Comments on Bratman, 9-2-09

I. I want to start with Boomer’s question: What does Bratman think is the test for getting his account of intention right? What would count for or against his theory?

Perhaps there is a mixture of empirical and normative considerations. First, he has certain ideas about what it would be like for a being to form and act on intentions. For example, a person who has an intention do something in the future will take that as a commitment, a fixed point, and will not form plans or act on desires that would conflict with the intention. But, he would conclude that we don’t have intentions, that they play no role in understanding human behavior, if we in fact don’t act as we would if we committed ourselves to plans in this way. He would conclude that we don’t have intentions if we did not decide on plans and take them to be settled. This is an example of an empirical test of his theory.

Second, there seem to be normative considerations as well. For example, he rejects the idea that intentions, just by themselves, give us reasons for doing things we decide to do on the ground that they are means to attaining the goals we intend. This would be unacceptable bootstrapping. It would mean that, just by my deciding to do something or to bring something about, I would make that goal rational for me (even if it frustrated my most basic desires or needs). But the objection here is not that people don’t ever treat intentions or decisions this way, but, rather, that they would be irrational if they did.

The upshot of all this is that Bratman seems to want to claim two things: On the one hand, we seem to be able to make better sense of how people behave and how they are able to do what they do if we assume that they have and act on intentions as Bratman tries to describe these. And, on the other hand, in the situation in which we find ourselves, the capacity for forming and acting on intentions is a rational capacity. It contributes to our being able to live as rationally as we do. Another way to put this is to say that, if we think of ourselves as (more or less) rational creatures, part of what makes us rational is just that we do form and act on intentions. We would not do as well if we simply always acted on the strongest desire of the moment. Nor would we do as well if we constantly stopped to reflect on what the balance of our desires (or other reasons) would dictate that we do. It is better if we can nail down some parts of our lives by fixing on certain goals and taking them as more or less settled, organizing features in our lives. But, if he is to make the case for this, Bratman needs to insist, for example, on certain rationality constraints on intentions. For example, the constraint that we not intend things which, given our beliefs, we clearly could not possibly succeed at.

II. Do we REALLY have intentions? A question James raised. One version of this question is whether we could explain, predict, or understand human behavior just as well if we had only the concept of a desire, but not the concept of an intention. Suppose we knew of someone what his desires were, but not his intentions, would we know as much about him? Davidson is an example of a philosopher who tries to analyze the idea of an intention in terms of desire. But even he realizes that, to do so, he will have to distinguish among desires. Only desires with a certain character can even begin to look like intentions. (Bratman responds to Davidson’s ideas in one of the essays in Faces of Intention.)

An other version of this question might be whether we would do as well, live as rationally or successfully if we lacked the capacity to settle on a course of action, or a plan, and have that decision govern our behavior in some way.

Yet another version treats the question as an issue in ontology. I’m not really the person to go to for questions about ontology. Around here, Josh Brown has that job. That said, how do we settle this question? Do tables really exist? Do rights really exist? We seem to think it is true that we sat around the table in 512 on Tuesday. And so it seems to be true that there is a table there. But, as philosophers of science sometimes say, the “table” is REALLY just a bunch of molecules. It is only a useful fiction to speak of tables. But then, if they are useful fictions, then they really are something.

I don’t think I am getting anywhere here. I prefer to skip heavy-duty ontology and ask just (just?) what are the truth conditions for statements like “I intend to finish these comments this morning.” I take it that such statements are sometimes true, and sometimes false. And it is sometimes useful to know whether they are true or false.

Comments on 9-8-09

Start with Kant. According to him, to act with a good will, one must be able to act from some motive other than a (mere) desire or inclination. [Already we might reflect on Frankfurt’s point that there are many different things we refer to as “wants” or “desires.” Is Kant wrong to assume that all “inclinations” are equally suspect from a moral point of view? As we already observed in class, not every desire is an egoistic desire. But I leave that to one side.]

