I. Our Humanity

Morality is what makes us human. One meaning of this common saying is plain enough. Refraining from injury to others, keeping our word, and helping those in need constitute the elementary decencies of society. If most of us did not observe these practices most of the time, or at least give one another the impression of doing so, no one would have the security to pursue a flourishing life. Even a life of basic dignity would be impossible if we found ourselves continually at the mercy of aggression, treachery, and indifference. Morality makes us human by providing rules of mutual respect without which there can be neither social cooperation nor individual achievement.

However, another meaning suggests itself as well. It has to do not with morality's function, but with its source. Other animals are like us in being able to show deference and feel affection, even to the point of sacrificing themselves for those whom they love. But morality, insofar as it involves looking beyond our own concerns and allegiances in order to respect others in and of themselves, lies beyond their ken. Does not then our very ability to think morally point to a peculiarly human power of self-transcendence, a power that we alone among the animals have of regarding ourselves from the outside as but one among others, and that finds in morality, if not its only, then certainly its most striking expression? This question engages our attention far less than it should. When people, philosophers included, wonder about the nature of morality, they tend to focus on what reasons there may be to be moral, what acting morally entails, or in what sense, if any, moral judgments count as true or false. All of these are important issues. But often the taken-for-granted deserves the greatest scrutiny. That we should be able at all to view the world impersonally, recognizing the independent and equal standing of others, involves an overcoming of self that is no less remarkable for having become largely second-nature.

Among philosophers, Immanuel Kant was one of the few, and certainly the most famous, to argue directly that morality is, in this sense, what makes us human. “Duty! Sublime and mighty name, what is an origin worthy of you?” Only a freedom, he replied, that “elevates man above himself as a part of the world of sense … a freedom and independence from the mechanism of all of nature.”Footnote ¹ Our moral consciousness, according to Kant, testifies to the freedom we alone have to rise above all that experience has made of us, so that we may act in accord with demands we understand as binding on us independently of our given interests and desires. I believe, like many, that Kant was on the track of an essential truth. But, like many too, I do not believe that the source of morality can be anything so extravagant (if morality itself is not to be an illusion) as a freedom unshaped by the course of experience. Freedom, in any form we can conceive, depends not only on external conditions proving conducive to our ends, but also on our having acquired, through training and effort, the abilities necessary for exercising control over ourselves and the world. After all, we have to learn how to think morally, which means developing a sense of social expectations as well as the self-discipline needed to distinguish the good of others from our own, or from what we wish their good would be.

Thus, Kant's intuition needs to be brought down to earth. One proposal might be that the capacity for self-transcendence to which morality gives expression consists in our nature as normative beings, responsive not merely to the causal impress of the environment, but to the authority of reasons as well. To see that we have a reason to think or act in a certain way is to see that we ought to do so, all other things being equal, and heeding an “ought”—as is clear with the moral “ought”—means holding ourselves to a demand we regard as binding upon us. Why is not this subordination of self the self-transcendence at issue?

Certainly morality is not possible except for beings that can respond to reasons. Yet the “ought” is not limited to the moral realm, any more than it is true that the only reasons for action we have are moral in character. Most importantly, the subordination of self involved in the recognition of reasons as such falls importantly short of the way that morality asks us to look beyond ourselves.

Suppose, for instance, that we are pursuing some interest of our own or the good of some people we hold dear, and doing so solely because that interest or those people matter to us. The reasons we then perceive to act one way rather than another are ones we would have to agree that anyone similarly disposed would have in such circumstances. All reasons are universal, binding on one person only if binding on all of whom the same conditions are true that make them applicable to the first. That is what it is for reasons to be binding, and why responding to reasons means holding ourselves responsible to the authority of an “ought,” distinct from our individual will. Nonetheless, this subordination of self is not the self-transcendence that morality demands. For the reasons in question only apply to us because we care about those ends. Their authority, though real, remains conditional.

Imagine, however, that we think we ought to help someone simply because that person is in distress. Perhaps it is someone whom we happen to hold dear, but we are attending, not to the love we feel, but only to the person's suffering. Reasons of this sort are not just universal, as all reasons are. They are impersonal as well, applying to us in abstraction from our own interests and affections. Or, more exactly, their applicability does not depend on any such personal factors insofar as the latter are identified as ours, since it is possible to have an impersonal reason to satisfy a desire of ours if the desire, understood solely as belonging to a person like any other, turns out to warrant attention.Footnote ² In short, impersonal reasons derive from the way things are in and of themselves—as, in the imagined case, from another person's actual condition. And it is just this impersonal element that constitutes the hallmark of moral reasons for action. Thus, acknowledging the force of reasons is not by itself sufficient to impel us to look beyond ourselves in the way that morality requires. Nor is it, incidentally, a distinctively human capacity. Many of the higher animals also have, or can be trained to have, a sense of how they ought to behave. Yet moral thinking, the grasping of impersonal reasons, remains beyond their reach.

This last point suggests the direction of a more promising approach. Though other animals can respond to reasons, they appear for the most part unable to reflect about what they ought to do, evaluating and weighing the reasons they see to favor one option or another. Reflection seems basically beyond their reach, if by “reflection” is meant standing back so as to regard ourselves from the outside, as we regard someone else, in order to figure out who we are or what we really ought to think or do. Why is it not then our capacity for reflection to which morality gives special expression? Are we not better able to consider others as persons in their own right, apart from the personal concerns that color our perception of the world, the more reflective distance we achieve toward ourselves? Of course, as that question implies, reflection need not always be impersonal. Sometimes we reflect by asking ourselves what someone close to us would think about our situation. But herein lies a strength of the approach I propose. For now the freedom that is integral to morality becomes intelligible. It is the ability to overcome the hold that a natural absorption in our own affairs exercises over our thinking in order to see the intrinsic value in the good of others, and it manifestly grows out of the less impartial ways we have of standing back from ourselves. Impersonal reflection is a creature of experience, and a more sensible answer than Kant's to the question of what constitutes our peculiar power of self-transcendence.

Morality is not, of course, the only example of impersonal reflection (any more than it could be the only example of the nonempirical freedom that Kant postulated). When we reflect upon what we should believe about a certain matter, and ask not what we are inclined to suppose or what our friends or community would conclude, but rather what the facts themselves require, we are reflecting impersonally. For our target then consists in impersonal reasons for belief—reasons (in parallel to impersonal reasons for action) that we grasp as binding on us independently of our desires and loyalties in virtue of being based on what we know to be true about the subject matter itself. Morality, however, provides a privileged illustration of our capacity for impersonal reflection. For what could be a more conspicuous expression of our ability to stand back from our own attachments, as though we were merely one person among many, than to consider other people's good as of equal moment with ours, particularly when it is to our personal disadvantage?

In this essay, I want to examine more deeply the way reflection serves as the source of our moral thinking. How is it that by viewing ourselves from without, we can learn to see others as having an equal claim on our attention? Reflection, as I have said, need not always assume an impersonal form. What, then, is involved in its coming to exhibit the sort of self-transcendence that morality demands? Obviously, the place to begin is the nature of reflection itself.

