Political Philosophy

What is Political Philosophy?

Charles Larmore, Brown University

1. Two Rival Conceptions 
The question in my title receives far less attention than it deserves.  Often the domain of political philosophy is defined by a series of classic texts (running from Aristotle’s Politics, past Hobbes’ Leviathan, to Rawls’ A Theory of Justice) along with a conventional list of the problems to be addressed – the acceptable limits of state action, the basis of political obligation, the virtues of citizenship, and the nature of social justice.  Precisely this last problem, however, shows why the question “What is political philosophy?” ought to have a greater urgency.  For justice is a topic that also belongs to moral philosophy.  How therefore are moral philosophy and political philosophy to be distinguished?  Both have to do with the principles by which we should live together in society.  How exactly do they differ?  If justice – to invoke a traditional tag as indisputable as it is uninformative – means giving everyone his due (suum cuique), then what is it to fill in the import of this phrase as a moral philosopher and to do so instead from the standpoint of political philosophy?

            These questions are not motivated by a general love of intellectual hygiene.  I do not assume that all the different areas of philosophy need to be cleanly demarcated from one another, in order to avoid contamination by alien concerns and influences.  In my view, disciplines arise in response to problems and the boundaries between them have whatever rationale they possess to the degree that different problems can be handled separately from one another.  The difficulty lies in the problems themselves with which political philosophy typically deals, particularly when the idea of justice comes into play.  Let us say, in again a rather vacuous phrase, that political philosophy consists in systematic reflection about the nature and purpose of political life.  Nothing puzzling in that, it would seem.  Yet political philosophers have tended to tackle this subject in two quite different ways, depending on how they position themselves with regard to the domain of morality.

The one approach understands moral philosophy to be the more general discipline, dealing as it does with the good and the right in all their manifold aspects, and not just in the realm of politics.  Political philosophy forms part of this larger enterprise, focusing on the class of moral principles that have to do, not with our special relationships to others, but with the shape our social life should have as a whole.  One of its primary themes is therefore justice, and justice regarded as a moral ideal, conceived in abstraction from the exigencies of practice.  The aim is to specify the relations in which we ought ideally to stand to one another as members of society, possessed of the appropriate rights and responsibilities.  Only once this basis is secured does political philosophy move on to take into account existing beliefs, motivations, and social conditions.  For then the ideal must be adjusted to reality, particularly given the limitations, both empirical and moral, on what may be achieved through the coercive power of the law.  None of this changes, however, the standpoint from which political philosophy begins and must judge these very concessions, namely the moral ideal of the good society.

The other approach sees political philosophy as an autonomous discipline, setting out not from the truths of morality, but instead from those basic features of the human condition that make up the reality of political life.  People disagree and their disagreements extend from their material and status interests to their very ideas of the right and the good, so that society is possible only through the establishment of authoritative rules, binding on all and backed by the threat or use of force.  These are the phenomena on which political philosophy must always keep its eye.  Certainly it has a normative aim, seeking to lay out the fundamental principles by which society should be structured.  But it carries out this project by asking in the first instance what principles ought to have the force of law.  Though these principles may well coincide with part of morality, that is not in itself their justification.  For political philosophy, their validity has to be judged by how well they handle the distinctive problems of political life, which are conflict, disagreement, power, and authority.  On this view, the very heart of justice lies in determining what rules may be legitimately imposed on the members of society.

I mentioned at the outset what appears a rather empty definition of political philosophy:  systematic reflection about the nature and purpose of political life.  But perhaps this is not such a platitude after all.  For the difference between the approaches just outlined seems to turn on which of the two terms receives the greater weight.  Should political philosophy look first and foremost to the purposes that ideally political association ought to pursue?  Or should it set out instead from the nature, that is, the reality, of political association, which is that interests conflict, people disagree, and without the institution of law and the exercise of state power no common existence is possible?  Depending on the point of departure adopted, political philosophy becomes a very different sort of enterprise.  Either it forms a branch of moral philosophy, working out what ideally the good society should be like, or it operates by principles of its own, propelled in no small part by the fact that moral ideals themselves prove politically divisive.  The difference, I insist again, is not that the second approach is any less normative by virtue of taking as its starting point the permanent features of political life.  For it understands these givens as constituting the problems to which political philosophy must work out the appropriate solution.  However, the principles serving to determine that solution are essentially political in character, defining the legitimate use of power.

The opposition between these two approaches is not unfamiliar.  Sometimes philosophers feel moved to endorse what is effectively the one line and to decry the other.  But their professions of faith are seldom accompanied by much argumentation or by any attempt to analyze the supposed errors in the contrary view.  Two recent exceptions are G.A. Cohen and Bernard Williams, advocates of rival sides of the issue, who expounded their positions at some length (though without, unfortunately, ever mentioning the other).  “We do not learn what justice fundamentally is,” Cohen declares, “by focusing on what it is permissible to coerce…. Justice is justice, whether or not it is possible to achieve it.”[1]  For Williams, by contrast, “political philosophy is not just applied moral philosophy, which is what in our culture it is often taken to be…. Political philosophy must use distinctively political concepts, such as power, and its normative relative, legitimation.”[2]  How should political philosophy approach the notion of social justice:  following Cohen as a moral ideal that is independent of questions about legitimate coercion, or following Williams as a political ideal that is inseparable from such questions?

I shall examine in some detail the views of these philosophers as I go on to pursue the question before us.  But one fact I want to underscore already is that for both Cohen and Williams, as for many others, the choice between the two conceptions appears stark and inescapable.  Political philosophy, they presume, cannot avoid deciding in the end which of the opposing paths it will follow.  This is a mistake.  The contending standpoints I have been sketching – “moralism” and “realism”, as Williams termed them to his own advantage – are not the only options.  Both of them contain an important element of truth, but both are also unsatisfactory, and to remedy their failings political philosophy needs to move beyond this habitual opposition.  Its subject must indeed be the characteristic problems of political life, including the prevalence of moral disagreement.  Yet it cannot determine how these problems are to be addressed except by reference to moral principles understood as having an antecedent validity, since serving to determine how the authoritative rules of society are to be established.  Political philosophy must be a more complex enterprise than either of the customary positions assumes, if it is to combine these two dimensions.

2. Philosophy and History

Before going on to develop this conception, I must pause to say something about the nature of philosophy in general.  It is a topic I broach with mixed feelings.  Though undoubtedly important, it lends itself all too easily to the pursuit of ulterior ends.  Typically, a definition of philosophy comes to little more than an expression of specific preoccupations and commitments, themselves quite questionable on philosophical grounds, but disguised as an impartial demarcation between what is “really” philosophy and what is not.  Think of the idea that the object of philosophy is the conditions of possibility for experience, or the idea that it consists in conceptual analysis.   I am myself, to be sure, engaged in explaining how one ought really to do political philosophy.  But my intention is not to suggest that the positions I oppose fail to qualify as “philosophical”; the claim is instead that they fail to get it right about the “political”.  Still, understanding aright the general goal and method of philosophical reflection helps to explain the particular view of political philosophy I adopt.  It is probably not possible to talk about the nature of philosophy without saying in the end something philosophically controversial.  I shall begin at least on neutral ground.