In addition to his point about the nature of a good will, Kant also thinks moral laws apply to us—we are subject to moral laws—only if we are able to do what they require despite contrary desires or inclinations. And, if there are no such laws that apply to us, then, for all intents and purposes, there is no morality and there are no moral imperatives. What is the point of saying how people ought to act if they are unable to choose to act that way? For Kant, the big question of moral philosophy is how are moral imperatives possible? In order for morality to be possible, we must be free from determination “alien” desires or inclinations that might conflict with morality’s demands. We must be able to decide whether or not to acquiesce in such desires.

In the Groundwork, Kant wonders how we could be guided by anything but desires. He thinks (432) that all previous attempts to explain the foundations of morality fail, because they propose motives for morality (like fear of divine wrath, an example he cites later), fail because they make the motive an alien motive of just the sort Kant thinks is excluded by the idea of a good will. But, he suggests, if we think of morally good agents as agents formulating their own maxims (plans?) for themselves and guiding themselves by those plans, then we see how those maxims or principles could control their behavior without being based (merely) on ‘alien” desires.

Now, the idea that people can govern themselves by their own decisions, commitments, or intentions is an idea we also find in Bratman. Perhaps you will say it is no more, and also no less, puzzling how this is possible in Bratman than in Kant. Or, if you find the idea of governance by one’s own intentions a natural idea, than Kant’s idea might seem to be less puzzling.

Of course, there is the added idea, in Kant, that one successfully and rationally guides one’s life by issuing oneself imperatives (specific ones like “I must help those in need,” or general ones like “I will always treat humanity as an end.”) only when one decides on an imperative by following the universalization procedure and thereby gives specific content to this general, procedural imperative. But it is not clear that Kant has a good justification for this added idea. And it is certainly true that Bratman does not argue that intentions are rationally settled upon only when they are chosen in accordance with such a principle. Frankfurt, also, makes no claim that the second order volitions underlying choice must have any moral content. He would not say our wills are free only if they are chosen on the basis of moral principles. But, the converse could be true: that we can behave morally only if we are able to distance ourselves from our first order desires, reflect on them, and decide on some further basis whether or not to adopt and act on a plan to satisfy them.

Potentially, there are two issues here. First, there is the question of what kind of self-governance we must be capable if we, as agents, are to be capable of acting morally. Second, there is the question of what capacities for self-governance we, as members of a community, have reason to expect others to have and to exercise.

Comments, mostly on Velleman, 9-16-09

As usual, I write these comments largely as a way of clarifying for myself what we have or have not accomplished and how the various pieces fit together. In this case, I am also going to try to come up with a solution to the problems James and others raised about Velleman’s theory.

First a little stage setting: Bratman’s work can look a bit superficial when compared to Velleman’s. Bratman, at least in part, wants simply to offer a kind of inventory of some of the attitudes or mental states that play a role in guiding our behavior. He mentions, specifically, desires, beliefs and intentions. These are different from one another, and he is especially interested in intentions. He characterizes these by reference to the functions they perform and argues, in particular, that they play an essential role in enabling us to carry out more or less long-term plans and projects and to coordinate our action with what others do. He also tries to give an account of how intentions manage to do this, e.g., by behaving as something like (revocable) commitments. He wants to show how they contribute to our living rational lives.

Velleman (and also Frankfurt) can be seen as addressing a question Bratman does not get to, at least in the work we have so far talked about. The question is, if we make the commonsense assumption that agents, as agents, play a role in deciding or settling on what intentions we form and how we carry out our intentions, how do they do this? What processes or events constitute their doing this? Velleman, as his title suggests, means his question to be very general. That is, for any action a person performs, what does the performance of this action consist of? This causes a problem.

