2002-2003 Lecture Series

[00:00:01] HOST:
Welcome to the second of Derek Parfit’s Tanner Lectures. The title of today’s lecture is Kant’s Formula of Universal Law. Once again, uh, there will not be a break after the lecture. We will proceed directly to comments by Professor Susan Wolf. Derek?

(unintelligible)

[00:00:30] DEREK PARFIT:
The rightness of our acts, Kant claims, depends on our maxims, by which he means roughly our policies and underlying aims. Some of Kant’s examples are: hasten my death to avoid suffering, let no insult pass unavenged, increase my wealth by every safe means, and the maxim of self-love or one’s own happiness. Well, as those examples show, maxims can range from intentions on which we could act only once to policies on which we could act throughout our lives.

Our acts are wrong, Kant claims, if our maxims could not be universal laws or we could not rationally will them to be such laws. On the first of those claims, which I can call Kant’s stated criterion of strict duties, it’s wrong to act on maxims that could not be universal laws. Well, that needs to be explained.

In some passages, when Kant supposes that some maxim is a universal law, he’s supposing that we’re all permitted to act on the maxim. And that may suggest that Kant’s criterion is, A, it’s wrong to act on some maxim unless we could all be permitted to act on it. But Kant writes nothing that supports that reading, nor would it be a helpful criterion since it assumes that we have another way of knowing whether some act is permitted.

In other passages, Kant supposes that some maxim is a universal law of nature, like the laws that govern the behavior of other animals. That may suggest that Kant’s criterion is B: it’s wrong to act on some maxim unless we could all accept it and act upon it. But I think that can’t be what Kant means.

When Kant claims that certain maxims could not be universal laws, he appeals to what would happen if everyone accepted those maxims. And he never suggests that even if we all accepted these maxims, we couldn’t all act on them. And besides fitting Kant’s text, B would be an implausible criterion.

There seems to be only one other possibility. Kant, it seems, must mean C. It’s wrong to act on a maxim unless we could all successfully act on it.

Now, the word everyone here refers to all the people to whom some maxim applies. Thus, the maxim, give up smoking, applies only to smokers. Of those who assume that Kant’s criterion is C, some defend that claim.

Thus, Onora O’Neill argues that by appealing to C, we can show the wrongness of coercion. We could not all successfully coerce others, O’Neill claims, because there’d be no one left to be coerced. Well, that claim overlooks mutual coercion.

I might coerce you by making one credible threat, while you coerce me by making another. But O’Neill is right to claim that if we appeal to C, it is, in her account, remarkably easy to derive significant moral conclusions. Many wrong acts are condemned by C. Consider the maxim: kill other people whenever that would benefit me.

Well, we couldn’t all successfully act on that maxim. Some attempted murders would fail, some successful murderers would be caught and punished. So Kant’s criterion condemns self-interested killing.

Such arguments, however, are too easy. C also condemns many permissible acts. Consider the maxims: give more to charity than the average person gives, become a doctor, and understand Kant’s philosophy.

Well, though we could all try to do these things, it’s either logically or practically impossible that we could all succeed. C implies falsely that on acting on these maxims, we act wrongly. Besides condemning many innocent or good acts, C has no plausibility.

Why should we think that if we couldn’t all successfully act on some maxim, no one should act on it? Innocent or worthy aims can be hard to achieve, nor is it wrong to make attempts, some of which are bound to fail. Besides having no plausibility, C is not the criterion to which Kant himself appeals.

It’s what he announces, but not what he appeals to. When he condemns the maxim of making lying promises, he doesn’t claim that we couldn’t all succeed in acting in that way. He claims that if we all accepted the maxim, or were all permitted to act on it, none of us could successfully act on it.

Now, when he appeals to the effects of our being permitted to act in some way, he’s really appealing to the effects of our believing that such acts are permitted. So I shall say that on Kant’s actual criterion of strict duties, it’s wrong to act on maxims whose being universally accepted or believed to be permissible would make it impossible for anyone successfully to act on them. Well, consider first the maxims kill, injure, and coerce others whenever that would benefit me.

If we all accepted and acted on these maxims, that wouldn’t make it impossible for any such act to succeed, so Kant’s criterion doesn’t condemn such acts. Turn next to lying. Kant’s criterion, Barbara Herman writes, seems adequate for maxims of deception.

Universal deception would be held by Kant to make speech, and thus deception, impossible. Korsgaard similarly writes, “Lies are usually efficacious in achieving their purposes because they deceive. But if they were universally practiced, they wouldn’t deceive.”

On Kant’s view, however, the wrongness of an act depends on the agent’s maxim, and few liars act on the maxim, “Always lie.” Most act on the maxim, “Lie when that would benefit me.” Now, Kant’s criterion condemns such lies only if in a world of self-interested liars, no such lies could succeed, and that would not be true.

It would seldom be in our interest to deceive others, and of the small proportion of cases in which deception would benefit us, there’s only a small proportion in which lying would be likely to deceive. There are, in contrast, many cases in which we benefit from speaking truly. So even if we were all self-interested liars, most of our statements would be true, and since we couldn’t always tell when other people were lying, some lies would be believed and would achieve the liar’s aim.

To explain why theft is wrong, Kant claims, were it to be a general rule to take away his belongings from everyone, mine and thine would be altogether at an end, for anything I might take from another, a third party would take from me. But as before, the relevant maxim isn’t always steal. Most thieves act on the maxim, steal when that would benefit me.

If that maxim were universally accepted, it wouldn’t produce a world in which theft could never achieve its aim. There’d still be property which wouldn’t always be successfully protected. So Kant’s criterion, I’ve argued, fails to condemn most of the acts that are most clearly wrong.

It doesn’t condemn self-interested killing, injuring, coercing, lying, and stealing. That may suggest that Kant’s criterion condemns nothing. But I haven’t yet considered Kant’s best example.

This is the maxim: Make lying promises and break my promises whenever that would benefit me. Kant claims that if such lying promises were universally believed to be permissible, that would make them impossible. The universality of a law that everyone could promise whatever he pleases with the intention of not keeping it would make it impossible since no one would believe what was promised him, and so on.

Well, many writers take that claim to be justified. Korsgaard writes, there’d be no such thing as a promise in the world of the universalized maxim. The practice of offering and accepting promises would have died out under the stress of too many violations.

Would there be too many violations? That’s not clear. Promises are efficacious, Korsgaard writes, only because they’re believed, and they’re believed only if they’re normally true.

But even if everyone accepted the self-interested promises maxim, promises would normally be true. This person’s maxim is in part, break my promises whenever that would benefit me. And even if we all accepted that maxim, it would be better for us if we kept most of our promises.

We’d benefit from being known to be promise keepers, since others would then admit us to mutually advantageous agreements. Well, those claims, it may be objected, apply only to the actual world in which there are many people who are truly trustworthy, since they keep their promises even when that’s worse for them. When there are many such people, there can also be parasites who keep most of their promises to preserve their appearance of trustworthiness.

But if no one was truly trustworthy, since everyone’s behavior with respect to promises was wholly self-interested, the practice of promising, it may be said, could not survive. Well, that objection, I think, is partly right. Let’s call promises trust-requiring if they would– wouldn’t be believed unless the promises were assumed to be trustworthy.

If we were all known to be self-interested, trust-requiring promises would never be believed. Many promises, however, are not trust-requiring, since the promisees assume that these promises would be kept for self-interested reasons. We can call such promises non-moral.

Even if we were all known to be self-interested, such promises would still be made and often believed. Well, it may be objected that if such promises were never kept for moral reasons, they wouldn’t really deserve to be called promises. And that may seem to justify Kant’s claim that if his lying promises maxim were universally accepted, the practice of promising wouldn’t survive, making it impossible for anyone to use lying promises to achieve their aims.

Well, in that form, I think the objection misunderstands Kant’s argument. Kant’s lying promisor has this maxim: When I believe myself to be in need of money, I shall borrow money and promise to repay it, even though I know that will never happen. Suppose first that this person’s promise isn’t trust-requiring.

Most money lenders rightly assume that in most cases it’s in the borrower’s interest to repay. The practice of making and keeping such promises would succeed, and it’s irrelevant whether those acts would be deservedly called promises, because whatever these acts should be called, they’d sometimes achieve their aims. Well, suppose next that Kant’s imagined person makes a trust-requiring promise.

Well, Kant could claim that if we were all known to be self-interested, no such promises would be believed. That doesn’t yet show that on Kant’s criterion, his lying promiser acts wrongly. Most lying promises act on a maxim that covers all promises, including those that are not trust-requiring.

However, Kant’s imagined person might be an exception. If that person’s maxim covered only trust-requiring promises, perhaps because he especially enjoyed exploiting other people’s trust, Kant’s criterion would condemn his acts. Well, now that we’ve found one kind of act that Kant’s criterion condemns, we can ask whether the criterion is plausible.

This criterion is in Part D: it’s wrong to act on maxims whose being universally believed to be permissible would make it impossible for anyone successfully to act on them. Now, D condemns those acts whose success depends on other people’s refraining from such acts because they believe them to be wrong. And D may seem to condemn these acts for the right reason.

These acts are wrong, we may think, because they exploit the conscientious self-restraint of others in ways that, if universal, would undermine the existence of valuable practices such as that of trust requiring promises. Kant’s criterion, however, seems more plausible than it really is. D applies to acts that many people believe to be wrong, and of the acts that are widely believed to be wrong, many are wrong.

But Kant’s criterion isn’t intended just to appeal to our moral beliefs. To judge this criterion, we should turn to actual or imagined cases in which people’s moral beliefs are mistaken. Suppose that because my nation’s tyrannical ruler is waging an unjust war, I adopt the maxim, kill this tyrant to end this war.

Most of my fellow citizens, we can suppose, would be appalled by such an act, since they accept Kant’s view that we should never try to overthrow any established government. Because this tyrant’s bodyguards know that nearly everyone accepts this view, they don’t expect any attempt on the tyrant’s life. That makes them inattentive, which allows me to kill this tyrant, thereby ending this war.

Now, if everyone believed that my maxim was permissible, the bodyguards would be more alert, making it impossible for any such an attempt to succeed. On these assumptions, D condemns my act. Well, I think this example counts against this criterion.

First, though Kant believed that killing our nation’s ruler is always wrong, that belief is false. It would have been right for a German to kill Hitler during the Second World War. Second, more important, even if tyrannicide were always wrong, D couldn’t provide the reason why.

The objection to tyrannicide couldn’t be that if we all believed that tyrannicide could be justified, that would make it impossible for any such act to succeed. Suppose next, and this is the more important case, that during this war, some German civilian knows that Jews are being rounded up and killed. This person acts on the maxim: Tell lies to the police when I could thereby help any Jews to escape.

We can plausibly suppose that if all Germans had believed that such lies were permissible, that would have made it impossible for anyone to help Jews in this way. German policemen would not have believed what civilians told them about the whereabouts of Jews. On these assumptions, D condemns this life-saving act.