Philosophy, I believe, following Wilfrid Sellars, is the effort “to understand how things in the broadest possible sense of the term hang together in the broadest possible sense of the term”.[3]  Philosophy aims at comprehensiveness of vision, at making sense of the connections between the way the world is and our various dealings with it, including the very project of rendering these interconnections intelligible.  Its ambition, we might say, is to be maximally reflective:  philosophy differs from the other kinds of human inquiry in that it seeks to uncover and evaluate the background assumptions on which they implicitly rely.  Even when it concentrates on some specific area, as in the philosophy of art or indeed in political philosophy, the concern is with the very makeup of this domain, its fundamental features and the human purposes it engages.

This definition is, of course, extremely general.  It tells us little about the direction in which such reflection should go, and different philosophers will proceed differently, in accord with their particular views and interests.  However, I want to mention one way the practice of philosophy takes on concrete form that philosophers themselves are prone to overlook, in a kind of endemic self-misconception – though here I am obviously turning toward the philosophically controversial.

In endeavoring to make sense of how things hang together, either overall or in a particular area, reflection has to find some footing.  It needs to draw upon existing knowledge and past experience, if it is to have any grasp of the problems it must handle and the avenues it should pursue.  The same thing holds when the philosopher turns to challenge some widespread assumption, arguing that it is actually unfounded or less fruitful than commonly presumed.  The resources for criticism have to come from what can count as settled in the matter under review.  Philosophy is therefore always situated, shaped by its historical context, even as it aspires to make sense of some subject in as comprehensive, as all-encompassing, a way as possible.

Now this inherent historicity of their enterprise is something that philosophers have a hard time acknowledging, despite its being what one should expect, given that reflection, however broad its scope, needs somewhere to stand if it is to see anything at all.  The resistance stems from an easy misunderstanding of the philosophical ambition to see things whole.  Here is how it arises.  First one observes that human inquiry, in the problems it tackles and the solutions it devises, always bears the mark of its time and place.  The modern natural sciences are no exception.  Though they develop through the testing of hypotheses against evidence, hypotheses and evidence alike reflect the theories of the day, the experimental procedures available, the course of previous inquiry, and sometimes the influence of wider social forces.  The sciences, however, give little heed to their historical context; their attention is directed toward the problems to be solved.  Philosophy is different.  It does not aim merely at solving problems.  Devoted as it is to being maximally reflective, philosophy asks why certain problems have come to constitute problems at all.  And thus it sets about tracking the ways that history shapes the thought and action characteristic of whatever domain it is endeavoring to understand.  Nothing wrong in that, to be sure.  But the temptation is to imagine that turning the weight of history into an object of reflection must mean aspiring to escape its sway.  In yielding to that notion, philosophy presents itself as the effort to discover what Reason itself, addressing us simply as rational beings independent of historical context, requires us to think and do.

I need not detail the different forms this aspiration has taken over the centuries.  It is a common refrain in philosophy and yet, for all that, an illusion.  There can be no transcendence of history, however comprehensively we reflect.  Still, the appropriate response to this fact is not to conclude that philosophical reflection must surrender its goal of understanding how everything hangs together.  That reaction is equally flawed, since it perpetuates the underlying mistake.  The mistake is the supposition that to the extent that our beliefs are historically conditioned, in philosophy or in any systematic inquiry for that matter, they fail to be reliably geared to the world as it really is.  The contingencies of history are not necessarily obstacles.  They are the very route by which finite beings like us gain a systematic access to truth, and that means to timeless truth, for there is no other kind.  One all too easily, but wrongly, assumes that only if we make our thinking itself timeless can it mirror faithfully the timelessness of truth.  Standing back from history, if we could manage to do it, would simply leave us nowhere to stand.  In reality, a sense of our historical situation, freed from the supposed antithesis between history and truth, gives philosophical reflection a clearer hold on the actual resources at its disposal.

I shall not go any further here in defending these views.[4]  They will guide me, however, in the arguments that follow.  The reason for this detour has been to explain why I shall not hesitate to suppose that what we have learned through history about the nature and purposes of political life may prove importantly relevant to determining how political philosophy itself should proceed.

               

3. Two Pictures of Political Society

 There have been, I observed, two competing conceptions of political philosophy.  The one sees it as that part of moral philosophy whose aim is to lay out the principles of the ideal society, while the other regards it as centered on those enduring features of the political realm – conflict and the need for authority – that stem not solely from divergent interests, but also from the right and the good being themselves a constant object of disagreement.  I also suggested that fueling this dispute have been opposing ideas about whether the purpose or instead the nature of political life should provide the point of departure for philosophical reflection.

That was, however, a rather superficial remark, at best a first approximation to what is really at stake.  For one thing, the nature of any human association, its typical activities and relationships, involves the way it actually pursues some set of purposes.  But in addition, we cannot determine the purposes it ought to pursue except by relying on some such notion of its nature.  I am not alluding to the sort of consideration that often goes under the name of “ought implies can”, a principle that (in some of its many meanings) I in fact reject:  are not the basic demands of morality binding on all even if people may at the time lack the motivational capacity to comply, and does there not figure among these demands some that no one can fully satisfy, but only honor more or less well?[5]  However, my concern here is that without an idea of the aims and practices some association embodies, we would not know the kind of association it is and would thus be in no position to pronounce on the purposes it ought to have.  Unless you know what normally goes on in banks, you cannot say what a bank, as opposed to a supermarket, ought to do.  Even when we are imagining an association that does not yet exist but would, we believe, serve to realize some desired end, we lean on assumptions about how it would function in practice.  Otherwise, we would have no basis for thinking that it would be such as to achieve the goal in question.

All this goes to show that the idea of political philosophy as devoting itself to the moral ideal must still presuppose some picture of what political life is like, though it is bound to be very different from the one assumed by the rival conception.  These two underlying pictures offer, in fact, a useful basis for tackling the theoretical debate that I have sketched.  They serve to orient the different conceptions of political philosophy, and where they prove defective, doubts must arise about those conceptions themselves.  Moreover, they make up in their own right another well-known opposition.  Often they are identified simply by the names of the thinkers who have provided their canonical formulation.  On the one hand there is the Aristotelian view of politics, and on the other the Hobbesian or Weberian view.  The divergent associations such phrases evoke show how familiar this dispute too has become, and thus I can rehearse the main features of these two views of political association by reference to the figures I have just mentioned.  As we proceed, however, shortcomings will emerge in both models, and they will indicate why neither of the rival conceptions of political philosophy is ultimately satisfactory.