As may already be apparent, and will become plainer as I proceed, reflecting about what to believe does not, to my mind, differ substantially from reflecting about how to act—except, of course, in subject matter. Grand distinctions are often made between theoretical and practical reason, particularly in the Kantian tradition. But they are largely overdrawn. (In this regard too, as in many others, the rationalism I espouse departs radically from Kant's.) Reason is best understood as the ability to respond to reasons, be they reasons for belief or for action, and the point of reflection, as I shall explain, is to consider explicitly what we have reason to think or do in regard to some problem that has disrupted our settled routines. Though the relation between reflection and morality is my ultimate concern, the next two sections will therefore look at the nature of reflection along quite general lines. I will not be losing sight, however, of what is involved in reflecting about how to act. It is precisely this comprehensive approach that reveals the true character of practical reflection, the common structure and function it shares with reflection about what to think and believe. So far from the practical and the theoretical constituting two disparate realms, all reflection is essentially cognitive in nature, aiming at a knowledge of the reasons there are. This general account will guide the more specific analysis of moral thinking to which I then return in Sections IV and V. There my concern will be to show how our capacity for impersonal reflection shapes the makeup of the moral point of view, its preconditions and implications. Because this capacity sets us off so dramatically from all the other animals, morality can indeed be said to form a signal expression of our humanity.

II. The Nature of Reflection

It is characteristic of the human condition that we rarely exist at one with ourselves. Often we feel torn between competing commitments. Or no sooner do we make some decision of moment than we recall our doubts about whether a different option might not be better. To deliberate honestly is to risk having a mind divided, for our inner conflicts seldom amount to mere confusions, vanishing in the wake of a more careful scrutiny. Generally, they mirror the real complexity of our situation, the multiplicity of demands that rightly exercise a hold on our attention. Even when we conclude that we have good reason to take one path rather than another, we leave behind something of ourselves in the possibilities we reject.

Yet conflict is not the only or the deepest way in which we fail to coincide with ourselves. Even more fundamental is our continual alternation between doing and reflecting. We move back and forth between two standpoints, the view from within and the view from without. Because each of us has a life that is ours alone to live, we naturally approach the world in the light of the interests and allegiances that happen to be ours. So long as everything goes its customary way, we think and act from within our own perspective. Yet we are also often moved to reflect on our thought and action—not for the pure pleasure of doing so, but because some problem has emerged that puts into question the way we have been proceeding. We are no longer clear about the sort of person we are or about what we should believe or do with regard to some other matter. Reflection is the response to a problem, the attempt to reestablish a fit between expectation and reality.

Reflecting also means, however, ceasing to live our lives from within in order to look at ourselves from without. To be sure, we stand back so as to examine the particular difficulty that has arisen. But even if it is something quite circumscribed (some trait of ours and not the shape of our life as a whole), and even if it concerns an object distinct from us (the real character of someone else or the confusing nature of some natural or social phenomenon), it is still ourselves, and not merely the difficulty, that we regard from the outside. As a response to a problem that has disturbed our routine, reflection is always a turning back upon ourselves, since it aims to figure out what we are to do about the difficulty before us. It requires us, moreover, to look at ourselves from without, since ascertaining what we ought to do means discovering what solution we have reason to adopt. Because reasons are necessarily universal (if they are binding on one, then they are binding on all under similar conditions), we cannot determine how we should proceed except by considering what anyone in our shoes ought to do. Naturally, we do not forget that in the case at hand we are the target, since our question remains, “What should we do?” But we approach this question (in effect, if not in so many words) by asking ourselves how anyone like us ought to think or act. When we simply respond to reasons, without reflecting, we do not take this sort of detour. But deliberating about what reasons we may have is a different matter. That is the essence of reflection, and it requires us to regard ourselves as one person among others—from without, though not necessarily, as will become clear, from an impersonal distance.

Reflection, therefore, always involves self-distancing. It requires us to adopt a third-person attitude toward ourselves, as though we were a “him” or a “her.” This fact, incidentally, points to an important truth about the pursuit of self-knowledge. Reflection need not have self-knowledge as its aim, of course, since often we reflect in order to discover what we should do or what we should believe about other matters in the world. When we do seek to know ourselves better, it is because some conflict has appeared between our conception of who we are and the behavior we notice, the desires we feel, or the unsettling remarks that others have made. And since we can only handle this problem by considering how someone like us ought to revise or deepen his self-understanding, we have no choice but to study ourselves as we would any other person, by collecting evidence and drawing inferences. Our knowledge of ourselves is then built up in the same empirical, trial-and-error way as our knowledge of others. We have no privileged access to the makeup of our own minds, which is why others can sometimes know us better than we do ourselves.Footnote ³

There are, in fact, two distinct dimensions along which we objectify ourselves whenever we reflect. First, in order to get hold of the problem we have encountered, we must adopt toward our own person the same sort of observational stance we occupy toward others when, living our lives from within, we regard their thoughts and feelings as among the elements of our environment. We look at ourselves as someone with a certain problem to solve, and we may have to study its ins and outs with care. But, second, we cannot determine what we, as such a person, ought to do in response except from a standpoint of evaluation that consists, not in everything we happen to believe and want, but in the standards and assumptions we think ought to govern our decision. This is as much as to say that, in deliberating about how we are to proceed, we examine ourselves through the eyes of someone we imagine as embodying just such an evaluative standpoint. In effect, we ask what such a person would conclude that we should do. The two kinds of self-distancing are quite distinct, of course: in the one, we describe ourselves as we are; in the other, we figure out what we are to do. It is in the way they work together that reflection functions as it does. To reflect is not simply to contemplate our own person, as though gazing at our reflection in a mirror, which is what the first kind of self-distancing alone would entail. We turn our attention toward ourselves in order to handle some obstacle that has disrupted our relation to the world.

Despite these differences between living our lives from within and reflecting on them from without, it would obviously be wrong to suppose that reflection stands opposed to life, or that it constitutes a luxury we might choose to forgo. We become the persons we are through the problems we confront. There is no end to the need for standing back, and we live as much outside ourselves, thinking about what we should do, as within the various activities we do pursue. Indeed, little in the way we see the world around us fails to show the mark of what we have learned by reflecting. That is why a concern for others, though it has its roots in reflection, forms an important part of our everyday lives. Internalizing what we have learned, we come to act with an eye to the well-being of those—family, friends, and associates—about whom we care enough to have considered how they feel. Similarly, we incorporate into our dealings with others moral principles we have acquired by reflecting impartially on how one should, in general, treat one's fellow man. What I am calling the view from within is not essentially self-interested.

Nonetheless, the business of living easily keeps us in the pull of our own orbit. This we often discover if we happen to reflect anew. For reflection knows no inherent limits. It allows us to make out the ways in which the very habits of mind we have acquired by reflecting still remain tied to the particularities of our life. Thus, we recognize, on taking a broader view, that our various loyalties and loves are likely to matter little to others with their own lives to live, their own ties and causes. Even the moral principles we espouse may appear a bit parochial or biased, once we consider the extent to which they have been shaped by culture or class.