We might start by taking Velleman to say this: When I leave work to get home by 7:00 PM, I first form an intention to get home at 7:00 PM. I form the intention by deciding to get home by 7:00, and I do that by figuring out that I have reason to get home then (because it makes sense to get home then) and deciding on the basis of my underlying desire to do what makes sense. But all this explains my action in leaving work by other actions, namely forming an intention, deciding, and figuring out. How do we explain or understand these further actions? We were trying to understand, in general, what happens when someone acts, but we have left some actions unexplained. What happens when someone decides, or when someone “figures out?”

Let me suggest something that Velleman might say in response to these questions. Velleman postulates that we have a fundamental desire to do what we have reason to do. This is the “motive that drives practical thought itself.” Let us say that we also are aware of our situation—of what desires we have (first order and higher order), of our intentions, of our beliefs about our needs and our environment, and of how we have acted in the past—and let us suppose that, in virtue of this awareness, we come to have beliefs about what we have reason to do, or what makes sense for us to do. (I imagine that the process is similar to the process by which we come to have beliefs about the world in virtue of the inputs of our senses.) The desire to do what makes sense together with our beliefs about what makes sense then cause us to have an intention to do whatever that is. Indeed, this process works both in bringing about an initial intention and in bringing about a subordinate intention to do what is needed to implement the basic one.

What we think of, or describe as, decision or reflection are just the events and processes I have described. Of course, some of what Frankfurt talks about, specifically first and second order desires, play a central role in this story. They affect our beliefs about what makes sense, but what motivates a decision, if the decision is rational, is the desire to make sense, together with our beliefs about our reasons. It is in virtue of having this desire, and sometimes even acting out of it, that we are rational beings. Of course, we don’t always act from this desire, and we sometimes get it wrong about what reasons we have.

We might well suppose, and I guess I suppose, that a causal story explains our having the disposition to form beliefs about our reasons as a result of our awareness of our situation and that a causal story explains our having a desire to do what we have reason to do. Perhaps the story is evolutionary and genetic. And also the particular beliefs we form are formed by a causal process. But I don’t see that the assumption of such a causal history undermines the claim that persons have at least the capacity to act rationally and to govern their actions by reason and reflection.

Why isn’t this enough freedom for Kant, or a Kantian?

Comments on Scanlon, 9/25/09

We gave Scanlon a rather bad time, and I think I should have done a better job of saying what I take to be good, or at least interesting, about what he says. My own view is that, while he has not succeeded in establishing his conclusions, he has also not been shown to be wrong. What are his conclusions? First, he claims that (mere) desires do not justify decisions. With this, many philosophers actually agree. What does the work of justifying, for an agent contemplating a decision, is facts—considerations expressible by “that” clauses. Desires themselves stand in need of justification. Second, and more controversially, desires are not needed in the explanation of actions, and they seldom play the role in explanation normally assigned to them. What normally does that work, again, are reasons. But what that means, as we saw, but only belatedly, is beliefs, such as the belief (true or false) that something is the case and is a reason, explain actions.

There are a number of tricky points and important distinctions that make it hard to read Scanlon correctly and also make him seem vulnerable to objections to which he might really be immune. One is the point just noted: it would seem odd to say that the mere fact that there is a reason automatically motivates (or explains) an action. The consideration has to be perceived and seen as a reason to do so. Also, the mere fact that there is a reason might well be relevant to what it is, in fact, most rational to do, but it is not relevant to whether an agent is irrational in not doing it. That depends on what the agent takes to be a reason. Finally, the fact that the term ‘desire’ is used in a variety of ways makes it hard to state a thesis like Scanlon’s in a way that won’t be misunderstood.

The question is whether the traditional view is actually true—that desires are basic to explanation and that no belief can motivate or explain an action except when there is a desire in the background. One might consider simple cases, like the belief that something will be pleasant or relieve pain, or the belief that the alternative will cause pain. In fact, people have argued for decades (Philippa Foot is a relatively early example) that such beliefs can motivate. Of course, some will say they motivate only because of an underlying desire to avoid pain. But, at a certain point, this can sound like the mere repetition of a dogma. Why believe it? Part of Scanlon’s strategy is simply an attempt, by making distinctions and offering persuasive examples, to shift the burden of proof to the person who insists on the desire-belief model.