            According to the one view, then, political life is the highest, most comprehensive form of human association since its principal aim is to promote the ultimate end of all our endeavors, the human good itself.  Such is the position we encounter in the opening pages of Aristotle’s Politics.[6]  None of us, he remarks, can live well by living alone, for we have not the self-sufficiency of gods.  Only in society are we able to obtain and make use of the means indispensable to a flourishing existence (eudaimonia) – the material resources needed to sustain our different activities, the education that steers us in the right direction, and the public space in which to deliberate together about how best to organize and pursue this collective enterprise.  The various areas of social life, such as the family or the economy, are oriented toward attaining different parts of the human good.  Political life is not simply one more kind of association alongside the rest, devoted to yet another particular goal.  It encompasses all the others (pasas periechousa tas allas) as the most authoritative (kuriotate) kind of association since its task is to ensure, through the just distribution of the necessary resources and opportunities, that our lives flourish as a whole.  This supremacy of political association comes to expression in the fact that the rules of justice it establishes (as opposed to those that may obtain within more limited social groups, such as the family) take the form of laws, coercively binding on all.  However, the nature of justice, precisely because it is an ingredient of the human good, represents a prior standard to which law, so far as possible, aims to give institutional shape.  This, it should be clear, is the picture of political society presupposed by all those, from Aristotle on, who have held that political philosophy needs to house itself within the larger enterprise of moral philosophy.

            Very different is the view of political society we find in the writings of Hobbes and Weber.  I begin with Weber since his account provides the starkest contrast, rejecting all reference to ends and defining the political in terms of means alone.  A group, he begins by observing in Wirtschaft und Gesellschaft, may determine its membership either through voluntary agreement or by imposition, that is, by stipulating which individuals are subject to its rules independently of any consent on their part.  Imposition (Oktroyierung) need not be by way of coercion.  Some religious organizations (the Catholic Church, for instance) assert authority over their flock from the very moment of birth even though, at least in the present day, their authority is solely spiritual and lacks the means of enforcement.  But when, Weber adds, the group does impose its rules on individuals by the use or threat of force, it becomes a political association, and if it successfully claims a monopoly on the legitimate use of force within a given territory, it acquires the particular form of a state.[7]  There are different ways that states have sought to legitimate the power they exercise, including the appeal to ideals of the human good that they may purport to be advancing.  But what makes them political in character and distinguishes them from other groups pursuing such ideals is the possession of the coercive means to implement whatever goals they happen to adopt.

            Now though this view of the nature of political society is quite familiar, Weber’s focus on means to the exclusion of ends overdraws the difference with the Aristotelian view.  In reality, if only implicitly, Weber is attributing to the state a particular end in portraying it as an association that claims a monopoly on the legitimate use of force.  The fundamental business of politics, he is assuming, lies in the establishment of order, securing through the rule of law the conditions for civil peace and social cooperation.  Nonetheless, this emendation scarcely narrows the gulf that separates the Weberian picture from the idea of political association as aiming essentially at justice and the human good.  To the extent that such ends become the object of state action, the form they take, according to this picture, is determined by the primary political goal of creating and maintaining social order, through coercion if necessary.  Justice as such cannot be the state’s concern.  It has to be justice insofar as it can be part of an authoritative set of rules that will be binding on all.  If this view of political society regards the establishment of order as the paramount end, the reason is not hard to divine.  It sees the potential for conflict everywhere in social life:  in the clash of interests, to be sure, but also in people’s differing notions of the right and the good.

Such is, of course, the idea of the political that animates the second conception of political philosophy.  It should also be plain how widespread this idea is, and who its premier theorist has been.  That is Hobbes, for whom the business of the state is, as he said, the “safety of the people”, guaranteed by a “common power to keep them all in awe”.[8]  Hobbes’ writings are unsurpassed, moreover, in the clarity with which they identify where these two pictures of political society differ.  The ultimate point of contention, he explained, is the kind of norms that are most important in shaping its activities and institutions.  The one view looks to ethics, the other to law.  “It is characteristic of man,” according to Aristotle, “that he alone, among living beings, has a grasp of good and evil, of the just and the unjust, and association based on these things makes a family and a state”.  To which Hobbes rejoined, “Where there is no common power, there is no law; where there is no law, no injustice”.[9]  For people, he insisted, tend naturally to disagree about what is good and evil – “many men, many different rules for virtue and vice” (quot homines tot virtutis et vitii diversae regulae) – so that law alone can provide a “common standard” (communis mensura).[10]  These, I believe, are the different perceptions of its subject matter that steer political philosophy in the opposing directions we have distinguished:  toward the pursuit of moral first principles or toward the need for order and authority.

4. The Centrality of Authority

Historically, the Hobbesian idea of political society has come to prominence in modern times, in reaction against the other, more idealized picture.  So an important question is why this has been so.  One sort of answer is very common.  We meet it, for instance, in the essay that Leo Strauss published under the same title as my own, though the answer is not limited to the circles that tend his shade.  The Hobbesian view has prevailed, Strauss claimed, because of a lowering of expectations and standards that is typical of modernity as a whole.[11]  The aim has been to take a more realistic approach, to describe political life as engaged in replacing conflict with authority instead of fostering virtue and the good life, in order that it may more easily live up to its vocation.

No doubt this answer captures part of the truth.  Yet it misses another motivation of enormous importance.  I mean the recognition that disagreement about the nature of the right and the good is not merely a recurrent phenomenon and not solely the expression of inexperience and error, but the natural outcome of people reasoning freely and conscientiously about how one should live.  The realization that moral disagreement is likely to be the outcome when reasonable people discuss among themselves the ultimate questions of life, or when a single person ponders them in his own mind, has been one of the seminal experiences of modernity.  “By sowing questions and dividing them up,” Montaigne wrote, “one makes the world flourish and teem with uncertainty and disputes… Difficultatem facit doctrina [learning creates difficulty]… Never have two men judged similarly of the same thing, and it is impossible to find two opinions exactly similar, not only in different men, but in the same man at different times”.  Or as Hobbes himself declared (echoing Montaigne), “Divers men, differ not only in their judgment, on the senses of what is pleasant, and unpleasant to the taste, smell, hearing, touch, and sight; but also of what is conformable to reason, in the actions of common life.  Nay, the same man, in divers times, differs from himself, and one time praiseth, that is, calleth good, what another time he dispraiseth, and calleth evil”.[12]

The important element in this outlook is not the mere realization that opinions vary.  That people often disagree about ethical issues is a fact of life known from time immemorial.  Aristotle himself began his Nicomachean Ethics with a survey of the competing notions of the good life (pleasure, honor, wealth, virtue, knowledge, and the various combinations of these).  What Aristotle did not envision, however, and has become a leading theme in modern thought is that reasonable people, exercising their general capacities of reason on questions having to do with how the individual should live and society be organized, tend to come to different conclusions – not through any defect of reason, but instead because of their different backgrounds, their different senses of what is salient, their different ways of weighing opposing considerations.  It is the realization that reason does not lead naturally to unanimity on these questions that represents the break with the past.  Aristotle understood that political society must be organized so as to settle the various conflicts that inevitably arise among its members.  Yet for him these conflicts were generally conflicts of interests, sometimes conflicts of opinions, but not conflicts in which reasonable people find themselves at odds about the nature of the right and the good.  That is why he could so confidently look to ethics as the source of the principles of political life.  Hobbes, by contrast, had a broader sense of the sources of social conflict:  reason itself can lead to divergent views.  That is why he and those following his lead have turned to law, not to ethics, as the foundation of political authority.