As I have previously remarked, reflection can proceed from more than one type of standpoint of evaluation. It need not aim at being impersonal—that is, at judging how we ought to think or act irrespective of our own interests and attachments. We may, for instance, base our evaluation of the options before us on what we imagine some individual (real or fictional) whom we hold dear would do in our place, or would want us to do. Philosophers tend to neglect this mode of reflection, perhaps because they believe themselves to be above it, but they are certainly wrong to do so. All of us lean from time to time on various exemplars, internalized heroes and idols, to figure out how we ought to think and act. Moreover, modeling ourselves on others is not in itself a vice, as though the proper course were always to think on our own. Everything depends on the worth of the models to which we appeal. Though it is often simple chance, or some special allure, that has led them to represent for us a kind of person we would like to be, this too is not necessarily deplorable. We may not have the time or the means to check their reliability. And even when we do, some particular trait of theirs may later prove instructive in ways we could not anticipate when confirming their general value.

So little is identifying with others an appropriate object of disapproval, it plays an essential role in our coming to grasp the impersonal point of view.Footnote ⁴ The capacity for reflection does not spring full-blown from anyone's head. It develops over time. In early life, reflection upon who we are and what we should do consists, quite naturally, in imagining what those who are close to us would say. As our horizons broaden and we discover that parents and friends disagree, we find ourselves impelled to devise more abstract angles of evaluation: thus, we come to examine ourselves by the standards of some larger community to which we feel bound. But the same factors tend to push us beyond that perspective as well. And thus we may eventually fashion the idea of a fully informed and perfectly rational standpoint, transcending the limitations in the attitudes of particular individuals and societies—though, even then, identification is not at an end, since generally we do so by looking in our own culture for exemplars of such a standpoint to take as our models. Only through this sort of process do we learn what it is to hold ourselves accountable to an impersonal standard of thought and action. Only thus can we come to grasp the specifically moral point of view, which is to see in another's good, separately from our own interests and attachments, a reason for action on our part.

Nonetheless, whether we assume the outlook of some other individual, or reflect instead from an impersonal standpoint, certain elements remain constant. One is that reflection aims, in response to a problem, at determining what we have reason to think or do. Reasons, as I have pointed out, are universal in character: if they are binding on us, then they are binding on all who find themselves in conditions similar to those that make them applicable to us. It follows that even when we reflect, not impersonally, but by identifying with the standpoint of someone we esteem, our conclusions are still ones we must assume that anyone like us ought to endorse in such a situation. That is not an awkward implication. For would we adopt that standpoint if we did not presume it to be attuned to how people should really behave who have interests like those that impel us to take it up? If I pattern my wardrobe on what some movie star wears, I am assuming that he knows how people who want to be cool ought to dress.

Another constant is that, in reflecting, we have to rely on our current understanding of the world. We cannot regard ourselves from without except by continuing to think, at least in part, from within the perspective we presently occupy. Precisely because reflection is the response to a problem, it is always situated: only in the light of our existing views can we so much as identify the problem before us, and we cannot hope to handle it except by drawing upon the relevant information at our disposal. When we reflect by imagining what someone important to us would say, we make use (for example) of what we understand to be that person's characteristic habits of mind. So too, when we consider impersonally what we ought to think or do, we base our reasoning on the knowledge we have acquired, not only of the matter before us, but also, more broadly, of how belief should be proportioned to evidence or of how people are to be treated fairly. The impersonal point of view is not the view from nowhere. It always bears the mark of our time and place.

And yet, I must add, its claim to being impersonal does not thereby show itself to be a sham. Our access to reality as it is in itself is always mediated through the contingencies of history. This is not the place to enter into a discussion of epistemological questions, but the general position to which I am alluding is easily summarized.Footnote ⁵ Impersonal reasons, I have said, are reasons that stem solely from facts outside of us, uncolored by our own interests and attachments. Though reflecting on what impersonal reasons we have certainly entails appealing to our existing conception of those facts, our conclusions are still valid precisely to the extent that the beliefs on which we rely, for all their rootedness in the prior course of our experience, constitute knowledge of the matter at hand. To be sure, we may be wrong about the reliability of these beliefs. But that just means that our ideas concerning what impersonal reasons we have are always revisable, which is neither surprising nor ruinous. Impersonal reflection is, after all, an achievement, and, consequently, it is the subject of constant scrutiny for lingering traces of bias or distortion.

At this point, however, there arise some crucial questions. Does reflection, impersonal or not, really aim at truth? Is it an organ of knowledge, and if so, what can it provide knowledge of? These questions would appear to admit of a ready answer. We reflect in order to be better able to discern how we ought to think or act in the given circumstances, and that seems clearly to count as an object of knowledge. For it is something of which we begin by feeling ignorant and seek, by reflecting, to gain a correct grasp. What we ought to do is tantamount to what there is reason for us to do. So reflection, in essence, aims at knowledge of reasons for belief and action.

Such a conception results from taking literally the way we ordinarily talk about reflection, and in my view it ought to be no cause for alarm. But many disagree. In their eyes, it entails an untenable metaphysics and is, in any case, untrue to the nature of reflection. To suppose that reflection's relation to reasons is one of knowledge means not only attributing to the world a domain of irreducibly normative fact, but also imagining that knowledge by itself could ever move us to action as reflection manifestly does. Belief by itself is motivationally inert, and can lead us to act only in conjunction with some additional, conative element of the mind, such as a desire or a commitment. Reflection, it is therefore claimed, is a practical rather than a theoretical enterprise: when we reflect on what we have reason to do, our purpose is to settle how we want to live, not to discover some fact about the world. To this set of claims I now turn.

III. Reflection and Knowledge

Reasons for belief and action are essentially normative in character. What we have reason to do is what we ought to do, all else being equal. There is no explaining what is meant by reasons except by appeal to this or similar ways of speaking. Reasons cannot therefore be equated with any features of the natural world, physical or psychological, even though they certainly depend on the natural facts being as they are. That is why many philosophers balk at allowing that reasons can properly count as objects of knowledge. If knowledge is of what is the case independently of our coming to hold a view about it, then supposing that reasons figure among the things we can know entails that the world, as the totality of what is the case, must include normative facts about how we ought to think and act. Such a view runs counter to the naturalism that constitutes the reigning philosophical orthodoxy of our day. All that really exists, it is said, belongs to the domain of the natural sciences, the realm of physical fact or of psychological fact too, if the latter is not further reducible to the physical. As should be plain, I reject this sort of naturalism. But my main business here is to clarify the character of reasons and to explain why we should conceive of reflection as a way of acquiring knowledge about them.Footnote ⁶

Consider then, first, why it is that a reason cannot consist in a physical state of affairs. We sometimes say that the rain, or the fact that it is raining, is a reason to take an umbrella when leaving the house. Strictly speaking, however, my reason to do so is not the rain itself, but rather a certain relation that the rain bears to my possibilities of action. After all, one might agree that it is raining, yet dispute that this fact gives me a reason to take an umbrella. Only insofar as the rain justifies or counts in favor of carrying an umbrella does it take on the status of a reason, and this status—the relation of counting in favor of—is not a physical quality of the rain. It is a normative attribute. “The rain is a reason to take an umbrella” means the same as “Given the rain, I ought to take an umbrella.” To be sure, other things may count in favor of not taking one, of wearing a hat instead, or of just getting wet. As a rule, reasons are in themselves pro tanto, as are the “ought”s to which they correspond. Only if they are not outweighed by contrary considerations do they indicate what we overall have reason or ought to do. The point is that, whether pro tanto or decisive, reasons are not identical with the physical facts that may give rise to them.

