to be legitimate either because of its pedigree — its historical origins or its institutional form — or because of what it provides for the people it governs — peace, prosperity, social justice, and so on. Usually these two forms of legitimation are combined. Now Williams certainly accepts as a necessary condition of legitimacy that a state should provide the basic “Hobbesian” goods, but output legitimacy of that sort is plainly not sufficient, otherwise the questions that he asks about what might be needed to satisfy the BLD under different historical conditions would not arise. Williams must envisage that various political set-ups might succeed in providing the basic goods, but not all of these would satisfy the BLD for a particular people. So what is to prevent a (non-realist) political philosopher presenting his theory as an answer to the BLD for the people he takes himself to be writing for? Consider Rawls, for example: Might he not simply rebrand his theory of justice as a theory of legitimacy for modern liberal societies? ^{Footnote 24} The claim would be that in such societies a political system is legitimate if and only if its institutions, laws, and policies come close enough to embodying the famous two principles of justice. What then becomes of Williams’s critique of “political moralism”? He might of course continue to believe that Rawls has the wrong theory, but that is different from saying that Rawls has the wrong type of theory — that it is not political in the way that it needs to be.

It is in fact hard to see exactly what Williams finds to object to in Rawls. He notes that in Political Liberalism Rawls describes his “political conception of justice” as “a moral conception worked out for a specific kind of subject, namely, for political, social, and economic institutions,” but Williams omits the accompanying footnote in which Rawls explains that “in saying that a conception is moral, I mean, among other things, that its content is given by certain ideals, principles and standards; and that these norms articulate certain values, in this case political values.” ^{Footnote 25} It seems that Rawls in this passage is using “moral” in what is perhaps a regrettably loose way to cover the whole domain of the normative (“ideals, principles and standards”) such that “political values” count as simply a subclass of “moral values.” Elsewhere Rawls says explicitly that as a political conception of justice, “justice as fairness is not applied moral philosophy.” ^{Footnote 26} Another of Williams’s complaints about A Theory of Justice is that it overreaches itself by aspiring to offer a “universal” theory of justice, but Williams recognizes that by the time of Political Liberalism this ambition had been reduced to providing an account relevant to the citizens of modern constitutional democracies, and in that respect what Rawls offers is consistent with Williams’s distinction between legitimation tout court and the form of legitimation appropriate to liberal societies. ^{Footnote 27} Is the objection, then, that Rawls does not address the “first political question” directly? But Rawls might reply to this that even if his theory of justice is broader than a direct response to the first political question, it does at least inclusively solve it, insofar as a society that complies with the two principles of justice must by definition be one that supplies the basic goods — order, protection, and so on — that according to Williams it is the fundamental purpose of politics to provide. And Rawls is certainly preoccupied with the problem of stability ^{Footnote 28} — with whether a society that organizes itself on the basis of his principles will continue to generate support for those principles among its own members — so he cannot be accused of ignoring Williams’s reminder that solving the first question is not a once and for all matter, but an ongoing challenge.

The real charge against Rawls must, I think, be that his theory of social justice, understood now as an account of political legitimacy for liberal societies in general, is too ambitious. It goes beyond what is needed to legitimate a liberal order, and in doing so encompasses principles that many people will find ethically contestable. The principle of fair (as opposed to formal) equality of opportunity might be an example here. As the continuing debate over affirmative action shows, many people take the view that as long as jobs and other positions are given to the best-qualified candidates, the relevant requirements of justice are satisfied. They do not demand that the background conditions that affect whether people have the necessary qualifications are fair in a deeper sense. Now it is a good question whether Rawls’s theory is too “liberal” for its intended audience in a way that might make “general” reflective equilibrium — where everybody can be brought to acknowledge that the theory adequately systematizes their own considered judgements about justice — impossible to achieve. ^{Footnote 29} This would defeat Rawls’s own hopes for his theory. It does not follow that the theory must fail as a response to the Basic Legitimation Demand.