Well, Kant might have accepted that conclusion since he condemned lying to a would-be murderer about where his intended victim is. That’s very similar. But such lies are clearly justified.

And in this example, D has no plausibility. It’s no objection to this way of saving people’s lives that if we all believed such acts to be permissible, that would make them impossible. Now, that example is intentionally similar to that of the lying trust-requiring promise, Kant’s best example.

The lying promises act succeeds because there are many people who can be trusted to keep their promises since they believe that breaking promises is wrong. If everyone was known to believe that lying promises are not wrong, that would make it impossible, Kant claims, for anyone to act successfully on this lying promises maxim. In the same way, when this German civilian lies to help some Jews to escape, this act succeeds because there are many people who can be trusted not to lie to the police since they believe that such lies are wrong.

If everyone was known to believe that such lies are not wrong, that would make it impossible for anyone successfully to act on this life-saving maxim. Well, the difference between these maxims is only in what these lies are intended to achieve, and that difference is ignored by D. Suppose next that German soldiers of this period could be relied upon to obey orders because they believed that disobedience would be wrong. That might have allowed some soldier to act on the maxim, disobey orders when that would help any Jews to escape.

It might be true that if all German soldiers had known to believe that such disobedience was permissible, their officers wouldn’t have given orders whose being disobeyed would allow Jews to escape. D would then mistakenly condemn this soldier’s act. Well, as these cases showed, D is wholly unacceptable.

It criterion– it does condemn some acts that are wrong, but for the wrong reason, and it condemns some acts that are clearly right Kant’s criterion is also in part, it’s wrong to act on maxims whose being universally accepted or acted on would make it impossible for anyone successfully to act on them. Now, there are many maxims which even if they were universally believed to be permissible, wouldn’t be universally accepted. And according– acting on such maxims, though it wouldn’t be condemned by the previous criterion, would be condemned by this.

Now, according to some writers, Kant’s criterion mistakenly condemns several good maxims, such as give to the poor and refuse to accept bribes. If those maxims were universally accepted, that would make it impossible to act on them since there’d cease to be any poor people and no one would offer bribes. Well, that couldn’t show these objectors’ claim that acting on these maxims is wrong.

Korsgaard, however, answers that objection well. If we accepted the maxim ‘Give to the poor,’ our aim would be to abolish poverty. If our maxim’s universal acceptance made it impossible to act on it because poverty would be abolished, that wouldn’t defeat but achieve our aim.

There are other cases, though, in which this reply would fail. Suppose we’re living at a time when many people accept some code of honor, like the one that led Pushkin to his death. Consider the maxim, fight duels to preserve my honor, but always shoot to miss.

Now, if everyone accepted that maxim, the practice of dueling would become farcical and would not survive. That would defeat this maxim’s purpose. Well, it may seems that E is right to condemn this maxim since dueling is wrong.

But E doesn’t condemn the maxim: fight duels to preserve my honor and always shoot to kill. And of these maxims, the second is clearly worse. Well, as that shows, E condemns the first maxim for the wrong reason.

It’s no objection to this maxim that if it were universally accepted, the practice of dueling wouldn’t survive. Turn next to the maxims: never take the first slice, don’t speak until others have spoken, and when you meet another car on a narrow road, stop and wait until the other car has passed. If we all acted on these maxims, none of us would achieve our aims.

Cakes would never get eaten, conversations never get started, and journeys on narrow roads would never end. That doesn’t show that acting on these maxims is wrong. For a more serious example, consider the maxim, have no children so as to have more time and energy to work for the future of humanity.

If we all acted on that maxim, that would make it impossible for anyone s-successfully to act on it since humanity would have no future. So E condemns that maxim in a way that’s clearly mistaken. Well, it may next be said that in some of the claims that I’ve quoted, Kant’s fault is only to exaggerate since it goes too far to say that if we were all self-interested, no promises would ever be believed, speech would be impossible, and there’d be no property.

Kantians might say that these claims, even if overstated, remind us that some valuable practices and institutions depend in part for their full effectiveness on moral motivation. This reply, however, cannot defend Kant’s criterion. Since Kant condemns maxims that couldn’t be universal laws, his criterion can’t be given a more moderate reading.

Now, that may be why Kant moves between claims at opposite extremes. Herman writes that Kant’s criterion seems adequate for maxims of deception and coercion. But while she condemns maxims of deception with the claim that if they were universally accepted, no one could successfully act on them, she condemns maxims of coercion with the claim that we could not all successfully act on them.

Now, since the word universal can cover both all and none, those claims may seem to appeal to the same criterion. But the first appeals to his stated criterion, the second to his actual criterion, and as I’ve said, those are very different. Most important maxims, whether good or bad, come in between those two extremes.

These are maxims on which even in the most favorable circumstances, we couldn’t all successfully act. But there are also maxims on which even in the least favorable circumstances, some of us could successfully act. So almost all important maxims are condemned by the version of Kant’s criterion which requires universal success.

Hardly any important maxims are condemned by the version which appeals to universal failure. So Kant’s criterion is either much too weak or much too strong.

[00:24:26] AUDIENCE MEMBER:
Mm-hmm. Yeah. Mm-hmm.

[00:24:33] DEREK PARFIT:
In neither version, moreover, does Kant’s criterion appeal to a morally relevant idea. For another illustration of this point, we can imagine a version of Kant’s criterion that tries to occupy the middle ground. This can be F.

It’s wrong to act on maxims whose being universally accepted or acted on would greatly reduce the number of people who could successfully act on them. Well, by appealing to F rather than D or E, we would be able to condemn a larger number of wrong acts. Thus, if everyone accepted the maxim of a self-interested liar, that might reduce the number of people who could lie successfully if they tried.

But F couldn’t explain why lying is wrong. There are many innocent maxims that F mistakenly condemns. It wasn’t wrong for romantic poets to try to have the experience of being the only human being in some natural wilderness, nor is it wrong to buy only second-hand books or give surprise parties.

Well, we can now turn different subject. Kant’s famous formula of universal law. It’s wrong to act on some maxim unless we could also rationally will it to be true that this maxim is a universal law.

Kant sometimes suggests that if some maxim fails this test, that gives us only an unstrict or imperfect duty not to act on it. We are sometimes permitted not to. On this reading of Kant’s formula, we’d be permitted, I say, to sometimes to act on these maxims, but we should ignore that reading, as Kant himself often does.

As we’ve seen, Kant’s criterion of strict duties fails to condemn nearly all of the acts that any adequate criterion must condemn. So we should ask whether Kant’s formula can fill this gap by implying that some kinds of act are always wrong, or at least always prima facie wrong, even if that might be overridden. Well, Kant often claims that in applying his formula, we should imagine that our maxim would become a universal law of nature.

On this version of Kant’s formula, which I should call the law of nature formula, it’s wrong to act on some maxim unless we could also rationally will it to be true that everyone accepts this maxim and acts on it. As before, everyone applies to all those people to whom the maxim applies. In other passages, Kant appeals to what we can call the permissibility formula.

It’s wrong to act on some maxim unless we could also rationally will it to be true that everyone is morally permitted to act in that way. Well, Kant assumes that if it were permissible to act on the maxims he discusses, people would be more likely to act in these ways, and these effects would be produced not by its being permissible to act in these ways, but by people’s believing that. So Kant is really appealing here to what I call the moral belief formula.

It’s wrong to act on a maxim unless we could also rationally will it to be true that everyone believes such acts to be permissible. Now, Kant remarks that he’s not proposing a new principle, but only a more precise statement of the principle which common human reason has always before its eyes. That remark understates Kant’s originality.

But these two versions of Kant’s formula can be claimed to develop the ideas that are implied by two familiar questions. What if everyone did that? And what if everyone thought like you?

The wrongness of an act, Kant claims, depends on the agent’s maxim. Now, as I say, Kant sometimes uses maxim to refer to the policy on which someone acts. On a second use, the maxim consists both of the policy and of the person’s underlying aim.

Suppose two merchants both act on the policy, never cheat my customers. One merchant acts in this way because he believes it to be his duty, while the other’s motive is to preserve his reputation and profits. These merchants, we can say, both have the same policy maxim but have different deep maxims.

Kant’s formula shouldn’t, I believe, appeal to either kind of maxim. Consider some egoist who has only one policy and one underlying aim. Do whatever would be best for me.

Now, this man couldn’t rationally will that his maxim be universal. Egoists have strong reasons n- to want other people to accept and follow not their egoistic maxim, but various moral principles. Egoists suffer from the egoism of others.

Since this egoist maxim cannot be rationally willed to be universal, Kant’s formula implies that whatever this man does, he acts wrongly. He acts wrongly not only when he steals, breaks promises, and harms other people, but also when, for self-interested reasons, he acts honestly, keeps his promises, and helps other people. Well, those are unacceptable conclusions.

When this egoist saves a drowning child because he hopes to get some reward, he isn’t acting wrongly. Turn now to someone who has mistaken moral beliefs. Our example can be Kant himself during the period when he accepted the maxim, never lie.

Well, that maxim is condemned by Kant’s formula, since Kant could not rationally will that no one ever tells lies, not even to prevent the would-be murderers from finding their victims. So Kant’s formula implies not only that Kant would have acted wrongly if he told such a murderer the truth, but also that he did act wrongly whenever he acted on his maxim, never lie. Since this maxim can’t be universalized, Kant acted wrongly whenever he told anyone the truth.

That’s another unacceptable conclusion. According to some writers, Kant distinguishes between some acts being wrong and its being contrary to duty. On this account, when my egoist keeps his promises and saves people’s lives, Kant’s formula implies that these acts are wrong without implying that they’re contrary to duty.

Now, Kant, I believe, doesn’t draw that distinction. He often claims that by applying his formula, we can answer the question whether some act would be contrary to duty or whether it’s in conformity with duty. Which is what we would mean by wrong, contrary to duty.

For Kant’s formula to answer such questions, it must be revised. An act’s wrongness depends not on the agent’s policy or underlying aim, but on what this person is intentionally doing. In the morally relevant description of someone’s act, we should include what this person intends her act to achieve and the other morally relevant effects that fe– she foresees.

It’s often irrelevant, however, on what policy this person acts. When Kant told someone the truth, it’s irrelevant that he was acting on the policy, never lie, so that he would have told the truth even to the would-be murderer. And when my egoist saves someone’s life, it’s irrelevant that his underlying aim is only to get some reward.

As Kant would claim, this man’s act is not wrong, though it has no moral worth. To meet these objections, Kant’s formula could become, it’s wrong to act in some way unless we could also rationally will it to be true that everyone does whatever in acting in this way we would be intentionally doing, or that everyone believes such acts to be permissible. Those revise the law of nature and moral belief versions.

Well, to save words, I shall continue to discuss the unrevised version and Kant sometimes uses maxims in a thinner sense, as when he discusses the maxim, hasten my death to avoid suffering. That maxim isn’t a policy or a principle, and when someone acts on this maxim, what he’s intentionally doing is hastening his death to avoid suffering. So in most of what follows, when it makes a difference, we could understand maxim in that thin sense.