So too, the reason people have to believe this or do that is not the same as any psychological state in which they find themselves. In particular, it is not, contrary to a widespread view, a combination of belief and desire.Footnote ⁷ My reason to take an umbrella, some would say, consists in my belief that it is raining conjoined with my desire to keep dry. But that cannot be right. It may be true that I would not have such a reason unless I wanted to stay dry, so that in this respect the reason is indeed conditional, binding only insofar as I have that desire. (Though it can also make perfect sense to say that I ought to take an umbrella against the rain whether I want to keep dry or not.) Even so, the reason does not consist in the desire as such, but rather in the fact that the desire counts in favor of taking an umbrella. After all, I might have the desire and still have no reason to take one, if, for example, the rain—unbeknownst to me—has ceased. Moreover, what makes it true that the desire favors that action is not my belief that it is raining, as though the reason were constituted by that belief conjoined in some appropriate way with the desire. The reason to take an umbrella depends in no way on my believing that it is raining. It is the fact that it is raining that gives rise to the reason. If you learned that the weather had just improved, you would inform me that I really have no reason to fetch my umbrella, even though clearly I still believe it is raining. In short, a reason is the possible object of a belief and not itself a mental state.Footnote ⁸

Recognizing that the reasons we have cannot consist in anything psychological, some have nonetheless supposed that the reasons that move us must belong to the mind. How else could it be said that reasons cause us to act? Whence a common distinction between “normative” reasons, which serve to justify an action, and “motivating” reasons, which are invoked to explain it.Footnote ⁹ Yet there are not in reality two different kinds of reasons. Talking about reasons as though they were motivating states in the mind is a bit of shorthand. Strictly speaking, the reference is to our view of the reasons we have, since in such cases only this attitude of ours, not the actual existence of those reasons, is thought relevant to the sort of explanation being sought. But the reasons we believe there to be are the reasons that appear to us to justify what we do. So-called motivational reasons, understood as psychological states moving us to act, are no more than our conception of the normative reasons we have. Moreover, sometimes there actually being such reasons, and not just our thinking they exist, does play an explanatory role—as when it turns out that we did the right thing for in fact the right reason. In such cases, reasons themselves lead us to act, via our correct ideas of them.

It is, then, their essentially normative character that precludes identifying reasons with any physical or psychological facts. As a consequence, they can have no place in the world itself as naturalistically conceived. This, as I have said, is one main line of argument leading many philosophers to deny, either implicitly or as a matter of doctrine, that reflection about what we ought to think or do can rightly count as a mode of knowledge. Conclusions about the reasons we have are not, it is held, really descriptive in intent, but serve instead to embody or announce our commitments for dealing with the world around us. After all, and this is meant to be a corroborating argument, knowledge alone cannot move us to act, as reflection is geared to do. The judgments we make about what we ought to think or do are practical, not theoretical, in nature.

It is a sign of how thoroughly modern philosophy has absorbed the naturalistic worldview that the otherwise rival schools of Hume and of Kant, which occupy so much of the terrain, are each impelled to embrace this view of reflection. Expressivists in the Humean tradition construe normative judgments of the form “A ought to X,” not as aiming to describe correctly the reasons A has, but as simply expressing our acceptance of a norm entailing that A conduct himself thus in the given circumstances.Footnote ¹⁰ Reflection is seen as charged with making clear to ourselves what norms we want in this sense to endorse, or what the norms we do endorse entail in a given situation. Kantians may differ from Humeans in their eagerness to deduce internal constraints governing the sorts of rules we can in this way coherently make our own—constraints that supposedly amount to the basic features of morality. But they agree that reasons cannot form part of reality itself. What we have reason to think or do is determined by principles of thought and action that we impose upon ourselves. Only through our own “autonomy,” that is, self-legislation, can the value-neutral facts of experience acquire significance for our conduct. “The ethics of autonomy is the only one consistent with the metaphysics of the modern world.”Footnote ¹¹

This widespread view of reasons and reflection is indeed driven by metaphysical assumptions. For naturalism, one should remember, is not a conclusion of the modern natural sciences themselves, which do not claim that they exhaust all that can count as real. Global assertions of that sort are the province of philosophy. Now I have no objection to metaphysics as such. But I do complain when metaphysical preconceptions contradict fundamental aspects of our everyday self-understanding. Such is the case here. The notion of reasons espoused by expressivists and Kantians alike can make no sense of what it is to reflect or even to reason.

Expressivists often focus on morality alone, even though their task is to explain the nature of normative judgment in general. It may seem plausible that moral judgments do no more than express a certain kind of approval or disapproval of their object, given the ease with which some people can persuade themselves that moral distinctions are ultimately of our own creation. But expressivism looks far less palatable when stated in its properly universal form. Can we really believe that the canons by which we judge whether a scientific theory is true—its fit with experimental data, its coherence with existing doctrine, its performance on severe tests—have only the authority we bestow on them? If we hold that there is good reason to adhere to such principles, are we only expressing our endorsement of a norm to the effect that everyone (ourselves included) ought to abide by them whether one happens to want to or not? Does our endorsing such a norm explain what it is to regard those principles as valid? If we were not constrained by an allegiance to naturalism, we would surely suppose that the order of explanation is the other way around. It is the perceived validity of the principles that accounts for why we think anyone ought to adhere to them and to endorse a norm to this effect.

Denying that we reflect in order to discover something we do not yet know—namely, what there is reason to think or do in the given circumstances—renders the process of reflection unrecognizable. In fact, the underlying refusal to countenance the reality of reasons distorts the very nature of reason, as we can observe if we look at the parallel Kantian notion of autonomy. Kantians think of reason as self-legislating, as giving itself principles of conduct, intellectual and practical, in the light of which alone facts in the world can acquire the status of being reasons to think and act in certain ways.Footnote ¹² But reason, like any faculty of mind, is to be defined in terms of the activity that is its characteristic exercise. What then could be more obvious than that reason is the faculty of reasoning? And how else can the activity of reasoning be understood except in terms of its being responsive to reasons? The most straightforward account of the nature of reason presumes, therefore, that reasons are a reality we discover, not an artifact of our ways of thinking. To be sure, we do sometimes impose rules on ourselves whose authority is of our own making. But we do so precisely because there appear to be reasons that warrant this step. If I give myself a rule never to borrow, it is in virtue of both knowing how prone I am to borrow more than I can pay back and accepting the antecedent authority of the principle that debts are to be repaid. Autonomy makes no sense as a global account of reasons.