What might a less demanding theory of liberal legitimacy look like? If we turn to Williams for illumination, we shall not find much. In a discussion of Judith Shklar’s “liberalism of fear,” ^{Footnote 30} to which he is broadly sympathetic, Williams emphasizes the familiar devices that liberal states use to control the exercise of power by state officials. ^{Footnote 31} But he recognizes that “preventing oppression by the state” cannot be all that there is to liberal legitimation. People fear other things besides state oppression — they also fear poverty, unemployment, ill health, and so forth. Once the political tools are available to combat these evils, it will be natural to add success in eliminating them as a condition of the state’s legitimacy. Moreover a liberal account of legitimation must, says Williams, also take into account the following two theses “rationalizations of disadvantage in term of race and gender are invalid” and “hierarchical structures which generate disadvantage are not self-legitimating.” ^{Footnote 32} The implication here is that in order to redeem itself, a liberal state must be seen to pursue egalitarian policies in certain domains: it must take steps to outlaw discrimination on grounds of race and gender, and it must call into question social hierarchies that cannot be justified by appeal to their beneficial consequences, especially consequences for those relatively disadvantaged by the hierarchies. A liberal state that met these conditions might not need to go as far as implementing Rawls’s second principle (fair equality of opportunity plus the difference principle) fully, but it would be moving in the same direction. It seems that for Williams, too, a commitment to social justice must play a large part in the legitimation narrative of a modern liberal state, and his disagreement with Rawls is much less about matters of substance than about how we should understand the status of that commitment.

Of course it would be possible to stop at the point suggested by the original version of the liberalism of fear. The liberal state’s legitimacy claim would just be that it provides the most secure and stable way of controlling political oppression and domination. But would that actually be enough to satisfy the BLD in modern societies, especially in light of the very large inequalities in life prospects that they generate? How can such a political order be justified to those who find themselves near the bottom end of the distribution? One possible way of presenting Rawls’s theory of justice as a theory of legitimacy for liberal societies is to say that it represents a compromise between social classes: the rights covered by the first principle — private property, free speech, and so forth — are of greatest value to members of the economic and political elite and protect their essential interests; whereas by restricting the range of inequality, the second principle provides the non-elite with assurance that the social system will work to their advantage too. ^{Footnote 33} This is not of course the way in which Rawls himself envisages the two principles being justified — he would say that to do so “makes political philosophy political in the wrong way” ^{Footnote 34} — but here I am presenting its “realist” credentials as a theory of legitimacy, and suggesting that from this point of view it is in fact more realistic than minimalist forms of liberalism (such as Nozick’s, for instance) that if implemented would fail to provide the disadvantaged with adequate reasons to support a liberal order.

I have suggested that none of the punches Williams throws at Rawls really lands on target. Although Williams seeks to make political philosophy more “political” by making legitimacy — primarily a property of political institutions — rather than justice its central concept, when we look at the actual substance of these two ideas, legitimacy according to Williams and justice according to Rawls appear to be close neighbors. Legitimacy may have other components that are more society-specific: if we ask why presidential government is legitimate in the United States and parliamentary government is legitimate in the United Kingdom, the answer will involve supplying a historical narrative. But this is not something that philosophy can address directly. ^{Footnote 35} All that philosophy can do is point to the common elements — the rule of law, representative government, individual rights, and so forth — that potentially qualify both of these systems as legitimate forms of government for a liberal society.

To say that Williams lands no punches on Rawls is not, however, to say that there is nothing to his critique of “political moralism.” He may have identified a tendency in contemporary political philosophy that Rawls’s work has encouraged even if Rawls himself is innocent of the charge as stated. ^{Footnote 36} As we have seen, Rawls does not help his own case by a somewhat cavalier use of the term “moral.” So what would it mean to treat political philosophy as “applied moral philosophy”?

V. Why Political Philosophy is Not Applied Moral Philosophy: The Discontinuity Thesis

Someone who believed that this was the right way to proceed would start with principles that apply to all human conduct, regardless of whether the conduct takes place in a political setting, and use these principles to evaluate political institutions and the laws and policies that they produce. Candidates here might be utilitarianism and libertarian rights theory: these purport to supply guidance to (private) individuals when deciding how to act, but also to provide standards by which institutions and public policies can be assessed. A question might be raised, though, about the direction of travel: If the principles in question are held to be perfectly general in their application, why assume that they treat political philosophy as applied moral philosophy, rather than the other way round? It is sometimes suggested that utilitarianism, for example, should be regarded primarily as a legal/political philosophy and only secondarily as a moral philosophy, on the grounds that laws, policies, and institutions can be meaningfully assessed by reference to their overall consequences for human welfare, whereas in the case of individual actions this is impossibly demanding, both cognitively and motivationally (rule-utilitarians of course have a response to this challenge). It is true, nevertheless, that when philosophers present these doctrines they usually do so by starting with individual-level cases and then going on to argue that the same principles must apply when institutions are being evaluated. Thus, libertarian rights-theorists typically begin with state-of-nature examples involving individuals in order to establish the general shape of the theory, and then proceed to investigate what kind of state (if any) might be consistent with the principles so established. But does this show that the individual-level cases are more fundamental from a normative perspective, or is it more a matter of being able better to capture the intended reader’s intuitions by presenting the simpler cases first?