While a number of philosophers have argued, with Scanlon, that “facts” can be reasons, and both justify and, if believed, explain actions, philosophers have been reluctant to believe that moral considerations motivate. They may think this will hurt me can motivate, but not this will hurt someone. In short, it is one thing to defend the general idea that some “facts” are reasons, another to defend various specific, substantive accounts of what reasons we have.

What if Scanlon were correct? For one thing, It would mean there are things called reasons, independent of desires, and we could plug Scanlon’s account of reasons into an account of intention and decision-making like Bratman’s, interpreting the latter’s references to reasons in Scanlon’s terms. It would give us a possible reading of Bratman’s ideas, though perhaps one he would not accept. Velleman might have to revise his account of action, or at least face the question of whether his ideas about reasons are compatible with Scanlon’s.

Scanlon in fact argues, in later parts of his book, that, among the reasons we have are reasons whose content corresponds to certain demands associated with morality. This is a substantive claim, of course, but, in defending it, Scanlon is more cautious—has a more limited conception of what morality, and reason, demands—than many writers. If he is right, though, here and in the rest of his theory, that means that the requirements of morality are rationally justifiable and people who attend to the reasons they have will be naturally motivated to act morally. A perfectly rational person will act morally—a conclusion Kant also endorses.

While I have long found Scanlon’s approach attractive, I am intrigued by the rather weaker claims both Velleman and Frankfurt make about human nature and morality. Both, I think, want to claim that various beliefs and motives we might associate with morality are supported by a proper understanding of human nature and human reason. But perhaps each of them has a less elevated, less “fancy,” conception of both human nature and morality than does Scanlon.

Comments on Velleman, 9/30

Our discussion yesterday was a little chaotic, but that did not prevent (maybe it facilitated) the emergence of a number of interesting issues. The big question is whether Velleman’s ideas undermine various ideas associated with the objectivity of values (or of morality).

In thinking about this, I am put in mind of a comment of David Lewis to the effect that the study of philosophy is the study of the costs of different theories. Many, if not all of you thought giving up the idea of objectivity an unacceptable cost. The issue, of course, is what (philosophical) price we have to pay to avoid this.

So what is Velleman’s idea? He, like a lot of philosophers, thinks that what is good (or good for a person) is what it is rational to desire, admire, etc. But, as he emphasizes, he dies not mean by this what most philosophers will first think. The issue is that, for him, what is rational, or makes sense, depends on what is “in character” for the person in question. This is a matter of what the person would knowingly desire given suitable reflection on her beliefs, desires, values etc.

This seems to imply that, if a given attitude like desire or admiration, really does make sense for you, then there is no room in the theory, no basis, for a claim that the thing or person in question is not really desirable or admirable. The theory recognizes no further criterion for admirability or desirability. The theory seems to recognize no such thing as something’s being desirable period, as opposed to desirable for this person or from this person’s perspective. [Does Velleman claim the contrary on 426-430? What is his argument?]

Leaving aside the last interpretative issue, it is objected to what seems to be Velleman’s main view that we do, in fact, think that there is such a thing as what is admirable, period; that this is the basis for assessing and sometimes overriding the judgments of individuals; and that Velleman owes us an account of this notion, or his theory is invomplete.

There are various replies to this. First, Velleman (430) suggests a rational mechanism which will naturally lead us to think attitudes correspond to real, external properties, even if they don’t. The suggestion is that “real” admirability is an illusion. Second, don’t many of us tend in fact to think that at least some questions of value (as opposed, perhaps, to rightness) are properly questions for individuals to resolve for themselves. That doesn’t mean we can’t disagree. “What’s so good about that? How could you admire that?” we might ask, or at least think. But even there, we probably see ourselves more as wondering about the character of a person rather than wondering about the correctness of a judgment. But, much of the time, we think it is none of our business. (We might also think that other people misunderstand themselves.)