Yet what of the counterargument that figuring out what reasons we have cannot consist in discovering some further facts about the way things are? Beliefs purport to represent the facts as they are, but beliefs alone, so it is held, cannot move us to act, as the results of reflection typically do. Thus, when we conclude that we have reason to act in a particular way, we are not reporting a discovery but rather expressing our confidence that the conduct in question satisfies or promotes our existing needs, interests, or commitments: “the word ‘reason’ refers to a kind of reflective success.”Footnote ¹³ Reasons, it is said, are the outcome of reflection when carried out well, not a reality to which reflection responds: what we have reason to think or do is what we can coherently endorse as we put our thinking in order.

This line of argument is a mainstay of both the Kantian and Humean traditions. Yet for all its popularity, it is rather poor. First, one might note that people (and other animals too) often act for reasons without reflecting. A more telling objection is that endorsing some option on the basis of our existing convictions depends on supposing that they justify this decision—that they indeed give us reason to endorse the option. Reasons cannot be simply the output of reflective success, since success in reflecting consists precisely in how well we respond to the reasons that there are.

However, the underlying error lies deeper still. It is a flawed conception of belief. Conclusions about reasons certainly serve to guide our conduct. Yet this does not mean that they cannot have the character of beliefs. For belief by itself is not in fact motivationally inert. Consider what it is, in general, to believe this or that. A belief is not a feeling (say, the particular vivacity with which an idea stands before the mind), nor is it the act of assenting to some proposition. It is a disposition, and one intimately connected with both the presumed truth of what is believed (since to believe that p means to believe that p is true) and with the behavior of the person whose belief it is (since we attribute beliefs to people in order to explain what they do). To believe that p is to be disposed to think and act in accord with the presumed truth of p. A person who has said he believes that the cat is on the mat, yet walks across the mat as though nothing were there, will not normally be held to believe what he has said he believes. To be sure, the specific things a belief disposes someone to do depend on other elements in his outlook, just as it is true that one sometimes fails to heed a belief one has. So the person in question may have his reasons for ignoring the cat or may be tramping across the mat out of inadvertence. But if he believes the cat is there, then he will be moved, all other things being equal, to comport himself compatibly with what, in his view, the truth of that belief implies. Beliefs do not merely represent the way things are or consist in holding certain facts to be true. They are at once descriptive and prescriptive. Being disposed to draw appropriate inferences from what we believe is an inherent part of believing. In other words, beliefs are commitments in their own right, commitments to think and act in accord with the presumed truth of what is believed.

Thus, reflection can very well aim at discovering what we have reason to do at the same time as it serves to guide us in how to live. Its goal is to arrive at correct beliefs about the reasons we have, beliefs which themselves, like all beliefs, commit us to conducting ourselves accordingly. Reflection is inseparably both theoretical and practical.

IV. The Reality of Reasons

So far, I have been arguing, negatively, against the idea that reflection cannot be an organ of knowledge. But I have yet to explain what sort of reality is constituted by the reasons for thought and action which are the objects of reflection. Here I must be brief.Footnote ¹⁴ If reasons belong to the fabric of reality, they are not, to be sure, some sort of independent entities, hovering alongside the more down-to-earth things we see and touch. As I have already suggested, reasons consist in a certain relation—the relation of counting in favor of—that features of the natural world, the physical and psychological facts making it up, bear to our possibilities of thought and action. This is a normative relation, and thus reasons cannot be equated with anything in nature. Yet, being relational, they manifestly depend on the existence of what they relate, and that means they depend on the natural (physical and psychological) facts being as they are, as well as on our having possibilities of thought and action. In this sense, therefore, reasons exist only because we do too, and it would certainly be bizarre to think otherwise. It does not follow, however, that we are the authors of the reasons there are, that they amount to the significance we bestow upon those facts. That one thing counts in favor of another is a relation (a normative relation) that we discover, that we can be right or wrong about, not one we institute ourselves. Relations are in general no less real than the things they relate when they enjoy this sort of independence from our beliefs about their existence and nature. Thus, reasons, too, form part of the world, understood broadly as the totality of what exists.

This conception may rightly be called “Platonistic,” since it holds that reasons, like Plato's Forms, constitute an intrinsically normative order of reality. But it is not an extravagant kind of Platonism. It does not suppose that reasons dwell in some Platonic heaven, unsullied by the contingencies of the world here below. On the contrary, this account follows closely our ordinary sense of what reasons are. When we suppose, as we ordinarily do, that how we ought to think and act is a matter of how the facts bear on the options before us, what we mean, in effect, is that reasons for belief and action are both relational and real. And when we conclude, in particular, that we have an impersonal reason to do something, what we mean is that the facts in and of themselves count in favor of that option, apart from our own interests and attachments. Theories that reduce reasons to an expression of our commitments have to devise formulas to mimic these ways of talking without taking them literally, and it is not surprising that the simulation never quite succeeds.Footnote ¹⁵

Not only do reasons not exist independently of the natural world or of us in particular, they also do not exist independently of one another. Reasons bear various logical and evidential connections to one another. Sometimes the reason we have to do one thing is contingent on our having a reason to do another, as when, thinking we have reason to believe the noise behind the bushes is a wild boar, we conclude that we have reason to climb up a nearby tree. And sometimes the reason to do a certain thing is, in essence, the application to particular circumstances of a far broader and standing reason—in other words, a principle—governing how we ought, in general, to comport ourselves. Thus, the reason we have to save part of our income, perhaps to buy a second home or to retire in some comfort, depends on our having more generally reason to be prudent and think about our long-term good.

What, then, about the relation between moral reasons and the other sorts of considerations that move us to act? It has often been thought, both in everyday life and in moral philosophy, that acting morally needs to be shown to be ultimately in the agent's own best interest. The idea goes back to Glaucon's appeal to Socrates at the beginning of Book II of Plato's Republic (357b): “Prove to us that it is better in every way to be just rather than unjust.” Only if the reason to concern ourselves with another's good can be derived from the presumably more basic reason we have to pursue our own good can morality really be authoritative for our conduct. The philosophical proofs have taken many different forms, depending on how both the agent's good and practical rationality have been conceived. Plato had Socrates respond to the challenge by arguing that “justice” or morality alone ensures the harmony of the soul to which we each ultimately aspire. The more common version in our own day has followed the lead of Thomas Hobbes, arguing that the adoption of principles of mutual restraint and cooperation is the most efficient means for each individual to satisfy his own interests over the long run, whatever they may be, given the essentials of the human condition.