To apply Kant’s formula, we s- must make assumptions about what we could rationally will. Now, on one view, it’s always irrational to act wrongly. But that view, even if true, is irrelevant here.

For Kant’s formula to succeed, it must provide a criterion for the wrongness of acts which doesn’t itself assume that the acts are wrong. It would be pointless to claim both that our act is wrong unless we can rationally will that everyone acts in this way and that we couldn’t rationally will that everyone acts in this way because such acts are wrong. And Kant doesn’t do that when he appeals to his examples.

According to another view, we’re rationally required to give significant weight to other people’s well-being. It’s irrational to act wrongly when our acts, though benefiting ourselves, impose certain great burdens on others. While that view, even if true, is also irrelevant here.

Kant’s formula doesn’t assume that such acts are irrational. The main idea behind his formula is that though it may be rational for us to act in ways that are wrong, we couldn’t rationally will either that everyone acts in these ways or that everyone believes such acts to be permissible. So when we apply Kant’s formula, we shouldn’t appeal either to the irrationality of acting wrongly or to the view that it would be irrational to give little or no weight to other people’s well-being.

We should appeal to claims about non-moral and non-altruistic rationality. On one such view, we couldn’t rationally will that everyone acts in some way if such a world would be bad for us. Now, that may seem not to be Kant’s view, since he calls the principle of prudence or self-love a merely hypothetical imperative which applies to us only insofar as we care about our future well-being.

But if we care about our future, as Kant assumes we all do, it would be instrumentally irrational for us to will that other people act in ways that would be bad for us. Some Kantians prefer to claim that we couldn’t rationally will a world in which our true needs would be worse met, or our rational agency would be frustrated, or our aims and purposes would be harder to fulfill. Well, I shall appeal to the claim that we couldn’t rationally will what would be worse for ourselves, but most of my arguments could be restated in those other ways.

I shall also appeal to claims about what would be likely to be good or bad for us. That just helps the formula. Kant’s formula works best when it’s applied to maxims or acts of which three things are true.

It would be possible for many people to act on this maxim or in this way. Whatever the number who act in this way, the effects of each act would be the same or roughly similar, and these acts would be randomly or roughly equally distributed between different people. These claims apply to many of the acts that are most clearly wrong, such as acts of self-interested injuring, coercing, or deceiving.

Most people could often act in these ways. Whatever the number who act in these ways, most of these acts would have similar effects since they’d benefit their agents but impose greater burdens on others. And in many cases, these burdens would be likely to be randomly or roughly equally distributed.

In such cases, it would be likely to be worse for most of us if everyone rather than no one acted in these ways. Even if each of us would gain from acting in these ways, each would be likely to lose more from the similar acts of others. Kant’s formula condemns these acts since we couldn’t rationally will that such acts be universal.

Well, in some of these cases, though any such act would impose burdens on others, it’s also true that since these burdens would be spread over many people, each act’s effects on each other person would be either trivial or imperceptible. Some examples could involve pollution, soil erosion, overfishing, overgrazing, overpopulation. In such cases, if each of us considers only the effects of our own acts, we may believe that we’re not acting wrongly.

Effects on everyone else is– each person is very trivial. When applied to such cases, Kant’s formula is much more successful than most other relevant principles, such as the act utilitarian principle and ordinary principles about harming others or the Golden Rule. Though each of these acts would impose only trivial burdens on others, we couldn’t rationally will that everyone acts in these ways, since these acts would together impose on everyone, including us, great burdens.

So these aren’t the cases that Kant had in mind, but they count strongly in favor of his formula. I think these are the cases where the formula really works best. When the first three conditions, though, are not all met, the formula works less well.

It here faces several objections, of which I have time to discuss only one. There are some wrong acts whose bad effects are not randomly or equally distributed between different people. These acts impose burdens only on the people who are in certain groups.

The golden rule requires impartiality by telling us to treat others as we would want or will that others treat us. Kant’s formula makes us impartial in a less direct way. When we apply this formula, rather than asking, “What if they did that to me?” we ask, “What if everyone did that?”

Now, this kind of impartiality has great importance. And if it’s permissible for us to act in some way, it must be permissible for everyone else to act in the same way as us in the same circumstances. When we act wrongly, as Kant points out, we often make unfair exceptions for ourselves, allowing ourselves to do things that we wouldn’t want or will other people to do.

And Kant’s formula, it’s wrong to do what we could not rationally will everyone to do. But this kind of impartiality is not, however, enough. Like the Golden Rule, Kant’s formula applies best to those wrong acts with which we benefit ourselves in ways that impose greater burdens on others.

We couldn’t rationally will that other people do these things to us, since we would then have to bear these greater burdens. But on Kant’s formula, we don’t ask whether we could will that other people do these things to us. We ask whether we could will that everyone does these things to others.

And we may know that even if everyone did these things to others, no one would do these things to us. Kant’s formula may then fail to condemn these wrong acts. I call this the impartiality objection.

If Kant’s formula can’t condemn those acts, it doesn’t ensure the kind of impartiality that, as Kant assumed, moral reasoning requires. Consider first some white racist in the age of segregation. This man might have claimed to be following the two versions of Kant’s formula of universal law.

He might have said, “When I exclude Blacks from my hotel,” I could rationally will that everyone acts in this way. Around here, everyone does act in this way. Every hotel owner excludes Blacks.

And I could rationally will that everyone believes such acts to be right. That’s what most of us do believe, and if the Blacks and commies changed their mind, that would be fine with me.” Well, in making these claims, would this man have misunderstood Kant’s formula?

I’m not asking whether he would have misunderstood Kant’s moral theory. Kant was in some ways remarkably egalitarian, and there’s much in Kant’s views that would condemn such racist attitudes and acts. My question is only what’s implied by Kant’s formula.

Well, Kant didn’t consider cases of this kind. When he imagined some wrongdoer asking, “Could I will that my maxim be a universal law?” Kant assumes that this person’s maxim isn’t such a law.

But in some cases, like that of this racist, a wrongdoer’s maxim may already be universal since it may already be acted on by all those people to whom it applies. Now, Kant’s formula permits these people’s acts if they could rationally will that they and others continue to act as they are now doing. Now, if it’s bad for these people that others are acting in the same way as them, as would be true, for example, in some anarchic war of all against all, these people couldn’t rationally will the continuation of the status quo.

But if the status quo is good for these people, we may face the following problem. The status quo may be good for these people precisely because their bad maxim is universal. Those to whom some maxim applies may be some powerful and privileged group who are oppressing other people.

Well, Kant’s formula condemns these people’s acts only if they couldn’t rationally will that they keep their privileged position. And for the reasons given above, we can’t defend such claims in ways that assume that these people’s acts are wrong, nor can we appeal to the claim that they are rationally required to give significant weight to other people’s well-being. When we apply Kant’s formula, we must claim that it would be non-morally irrational for such people to will that they keep their privileged position.

And such claims may be hard to defend. Nor would it help to turn to the moral belief version of Kant’s formula. If these people could rationally will that everyone acts in the same way as them, they could rationally will that everyone believes such acts to be permissible.

They’d have no relevant reason to prefer that everyone believes their acts to be wrong. Consider, for example, those men who treat women as inferior, denying them various rights and privileges, and giving less weight to their well-being. On Kant’s formula, such a man acts wrongly if he couldn’t rationally will it to be true that everyone treats women as inferior to men.

Now, that’s not a useful claim. If such a man could rationally will that he treats women as inferior, it doesn’t help to ask whether he could rationally will that everyone acts in this way and that everyone believes such acts to be justified. For most of history, most people, including most women, have treated women as inferior to men.

And most people, including most women, have believed such treatment to be justified. On Kant’s formula, it’s not wrong for men to treat women as inferior if these men could rationally will that they regain their old supremacy, and they could also rationally will that feminists who now object to male supremacy come to believe their objections are mistaken. And slave owners did not act wrongly if they could have rationally willed that they keep their slaves and willed that everyone, including all the slaves, believe slavery to be justified.

We would do better here to appeal to the golden rule, which Kant contemptuously dismissed. Men and slave owners would not will that they be treated as inferior or as mere property if they supposed that they themselves were going to be women or slaves. For another example, consider global inequality.

On any plausible moral view, those who control most of the world’s resources ought to transfer some of their wealth to the billion poorest people in the world, those whose daily income is less than two dollars. Many rich people now transfer nothing to these poor people. Kant’s formula doesn’t condemn these people’s acts if they could rationally will it to be true that all rich people act like M and that everyone, including the poor, believes such acts to be justified.

Well, as before, Kant’s formula here achieves nothing. When Korsgaard discusses Kant’s formula of universal law, she writes, “The kind of case around which the view is framed and which it handles best is the temptation to make oneself an exception: selfishness, meanness, advantage-taking, and disregard for the rights of others.” It’s this sort of thing, not violent crimes born of despair or illness, that serves as Kant’s model of immoral conduct.

I do not think we can fault him on this, for this and not the other is the sort of evil that most people are tempted by in their ordinary lives. Well, what Kant’s view handles best is not, I have argued, all kinds of selfishness or advantage-taking. Kant’s formula fails to condemn many of the acts with which some people take advantage of others, as when men, the rich and powerful take advantage of women, the poor and the weak.

And since Kant presents his formula as the supreme principle of morality, we can fault his formula for its failure to condemn such acts. These kinds of selfishness and advantage-taking are precisely the sorts of evil that the rich and powerful are tempted by and often commit in their ordinary lives. Some may think that in presenting this objection, I have misinterpreted Kant’s formula.

Thomas Nagel suggests that when we ask whether we could rationally will that everyone acts in the same way as us, Kant intends us to imagine that we ourselves are going to be in everyone else’s position. That suggestion makes Kant’s formula more like the golden rule. But none of Kant’s claims about his formula support Nagel’s reading, and there are contrary passages, such as Kant’s discussion of the self-reliant man who has the maxim of not helping others who are in need.

When he explains why this man could not rationally will that his maxim be a universal law. Now, Kant writes, “Many cases could occur in which, by such a law of nature arisen from his own will, he would rob himself of all the assistance he wishes for himself.” Well, if he intended this man to imagine that he would be in the positions of all the people who would need help, it would be baffling why he doesn’t say that here.

Nagel defends his reading with the claim that if Kant didn’t intend us to imagine being in everyone else’s position, Kant’s formula would be open to serious objections. But even the greatest philosophers can overlook possible objections. We shouldn’t assume that when great philosophers seem to make some mistake, they cannot have meant what they wrote.

Rawls proposes another reading of Kant’s formula. When we apply this formula, Rawls suggests, Kant intends us to imagine that we don’t know anything about ourselves or our circumstances. We should ask what we could rationally will if we were behind a veil of ignorance, not knowing whether we are men or women, rich or poor, fortunate or in need of help.

Like Nagel, Rawls supports his reading with the claim that it seems needed to defend Kant’s formula from objections. Rawls writes, “I believe that Kant may have assumed that our decision is subject to at least two kinds of limit on information. That some limits are necessary seems evident.”