To make the “political philosophy is not applied moral philosophy” charge stick, therefore, it would be necessary to show that there is something fundamentally wrong with unified normative theories of the kind referred to in the previous paragraph. One would need to establish what I will call a discontinuity thesis, which holds that political evaluation involves standards that are distinct from, and irreducible to, the standards that apply to individual conduct. We can throw some light on this possibility by looking at the concepts and principles that are used in practice to assess political institutions and policies. They can be divided into three classes. First, there are those that cannot be applied (except perhaps in a metaphorical sense) to individuals — they identify properties that only institutions can have. “Democratic,” “representative,” and “legitimate” are examples of evaluative terms that fall into this first category. ^{Footnote 37} Second, there are terms that can be applied both to individual behavior and to institutions and policies, but it is plausible to think that their meaning changes between the two contexts of application. This, for example, is Rawls’s claim about the kind of justice that chiefly concerns him when he describes it as a political value that applies to the basic structure of society: he is not denying that there are other forms of justice that should govern individual behavior. ^{Footnote 38} Something similar might be claimed about “freedom as a political value” as opposed to “primitive freedom.” ^{Footnote 39} Then third, there are principles that apply across both domains without any apparent change of meaning: utility, for example, or rights understood as moral claims of individuals.

To establish the discontinuity thesis, one needs to show both that principles in the first two categories play an essential role in political philosophy and that these principles are not derivative from more fundamental principles that belong in the third category. An illustrative argument of this kind might say that institutions must be evaluated at least in part by reference to their democratic character, while the value of democracy is not merely instrumental to some third-category value such as utility. There is a danger, however, that this may appear to turn the question into a simple first-order dispute about values: we imagine the democrat insisting that it is intrinsically valuable for an institution to have a democratic character, and his utilitarian opponent arguing that democracy is indeed relevant, but only because it tends to deliver policies that promote the greatest happiness. The issue, however, is not over whether democracy is more or less important than utility as a value. The issue is about the kind of principles it is appropriate to use in political philosophy. To establish discontinuity, we have to be able to show that there is something inappropriate about assessing institutions by appealing solely to principles that also apply to individual behavior; it is not necessary to prove that principles such as utility have no place at all in political evaluation, only that principles in the first two categories play an essential and irreducible part.

What features of politics might we cite in support of the discontinuity thesis? Political institutions and the laws and policies they produce affect the lives of large numbers of people in very significant ways, these people generally have no choice but to experience their effects, and are obliged to comply with what is enacted politically whether or not they agree with it. ^{Footnote 40} The combined effect of these features is to make political justification more demanding than moral justification in general; the justification given must be one that the people who are subject to political rule must be able to accept. As we have seen in relation to Williams’s “basic legitimation demand” (and the same applies to Rawls’s idea of an “overlapping consensus” on principles of justice), pinning down the exact meaning of “are able to accept” here is a difficult task. It hovers uneasily between “will in fact accept” and “ought to accept” (where the latter appeals to something outside of the subjective beliefs of the people in question). We know that as a matter of fact there will always be people who reject even the best-constructed forms of political authority. We want to be able to say that nevertheless they have good reason to accept these forms as legitimate, which implies that the justification we give hooks on to other things that they believe — so in rejecting the authority (or its outputs) they are guilty of an inner contradiction. I return to this dilemma briefly below. The relevance of the more demanding nature of political justification to the discontinuity thesis is that distinctively political principles (first and second categories in my schema), as opposed to general moral principles, are designed to play this justificatory role. They are principles that people ought to be able to accept given the general circumstances of politics outlined above.