This is not the place to evaluate in detail these various attempts to show how we may reason ourselves into the moral point of view from a standpoint located outside it and presumed to be more basic.Footnote ¹⁶ Let me observe here, not only that they all fail (for that is well known), but that they fail in two distinct ways. They do not succeed in explaining how the moral “ought” derives from the “ought” of individual prudence, duly enlightened. Even more importantly, the very understanding of morality on which they rely is typically defective, since it is slanted to fit the extra-moral starting point they invoke.

Consider the Hobbesian approach, as developed in our time by a great many writers. It conceives of morality as a set of rules for social cooperation, founded upon mutual advantage. The limits these rules place on the pursuit of our own ends are held to maximize our good in the long run through the interactions they permit with others, provided that others, too, comply with these restraints. Yet such a view excludes from the domain of moral concern two groups of people whom it is the business of morality to move us to treat better than we would otherwise be inclined to do: those unlikely to come into frequent contact with us and thus ever to contribute to our good, and those with neither the power nor the ability to enhance or jeopardize our interests in any way, however often we may cross their path. Morality conceived as a cooperative scheme for mutual benefit fails to embody so fundamental a norm as the respect we owe to strangers and to the weak. This is not, moreover, a failure to account for some marginal phenomena, to be remedied by further refining the theory. For surely it belongs to the very core of our moral thinking that we are to show equal respect to those who, through circumstance or misfortune, may never be in a position to benefit us in return. Indeed, we act toward the powerful and the useful in a moral and not merely prudent fashion, when we behave as we would even if they lacked those assets that make them of interest to our own endeavors. We aim then to treat them with the respect they deserve as human beings, not with the sort of circumspection we exercise in dealing with the various natural forces for good and ill in our environment.

The Hobbesian idea of morality reflects the notion of the rational agent, concerned to pursue his own interests as efficiently as possible, that Hobbesians take as their point of departure for showing how we can reason ourselves into morality. That is why this idea is so distorted. For if, as I have assumed from the outset (and will explain more thoroughly in the next section), the moral point of view consists in seeing in another's good, in and of itself, a reason for action on our part, then how could it ever come within reach of a person whose reasoning focuses essentially on satisfying his own interests? Let these interests be other-directed as well as self-directed; it does not matter. There is all the difference in the world between respecting or fostering another's good because one has among one's interests a particular attachment to that person, and doing so simply because that person's good is in question. It is the latter, the impersonally motivated act, that embodies the properly moral attitude. Every attempt to bring the moral point of view within the orbit of the rational pursuit of one's own concerns is bound to erase its most distinctive feature, which is the concern we conceive for another's good just because it is his or hers.

We cannot, then, reason ourselves into an appreciation of what it is to act morally, beginning from some location outside the moral point of view. Morality has to speak for itself. It constitutes a class of reasons for action that is sui generis, unintelligible in terms of any supposedly more primitive class of reasons. This fact ought not to be disturbing, though it does have an important philosophical consequence.

To understand why it does not make a mystery of moral thinking, consider the case of prudence—precisely the form of reasoning in which philosophers (and no doubt others too) feel a hankering to ground morality. How would we go about changing the mind of someone who perceives no reason to be prudent, no reason to take into account, in deciding what to do, the desires that he will predictably have later even if they are not his at the moment? We might have some luck by exposing the erroneous assumptions that keep him perhaps from recognizing the value of prudence (such as the notion that reasons for action can be but the expression of one's given desires). But if negative arguments of this sort do not work, what else can we do? There can be no prospect of reasoning him into an acceptance of prudential reasons by appealing to some deeper set of motivations he could be presumed to possess. What would such an argument look like? That it is in his interest to be prudent? That is hopelessly circular. The fact is that prudence is just like morality: it constitutes a class of sui generis reasons for action, underivable from any more basic sorts of considerations.

Herein lies an important philosophical lesson, not just about morality, but about reflection in general. If we are to move the person who has been stubbornly refusing to listen to the voice of prudence, we can only say something like: “Think about what you will surely want one year from now. Don't you see that you ought to act accordingly?” In the end, people just have to acknowledge that there exist reasons of prudence. There is no way they can be led to reason themselves from some external standpoint into a sense of the value of prudence. So, too, with morality. We simply have to see that another's good is, in itself, a reason for action on our part. Reflection does not always proceed, then, by inferring some reasons from others with which are already familiar. It also includes the power of acknowledging the force of reasons that speak for themselves. Such is the kind of reflection that deep shifts in our thinking typically require.

This conclusion may seem to contradict my earlier insistence (in Section II) that our capacity for reflection develops as we mature. But that is not so. When we are young, as I said, reflecting means taking up the standpoint (or what we imagine to be the standpoint) of those to whom we feel close, judging ourselves and our possibilities accordingly. As our horizons broaden and we encounter conflicts among our various loyalties, we are moved to work out less partisan standards of evaluation. Thus, we develop an ability to reflect impersonally on what we ought to believe and do. Yet though this process takes place step by step, it does not consist in grasping how our various interests and attachments, along with the desire to handle the tensions among them, point us to reasons for belief and action that are impersonal in character. The capacity of reflection is one thing; its objects are another. We learn to stand back from our individual commitments so as to view the world as it is in itself. But this achievement by itself leaves open what reasons, if any, we then discover about how we are to think and act. To see what reasons there are, we must actually reflect. After all, when we realize that our own good does not, absolutely speaking, matter more than the good of others, we might perhaps conclude that therefore neither theirs nor ours matter at all. Being impartial has been known more than once to produce indifference.

Some will urge that if we are not to remain indifferent, we must care about how people in general fare. This is true. Yet the ready appeal to feelings does little to illuminate the matter. For why would we care, if not that we see some reason to do so? The point at issue is precisely how we come to discern such a reason. It is not, I have argued, by determining what in the long run or in the deepest sense will serve our own good. If there is a reason to care about the good of others in and of itself, it has to be one that speaks for itself when we view the world impersonally. But it is also true that impersonal reflection must look in the right place. We need, in particular, to focus not on the fact that from the standpoint of the universe no one's good matters much at all, but on the fact that each person's good matters enormously from his own perspective.

V. The Moral Point of View

Such then are the basic features of reflection, its modalities and aims. We reflect in order to figure out what we ought to think or do in response to a problem that has disrupted our usual ways of dealing with the world. To find a solution, we stand back from how we have gone on before, looking at ourselves from the outside and appraising our options from a standpoint we consider authoritative for the problem before us. The explicit object now before our mind is what we have reason to think or do, and the reasons we seek to discover are by definition universal: if they are valid for us, then they are valid for all under similar circumstances. They are, in addition, impersonal if they are reasons we can grasp as binding on us independently of our own interests and attachments.

With this framework in place, it is time to return to the main theme. We are moral beings, I have said, because of the remarkable power of self-transcendence that reflection makes possible when it becomes impersonal. We can learn how to stand back from our own concerns so as to be able to see in another's good, in and of itself, a reason for action on our part. The preceding analysis has shown what is involved in reflection taking on an impersonal form. But much more needs to be said about the role that impersonal reflection plays in our moral thinking.