But as before, even if Kant ought to have made such an assumption, that doesn’t show that he did. In his discussions of his formula, Kant never suggests that we should imagine being behind a veil of ignorance. Scanlon proposes a third reading.

Kant writes, “I ought never to act except in such a way that I could also will that my maxim be a universal law.” Scanlon suggests that when we apply this test, Kant intends us to ask whether everyone could rationally will that our maxim be a universal law. And to defend this reading, Scanlon writes that if Kant were merely telling us to ask ourselves what we ourselves could rationally will, he would be wrong to claim that his different formulas would have the same implications.

Well, like the other two proposals, Scanlon’s proposal cannot, I believe, be what Kant meant. Kant gives nearly twenty different statements of the formula of universal law, none of which refer to what everyone could will. These proposals are best regarded not as interpretations, but as ways of revising Kant’s formula so that it avoids the impartiality objection.

Now, these and other such proposals are as shown on the diagram, which is unreadable in the handout, and I shan’t take you down these swerving tracks here. Um, according to one, to decide whether some act is wrong– Well, I’ll let you look at it briefly.

You get the idea. It’s just meant to cover all of the main possibilities for this kind of view. According to one, to decide whether some act is wrong, it’s enough to ask whether we could rationally will it to be true that we act in this way.

Now, as I’ve said, even Kant doesn’t make that assumption, And he thinks that it is always irrational to act wrongly, but you don’t argue for the wrongness of the act by claiming that it’s irrational. According to four, the law of nature version of Kant’s formula of universal law, we should ask whether we could rationally will that everyone acts in this way, and according to seven, the moral belief version, we should ask whether we could will that everyone believes such acts to be permissible. Now, these formulas I’ve argued fail.

Many wrongdoers could rationally will both that everyone acts like them and that everyone believes such acts to be permissible. These people may know that even if everyone did to others what they are doing, no one would do these things to them. Many men, for example, could will that everyone treats women as inferior.

Two, the golden rule avoids this objection. These men couldn’t rationally will such treatment if they were going to be women. According to five, Nagel’s proposal, Kant’s law of nature formula becomes like the golden rule.

We ask what we could rationally will everyone to do if we suppose that we ourselves were going to be in everyone’s position. Eight is a similar re-revision of Kant’s moral belief formula. Uh, the numbers don’t matter.

When revised in this way, these formulas avoid the impartiality objection. It’s hard, however, to imagine ourselves in the position of every other person, all ten billion of them. And when Rawls discusses Hare’s version of this view, he suggests another objection.

In imagining ourselves in all these people’s positions, we shall think of ourselves as living all of these lives. That’s what Hare tells us to do. And that may lead us to ignore the separateness of these lives and the fact that one person’s burdens can’t be compensated by benefits to other people.

So we may thus be led to ignore the grounds for accepting principles of distributive justice. Rawls suggests that to avoid this objection, we should imagine that we’re behind a veil of ignorance. We should suppose that rather being in– rather than being in everyone’s position, we shall be in one person’s position, but we don’t know whose.

When revised in this way, Kant’s law of nature and moral belief formulas become six and nine on the chart. I should drop those entries. Scanlon suggests that rather than asking what we ourselves could rationally will, we should ask what everyone could rationally will.

Now, according to Kant’s law of nature formula, our act is wrong unless we ourselves could rationally will it to be true that everyone acts in this way. On Scanlon’s proposal, that would become, an act is wrong unless everyone could rationally will that everyone acts in this way. Well, it’s worth making a further revision.

If we appeal to what everyone could will, that’s enough to achieve impartiality. We needn’t ask whether everyone could will that everyone acts in some way, and there are some acts that are right, though we couldn’t rationally will that everyone acts in these ways. Kant did not act wrongly, for example, in having no children.

So we do better to turn to what I call the formula of universally willed acts. An act is wrong unless it could be rationally willed by everyone. Well, that’s a wider version of Kant’s consent principle.

On that principle, we ought to treat everyone only in ways to which they could rationally consent. On this wider formula, an act is wrong unless everyone, if they had the choice, could rationally choose that this act be done. According to Kant’s moral belief formula, our act is wrong unless we ourselves could rationally will it to be true that everyone believes such acts to be permissible.

On Scanlon’s proposal, this becomes: an act is wrong unless everyone could rationally will it to be true that everyone believes such acts to be permissible. In Scanlon’s words, to answer the question of right and wrong, what we must ask is what general principles of action could we all will? Now, that formula is strongly suggested by several of Kant’s claims about his other two main principles, the formulas of autonomy and of the realm of ends.

For example, Kant refers to the concept of every rational being as one who must regard himself as giving universal law. Kant never explicitly appeals to what everyone could rationally will. The phrase just quoted, for example, ends, “who must regard himself as giving universal law” through all the maxims of his will.”

Well, if each person regards himself as willing universal laws, he’s not asking which are the laws that everyone could will. At several other points, when Kant seems about to appeal to what everyone could will, he returns to his formula of universal law, telling us to appeal to the laws that we ourselves could will. Now, I think he doesn’t make that further step because he thinks it’s not necessary.

They’d have all the same implications. But as I’ve argued, the form of universal law needs to be revised, and this is the revision that seems closest to Kant’s own view. Now, we might call it the formula of universally willed moral beliefs, but I restate it and give it a shorter name as Kant’s contractualist formula.

We ought to act on the principles whose universal acceptance everyone could rationally will. Finally, on my chart, there’s one other view that differs from Kant’s formula by including a veil of ignorance. On the best-known version of such a view, or Rawls’s formula, we ought to act on the principles that it would be rational for everyone to choose as the principles that we’d all accept if no one knew anything about themselves or their circumstances.

Now, of these formulas, the best, I believe, are the formula of universally willed acts and Kant’s contractualist formula. Uh, if they’re the best, it’s also pleasing that they’re the simplest. These formulas both appeal to what everyone could rationally will.

The first formula applies this test directly to particular acts. The second applies it to the principles on which we act. In tomorrow’s lecture, my main subject will be Kant’s contractualist formula.

If we combine this formula with the right view about reasons, we shall reach what may be the best version of contractualism, and we shall also be led, I think, to some surprising conclusions.

(applause)

[01:02:03] HOST:
Today’s commentator is Professor Susan Wolf. Professor Wolf has just taken up, uh, this year a new position as the Edna J. Koury Distinguished Professor of Philosophy at the University of North Carolina at Chapel Hill, after having taught for many years at Johns Hopkins University, where she was chair of the Philosophy department, as well as Duane Peterson Professor of Ethics. Prior to that, she taught at the University of Maryland and at Harvard.

She has written extensively on topics in ethics, metaphysics, and the philosophy of mind. Her book, Freedom Within Reason, is one of the most influential works about the freedom of the will to have been written in recent decades. It’s a great pleasure to welcome Professor Wolf to Berkeley.

(applause)

[01:02:53] SUSAN WOLF:
Thank you. I didn’t realize until it– Can you hear me? Is this…

I didn’t realize until it was too late to revise my comments, um, that the lecture on the universal, on the formula of universal law, uh, was to be given on Election Day. That’s especially fitting since the ethics of voting or not voting, um, is a perfect example by which to consider the formula of universal law. Uh, but except for a reference to an elected official at the end, uh, an imaginary elected official at the end of my comments, I’m not working that in.

Uh, nonetheless, if there’s anyone here who hasn’t voted and for whom, uh, leaving this lecture early or, uh, would enable them to vote, uh, I do believe it passes the formula of universal law and- I could give my rational consent to it.

(laughter)

Yesterday, Derek Parfit’s lecture focused on Kant’s formula of ends in themselves, or as Allen Wood referred to it as Kant’s formula of humanity. He rightly called this Kant’s best-loved statement of the moral law, especially if by best-loved, he meant best-loved by scholars, and especially by friendly scholars of Kant. The formula that is the topic of today’s lecture, the formula of universal law, has some claim to being best loved by Kant himself.

At least Kant wrote of it that of all the formulations of the moral law, one does better if in moral judgment, he follows the rigorous method and takes as his basis the universal formula. Parfit’s lecture today shows us many reasons why we should not take Kant’s advice on this score. I share many of Parfit’s reasons for being dissatisfied with this formula, as I hope my remarks will make clear.

Still, Kant also thought that each formulation had something distinctive in it to offer, that each brought out some aspect of the moral law that was not highlighted by the others. If Kant is right, then there may be something important and distinctive about the universalization formula

(coughs)

that Parfit’s discussion fails to bring out. In seeing what this might be, it will be helpful to distinguish, as Parfit does, between two ways in which, according to Kant, a maxim can fail to be universalizable. A maxim may fail first because a world in which everyone acts on the maxim is impossible to conceive.

The idea of people acting on the maxim in question self-destructs when we try to imagine a world in which everyone so acts. When we ask whether the universalization of a maxim can so much as be conceived, we put our maxim through what John Rawls refers to as the contradiction in conception test. Kant believed that if a maxim failed this test, then acting on it would be wrong.

The formula of universal law, however, requires that it meet more than this contradiction in conception test. Of those maxims on which we can imagine everyone acting, we must ask whether we can will a world in which they do so. Could we, consistently with everything else that we will, will that everyone act on the maxim on which we are considering acting?

If not, the maxim fails what Rawls calls the contradiction in the will test. In that case, too, according to Kant, acting on that maxim would be wrong. Now, the contradiction in conception test has been much talked about by students of Kant, not just in scholarly treatises and conferences, but in introductory ethics classrooms as well.

Partly, I think this is because if it were true that one could separate right actions from wrong on the basis of whether one could so much as conceive of their maxims being universally adopted, that would be a really neat result, and because it is sort of fun to think about whether and why various types of action would self-destruct upon being universally performed. If one can find anything really neat or fun when studying or when teaching Kant, one should make the most of it. Partly also, I think the test has been taken seriously because it seems to yield some interesting and plausible results.

Parfit admits that the test succeeds in ruling out the abuse of what he calls trust-requiring promises. And though he speaks of these as being just one rather specialized species of promise, I think they rather constitute the core of promising. When one says, “I promise to meet you for lunch, to return the book tomorrow, to keep your secret,” one commits oneself to following through regardless of whether it is or continues to be in one’s interest to do so.

The usefulness of promising depends on our being able to rely, that is, to trust people to live up to such commitments. Without trust-requiring promises, we would still be able to make contracts and other less formal sorts of agreements. These would give us some of the benefits we get from promising, but probably not all of them.

Moreover, it seems to me that the quality of human interaction in the one case is different from what it is in the other. A world in which agreements were all made by contract and none made by promises would be a colder and more desolate world. A similar line of thought might be applied to the practice of truth-telling.

That is, it seems plausible to say that at the core of ordinary human communication is the understanding that people will tell you the truth whether it is in their interest to do so or not. That, in other words, you can trust them to be honest. It is against the background of this assumption that I believe what you tell me simply because you tell me, and not, for example, because you tell me and I have reason to believe that it would not be in your interest on this occasion for you to tell me something that is false.