Even sticking with questions of value, as opposed to questions about the right, Velleman would surely say that it makes sense for most of us, to care, up to a point, about what makes sense to others—about what they find desirable or admirable. (Compare the issue Scanlon raises at the end of Ch. 1 concerning “other peoples’ reasons.”) More, if each of us cares about the attitudes of others and each knows that each cares, then, in certain areas of life, there will be pressures either to come to agreement, or to agreement to disagree. To call attention to these things, however, is not to claim that there are properties of “real” admirability, out there.

To what extent can we understand our moral and evaluative attitudes on the basis of an understanding of how we, as rational beings, act and decide? Are moral beliefs and commitments things that we will naturally arrive at to the extent that we exercise our powers of rational choice and reflection? I think this is Velleman’s question. An answer to this question is an answer to at least one version of the question whether morality is natural, objective or rational. His answer to this may be: not as natural as you might wish.

If you don’t like this answer, you have at least these options: (1) Argue that he is wrong about the nature of rational, reflective decision. (2) Look for objectivity somewhere else than in our powers of rational decision.

Bratman on Autonomy, 10/15

These should be short and quick.

1. One possible connection between moral issues and an account of autonomy is that the latter could contribute to an understanding of when agents can be held responsible (blamed, for example) for what they do. Since Bratman sees autonomous agency as consistent with a naturalistic, causal understanding of agents, this might also suggest an account of blameworthiness that is compatible with naturalistic causation, if we think that autonomy is necessary and sufficient for blameworthiness.

Bratman proposes that an agent acts autonomously when the agent is moved by higher-order policies (intentions) that serve to organize and manage the rest of the agent’s desires, intentions and so on. But it seems, I said, unreasonable to think that a person can be blamed only when so moved. Borrowing on some of Scanlon’s work, we might say that blame is appropriate only when an agent manifests, through actions or attitudes, a kind of failure of self-governance. In particular, the person acts in a way that violates relevant standards of conduct, having failed to subject his attitudes to control by appropriate policies of self-governance. (Note: the ebola virus case is not a case of a faulty action or attitude.) One way this could happen is that the person acts autonomously, but governs himself by immoral standards. But another way is that the agent simply lacks the kind of integrity needed for autonomy, and so fails to manage lower order attitudes at all. The person is, to some degree, out of his own control.

2. Notice, that one way of acting badly, and so being to blame, is to be autonomous, and act autonomously, but to have a bad character. That does seem to be a possibility, on Bratman’s view. Autonomy does not entail morally good motives. It does entail a capacity to manage or control desires and impulses—something Kant thinks essential to moral action. But it does not entail that this power will be exercised well. Indeed, neither Bratman, nor Velleman, nor Frankfurt thinks that rationality—and its associated powers that raise us “above” brute animals or give us some kind of capacity for freedom or autonomy—leads in any automatic way to our being morally good. Still (and this is something we will look at further in Velleman’s recent work), there may be reason to think that rational, self-governing persons, under certain empirical conditions, will naturally have some motive to adopt and act on values and principles of the sort we associate with morality. Understanding rationality naturalistically, then, we would be led to a contingent, naturalistic account of (some aspects of) morality.

3. I speak of understanding rationality naturalistically. I earlier said Bratman and the others see their accounts of autonomy, freedom and the like as compatible with a causal understanding of persons and the world. When I started the semester, I found myself thinking these theories would hold an action, decision, intention or other attitude to be free, autonomous, or “the agent’s own,” just when it had the right cause or explanation. And so, I read looking for the cause that was supposed to make an intention, say, one’s own (or” internal,” or “identified with”). I now think I was misunderstanding a lot of what was going on. At least in the case of Frankfurt and Bratman, I think the states that represent the agent, or the agent’s point of view, are not the ones that were caused or chosen in a certain way, but rather the ones that play a certain functional role in the agent’s subsequent life. In Frankfurt, for example, the attitudes one identifies with (wholeheartedly) are not just attitudes caused or approved by the agent’s higher order desires. (Down this path we get Watson’s questions about why those desires have authority or represent the agent’s point of view.) Instead, I now think, the desires that are “really yours,” or “you,” are the ones (reading in Bratman’s interpretation) that you, however this came about, are completely satisfied to treat as reasons. They are the desires that therefore play a significant role in organizing and controlling your subsequent behavior.