I have not, for instance, explained what sort of being this “other” is whose good appears from the moral point of view as an intrinsic object of concern. Is it any human being as such, or is it more broadly all living beings, animals and even plants included, that may be considered to possess a good? How widely does our moral responsibility extend? I have been assuming that it encompasses every human being, and that is the assumption in all that follows. But I do not suppose that our moral concern should stop at the boundaries of humanity. As to how far it should extend, I have no systematic answer, and I believe moreover that we should not be in a rush to devise one. Uncertainty about the scope of our moral responsibility seems to me in keeping with the present intellectual situation. For plainly those traditional answers will not do that equate the objects of moral concern with those who are themselves able to take up the moral point of view. Such is the approach adopted whenever one conceives of morality as a set of rules for social cooperation (as in the Hobbesian tradition) or as a system of reciprocal claims rational agents make upon one another (as in the Kantian tradition). And the approach fails because it cannot even attribute to all human beings a moral standing in their own right—not to infants, not to the mentally disabled, not to the doddering elderly, which with the advances of modern medicine we all have a good chance of becoming. At best, it can award such people only a derivative status, dependent on what is owed to their morally functioning trustees. Surely, then, the domain of moral concern must extend beyond the domain of moral agents, and how far it reaches is a question at the frontiers of moral thinking today.Footnote ¹⁷

What I do want to discuss in some detail here is the relation itself between impersonal reflection and morality. On several different scores, one might in fact dispute whether so intimate a connection exists between the two as I have been asserting. Does the essence of the moral point of view really consist in recognizing that the good of another, purely by virtue of being a person's good, is of equal moment with our own?

It will help to review the position I have assumed. There is, in general, more than one basis on which we can see in another's good a reason for us to act in some way, to refrain from doing what might otherwise prove attractive or even to do what we can to help him to achieve his good. We may, for instance, hope that thereby the person, or others happening to witness our behavior, will treat us well in return. Or we may feel a special bond of affection for the person—a relative, friend, or coreligionist perhaps—and on this basis put his good at the center of our attention. Finally, however, we may act so as to respect or foster another's good without an eye to our own interests and attachments but simply because someone's well-being is at stake. The person may not be anyone we know or suppose we will ever meet again, or he may be an individual to whom we do have some particular tie but whose situation and needs we are considering apart from that bond. In either case, the reason we then recognize to care about the other's good is impersonal in character, and this sort of thinking—such has been my refrain—constitutes the heart of the moral point of view. It can become a matter of habit, and that is quite desirable. Yet not only do we acquire this sort of concern for others by learning to reflect impersonally, but we need continually to reflect anew, exercising the same capacity of distancing ourselves from our own interests and attachments to survey the possibilities before us, if we are to monitor and refine our moral outlook on the world.

Now one might immediately object that this conception of morality is a highly contentious one. Does not its insistence on impersonality and the need to stand back from our own concerns signal an allegiance, for all my rejection of the idea of autonomy (in Section III), to the rigoristic distinction between duty and inclination typical of Kantian ethics? In large part, this objection is a misunderstanding. I have so far said little about moral feelings such as sympathy, but my aim has not been to deny their importance when we take an intrinsic interest in another person's good. Contrary to Kant, moral character does not show through most when one does what is right despite “being cold in temperament and indifferent to the sufferings of others,”Footnote ¹⁸ since there is nothing much left for moral character to be, given that sort of insensibility. Our thinking about right and wrong takes shape through experience in the way I sketched earlier (in Section II). We handle the conflicts between our various allegiances by learning to consider others from a wider perspective, and this process involves a generalization of the ability to care about how another fares that we first acquire in our relations to those who are close to us. But it is crucial to understand rightly the role feelings play. To sympathize with the pain or joy of others, at least insofar as morality is concerned, is not to feel their pain or joy ourselves, as though by a sort of contagion; it is to feel sadness at their suffering, delight at their happiness, a second-order feeling in which we express our sense that they have reason to feel as they do. Sympathy consists in being moved by the good of others. It therefore remains narrower than it might otherwise be, so long as it is filtered through our own interests and attachments. Feelings come to have a moral role precisely to the extent that they take us out of ourselves, and in this way they become indeed impersonal.

Nothing, then, in the conception of morality on which I have been leaning appears peculiarly Kantian. In fact, if I had to cite a single source of inspiration, it would not be Kant, but rather Scripture—though I hasten to add that my agenda is strictly secular. In both the Hebrew Bible and the New Testament, we meet the precept, “Love thy neighbor as thyself.”Footnote ¹⁹ It contains, I believe, a profound insight into the true character of moral thinking. All of us by nature love ourselves, caring immensely whether our desires are satisfied and whether our lives go as we would like. Obviously, this self-love often stands in the way of our caring about others as we should. Yet it also offers a paradigm of what a concern for others would be like if we were to look beyond our own sphere—beyond, that is, our individual interests as well as the interests of those who matter to us because of their special connection to ourselves. For the conspicuous feature of self-love is not so much its magnitude as its immediacy. The concern we feel for our own good is not channeled through other considerations: we do not care about how our own lives go because we are someone to whom we feel particularly close, or because we hope that we will secure from ourselves some benefit in return. Each of us cares quite simply because “it's me!” Suppose, then, that we care about another's good in the same direct, unmediated way, solely because it is his or her good. That would be, in the Biblical phrase, to “love thy neighbor as thyself,” or, in the more analytic language I have used, to see in another's good, in and of itself, a reason for action on our part. Such, I claim, is the core of morality.

Two other features of this conception should also be noted, again to avoid misunderstanding. The first is that it does not assume that the demands of morality are supremely authoritative. To see in another's good a claim on our attention no less direct than the claim made by our own does involve regarding these claims as essentially equal—differing only in the importance and urgency of what is at stake, and not in our happening to be one of the persons involved. The moral point of view consists in seeing oneself as but one person among others. Yet it does not follow that the morally best action is always the one that, all things considered, we have most reason to do. We become moral beings by learning to recognize in another's good, in and of itself, a reason for action on our part. But that means that we must, in given situations, weigh such reasons against the other sorts of reasons we perceive, reasons that may favor our own interests or attachments. There is, I believe, no general principle that can plausibly inform us how to handle every such case, certainly none that can convince us that morality has to override all other concerns. Philosophers in the Kantian tradition have often tried to show that morality possesses supreme authority, but generally they have proceeded by supposing that reason must give itself its own principles of action, arguing then that the conditions under which it can coherently do so amount to the self-legislation of basic moral norms. If instead, as I have argued (in Section III), reason consists essentially in a responsiveness to reasons, then there can be no such a priori guarantee that the claims of morality must prove paramount. We have to consider the particularities of the situation at hand.