Ordinarily, when I believe what you say, it is not because you’re saying it gives me evidence that it is true. Rather, I believe you, and in so doing, I trust you to be honest. Though the possibility of taking people’s statements as evidence of those statements’ truth would give us some of the benefits we get from simply being able to believe people, it would not, I think, give us all of them.

Moreover, the quality of human interaction in the one case is different from what it is in the other. A world in which whether to believe what people said depended on calculations of what it would be in their interest to say would be a colder and more desolate world. If I’m right, then trust-requiring promises and trust-requiring communication play a much bigger role in our lives than Parfit’s remarks suggest.

And so if, as I think, the contradiction in conception test rules out the abuse of these practices, that is a fairly significant result. Nonetheless, I think Parfit is right to reject the contradiction in conception test, and he is right because, as he says, it fails to capture a morally relevant idea. As Parfit points out, the fact that a policy is such that if everyone followed it, it would become impossible for anyone to follow it, does not in itself give us a moral reason not to do it.

If it did, as Parfit notes, it would be immoral to throw surprise parties or to adopt the policy of never taking the first slice. I love that example because it seemed as if it gave me some moral credit when I did take the first slice. But, uh,

(laughter)

if you think that through, that doesn’t actually work.

(laughter)

Also, though I believe that part of what is objectionable about making a false promise or lying is connected to the failure of these acts to meet the condition of universalizability, it is not, I think, specifically related to their failing to meet the contradiction-in-conception test. The second test, to which perf-Parfit turns in the second half of his lecture, seems to have more going for it. According to this test, an act is wrong if one not– cannot will its maxim to be universally obeyed.

Unlike the first test, this one does seem to capture something morally significant. Specifically, the idea that if I would not be willing to have everyone act a certain way, it would be wrong for me to act that way, expresses the idea that, in Parfit’s words, “It is wrong to make special exceptions of ourselves.” “This thought,” Parfit notes, “captures a kind of impartiality which it is plausible to think is fundamentally connected to a moral point of view.”

Parfit, however, believes that this kind of impartiality is not enough. For, he argues, the contradiction in the will test, or more generally, the formula of the universal law, could be passed or satisfied by a white racist hotel owner who wants to exclude blacks from staying at his hotel, or by men who want to treat women as inferior beings, or by a rich man who wants to keep his money rather than transfer any to the poor. According to Parfit, were a rich, white, racist male to ask himself whether he would be willing to have everyone act as he wishes to act, he might well answer yes.

If everyone were to act in ways that favor rich, white, racist men, that would be perfectly all right with him. Since, according to Parfit, these acts would not be excluded by the condition of universal, uh, the condition of universalizability or the formula of universal law, he concludes that what is wrong with these acts must be located elsewhere. To see what is wrong with these acts, Parfit argues, we should not focus on what the agent would be willing to have everyone do.

Rather, we should focus on what everyone, including the poor, the Blacks, and the women, would be willing to have the agent do. To make this shift is, in effect, to revert to a consideration of the formula of ends in themselves, or at least to Parfit’s interpretation of it as the principle of rational consent or the principle of universally willed acts. Does the Formula of Universal Law really allow the kinds of racism, sexism, and what, for lack of a better word, I will call classism that Parfit envisions?

Could a racist, sexist, classist sincerely affirm a world in which everyone acted as he is contemplating acting? As Parfit notes, the answer may depend on which description of the contemplated action or policy we choose as the relevant characterization of the agent’s maxim. The maxim of the racist, for example, might be formulated as, “Out of personal preference, I shall refuse to do business with blacks,” or alternatively, as, “Out of personal preference, I shall refuse to do business with members of a race other than my own.”

” More abstractly still, he may be described as wishing to refuse to do business with other rational beings on the basis of properties that are morally arbitrary. Depending on how the agent formulates his maxim, his answer to the question of whether he would be willing to have everyone act on that maxim may vary. Even if we accept the rather thick characterizations Parfit initially uses to describe the objectionable actions, we might questioning the reason– question the reasoning Parfit attributes to his rich man in the following way.

You think that you can will a world in which the rich don’t help the poor, but wealth is not an essential property of yours. You could lose your money like Job and become one of the world’s poor, And in that case, you wouldn’t want to have willed a world in which the rich treat you as you now want to treat the poor. Thus, one might argue that the rich man cannot rationally will that the rich don’t help the poor.

Unfortunately, analogous lines of reasoning in the case of the racist and the sexist are implausible. Still, the spirit of the formula of universal law seems contrary to racism and sexism, too. For if the core idea behind universalizability is that it is wrong to make an exception of oneself, a natural extension of that idea would recognize that it is wrong to make an exception of one’s race or one’s gender as well.

By appealing to the spirit of universalizability and to Kant’s views about which features of persons are morally significant, we may be able to supplement the formula of universal law with some rules for choosing the appropriate level of description for the maxims to be tested by it. And in that way, we may be able to yield prohibitions against the sorts of racist and sexist behavior Parfit imagines. If we could do that, then the formula of universal law might turn out to yield the same conclusions as the formula of ends in themselves, at least with respect to racism, sexism, and other forms of discrimination.

What I would be willing to have everyone to do would, in that case, when it came to treating one category of rational beings differently from another, would in that case arguably be the same as what everyone would be willing to have me to do. In that case, Parfit might think the two formulas were equivalent, or at least equivalent enough for the purposes of moral evaluation. Both, Parfit might think, would capture a single kind of impartiality, the kind that morality should embody.

I’m not certain, however, that the formulas are equivalent or that they would capture the same kind of impartiality. Nor am I sure that there is only one kind of impartiality that is morally significant. To give a sense of my doubts, I shall offer two examples.

First, let us reconsider the case of the German who wants to assassinate Hitler. Parfit claims that it would have been morally right for someone to assassinate Hitler, and though I am uncertain about this, it is easy enough to grant it for the sake of argument. Presumably, if it would have been right to assassinate Hitler, that is because Hitler was an evil man who, if not assassinated, would and did go on to bring about the deaths of millions of innocent people.

If w– if it would have been all right for someone to assassinate Hitler, it would presumably have been all right for anyone to have done so. The maxim, to prevent genocide, I will assassinate Hitler, meets the standard of universalizability. Does it also meet the standard of universally willed acts or of Parfit’s principle of rational consent?

Is it an act that could rationally be willed by everyone? This seems to me less clear. It is doubtful whether Hitler could have rationally willed that anyone assassinate him, or even that Hitler’s mother could have rationally willed this.

And to be honest, I am not sure that my judgment of whether it would be right to assassinate Hitler really depends on settling the question of whether Hitler could rationally have agreed with me. If it doesn’t, that would suggest not only that the formula of universalizability is not equivalent to the formula of universally willed acts, but that the formula of universally willed acts is not an absolutely necessary condition of moral permissibility. For my second example, imagine an elected official who is unusually smart, dedicated, effective, and incorruptible.

A number of government contractors have offered her gifts which she would like to be able to accept. Since she is incorruptible, she would not be moved by these gifts to reward these people with more or better contracts. And since her constituency, we may suppose, correctly, believe that she is incorruptible and want her to stay happy in her job, they have no objection to her accepting these gifts.

That is, everyone consents to her accepting gifts from government contractors. So the act of her accepting these gifts passes the tested of universally willed acts. Nonetheless, I am not sure that it would be all right for her to accept these gifts.

She would rightly not want others in her position to accept such gifts, and so, were she to accept these gifts, she would be making a special exception for herself. Here then, although the maxim, out of self-interest, I shall accept gifts from government contractors, appears to satisfy the condition of universally willed acts or of rational consent, it does not satisfy the condition of universal law or of universalizability. and this seems to make it morally questionable.

If it would indeed be wrong to accept these gifts, even if everyone consents, that again suggests that the formula of universalizability is not equivalent to the formula of universal- universally willed acts. This time, it would also suggest that the formula of universalizably– universally willed acts may not be a sufficient condition of moral permissibility, and that universalizability may be an independent, morally re-relevant consideration. In the end, we may find reason to decide that both formulas yield the same result in these cases.

If we supplement the formula of universally willed acts with a substantive theory of rationality, we may conclude that, after all, Hitler could have rationally consented to his own assassination, and that the official’s constituents could not rationally consent to their congresswoman’s acceptance of gifts. Alternatively, if we supplement the formula of universal law with a substance– a substantive theory of morally significant properties which guide us in finding the correct formulations of our maxims, we may conclude that after all, we could not will that anyone assassinate Hitler, or that we could will that anyone relevantly similar to our incorruptible official be permitted to accept gifts from government contractors.

[01:21:56] DEREK PARFIT:
Kant certainly thought that at bottom these formulas were equivalent, and something like this may be what he had in mind. He also thought, however, that there is a difference in the different formulas, which is subjectively rather than objectively prac-practical. I’m quoting him here.

Namely, each is intended to bring an idea of reason closer to intuition and thereby closer to feeling. These last remarks are offered in the spirit of that second thought and the idea connected with it that there may be more than one kind of impartiality that is morally significant. Um

(applause)

Susan Wolf describes and discusses, I think, seven different ways in which some act could be claimed to be wrong because its maxim cannot be universalized. I discussed four other versions of Kant’s formula, which I called the stated and actual criterion of strict duty, and the moral belief and law of nature versions of the formula of universal law. What Wolf doesn’t discuss those four principles, though she makes some claims which can be applied to them.

So between us, we discuss a total of eleven ways, in which some maxim might fail to be universalizable. You’ll be reliev- relieved to hear that I shan’t discuss all of these eleven principles. There are two main differences between the seven principles that Wolf discusses and the four principles that I discussed.

First, um, of the principles that, that Wolf discusses when she considers Kant’s formula of universal law, none refer to what we could rationally will. Thus, she writes, “I– If I would not want others to act in a certain way, I ought not to act in that way myself, and one should restrict oneself to doing only what one would be willing to have everyone to do.” And I think it’s crucial to appeal not to what one would want or will, but what one could rationally want or will.

Similarly, when Kant turns to my proposed revision of Kant’s formula at the end, she takes one of my suggestions, which she calls the formula of universal consent, to be one should restrict oneself to doing only what everyone would be willing to have you, the agent, do. Now, that’s what I call the veto principle, which I I think is quite unacceptable. We, we can’t appeal to what everyone would consent to, only to what they could rationally consent to.

The other main difference is that Wolff doesn’t discuss the moral belief versions of Kant’s various principles, and those, I think, are more plausible. We get better results if, rather than asking whether we could rationally will that everyone accepts or acts on our maxim, whether we could rationally will that everyone believes such acts to be permissible. Now, this reading, uh, which Scanlon proposed some years ago in unpublished lectures, I think is fully supported by Kant’s text.

He at least as often puts forward this as the law of nature than he does the law of nature formula. Well, I turn now to some of Susan’s more particular claims. I granted Kant’s actual criterion of strict duty one success, the case of trust requiring promises.