The expectation of reasonable disagreement is not, I should caution, the same as the doctrine often called value pluralism, which holds that the ultimate sources of the good and the right, objectively speaking, are not one but many.  Nor is it tantamount to skepticism, which claims that the proper response to intractable controversy is to suspend judgment or at least to regard one’s own position as more an article of faith than an item of knowledge.  Pluralism and skepticism are reactions to the pervasiveness of reasonable disagreement:  they are attempts that aim, respectively, at explaining or at evaluating the phenomenon, and consequently form prominent movements in the history of modern thought.  But they too have become subjects of dispute in turn.[13]

The crucial fact is then reasonable disagreement itself, in all its breadth.  For once one appreciates how it pervades the moral realm, the idea that the aim of political association is the collective pursuit of the human good must look out of touch with reality.  Its central task, so the Hobbesian outlook insists, is instead the construction of an authoritative order for the regulation of social conflict.  Again, no one would deny that adjudicating conflict is a political imperative.  But in the Aristotelian picture, this work takes place by reference to an ideal of justice whose principles are discernible by moral reasoning from the character of the human good.  If, however, the good and the right, human flourishing and justice, prove themselves to be an object of conflict among reasonable people, their import for the rules of political life cannot count as already settled by reason, at least not in terms that all accept.  Their political meaning has to be determined by law and the institution of authority becomes the paramount concern.  To be sure, the Hobbesian picture of political society gives no less attention to the conflicts that arise from material and status interests.  But it is the recognition of widespread moral conflict that constitutes its distinctive feature and impels its break with the Aristotelian model.

5. Beyond Moralism and Realism: A Critique of Williams

Experience has taught us, I believe, that this picture offers a far better portrait of political society, not just in modern times, but throughout human history.  We have learned to expect that in a free and open discussion reasonable people are likely to disagree about justice and the human good – if not about basic matters (the injustice of slavery, the elementary rules of morality such as keeping one’s promises and respecting the bodily integrity of others, the importance of both agreeable experiences and actual achievement in a life lived well), then certainly about more complex questions as well as about the underlying principles that justify these judgments and explain the nature of the right and the good.  The controversies about the makeup of human flourishing are notorious.  But justice is no less contentious a subject.  One need only recall the unending debates about whether individual desert, the general good, or the equal worth of each citizen should provide the basis for determining the distribution of material resources.  Consensus in these areas, we suspect, is likely due to people failing to have thought deeply enough, to have listened to what others are actually saying, or to have escaped external pressure or internal inhibition.  Whatever unanimity may have reigned in pre-modern societies seems likely to have arisen from the operation of such factors.

As a result, the corresponding conception of political philosophy, centered on the problems of conflict and authority, seems to me superior to the view that regards it as principally engaged in mapping out the structure of the ideal society.  The moral ideals to which the latter view appeals are bound to prove controversial, forming part of the problems of political life, rather than providing the basis of their solution.  Political philosophy cannot therefore consist in applied moral philosophy.  It has to be a more autonomous sort of discipline, dealing with a domain shaped by the deep disagreements to which moral thinking itself so often leads.  This conclusion depends, of course, on an understanding of what we have learned through history about the nature of political society, about what it has become in modern times and about what, as we now see, it would have been all along, had the conditions for free and open discussion prevailed.  However, as I argued earlier (§2), an historical sense does not compromise, but instead enhances the prospects of philosophical reflection.

Only up to a point, however, do I accept the Hobbesian picture of political society and the corresponding idea of political philosophy.  For both are incomplete in a crucial regard.  If political society rests on an authoritative order for the regulation of social conflict, what is it that makes such an order “authoritative”?  The direct answer to this question is that the order is authoritative if it enjoys legitimacy in the eyes of the society’s members, that is, if they generally regard it as justified.  (Recall that the state, in Weber’s definition, is a form of association possessing, not simply overwhelming force to exact compliance with its rules, but rather a monopoly on what is perceived to be the legitimate use of force).  But now the question arises whether any political system can secure the perception of legitimacy without justifying its rules or its power to make them by appeal to principles of justice it asks to be understood as having a validity independent of the political order itself and as prescribing the shape such an order ought to have.  I do not think it can.  For the conditions under which coercive power may rightly be exercised are surely one aspect of the concept of justice.  In this respect, the Hobbesian picture proves defective and political philosophy, as it sets about exploring how political life ought to be organized, cannot avoid taking its bearings from morality.

Consider, for instance, the way in which Bernard Williams lays out such a picture of political society, in order precisely to recommend his “realist” conception of how philosophy should proceed in this area (see §1 above).  “The first political question”, Williams asserts, is “in Hobbesian terms … the securing of order, protection, safety, trust, and the conditions of cooperation”, but no state can do so without, he adds, satisfying the “basic legitimation demand”, which requires that the state “offer a justification of its power to each subject”.[14]  This BLD, as he dubs it, is not to be misunderstood.  It does not necessarily require that the legitimation be one each person can see reason to accept in the light of an understanding of himself and others as free and equal members of the society.  That is the particular form the justification must take in what may broadly be called the modern liberal world.  But there have been other terms in which states have sought to legitimate themselves without addressing their subjects as free and equal citizens:  stories about mythical foundings, about the monarch being God’s representative on earth, about the leader being the voice of the Volk.  The point is that some account justifying the state’s power must be given to each of its members so that he or she will have a reason other than mere fear to obey its commands.  Otherwise, the person is not really part of the society, but instead a foreigner or even an enemy, happening to live within its borders (like the Helot population of ancient Sparta).  Over such people the state may be able to exercise its power, but not its authority.

To defend his “political realism”, Williams turns to the question whether the BLD is itself a moral principle and, if so, whether he has therefore failed to escape the “political moralism”, the supposed “priority of the moral over the political”, that he deplores.  His reply is that “if it is, it does not represent a morality which is prior to politics.  It is a claim that is inherent in there being such a thing as politics”.[15]  This answer misses the crux, however.  It is not so much the BLD as rather the justification of authority to which the BLD appeals, whatever it may be, that must express a “morality prior to politics”:  it has to embody an idea of what constitutes a just political order – in particular, of what constitutes the just exercise of coercive power – and that is a moral conception whose validity must be understood as antecedent to the state’s own authority because serving to ground it.  I do not say, obviously enough, that this idea must be correct, that the moral principles invoked must truly be valid.  But they must be regarded as being so if they are to be taken as justifying the state’s exercise of coercive power.

Williams likes to declare that the liberal notion of legitimacy (as I have very roughly characterized it) arises from the BLD working in conjunction with the historical conditions known as “modernity”.[16]  That is true enough.  But “history” is not a black box.  What it must in this case contain, among other things, is the conviction that political rule is legitimate only if those whom it is to bind, understanding one another as free and equal citizens, can see reason to endorse the fundamental rules by which it operates.  This conviction does have a history, and Williams is right to complain about a tendency among liberal philosophers to imagine that their notion of political legitimacy is a deliverance of Reason itself, instead of the result of historical experience.  Yet not only does its historicity do nothing to undermine its validity (see §2 above); it manifestly expresses a moral principle that serves to define the authority a state may rightly assert to govern the conduct of its citizens.