The fact that some desire or set of desires play a central role in a person’s life make them quite significant in an assessment of that person—in whether you would want your daughter to marry him, say. But, in line with what I said earlier, I don’t think that means that other (“external”) desires should play no role in our assessment of the person. But there may be a number of subtle shades of externality that affect how desires bear on assessment. However that goes, assessments need not, as I see it, depend primarily on how it came about that the agent has that attitude.

Notes on Frankfurt, 10/23/09

We left a couple of things unsettled at the end of Tuesday’s class.

1. In at least some cases, something is worth caring about only if it is independently important for the agent. Frankfurt seems to think that there is such a thing as importance (or does he?); but, in any case, we were left wondering what this might be. The very last sentence of “The Importance . . . .” seems to propose an answer. What does it mean? Does it help?

2. In “Unthinkable,” Frankfurt insists, against Hume, that some things just are, in themselves, unthinkable. It would be ‘lunatic,” “crazy,” to have these unthinkable desires. They are irrational in themselves, not just as a result of the falsity of some belief. But what is his account of irrationality, of craziness? Perhaps if one is tempted to agree with him, one could argue in terms of Scanlon’s ideas about reasons. (Another person to consider is Parfit. I suspect Phillips can give you good references to his work on reasons. I know this less well.) It is interesting, though, that Frankfurt seems not to be drawn to this kind of objectivist work.

Comments on Velleman, 11/11

Interacting-

I take this to be the first of several chapters in which Velleman explores the extent to which practical reason—being rational—supports being moral. To put it differently, these chapters discuss whether morality can be defended or justified by showing that its requirements are themselves requirements of practical reason. I believe the thought is that this or that fragment of (what we ordinarily and pretheoretically think of as) morality can be so supported—though perhaps different fragments in different ways.

The first point is that we have to deal with other rational persons, and that means we have to deal with persons who are trying to enact a life that makes sense for them. For them, and for us, what makes sense depends partly on what we think other people are like and what they are going to do. But what it makes sense for them to do, and so what they do, is not fixed independent of what we do.

Velleman seems rather quickly to propose that our best strategy will be to act in such a way that others can understand us correctly. That is, we should do what actually reflects our nature as we take it to be. But why assume this? There is a very large literature on conflict situations (the cold war and nuclear deterrence, for example), which emphasizes the importance of deceiving others. If we want to deter them, then we want them to believe, even if it is false, that we are crazy. To deter them, we have to convince that we will launch a retaliatory strike even though, once they have attacked, it is too late to deter them. The THREAT of a retaliatory strike was supposed to do that, but the strike itself simply does harm for no purpose. We may know and believe this, but they have to believe that we will do harm for no purpose. That is, that we are, relative to our own values, crazy.

Still, will it make sense for us, in ordinary situations, to act so as to reveal our true nature, or will it make more sense to conceal or disguise our real intentions and desires? Velleman thinks there are at least weak, but pervasive, rational pressures against deception. He refers to these as the costs of double bookkeeping. What are these? You may have to lie, and then you have to remember the lie and not reveal its falsehood. If you want to conceal your real desires, you may have to go without satisfying them, at least for a while, in which case you will have to endure some frustration. And so on. And if all this is true, then reason, other things equal, supports a kind of honesty. Hence, reason supports at least a part of what most of us take to be morality.