As I pointed out, Kant’s objection to his lying promiser wouldn’t condemn most actual lying promises because most lying promises don’t have the maxim of making trust requiring promises whenever that would be best for me. They have a maxim like, do whatever’s best for me. But as I argued, we must drop Kant’s appeal to maxims and should turn to what we intentionally do.

And on that revision, Kant’s criterion would condemn these promises. Uh, Wolf then claims that Kant’s criterion would also condemn other lies. The contradiction in conception test, she says, rules out the abuse of trust requiring communication.

That’s a fairly significant result. But I think that isn’t a result. Uh, Wolf says at one point that on Kant’s criterion, it’s wrong to act on certain maxims if one could not so much as conceive of these maxims being universally adopted.

Well, that criterion wouldn’t condemn even lying, trust-requiring promises. Everyone could accept some maxim, even if no one’s ever able to act on it. She also says that on Kant’s criterion, a maxim isn’t universalizable in one sense if we can’t conceive a world in which everyone acts on it.

I, strictly, that doesn’t condemn the trust-requiring promises, but it certainly doesn’t condemn lying. We can conceive a world in which everyone lies when they believe that would be better for themselves. Now, Kant’s actual criterion, I suggested, is that if everyone accepted a maxim or believed it was per-permissible, that would make it impossible for anyone successfully to act on it, and that doesn’t condemn lies.

We can’t claim that if everyone accepted the maxim of a self-interested liar, that would make it impossible for anyone to succeed in deceiving others to their own advantage. However, we don’t have a significant disagreement here since Wolf agrees that Kant’s criterion of strict duty in any of these versions fails to capture a morally relevant idea. So, Kant’s formula of universal law.

Wolf says, “Why should the fact that I would not be willing to have everyone act in some way be a reason to think that it would be wrong for me to act in this way?\” An answer suggests itself that it’s wrong to make special exceptions for ourselves. This thought may be expressed in the idea that if it’s all right for me to act in some way, it must be all right for anyone in my circumstances to do the same.\” Well, that sense of universalizability, though not unimportant, seems to me far from Kant’s formula, since it makes no appeal to claims about what we either would or could rationally want or will. It just expresses the truth that an act’s rightness supervenes on the properties that make it right, so that any similar act, relevantly similar act, must be right.

Wolf then makes claims which could be applied to the law of nature formula. Um, and she discusses the impartiality objection. Uh, she suggests that Kant’s formula might avoid this objection if we redescribe the relevant maxims.

Perhaps the white racist maxim might be described as refusing to do business with members of a race other than his own. But I think that doesn’t help. Um, Question.

That’s about it, real that blacks refused to do business with him by not

[01:29:17] AUDIENCE MEMBER:
Right.

[01:29:20] DEREK PARFIT:
trying to use his hotels, and he could certainly accept that. Um, more abstractly still, O’Neill continues, he may be described as wishing to refuse to do business with others on the basis of properties that are morally arbitrary. Well, if we’re talking about this man’s actual maxim, that wouldn’t be the right description of it.

But I’ve argued that we ought to revise the formula so that it appeals to what people are intentionally doing, and we might say what he’s intentionally doing is refusing to do business with others on the basis of a morally arbitrary distinction. However, I don’t think that’s the relevant description of what he’s intentionally doing. we may decide that in doing what he does, he’s giving the wrong weight to an arbitrary distinction.

But that’s our comment on what he’s intentionally doing, and a lot more would have to be said for this objection to work. There are countless morally arbitrary properties which it could be rational to take into account. Turning to the rich, Wood follows Kant, suggesting that the rich couldn’t rationally will that rich people don’t give to the poor, since wealth isn’t an essential property, and they could never be sure that they might not lose all their wealth and become one of the world’s poor.

Well, I think that’s an inadequate reply. First, the rich may know that the likelihood of that happening is extremely small. I mean, they can keep their money in different stocks and shares and so on.

Second, Wolf claims that if this unlikely event occurred, this newly poor ex-rich person wouldn’t want to have willed a world in which the rich give nothing to the poor. But that version of Kant’s formula, as Rawls points out, is much too strong. It assumes we couldn’t rationally will something if there’s a possibility we might later regret having willed it.

Wolff says that this kind of move, redescribing the maxim, can’t be made in some of other cases. No, this isn’t redescribing the maxim. That’s saying, be careful.

Yeah. Such as those involving differences of race or sex. Well, I think in case of sex, that isn’t clear.

It’s at least as likely that a man might discover he has transsexual tendencies and turn himself into a woman as that a rich man might become very poor. However, um, she continues, “If the core behind the idea of universalizability is that it’s wrong to make an exception of oneself, a natural extension of that idea would recognize that it’s wrong to make an exception of one’s race or gender as well.” Well, on Kant’s formula, acting on a maxim is wrong if the agent couldn’t rationally will that everyone acts in that way.

And on Wolf’s suggested extension, an act is wrong if it makes exceptions of one’s race or gender. And that, I think, really is far from Kant’s formula. Uh, she suggests other ways of rescuing the formula, um, but there are other simpler revisions.

Um, and I think the best, simplest revision is to switch to what everyone could rationally will. That’s very Kantian, and it meets the– this objection perfectly. Um, well, finally, she discusses the rational consent principle and the wider formula of universally willed acts.

Case of Hitler, um, she says, could Hitler or Hitler’s mother rationally willed that he be assassinated? Well, there is a question there, But I think the answer is yes. Um, helps to return to the question whether people could have rationally given in advance unconditional consent to being treated in some way, even if they later don’t consent.

And Hitler could have rationally willed in advance that if he became a mass murderer inflicting suffering on millions, he’d be killed before he’d do more harm. I mean, he prob– unlikely that he would have, but he, he had sufficient reason to will that. Uh, Wolf then continues, “Some of us might, if we being honest with ourselves, wonder whether our moral attitude to the prospect of Hitler’s assassination really depends on whether it could be rationally willed by Hitler himself.

And she, she suggests that if the answer is no, that would show that the formula isn’t an absolutely necessary condition of moral rightness. Well, I accept the first suggestion here, but I’m inclined to reject the second. Our attitude to Hitler’s assassination would not, I agree, depend on whether it could have been rationally willed by Hitler him-himself.

The rightness of the assassination could be plausibly defended in many other ways or by appeal to many other principles. Our question is only whether the consent principle here gives the wrong answer by condemning Hitler’s assassination, and I believe it doesn’t. Um, she ends with the incorruptible elected official.

And as she rightly says, such cases may suggest that the formula of universal consent may not be a sufficient condition of rightness. Now, I didn’t intend it to claim to be that. That’s why it says an act is wrong if people couldn’t rationally consent, not if and only if.

However, if we’re looking for a single principle that appeals to the idea of rational choice or consent, then I think we should turn from the formula of universally willed acts to the principles whose universal acceptance. And that’s what I shall discuss tomorrow. Uh, but I just end with one small point.

Um, Wolf doesn’t discuss my argument for the claim that Kant should give up his appeal to the agent’s maxims and appeal to what he intentionally does. And I thought the case of the egoist and Kant himself seems to me to show that. Now, this proposal isn’t in, in British parliamentary terms, a wrecking amendment.

I don’t think it weakens Kant’s view at all if he drops all references to maxims. Uh, and I wondered if she thinks that too.

[01:35:42] COMMENTATOR:
The, the tyrannicide example and the example of telling a lie in order to save the Jews and that sort of thing. And I suppose, uh, it wasn’t clear to me what we were supposed to be imagining in those thought experiments. It seemed at least as plausible that a world in which everyone believed that tyrannicide was permissible, uh, would be a world in which the guards also believe this.

Uh, and it seems to me at least possible that in some of the closest of those worlds that it is succ– uh, it is, um, for different reasons than it would be in the actual world, but it is possible successfully to act on your tyrannicide, tyrannicidal, uh, maxim or policy. It might be the guards would help you in that situation since they believe it’s permissible to do that sort of thing. And to make sense of that belief, you’ve got to assume enough else about the rest of their moral beliefs, um, to to make that a sensible thing for them to hold, you could say.

And when you hold all these things fixed, at least some of the closest worlds in which that condition is satisfied might also be worlds, it seems, in which I successfully, uh, act on my tyrannicidal policy. And, you know, you might say something similar for the case in which everyone believes it’s permissible to tell a lie in order to save the Jews. Um, that’s also a case in which, uh, you know, even the, the Nazis believe this, and well, I don’t, you know, it’s really hard to get your mind around those kinds of thought experiments.

But, uh, but, you know, I don’t see why we should rule it out so quickly that, um, it– some of the worlds in which, uh, that is universally belie-believed by everyone, including the, the Nazis, are worlds in which it is pos-possible for you by lying to, uh, to actually succeed in saving

(cough)

some, some Jews. It’s not necessarily the case. It’s not obviously the case anyway, that the Nazis, if they believe that that’s permissible, are going to be, uh, doubly vigilant about, uh, protect– uh, prohibiting you from doing what you’re doing.

Um, anyway, so I wanted to hear a little bit more about why, uh-

[01:37:38] DEREK PARFIT:
Well, I think I’d say that, um, there are different versions of the case in which the argument wouldn’t apply, but I’m supposing that the facts were as I’ve described. I mean, even if everyone, including the guards, believed that the tyrannicide was not wrong, it’s not that difficult to get guards who will nonetheless act as guards despite having that belief. So I mean, I think I’ll just say my examples, what I suppose to be true, That isn’t wildly implausible.

Actually, I think it’s fairly plausible. Now, you can describe other possible worlds that aren’t very distant in which it wouldn’t be true, but it’s enough to describe one world in which it would be, unless it’s really fantastic, and I don’t think that’s so. Um, I mean, I think if you look at the difference in what happened in Germany and what happened in Italy, uh, you get very striking differences of exactly this type, uh, in the second world war.

[01:38:34] COMMENTATOR:
Well, but, but the, the test asks you whether the, the, the universal acceptance– a belief that the acting on the maxim is pers- permissible would make it impossible for anyone successfully to act, act on that maxim. That seems to me to require that, uh, you know, any of the closed possible worlds in– that all of the closed possible worlds in which, um, um, the belief that it’s, uh, permissible to act on this maxim-

[01:39:01] DEREK PARFIT:
Well, well- well- that might be the case, but I don’t see why. Because what I’m saying is, look, even if in the world as it actually is or it actually was, even if in this kind of case, the formula would have gone obey, we can easily suppose a world not very different from the actual world in which this formula would have given clearly the wrong result.

Um, and I think that’s enough. I don’t have to say, you know, that in every possible world it would give the wrong result in this kind of case. But perhaps that’s wrong, but I assumed it was enough to get counterexamples, and this is a counterexample.

[01:39:44] COMMENTATOR:
Well, I just– I guess I’m, I’m still… The, the, the thought experiment we’re asking here when we’re– we ask that– we, we ima- try to imagine that the belief in the permissibility of some kind of action is universally accepted, and then ask whether its being universally accepted would make it impossible for us to act, uh, um, on the original maxim. It’s, it’s a very hard thought experiment to get your-

Thank you, chair.