None of this implies, however, that we find ourselves back with the opposing standpoint.  The moral principles on which political philosophy must rely do not consist in some comprehensive vision of the human good.  Nor does the conception of justice they must contain refer, as one may say, to justice as a purely moral ideal – that is, to justice as it would be apart from any concern for how it ought to be made authoritative in society.  Rather, these foundational moral principles have to do with the conditions of permissible coercion.  They form the basis of what may be called political justice – justice insofar as it can rightly have the force of law and shape the life of a political association.  The view of political philosophy I am outlining departs therefore from both the conceptions with which we began.  It sees reasonable disagreement about the right and the good as integral to the problem at the center of political life, so that political philosophy cannot simply be applied moral philosophy.  But it also understands that any solution to this problem has to anchor itself in undeniably moral principles governing the legitimate use of coercion.[17]

6. The Relation between Political and Moral Philosophy

Several aspects of this position call for clarification, in order to obviate some possible objections.

(i) I have asserted that the moral principles to which political philosophy must appeal belong, not to justice as a purely moral ideal, but to what I termed “political justice”, since they lay out the conditions under which it is proper to hold people subject to the coercive rules that make political association possible.  The emphasis is on the “must”.  I do not mean that political philosophy cannot go further, that it ought not to work out full-fledged theories concerning the liberties, opportunities, and resources that citizens should enjoy.  Of course it may, and I need not mention the works of this sort that are among the glories of the discipline.[18]  But such theories must include as a constraint on whatever else they affirm about justice an account of the conditions under which it can rightly count as authoritative and its principles may rightly be imposed on the members of society.  That is what makes these theories into theories of “political justice”.

(ii) Any understanding of justice has to say something about the proper province of the law.  For law is coercive in character, and the extent to which individual conduct may rightly be subject to coercion is undeniably a question of justice.  So long, however, as the focus is “justice as a purely moral ideal”, one need only say (with Aristotle, for instance) that laws are justly instituted if they serve the cause of justice, if they preserve or bring about that overall distribution of essential social goods which can antecedently be ascertained to constitute each person’s due.  It is simply a case of application.  To make the idea of “political justice” basic entails a more complex view.  Then, the conditions under which the rules of collective life ought to be made authoritative circumscribe from the outset the very nature of justice.  The relation of justice to law ceases to be one of mere application, since what justice is must now be defined in terms of how it may justly acquire the force of law.

Behind this approach, so I have argued, lies the modern realization that reasonable people tend naturally to disagree about fundamental aspects of the right and the good.  Once we acknowledge that reason does not speak with a single voice on these matters, we cannot regard the conception of justice that we ourselves see most reason to endorse in the abstract as necessarily the one that ought to define the terms of our political life.  However convinced we may be of that conception’s superiority, we have to reckon with the likelihood that it will prove controversial among other people equally reasonable.  We need therefore an account of the conditions under which a conception of justice may rightly become authoritative, an account that goes beyond our saying that the conception is the one that reason recommends, since others can say the same about their rival conceptions.  It has to explain, on the basis of moral principles determining the conditions of permissible coercion – principles that constitute, as it were, “second-order” principles of justice – why a particular conception of justice should enjoy the force of the law amidst reasonable disagreement about what justice involves.  Political justice is thus inherently reflexive:  it delimits what justice means, politically speaking, in the light of the fact that the very nature of justice is widely controversial.  The reason why in my view political philosophy ought not to take the form of applied moral philosophy is that it needs, despite its dependence on moral principles, to incorporate this sort of reflexivity.

(iii) I need to be clear, however, about what this approach does not entail.  First of all, it does not signify that thinking about “justice as a purely moral ideal” is a wrong-headed or futile enterprise.  Describing what ideally should be each person’s due, apart from considering what would make such a scheme binding, remains an important part of moral philosophy.  The point is that political philosophy needs to proceed differently:  justice as a political ideal must specify the conditions under which its terms can rightly be made authoritative.

Secondly, these conditions do not necessarily have to embody a liberal notion of political legitimacy of the sort I sketched earlier (§5).  There is no inconsistency in espousing a conception of political justice that denies that the fundamental principles of political life are binding only if they treat those subject to them as free and equal citizens.  The liberal conception does have a special relation to the modern experience of reasonable disagreement:  it affirms the value of people exercising their reason by their own best lights, even at the price of deep and widespread disagreement, inasmuch as it holds that basic political principles ought to be such that citizens understanding one another as free and equal can, despite their disagreements, all see reason to endorse them.  But this relation is moral, not logical in character.  One might very well believe that precisely because reasonable people disagree about the nature of justice political association should govern itself, not by respect for individual reason, but instead by conformity to God’s will.  Such a belief departs from the liberal perspective in virtue of its moral assumptions.   

(iv) In the last two sections §§5-6, I have often referred to the idea of legitimacy, but I should be more precise about what it means as well as about its relation to the idea of justice, since the two concepts are not synonymous.  First, by a state’s “legitimacy” I mean the fact that it is indeed justified in the coercive power it wields.  That is manifestly not the same as its exercise of power being generally perceived to be justified, a circumstance which suffices to make it authoritative, but not legitimate or, as I have also said, rightly authoritative.  Some theories of political legitimacy tend to confuse the two.  But when people accept a state as legitimate, and when a state claims legitimacy for itself, they are holding that the state’s exercise of coercive power really is justified, not merely that it is viewed as being so; that the appeal is to actual, not simply perceived, legitimacy is no less true when the justification consists in some version of the “consent of the governed”.  Equally clearly, the legitimacy story promulgated by a state and accepted by its members may be based on error, illusion, or intimidation, in which case the state’s power, though it may be authoritative, is unlikely to be actually legitimate.  Nonetheless, whatever its quality, the attempt to justify the state’s exercise of coercive power has to base itself on principles assumed to have a validity antecedent to, because purportedly justifying, the authority claimed for the state.  These are principles of justice, since they stipulate the conditions under which coercive power may rightly be claimed and deployed, and they accordingly embody a “morality prior to politics”.  Such has been my objection to Williams’ “realism”.

Legitimacy and justice are not the same thing, however.  Laws may be just (in their content) without having been legitimately enacted and legitimately enacted while failing to be just.  Still, the concepts do intersect at two points, each having to do, not with what it is for laws, but rather with what it is for a political regime to be legitimate.  First, as I have argued against Williams, no regime can claim to be legitimate without appealing to some conception of what constitutes a just political order.  (Naturally, such a regime may still proceed to institute some laws that happen to be unjust).  Second, it is clearly a part of justice to determine to what extent laws, being coercive in nature, may fairly be imposed on the members of society.  Since a state is legitimate insofar as it is justified in exercising its coercive power over its members, legitimacy in this sense forms an essential part of the idea of justice.  The connection between legitimacy and justice is therefore twofold:  to be legitimate, a regime must not only aim, with whatever success, to conduct itself in accord with a conception of justice; it must also actually conform to that part of justice that defines the conditions under which the state may rightly exercise its coercive power.