One might still wonder whether political justification, of the kind just outlined, can also be philosophical. It may sound too pragmatic, insufficiently foundational. When the death of political philosophy was being announced earlier on, one of the charges laid against past political philosophy, so-called, was that it was nothing more than ideology dressed up as philosophy. It had claimed to provide foundations for political principles of a kind that in the nature of the case could not be supplied. If valid political principles, for a particular society, are principles that people within it must be able to accept, this seems to make political justification too dependent on what people are as a matter of fact willing to regard as legitimate — whereas philosophy is supposed to be in the business of putting received beliefs under the critical microscope. So political philosophy might be seen as impossible, not for the general reason given back in the heyday of logical positivism that any form of practical philosophy was impossible, but for the more specific reason that the nature of political justification is such that philosophy cannot engage with it. However this depends on a thesis about the nature of philosophy itself that is open to contestation. It holds that philosophy must always operate sub specie aeternitatis, rising above the contingent belief-systems of particular societies and viewing them critically from afar. This is certainly a thesis that both Williams throughout his career and Rawls in the later stages of his were keen to reject. For Williams, historical understanding was essential to philosophy; in particular “philosophers cannot altogether ignore history if they are going to understand our ethical concepts at all. One reason for this is that in many cases the content of our concepts is a contingent historical phenomenon.” ^{Footnote 41} But this was no reason to stop engaging with these concepts philosophically:

The lesson I want to draw from this is that there is no contradiction between recognizing that usable political ideas must be ones that, in our societies, could serve to legitimate a political order — and which ideas pass that test will be a historically contingent matter — and subjecting those ideas to philosophical scrutiny. That in a nutshell is what political philosophy can be and should be.

VI. Conclusion

Let me now try to summarize the argument I have advanced. My aim has been to evaluate the charge that although political philosophy is no longer “dead,” the form in which it has been revived has turned it into a subordinate branch of moral philosophy — and this is a mistake. The contrary claim is that the very nature of politics prevents it from being studied (normatively) in this way. I first considered what I called the “hyper-realist” view that politics is essentially nothing but a struggle for domination, and argued that this is both excessively narrow, and actually misses what is most distinctive about politics itself. At least in standard cases, struggles for power occur within a framework that is generally recognized as having authoritative status, and that includes many forms of interaction besides bare coercion. ^{Footnote 43} Since alongside attempts to dominate by means of threats and so forth, politics also involves persuasion, argument, and the giving of reasons — for and against laws, policies, procedures, institutional designs, and so forth — it is amenable to normative assessment of the kind that political philosophy provides. I turned next to the more plausible form of realism advanced by Williams that tries to draw normative requirements out of the very idea of politics itself — politics again contrasted with brute domination or coercion. Williams is right, I believe, to make the issue of legitimacy central to politics, and therefore to political philosophy, but wrong to suggest that legitimacy can be understood without reference to moral values. On the one hand, the goods whose provision is a necessary condition of political legitimacy are identified through moral reflection on the basic needs of human beings. On the other hand, to explain why it matters that political power should be exercised through institutions that satisfy the “basic legitimation demand,” a moral principle such as respect for persons must be invoked. Moreover, Williams fails to show why satisfying the legitimation requirement for a liberal society cannot take the form of advancing a full-blown theory of justice such as Rawls’s, but must instead involve something more minimal such as “the liberalism of fear.” In the end, Williams’s complaint about Rawls — the accusation that he is a “political moralist” — seems to have less to do with the substance of Rawls’s theory than with the way he chooses to present it, particularly in his earlier writings. I suggested that the charge of “moralism” could only properly be levelled against philosophies that apply general moral principles directly to politics without taking into account the justificatory burden carried by exercises of political power — a burden that explains why distinctively political principles, shaped by a concern for legitimacy, are also needed for political evaluation.

Since we do indeed encounter such philosophies, the realist critique, in the form presented by Williams, and indeed to some extent by Rawls when he insists that a viable liberalism must be “political,” not (morally) comprehensive, has some bite. It is a valid criticism of much recent “political philosophy” that by ignoring questions of legitimacy altogether, ^{Footnote 44} it becomes merely aspirational at best; at worst it advocates policies that, if implemented, would make it difficult for liberal societies to reproduce themselves over time, for example by ensuring that enough of their members were committed to liberal values. ^{Footnote 45} Here, Williams’s reminders that the first political question can never be solved once and for all, and that liberalism in particular must be flexible enough to respond to the differing circumstances in which it may come under attack, remain salutary.