[01:40:07] DEREK PARFIT:
No. Well, we’re asking not whether in the actual world-

[01:40:13] COMMENTATOR:
Well, right,

[01:40:14] DEREK PARFIT:
something’s true.

[01:40:15] COMMENTATOR:
Yeah. We’re asking-

[01:40:15] DEREK PARFIT:
But whether if the world had been, as I suppose it to be, as long as it’s not wildly implausible, what would the criterion then have implied? Um, I mean to say, oh, but that’s an imagined case, it wasn’t actually true, doesn’t seem to be relevant.

[01:40:38] COMMENTATOR:
No, I, I’m sorry. That was not the, the complaint. I mean, the, the, the criterion isn’t in any case asking us to imagine, uh, what things would be like, uh-

Yeah, if a certain counterfactual condition were satisfied, namely that everyone believed that a certain kind of action is permissible, right?

[01:40:57] DEREK PARFIT:
Yeah. And okay. And so every, every application of this asks us to imagine about, uh, imagine what would be the case in a certain counterfactual situation, and, uh, in particular, to ask whether assuming a certain counterfactual condition to, to hold, namely that everyone believes, uh, a certain kind of action to be permissible, it would be impossible, uh, to successfully act on the original intention.

Uh, and that seems necessarily to be a thought experiment. Well, I mean, what I’m, what I’m assuming is that, is that what I’m— The question is now that there are possible worlds in which the counterfactual condition is satisfied. If the head of the Gestapo took some of the agents and said, \”Well, look, there are all these people who, you know, won’t tell you the truth about where the Jews are, so you shouldn’t just go and ask people.

You shouldn’t trust what they say.\” And then the executive said, \”Well, we won’t do it that way. We won’t, you know, if we ask if you’re keeping any Jews, and they say, ‘No,’ we won’t believe them.

We’ll go in and search.\” Now, I mean, why can’t I say that might have been what would have happened? And if it had been what would have happened, this criterion would condemn lying to the Gestapo about whether there was a Jew in your attic.

[01:42:09] SUSAN WOLF:
Right. Uh- Um. Well,

[01:42:10] COMMENTATOR:
I guess my worry is that what you s-argued suggests that some of the possible worlds in which everyone believes that acting in the maxim is impermissible are worlds in which you can’t successfully act under the original maxim.

[01:42:23] DEREK PARFIT:
It’s one side complication, which is clearly what I’m doing. I’m saying in one possible world, then certain truths can hold about what in another possible world, which is related to that one, would have happened. But then I think I’d say I’m building into my account of the possible world plausible assumptions which would have supported the claim that in the relevant counterfactual, it wouldn’t have worked.

Do you see what I mean? I mean, I see that, that there are two possible worlds. Two possible, right.

Right. Um, But at the moment, I, I think it’s okay for me to suppose that things might easily be— have been such that a certain counterfactual would have been true. That’s all I need.

[01:43:05] COMMENTATOR:
Yeah. Um— Well, I don’t wanna keep going too long about this, which isn’t anyway a major point, but it, it’s, it’s hard for me, for what it’s worth, to imagine changes in the actual world such that assuming those changes, uh, um, you know, the thought experiment, uh, fails. Um, but perhaps this is a lack of imagination on my part.

[01:43:26] DEREK PARFIT:
Robert? Well, but, but I mean— Why don’t we, why don’t we take some of this? Robert?

[01:43:31] ROBERT:
Um, if I could, I’d like to just pursue, um, Susan Wolf’s question about the description of a maxim. Suppose that, uh, we– the, the maxim of the sexist man is not oppress women, but rather oppress people based on their sex. Or the maxim of the racist is not oppress blacks, but oppress people based on their race.

Would that solve the problem?

[01:43:56] DEREK PARFIT:
I don’t think so. Why? Um, Um, because, um, it could be true of my actual man that his maxim isn’t treat people of the opposite sex as inferior, but it’s treat women as inferior.

And that could be true because he’s in a world in which everyone, including most women, believe that women should be treated as inferior. It’s just a misdescription. Now, you could say, and I argue, that you shouldn’t appeal to his actual maxim, but to what he’s intentionally doing.

Um, but I don’t think that’s going to help enough in those cases. Um, because when it’s, you know, the powerful and the weak and the rich and the poor, if the description is, um, that the powerful people are accepting the maxim, you know, use your power to… That doesn’t–

If the conch– I mean, I don’t think you’re gonna get a redescription

(cough)

that’s gonna work in those cases.

[01:44:54] ROBERT:
Well, I guess the question is when you say the actual maxim–

[01:44:57] DEREK PARFIT:
don’t, you know, don’t give donations whose financial position is very unlike yours. Right. Well, that will imply that the poor shouldn’t give anything to the rich. But that’s not going to worry the rich.

[01:45:09] ROBERT:
Well, if we take it back to the case of race for a second. Why do you say that it’s not the actual maxim? I mean, is that a psychological point?

[01:45:16] DEREK PARFIT:
No, that’s what that’s Kant’s own view. He– Kant is appealing to the agent’s actual maxim. I think he shouldn’t, um, but so long as we’re discussing Kant’s view, we have to appeal to what the agent’s actual maxim is.

[01:45:28] ROBERT:
No, but my question is, what counts as the actual maxim?

[01:45:31] DEREK PARFIT:
Oh, well, it depends on, uh, a combination of what the agent himself actually decides to do, what the policy is. But it also depends, ’cause often you don’t fully formulate it, on how you would have reacted if things had been relevantly different. Um, So that’s the way you find out what his policy actually is.

[01:45:56] ROBERT:
So in other words, I would talk to the sexist man, and I would say, uh, “Do you believe that you should oppress women?” “Yes.” “Do you– Does it follow that you should oppress people based on their sex?”

And if they say no, then it’s not their actual maximum. If they say yes, it is. That’s how it works?

[01:46:11] DEREK PARFIT:
Not quite as simply as that, because, uh, sometimes people will not understand what their policy is.

[01:46:18] AUDIENCE MEMBER:
Um,

[01:46:20] DEREK PARFIT:
but the prior question, as I say here, is whether I was right in claiming it. I mean, I think there’s no way of denying that there are many people who have only one maxim, do whatever is best for me. That’s their only policy.

And as far as I can see, that’s decisive because clearly you can’t say that when the egoist saves a drowning child, he’s acting wrongly. So I think you have to turn to what he’s intentionally doing. Um…

[01:47:01] HOST:
Alan?

[01:47:02] ALAN:
Oh. I agree with Susan that the way that this part of the groundwork has been read has been determined a lot by what people think is fun and neat. Or maybe we could be a, a little more dignified and say what is philosophically interesting and productive, and I think there is a lot that’s philosophically interesting and productive in the way of reading the, the discussions of the formula of universal law and so forth.

But I actually think that, um, this is a deeply mistaken reading of the text. That– And let me just sketch wha-how I think the text ought to be read, not that I can defend it at, because it would take a lot to do. But I think you ought always, in reading the second part of the groundwork, to keep your eye on the ball, which is that he’s trying to derive a system of formulations of the moral law.

The formula of universal law and its subsidiary formula of the law of nature are the first, most purest, most abstract, the earliest in a developing system of formulas, of which the last is the most adequate and the most universal. And for that reason, I disagree with one thing that you said that, um, that most people say. I mean, you’re not surviving it alone.

Maybe I’m the only one who thinks this, but I think when he talks about the universal formula, he’s not talking about the formula of universal law. And I think if you look at what the formula says, you will see that he’s talking about the formula of autonomy.

[01:48:54] DEREK PARFIT:
What happens when he says– sorry, you act on a maxim that can make itself– that can make itself. That’s different from– can you, can you will to be– How can a maxim make itself a universal law?

[01:49:06] ALAN:
Because the will, the, the, the, the rational will does will it, does legislate it. The will that formulates that maxim does at the same time legislate it.

[01:49:17] DEREK PARFIT:
Right. That’s the– Yeah, that’s how I read that. Yeah.

[01:49:21] ALAN:
And also, if you look where that, uh, universal formula appears, uh, it appears in the paragraph devoted to the realm of ends formula, which is the more intuitive version of the formula of autonomy. So he’s listing that formula in the appropriate paragraph. Moreover, the way I read the phrase universal formula is, what he means is the formula that is universal in the sense that it best sums up all five, and that’s the formula of autonomy.

Um,

[01:49:56] DEREK PARFIT:
so I mean, that, that’s the s-that’s the general strategy. And I see the, the four examples listed, uh, under the formula of the law of nature as a, an attempt. Kant was notoriously bad at popularizing what he wanted to say, and I regard this as an example of his failed attempt to popularize.

He was trying to give… toss a bone to the poor reader who doesn’t, you know, doesn’t see any relation between what the, what his abstract reasoning about the rational will and morality might have to do with ordinary moral judgments. He’s trying to give you some, you know, grasp on what’s going on. And so he enumerates four duties, and he’s assuming, he picks maxims that would be typical maxims to vio– on which to violate each of those duties.

And as he says at the end, these are cases in which we already do will as universal law, that’s where the duty comes from, the opposite maxim. And so you– he’s picked these maxims to be specific contraries of maxims that he takes to be universally legislated. And the point of the four examples, as I see it, It’s just that when you have these duties and these laws that you’ve already got, um, then you can, you can be brought fairly easily to see that making– trying to make, uh, an exception of yourself by adopting a maxim that is directly the contrary of these dutiful ones, uh, you, you can’t, you can’t consistently will both the law that you do will and the contrary maxim.

Which I admit sort of seems to… I mean, that takes all the fun out of it to read it if that’s the only point he’s making, but I think that is the only point he’s making. So

(coughs)

[01:51:52] HOST:
Do you want to respond, sir?

[01:51:54] DEREK PARFIT:
Um, well, that sounded like saying that he was only intending to make a small trivial point. I think he had– I think the supreme principle of morality was intended to do more than that.

[01:52:16] ALAN:
Sure. All five formulas. All five. Not, not the first, most abstract, poorest formula. It wasn’t intended to do everything.

[01:52:27] DEREK PARFIT:
Well, I mean, in, in some ways, my interpretation is quite close to yours, because– What I’m calling, uh, Kant’s contractualist form, um, it’s wrong to act on maxims or act in ways, um, that are disallowed by principles that everyone could rationally will to be universal. And it could take the form not disallowed by principles. That I think, uh, is what Kant would turn to if he’d realized that that differed from the appeal to the maxims that I could rationally will.

If he thought that different people could rationally will conflicting maxims, and he considered my case about the rich and the poor, the weak and the strong, the obvious move to make is the one that Scanlon suggests, switch to what everyone could rationally will. And now our point is this, that’s very close to the formula of Kingdom of Ends. He never quite says it, but I think you’re staying very close to Kant if you’re doing that.

[01:53:28] AUDIENCE MEMBER:
Mm-hmm.

[01:53:28] DEREK PARFIT:
And then the claim to make about the original form of that is that, yeah, no, it, it needs to be enriched, and this is the way to enrich it.