(v) I have argued in effect that political philosophy needs to understand justice in a way that satisfies not one but two conditions:  given the prevalence of reasonable disagreement about the right and the good, it has to recognize that determining what justice is means specifying the conditions under which any conception of justice may correctly count as authoritative, but at the same time it needs to realize that any account of these conditions must be rooted in moral principles delineating the legitimate use of coercion.  These two demands may seem to conflict.  Will there not always be some reasonable people who disagree with those foundational moral principles?

The apparent antinomy disappears once it is admitted that no political association, however legitimate, can be infinitely accommodating.  Any principle of legitimacy excludes some views about the human good or about justice, even if there are reasonable people – by which I mean (see §4 above) people exercising conscientiously their general capacities of reason – who happen to espouse them.  The distinctive feature of modern liberalism has been its effort to organize political life around a respect for the tendency to reasonable disagreement, a respect that requires that the fundamental rules of political life be acceptable to all who, whatever else they may believe, share a fundamental commitment to treating one another as free and equal citizens.[19]  Such is the principle of legitimacy on which a liberal polity relies as it guarantees basic individual freedoms and provides for democratic procedures to decide, among other things, the more complex and controversial aspects of social justice.  This principle, whose authority must be understood as antecedent to political association, governs how the state is to use its power in responding to the conflicts among its citizens.  It has been a liberal hope that citizens, despite their moral differences, can nonetheless agree on this principle of political legitimacy, a hope that liberal societies have done much to realize through moral and civic education and through the peace and prosperity they have achieved over the years.  But it is a hope, not a prediction or an assurance.  Many ideas of the good and the right, to which reasonable people have found themselves drawn, are excluded by this principle of legitimacy.  Consider once again the idea that the most important feature a political association must display is that it be pleasing to God, whether or not it meets with acceptance from its members – though this idea needs to be distinguished from the belief, itself tremendously important in the history of liberal thought, that God’s will is precisely that political life take shape in accord with principles its members can endorse as free and equal citizens.

7. Justice and Human Nature: A Critique of Cohen

With these clarifications in place, I turn to the objections that G.A. Cohen raised against any view of political justice like the one for which I have been arguing.  “We do not,” he insists in the declaration I quoted before (§1), “learn what justice fundamentally is by focusing on what it is permissible to coerce”.  The coercive power of the state, he concedes, may well be needed to deter malefactors and to assure citizens of one another’s compliance with the law.[20]  But legitimate coercion, according to Cohen, has to be understood as implementing a conception of justice that is already in place.  The conditions of permissible coercion, he assumes, cannot delimit what justice itself is since it is the nature of justice that tells us what those conditions are.  What Cohen fails to appreciate, as we now can see, is that if reasonable people disagree about the nature of justice, some account is necessary of what should count under these circumstances as the authoritative conception of justice.  Such an account is itself a part of justice, and since it consists in moral principles determining what may be legitimately imposed, justice in this respect does serve to define the conditions of permissible coercion.  However, these conditions circumscribe in turn the rest of what justice can be, politically speaking.  In this second respect, therefore, justice depends on what rules of social life may rightly have the force of law.  The combination of these two elements makes up the reflexive character of political justice to which I referred earlier (§6.ii), and which stems from a recognition of the depth of reasonable disagreement about the right and the good.

But let us look more closely at Cohen’s position.  Behind his idea of the relation between justice and legitimate coercion stands an allegiance to one of the two conceptions of political philosophy I reject, not to Williams’ realism to be sure, but rather to its opposite, the position that political philosophy ought in the first instance to lay out the nature of social justice as it is “in itself”, independently of any concern with the problems involved in making an idea of justice a political reality.  In Cohen’s view, there is a world of difference between “fundamental principles of justice” and what he calls “rules of social regulation”, in whose sphere belong problems such as that of permissible coercion.  The former define the fair distribution of the desirable elements of social life; the latter deal with how social life is to be arranged in the light of the facts of human nature and people’s beliefs and motivations.  Rules of regulation involve therefore the application of what are presumed to be fundamental principles of justice, and it is a mistake, Cohen claims, to confuse the two, as he argues that John Rawls, for instance, systematically does when allowing his theory of justice to be shaped by a regard for such empirical factors as the so-called “circumstances of justice” (limited scarcity, limited altruism, people’s conflicting ends and purposes).[21]

            Yet how solid is Cohen’s distinction between fundamental principles of justice and rules of social regulation?  I concur that there are some facts about what people want and believe on which it cannot be supposed that principles of justice depend.  No one should think that the fair distribution of resources has to respect some people’s exceptional greed, their wish to be richer than their neighbors, or their prejudices with regard to various religious or ethnic groups.  One may also wonder whether Rawls was right to let the terms of economic justice be shaped by the self-interest of the more talented and productive members of society, rewarding them (via his “difference principle”) with greater wealth so that they will have the incentive to engage in those activities whose effect is to make the least well off better off than they otherwise would be.  This was the starting point of Cohen’s critique of Rawls, and I do not deny its considerable force.  But Cohen pushes his critique to the extreme.  Can it really be true, as he asserts, that the principles of justice, as opposed to rules of social regulation, are not grounded in any facts at all about human nature?  According to him, their validity, or at least that of the most fundamental among them, is utterly “fact-independent”.  Such is the nature of “that elusive virtue discussed for a few thousand years by philosophers who did not conceive themselves to be (primarily) legislators and who consequently had a different project”.[22]

What illustrious philosophers of the past can Cohen have had in mind?  It is not surprising that someone such as myself, holding that the tendency of reasonable people to disagree about important moral questions must serve as a constraint on what justice can be, sees no basis for the dichotomy as he defines it between principles of justice and rules of social regulation.  But not even a champion of justice as a purely moral ideal such as Aristotle would have endorsed it.  He thought it laughable that the gods would care about justice (“making contracts and returning deposits”) since they face none of the practical problems that human beings do.[23]  In fact, the dichotomy collapses in Cohen’s own hands.

Equality, he holds, lies at the heart of social justice and in a sense that “endorses deviations from equality if and only if the unequally placed parties are relevantly responsible for that deviation”.[24]  But he is keen to add that justice also includes a “personal prerogative” entitling the individual to pursue his own interests to a certain extent:  “justice is fully served only if people’s access to desirable conditions of life is equal, within the constraint of a reasonable personal prerogative”, since we have “the right to be something other than an engine for the welfare of other people: we are not nothing but slaves to social justice”.[25]  These remarks are certainly sensible.  Yet note that Cohen attributes this “prerogative”, whatever its precise extent, to the very nature of justice; it is not invoked as another kind of moral consideration that moderates the pursuit of justice in the full sense.  What, then, is this conviction that justice should not require that we become its slaves, if not an admission that justice would not be justice if it did not specify the conditions under which, given various facts about human life and motivation (namely those justifying the prerogative), people can justly be expected to comply with its demands?  Angels, from what I hear, are of a nature that would make them quite happy and fulfilled to be nothing other than engines for the welfare of others.  Cohen’s laudable wish to avoid “moral rigorism” entails the dissolution of his grand distinction between fundamental principles of justice and rules of social regulation.