A couple of comments. First, as Velleman says, the pressures for honesty are not very strong and can often be outweighed. Partly, they can be outweighed simply by strong desires better satisfied by lying. Partly also they can be outweighed by the fact that others may not respond to our being honest in the way that nice guys would. They may not be nice guys. (There is a gigantic literature about the prisoners’ dilemma. Suppose you are disposed to cooperate and you go first. How do you think the other guy will respond? Many believe that the best long-term strategy over repeated plays of the dilemma involves cooperating with cooperators. If others believe this, then you do best to cooperate if you go first. But you are taking a risk that they may not be following the best strategy.) Still, I think Velleman is right that there are some advantages to honesty, even if they are far from overwhelming.

Second, I think the alleged costs of double bookkeeping are costs involved in a particular interaction or series of interaction with a given person. The examples discussed in class (I act one way with one person, another way with a different person) are not examples. So the fact that this is easy and advantageous does not mean that there are no pressures of the kind Velleman discusses against the kind of deception he has in mind.

But how much of a defense of MORALITY is this? How much of morality consists, merely, of refraining from deception? Well, that’s a question, and when Velleman speaks of skepticism about morality, perhaps he has in mind that only bits and pieces of what philosophers call ‘morality’ get supported.

One final point. There is an intriguing remark at the end of this chapter (which reappears, I think, on pp. 102-3 of the next chapter) about the fact that, in interacting with a person in a role, like teacher, lender, doctor, etc., we are always also interacting with the person who chooses to occupy that role. Velleman, and others he quotes, seem to see this as a particularly profound kind of moral relationship. What is he thinking? I am far from sure. But perhaps he sees the person himself, the person characterized simply as the person trying to make sense and to do what is supported by the best reasons, as, in the deepest sense possible, free. He is not limited by whatever role he has taken on at the moment; and he always has the opportunity—and, frighteningly, the responsibility—to reconsider. To the extent that we interact with that person, we interact with someone who can only think of himself as bearing a weighty responsibility for what he does.

Reflecting-

This chapter continues the process of trying to show that ordinary rationality will lead a person to care, to some extent, about morality. Specifically, it will lead us to care about our commitments (to ourselves and to others), it will lead us to respect volitional necessities, and it will speak in favor of adhering to a socially accepted way of life. That each of these things is actually an example of “being moral” is, of course, controversial. One might think that the socially accepted way of life is immoral or that some volitional necessities, for particular persons, are either not moral necessities or even stand in the way of our doing what morality requires. But perhaps that is just an example of what Velleman already thinks, namely that the boundaries of morality are vague and that reason’s relation to morality is ambiguous.

What is the distinctive idea in this chapter? The key notion is that reason requires an idea that corresponds to what people think of as a conscience. The image is that of a separate person, looking over one’s shoulder and issuing warnings or criticisms. But why do we need this? Well, here we are. We are persons motivated by a desire to do what makes sense to us, in a world complicated by the existence of other persons trying to make sense to themselves. They are variables somewhat out of our control, but we can gain some control over them by controlling how we appear to them, given our own understanding of what they are like, and so how they will respond to the way we appear to them . . ..

I can mess up in several ways. First, I can do a bad job of figuring out just what my aims, desires and character traits actually are and how I can most effectively act on them. Second, there may be a best way for me to appear to others, but I may, through poor self-governance, inattention, mistake, or inadvertence, fail to project the best appearance. (A failure of self-presentation.) Third, I may fail to do what others reasonably expect me to do and so prove myself an unreliable cooperator in a joint enterprise. If I need the cooperation of others and their respect, then failures of these latter sorts will affect me adversely. Conscience makes me aware of my actual or potential failures.

The image of conscience is the image of a separate person. But that is just an image. All we need to imagine is a capacity to be aware of what we are doing, how it might appear, and what our needs are, together with a desire to behave effectively, in light of what makes sense. If we have the capacity for reflective awareness plus the desire to act effectively, our awareness will lead us to steer away from pitfalls. When we do mess up, our recognition of this, together with our motives to succeed, will lead to shame (hiding, concealment) or guilt (a desire to make amends and regain the confidence of others).