Thus, to invoke an example made famous by Bernard Williams, if we can rescue from imminent death only one of two people and one happens to be our spouse, it would appear to be “one thought too many” were we to rush toward our spouse, not solely out of love, but also with the idea that it is morally permissible in such a situation to favor the people we love, or even (as a utilitarian might hold) that devotion to loved ones is generally the best means for each individual to maximize the general good.Footnote ²⁰ In cases such as this, requiring an impersonal justification for our response, of the sort provided by morality, is worse than superfluous. It casts considerable doubt on the genuineness of our love. Morality need not be our ultimate standpoint of evaluation. Sometimes other things rightly matter a lot more. Only the moralistic believe that being moral means always putting morality first.

A second misunderstanding would be to suppose that if the moral point of view consists in seeing everyone's good as of equal moment, acting morally must therefore amount to acting so as to bring about the most good overall. That is not so. Nothing inherently favors this sort of “consequentialist” position. On the contrary, the moral point of view is in itself mute about how exactly we are to advance the good of others and adjudicate the competing claims they make on our attention. Principles of moral reasoning have to be introduced to make the moral point of view operational.

One principle is aggregative: giving equal consideration to each person's good can be taken to mean that, in effect, we should consider each person impartially as we determine what will most increase the total good of all. That is precisely the consequentialist approach. But it is wrong to think, like many philosophers, that if we have a concern for other people's good, then wanting to bring about as much of it as possible is the only step that makes sense. We could instead resolve to respect each person's good as precious in its own right. Another principle is thus distributive: if each person's good is to be treated equally—at least in ways that protect and foster his very ability to have and pursue a conception of his good—then it ought not to be compromised or sacrificed even for the sake of ensuring a greater good for others. This is the underlying rationale of deontological theories, which define what we owe to a given person, the treatment we should (as in keeping a promise) and should never (as in doing violence) accord him, without regard to how we could otherwise act so as to benefit others or to what others might do in response to our action.

It is clear that these two ways of thinking are not the same and can lead to very different judgments about what to do. But it is also clear that sometimes the one and sometimes the other seems the natural approach to take. When a great number of people may be affected and their needs are particularly urgent, satisfying the needs of as many as possible—in other words, bringing about the most good overall—easily looks like the path to adopt. But when that is not the case, or when helping the many entails doing significant harm to some, respecting the integrity of the individual tends to square better with our conscience. Sometimes, of course, we find ourselves unsure about whether, in the given situation, the one or the other stance is more appropriate. I do not believe that we should try to settle, in a general way, which of these two basic principles of moral reasoning, consequentialist or deontological, defines the single correct theory of morality. Each of them is a plausible interpretation of what the moral point of view means in practice. That is why it can so often appear right, depending on the situation, to base our actions on the one instead of on the other. We do better, then, to recognize their common validity as well as their capacity to conflict.Footnote ²¹

What ought not to be lost sight of, however, is that both the consequentialist and deontological principles have their basis in the impersonal perspective in which each person's good presents itself as constituting an equal reason for concern on our part. It is not that they can somehow be derived from that perspective or that their content can be explicated in terms of that perspective alone. But they presuppose it and are not fully intelligible without it. Stephen Darwall seems to me to miss this truth when he argues in his recent book, The Second-Person Standpoint, that moral reasons for action are founded in basic relations of mutual accountability in which we all stand to one another and thus that the moral point of view is “intersubjective” rather than “impersonal.”Footnote ²² Here is his favorite example to illustrate the thesis. If someone steps on your foot, you might assert that he ought to get off for two different reasons—either because he is in a position to stop the pain you are undergoing, or because he has no business treating you in such a fashion. The one reason is impersonal or “agent-neutral,” since a third party would have the same reason to remove that person's foot so as to end the pain if he were able to do so. The other reason is “second-personal” or “agent-relative,” since it is one that the perpetrator alone can be said to have in this situation, a reason he has because of the relations of mutual accountability in which he stands with regard to you and others. Moral obligations, Darwall argues, rest essentially on reasons of this second, intersubjective sort.

Such reasons certainly exist. Moreover, in a strict sense (narrower than what has become common usage), the term “moral obligation” refers, not to the whole of what we ought morally to do, but solely to what we owe to others in virtue of the relation in which we stand to them, a relation that gives them a corresponding right to demand that we act in the appropriate way, such that if we fail to do so, we have not simply done wrong, but have wronged them.Footnote ²³ A clear example is the obligation to keep our promises, in contrast, say, to the fact that we ought to give to charity, which is not something we owe to the poor or something they have a right to demand. Yet is it not then plain that part of morality, at least, has nothing to do with agent-relative reasons? For surely giving to the poor, like rescuing someone in danger or indeed relieving someone's pain, figures among the things we ought morally to do and constitutes, in an appropriately broad sense, a “moral obligation,” even though such duties are ours for the same (agent-neutral) reason that anyone has to aid those who happen to be in distress. Imagine, for instance, that a person is in pain, not because of being treated negligently by someone else, but as a result of some natural disaster. People in a position to help ought morally to do so, despite the fact that the person has not been wronged by anyone. Darwall's theory offers no place for obligations of this sort.Footnote ²⁴

Furthermore, in the part of morality where agent-relative reasons do play an essential role, their validity depends on a deeper set of reasons that come into focus only within an impersonal point of view. Take again the case where someone steps on your foot. Only because it is an act that can cause pain, an ill which, as such, anyone has reason to put an end to, can you claim that that person owes it to you not to do such a thing and that you have a right to complain if he does. The agent-relative reason he has not to treat you in such a way rests on the agent-neutral consideration that pain is something to be prevented or stopped when it occurs. True, the person owes it to you to avoid stepping on your foot even if the act does not in fact cause you pain, since it expresses a lack of respect for your person. But it shows disrespect because it is an act that might well have caused pain, or that shows disregard for that possibility.

The same is true in the case of promising. Only because there is an impersonal reason to value the benefits that people derive from being able to trust one another, does the practice of promise-keeping acquire the authority it has and do we thus have the agent-relative reason to keep our promises to those to whom they are made. After all, do we not conclude that we should break a promise to a person if keeping it would do him or others a far greater harm, judged impersonally, than our breach of trust would cause him or them? It may well be true that this harm must consist not merely in some lessening of their potential happiness, but in our failing to honor a weightier duty we have to them (say, to avoid telling them a wicked lie or causing them bodily pain), if it is to license our breaking the promise.Footnote ²⁵ But how are we to determine that in the circumstances the other duty takes precedence, if not by considering impartially the good of all those involved?

In general, the kinds of respect we owe to one another depend on these relations of mutual accountability serving what is each person's good, considered impersonally. I do not mean that such practices have a claim on us because they tend to bring about the most good overall, for, again, the impersonal standpoint is not in itself consequentialist, any more than it is in itself deontological. The point is, rather, that only if each person's good is understood to be of equal moment does it become intelligible what is at stake in the kinds of respect we owe to others and why we owe such respect to anyone who happens to stand in certain basic relations to us. Darwall's account of mutual accountability is essentially a reconstruction of the deontological outlook.Footnote ²⁶ All the more reason, it seems to me, to recognize that morality, even where it displays this “intersubjective” character, draws upon an underlying impersonal point of view.