8. The Nature of Principles

Cohen presents something like a master argument for why there must be fact-independent principles, among which he places the fundamental principles of justice.  But can any principles of conduct really be fact-independent?  I do not think so, and the difference between my conception of political philosophy and the one I have associated with Aristotle turns, I believe, on which facts are attributed a grounding role.  That would have been Aristotle’s own opinion, but not Cohen’s, who champions therefore an extreme version of the conception I am rejecting.  However, I want to look closely at Cohen’s master argument, even if it is rather abstract, since its underlying mistake is instructive.

The argument goes as follows:

                        (1) Whenever a fact F confers support on a principle P, there is an

 explanation why F supports or represents a reason to endorse P.

                        (2) This explanation must invoke or imply a more ultimate principle, valid

 independently of F, and able to explain why F supports P.

                        (3) The sequence cannot proceed without end because our resources of

 conviction are finite.

                        (4) Thus, there must be some ultimate, fact-independent principles.[26]

This argument exhibits a well-worn pattern, and that alone ought to give us pause:  A depends on B, but that dependence must depend on C, which dependence must in turn depend on D … until we must come – for the regress cannot be infinite if it is to exist at all – to a prime mover, or to an ultimate fact-independent principle.  However, I do not propose to dismiss Cohen’s argument by means of guilt-by-association.  Nor shall I plead that the regress can go on unto infinity or resort to vague notions about justification needing no terminus since it is always a “holistic” affair of how a set of beliefs “cohere” together.  My basis for rejecting this argument is a lot simpler.

            First, consider what a principle is, in the sense under discussion.  Principles are general rules of thought and action, asserting that certain lines of conduct are what we have reason to adopt in the sorts of circumstances they stipulate.  Principles refer to what we may call standing reasons for how to think or act:  the principle that one ought to help anyone in distress is tantamount to the assertion that one has as a rule good reason to do so.  Thus, the question of whether principles are grounded in facts comes down to the question of whether reasons are so grounded, and that means to the question of what a reason is.  This question is notoriously a disputed area in philosophy, but here is how I see the matter.[27]

Reasons are both normative and relational in character.  They consist in the way certain facts in the world count in favor (a normative relation) of certain possibilities of conduct:  my reason to carry an umbrella consists, not simply in the fact that it is raining, but in this fact counting in favor of the option of taking an umbrella.  Reasons may sometimes turn out to be prima facie – presumptive demands that upon reflection we conclude do not really have a claim on us.  They are also usually pro tanto – if truly applying to us, still able to be outweighed by what may on balance be superior claims.  Reasons become standing reasons when they are such as generally to prevail over other competing considerations.  Yet whatever the case, reasons do not float free of the (non-normative) facts, but depend on the facts being as they are.  So one wonders how principles could be any different, how there could be such a thing as Cohen imagines, a fact-independent principle.

True, sometimes certain facts count in favor of a line of conduct – that is, they give one reason to adopt it – only in virtue of a general principle that bestows on them that status.  But this need not always be so.  We can see a reason to take an umbrella some rainy day without relying on any general principle that makes it a reason, such as that one ought always to stay dry, since at other times we may not mind getting wet.  Explaining why the present occasion is special may come to no more than detailing the various ways it differs from other rainy days (we are late for work, we are tired, the sky looks particularly dark, it has been raining for several days now, etc).  Though we would have to agree that, were another day to come along just like this one, we would also have a reason to take an umbrella, that acknowledgement does not express an allegiance to some principle.  One should not confuse the universality inherent in any reason as such with an underlying principle of the sort in question, namely with an explanatory principle.  Every reason is universal, in the sense that it applies in a given situation only if it applies in all relevantly similar circumstances.  Yet the statement “in all situations of type S, one has reason to do A” is not a principle that explains why a particular situation of that type gives one the reason to act in the way indicated.  It merely spells out what is involved in there being such a reason at all.  A reason just is certain facts in the world (e.g., those having the features that make them of type S) counting in favor of certain of our possibilities (e.g., A).

Now if all this is true of reasons, the same must apply to principles as well.  Some principles may be grounded in certain facts because of underlying principles that make that so.  But sometimes too, principles may be grounded in facts without there being any deeper principle that explains the grounding, if the standing reasons to which they refer have that sort of basis.  In either case, principles are always grounded ultimately in facts, because reasons are so.  Cohen, I suspect, holds otherwise because he supposes in general that the only way something non-normative, the facts, can generate something normative, a reason or a principle, is through the intermediary of some further normative element.[28]  But that supposition amounts to a misunderstanding of what normativity is:  it is the way that non-normative facts count in favor of possibilities of thought and action.  If one finds that puzzling and thinks it stands in need of explanation, the source of puzzlement must be the idea that there can exist such a thing as reasons at all.  Reasons are by their very nature fact-dependent, and principles cannot be any different.  The principle “one ought not to cause pain” is grounded in what the experience of pain is like.  We can say, if we like, that the principle would remain valid even if there were no beings capable of feeling pain.  But all that this statement means is that were any to exist, its ability to feel pain would give one reason to take the necessary precautions.  There is a fact about what pain feels like if it happens to occur, and this fact grounds the principle that one ought not to cause pain.  If pain did not feel the way it does, the principle would lose its point.[29]

It will not do to argue that whenever a principle P is grounded in certain facts F, the statement “if any situation contains facts of type F, then one ought to act in accord with P” – or more succinctly “in situations of type F, one ought to act in accord with P” – will hold independently of there actually being any such facts F and therefore expresses a principle, call it P′, that is, to this extent at least, fact-independent.[30]  For P′ is not a principle that explains why facts of sort F ground the principle P.  It is simply a statement to the effect that they do so.  Moreover, the sense in which P′ is fact-independent is that there need not exist any facts of sort F for it to be true that, were they to exist, they would ground principle P.  So this maneuver does nothing to disprove that principle P is ultimately grounded in certain facts of the sort F.   

An example of an ultimately fact-dependent principle may help to make my objection clearer.  Take the simple principle of prudence, which tells us (rather loosely formulated) to weigh together our long-term with our proximate good when deciding how to act.  Counting in favor of this principle are the basic facts of the human condition – that we are more than creatures of a moment, that life is full of unexpected changes, that present pleasure can lead to later pain.  If it is now asked why these facts ground that principle, what is there to say?  If it is insisted that they can do so only in virtue of some deeper principle, what could that principle be?  That one should pursue one’s good in a way that respects all the facts, both present and future?  This is simply prudence all over again.  No, either one sees, in the light of these facts, the standing reason to be prudent or one does not.  Nor should that be surprising if one remembers a general truth:  reasons, being the way that facts count in favor of possibilities, cannot be such that, for the facts to count in that way, there must always be some deeper reason to explain why they do so.

What I reject, then, is the very first premise of Cohen’s argument – that if a fact F supports a principle P, there must be an explanation why F represents a reason to endorse P.  Rejecting the initial premise is, in general, the best way to handle regress arguments of this sort.  To Aquinas’ argument that there must be a prime mover since, whatever moves requiring a mover, the chain cannot go back to infinity, the proper response is to object at the outset, with Galileo at our side, “But it is just not true that whatever moves requires a mover”.

9. Conclusion

            The facts therefore matter, even for philosophy.  There is no way that political philosophy can determine what principles ought to govern our collective life except with an eye to its enduring realities.  Yet it must also look to morality in order to judge how these realities are to be handled – though moral reasons too depend on certain facts being as they are.  When we put these two dimensions together in the way I have proposed, the nature of political philosophy looks quite different from either of the standard conceptions with which we began.  For though remaining rooted in moral philosophy, political philosophy cannot be profitably described as the application of that discipline to political reality.  That is because it has to adopt a more reflective stance than is usual in moral philosophy.  It has to figure out what is to be done in face of the fact that moral philosophy, and moral thinking in general, so often proves controversial and divisive.  Herein lies what makes for the autonomy of political philosophy, what makes it more than just a part of the supposedly more general discipline of moral philosophy.  You have your moral views, I have mine, and each of us is convinced that he is right, standing ready to show the other the error of his ways.  But once we confront the problem of how people like us are to live together, we enter the terrain of political philosophy.

---

* In the summer of 2009, several weeks before his sudden death, Jerry Cohen very kindly gave me an extensive set of comments on an earlier version of this paper.  We remained at loggerheads.  But I wish to express here my gratitude and sadness.

[1] G.A. Cohen, Rescuing Justice and Equality (Cambridge, MA: Harvard University Press, 2009), pp. 148 and 155.

[2] Bernard Williams, In the Beginning was the Deed (Princeton: Princeton University Press, 2005), p. 77.

[3] Wilfrid Sellars, Science, Perception, and Reality (London: Routledge & Kegan Paul, 1963), p.1.

[4] For a detailed defense, see my book The Autonomy of Morality (Cambridge: Cambridge University Press, 2008), chapter 1, “History and Truth”.

[5] For my doubts about the principle, see Patterns of Moral Complexity (Cambridge: Cambridge University Press, 1987), pp. 33, 57f., and The Morals of Modernity (Cambridge: Cambridge University Press, 1996), pp. 84ff., 149.

[6] Aristotle, Politics I.1-3

[7] Weber, Wirtschaft und Gesellschaft I.1.2.12-17

[8] Hobbes, Leviathan, Introduction and I.xiii.8.

[9] Aristotle, Politics 1253a15-20 and Hobbes, Leviathan I.xiii.13.

[10] Hobbes, De homine, XIII.8-9.  See too Leviathan IV.xlvi.32.

[11] Strauss, What is Political Philosophy? (Chicago: University of Chicago Press, 1988 [1959]), pp. 40ff.

[12] Montaigne, Essais III.13 (“En semant les questions et les retaillant, on faict fructifier et foisonner le monde en incertitude et en querelles… Difficultatem facit doctrina…  Jamais deux hommes ne jugerent pareillement de mesme chose, et est impossible de voir deux opinions semblables exactement, non seulement en divers hommes, mais en mesme homme à diverses heures” – ed. Villey [Paris: Presses Universitaires de France, 1999], p. 1067), and Hobbes, Leviathan I.xv.40.

[13] On the differences between reasonable disagreement, pluralism, and skepticism, see my book The Morals of Modernity, chapter 7.

[14] Williams, “Realism and Moralism in Political Theory”, op. cit., pp. 3-4.

[15] Ibid, p. 5.

[16] Ibid, pp. 8, 10.

[17] Here and throughout I am speaking of justice insofar as it has some import for social life, as either “justice as a purely moral ideal” or as “political justice”.  Sometimes, however, we talk about justice in a more expansive sense, as when we say that it is unfair that someone was not born in an earlier age more suited to her talents, or died before he had time to demonstrate his promise.  Cosmic justice is not the present topic.

[18] Yet there may also be reasons why political philosophy should hesitate to pursue such comprehensive theories, including the fact that reasonable people can disagree so deeply about what justice requires.  See Amartya Sen, “What Do We Want from a Theory of Justice?”, Journal of Philosophy CIII 5 (May 2006), pp. 215-38, esp. 223-25.  Instead of pursuing the “transcendental” approach of laying out the contours of “the” just society, one might do better, Sen argues, to proceed “comparatively”, examining what social arrangements are more just than others.  But in either case, one ought, I would add, to specify the conditions under which coercively binding rules may legitimately be imposed.

[19] I discuss this idea of liberalism in greater detail in my book The Autonomy of Morality, chapter 6.

[20] The quoted passage (Cohen, Rescuing Justice and Equality, p. 148) continues thus:  “if coercion is necessary only for deviance or assurance reasons”.

[21] For Cohen’s account of the difference between such principles and rules, see Rescuing Justice and Equality, pp. 269, 276ff.  Rawls is his target throughout the book, though I come in for glancing criticism at p. 148 (ftn. 65).

[22] Cohen, op. cit., p. 304.  For the claim that principles of justice are completely “fact-independent”, see pp. 278, 285.

[23] Aristotle, Nicomachean Ethics 1178b8-18.

[24] Cohen, op. cit., pp. 310-311 (ftn. 51).  This is his so-called “luck-egalitarianism”; I take no stand here on its merits.

[25] Ibid., pp. 181 and 10.  On p. 61, he says that only an “extreme moral rigorist” would deny such a prerogative.  Much of Cohen’s debate with his critics (see pp. 373-411 of Rescuing Justice and Equality) revolved around the scope of this prerogative.  What seems not to have been recognized is that to make it integral to justice as Cohen does undermines his dichotomy between fact-independent principles of justice and fact-dependent rules of social regulation.

[26] Ibid, pp. 236-237.  I have followed his wording closely.  It is important to see that Cohen intends this argument to focus on what is involved in a fact serving to ground or support a principle.  That is not the same as the extent to which facts may determine the content or the range of applicability of a principle.  Thomas Pogge’s critique of Cohen (“Cohen to the Rescue!”, Ratio XXI 4 [December 2008], pp. 454-475) seems to overlook this distinction.

[27] I present a more detailed account of the nature of reasons in The Autonomy of Morality, especially Chapter 5, §§6-7.

[28] That may well be because he also assumes that for one thing to ground another it must, in conjunction with other perhaps implicit premises, entail that conclusion.  See David Miller’s critique of Cohen’s argument on this score in “Political Philosophy for Earthlings”, pp. 29-48 (31-38) in David Leopold and Marc Stears (ed.), Political Theory: Methods and Approaches (Oxford: Oxford University Press, 2008).

[29] Cohen suggests (op. cit., p. 245) that this very principle is fact-independent, claiming that its validity does not depend on the fact that sentient beings exist.  He fails to consider the fact on which it is in reality grounded, namely the fact of what pain feels like.

[30] Sometimes Cohen argues in this fashion, as when (op. cit., p. 235) he declares that if the principle P₃, “we ought to express our respect for people”, is grounded in the fact that “people possess what are thought to be respect-meriting characteristics”, then only because of the “fact-free principle P₄”, that “one ought to respect beings, human or otherwise, who have the relevant characteristics”.