[01:53:35] AUDIENCE MEMBER:
Right. Yeah. Thanks. My question is precisely about how sympathetic is your revision to, um, to the original formulation. If I understand rightly, the, the big move in this chart is from our act is wrong unless we ourselves-

[01:53:52] DEREK PARFIT:
Yeah.

[01:53:52] AUDIENCE MEMBER:
-to our act is wrong unless everyone.

[01:53:53] DEREK PARFIT:
Everyone. That’s the big move.

[01:53:54] AUDIENCE MEMBER:
This is the, the big move.

[01:53:55] DEREK PARFIT:
Yeah, yeah.

[01:53:56] AUDIENCE MEMBER:
And so… And the reason why I’m wondering about how sympathetic the move is, is because I always thought that the, the, uh, the fundamental idea be-behind this Kantian formula is to, uh, ground moral decisions in what individuals, uh, can, uh, uh, reason or can by practical reason allow themselves to do.

[01:54:21] DEREK PARFIT:
Well, it’s– I mean, that maybe,

[01:54:22] AUDIENCE MEMBER:
and now this is the fundamental point behind the, the, uh, the Kantian ideas, then the question of what other people could rationally do is, is a great revision So I’m wondering how, how big of a revision it is or how sympathetic it is to the, to the Kantian idea.

[01:54:37] DEREK PARFIT:
Yeah, well, I mean, I think one way of interpreting it, um, is that in several of his formulations, not the second one, but in several of them, and in other contractualist views, you appeal to what people could rationally will or choose or consent to, and you do it in a way which achieves impartiality. And I contrasted three main ways of doing that. Um, you can appeal to what the agent himself could rationally will if he supposed that he was going to be on the receiving end.

That’s the golden rule. You could appeal to whether the agent could rationally will, not only that he acts in this way, but that everyone who’s able to acts in this way, and that everyone thinks that’s permissible. So that gets the impartiality because it says, “Look, you may be able to rationally will that you do it, but you couldn’t rationally will that everyone does it.

Now, that is a way of trying to get impartiality, and my claim is that it com– it fails very seriously in these main kinds of games.

[01:55:43] AUDIENCE MEMBER:
Right.

[01:55:44] DEREK PARFIT:
Well, so Nagel then suggests make it more like the golden rule. Uh, Rawls suggests pull down a veil of ignorance. The simplest move is just to switch to what everyone would rationally will. Now, if Kant’s aim was like the aim of some tough Hobbesian contractualists like Gauthier,

[01:56:04] AUDIENCE MEMBER:
Mm-hmm.

[01:56:04] DEREK PARFIT:
which is to show that starting from a non-moral rational self-interested agent, you can show that he himself couldn’t rationally choose that he act in a way that’s wrong. That would be more impressive, but it can’t be done.

[01:56:19] AUDIENCE MEMBER:
Um- So it is, so it is a- You have to- a great revision and not so sympathetic to that-

[01:56:24] DEREK PARFIT:
No, no, it’s not a great re– No, it’s not a great revision of Kant’s view because I don’t think Kant thought that you could show that it’s irrational to someone on non-moral grounds for him to act in a way that’s wrong. Now, Kant believed that you couldn’t rationally act wrongly, but rightly he didn’t say

(laughter)

, um, when he’s blind, could you rationally will it be universal? He didn’t say, “Oh no,” you couldn’t because you’d be acting wrongly, and you couldn’t rationally act wrongly.” I mean, that just would’ve had no interest.

Now, this is an attempt to work out what’s wrong by appealing to assumptions about what you could rationally choose or will. So it’s not a great revision of Kant to appeal to what everyone could rationally will. Um, he comes terribly close to doing it, and I think he only doesn’t do it ’cause he thinks it’s not necessary.

And I also think my other revision, dropping the appeal to maxims, I don’t think that that’s a wrecking amendment. Um, Onora O’Neill gives two main reasons why she thinks that the appeal to the agent’s maxim is an indispensable part of Kant’s moral theory. One is she thinks that it solves the problem of what the morally relevant description of the agent’s act is, and the other is that it enables Kant to claim that there’s a contradiction in people’s will when they act in a way that’s wrong.

Now, the point about the first is that it’s not true. Isn’t the morally relevant description of my egoist’s act that he’s doing whatever would be best for himself? That tells you nothing about relevantly what he’s doing.

Uh, as to the second, I don’t think Kant does think that you can show that it’s irrational for the agent to be acting wrongly as it stands because there’s a contradiction in his will. Uh, you have to ask the hypothetical question, what if everyone? So I don’t think either of those reasons for appealing to the agent’s maxim, um, have force, and I don’t think that Kant loses anything, uh, if the criterion of whether the act is wrong or contrary to duty appealed to what the agent intentionally did.

The agent’s maxim has great bearing on whether his act has moral worth and whether you should think of him as a morally admirable person, but those just need to be treated separately. That’s what I think.

[01:58:44] MODERATOR:
Susan?

[01:58:45] SUSAN WOLF:
Um, I just wanted to make a comment in connection with this last question. Um, if– Insofar as the question is, um, is this a sympathetic or unsympathetic revision of Kant’s view, where Kant’s view means Kant’s overall moral theory, then, um, perhaps it’s, it’s right to say, no, it’s not unsympathetic, and it’s not even a big revision to Kant’s view. Um, because in connection with your response to Alan, and, um, I would agree with this too, Kant’s view as a whole suggests that we look at all the different formulations of the supreme moral law and interpret each one in the light of the other as a unified whole, and then we get a view that brings all these in.

Insofar as the question is, look, in saying that a revision of the formula of universal law is, now let’s switch. From what I would will everyone do to what everyone would.

[01:59:44] DEREK PARFIT:
Well, I do.

[01:59:45] SUSAN WOLF:
That seems to me, as it does to the questioner, that’s a huge revision of that particular formula-

[01:59:52] DEREK PARFIT:
Oh, well I-

[01:59:53] SUSAN WOLF:
Because it basically reverts-

[01:59:54] DEREK PARFIT:
Yeah

[01:59:54] SUSAN WOLF:
to, uh, the formula of humanity. It, it changes the attention from what the agent is thinking to what other people-

[02:00:01] DEREK PARFIT:
Yeah,

[02:00:01] SUSAN WOLF:
could rationally will.

[02:00:02] DEREK PARFIT:
But, but I mean, what I’m treating as the plausible Kantian revision is not whether everyone could rationally will that I do what I’m doing. I mean, I had that on the list-

[02:00:14] SUSAN WOLF:
Right,

[02:00:14] DEREK PARFIT:
because it naturally comes out of the taxonomy, and it’s quite an interesting principle. Um, but the natural move from the formula of universal law, particularly in its moral belief version, isn’t to that. It’s to whether everyone could rationally will that such acts be believed to be permissible.

[02:00:32] SUSAN WOLF:
Oh, well, Okay. I had a different– I had a different suggestion, which you may still regard as a sympathetic one. Um, I left out the morally believed stuff because– Yeah, having not read Tim’s unpublished paper, I have never seen that in Kant’s writing at all, and never–

So I just never thought of– Oh, wait, no, wait, wait. Yeah.

I just had a different point– Yeah. Which was something you said at the end of your comments to me, which I thought was exactly right in your further move, which is to switch from a principle of universally willed acts that says, “Look at what everyone would will that I do on this occasion,” to a principle of what everyone what principles everyone would accept.

Yeah. Right? Um, by making that move, you’re building in both the attention to what others feel and the move that’s essential to the first.

formulation, universalization formula, which is to look at things at a principled level of what, Yeah, you know, what people generally do or as a re– And that’s, I think, the really important move that sort of brings the universal law formula into the formula that you’re eventually gonna be-

[02:01:45] DEREK PARFIT:
Yes, I agree. And, and the natural way to treat the formula of universal will with that is that that’s a broader version of the consent principle. And I mean, I, I do think that if you ask what can be made of, He cannot possibly agree to my way of treating him.

As I said last time, uh, O’Neill and Korsgaard think, uh, well, if he can’t possibly agree to it, that’s because we haven’t given him knowledge and power over the proceedings. And that suggests that we have to give him knowledge and power. We shouldn’t treat him in any way that he doesn’t choose to be treated in, the veto principle, and that I think is just far too strong.

[02:02:25] AUDIENCE MEMBER:
Well, if you switch to it, no, not— have we given him the veto, but if we gave him the choice, could he consent? Well, the answer is trivial. Yes, of course he could.

[02:02:42] DEREK PARFIT:
The question is whether he could rationally consent. So I do think that’s plausibly there. And now the formula of universal law is just the wider version of that.

But you’re quite right in saying that the natural inheritor of the formula of universal law is, uh, the contractualist formula that appeals to the principles that everyone’s rational about. Yeah. Um, I’m sure that’s right.

And I mean, one way of putting that is, is it that you could will to be a universal law. Now, that’s very plausibly a matter of the principles. Um, now, in the other version, it’s that you could rationally will everyone to do it, but actually that doesn’t work as well as that it’s a universal law about what people may do.

Uh, whereas the formula of the realm of ends, there’s no reference to universal law, and that’s why that more naturally leads to consent. Um, that’s my way of, um…

[02:03:37] HOST:
Um… Chris?

[02:03:42] CHRIS:
Uh, yeah, I was wondering if in a different respect, you might be too charitable to Kant. You say that the, um, with the respect to, uh, pollution cases, sort of the tragedy of the commons cases, his formula, Well, actually, you say it only works better, but– but, uh, it works better than other formulations. I’m wondering how well it works in those cases at all.

In the first place, where it’s going to work will depend on… So let’s say it– I think you assume it works where the costs of everyone polluting outweigh whatever gain there is to my own, um–

[02:04:11] DEREK PARFIT:
Yeah, my own polluting. And, and then– it’s a gain in something, but imposes a much greater– cost than others. But it’s so thinly spread that the effect on each other person is trivial or imperceptible.

[02:04:19] CHRIS:
But so even in those cases, uh, it’s going to depend upon a very narrow conception of rationality, a narrow conception of cost-benefit rationality.

[02:04:25] DEREK PARFIT:
No, I don’t, I don’t see why that is so.

[02:04:27] CHRIS:
Well, sorry, hold on one second, mate. Sorry. Um, but the main problem is that it won’t work in a lot of cases that intuitively still count as sort of morally impermissible free riding, namely all those cases in which it’s in my interest to pollute

and it’s better yet when other people don’t pollute, but it’s still in my interest when others do pollute or in, or in, you know, extract some benefit from the commons. And we want a principle that’s gonna exclude those cases.

[02:04:50] DEREK PARFIT:
Well, no, but you’re- Yeah. This doesn’t.

Well, I mean, in a fuller statement of it, the point is that the question is, could I rationally will that everyone rather than no one acts in this way? And these are all cases in which if no one did it, that would be better for everyone than if everyone did it.

[02:05:04] CHRIS:
Right.

[02:05:04] DEREK PARFIT:
Prisoners’ dilemmas. Right. The problem is that however many other people are doing it, it can be better for me if I do it.

[02:05:11] CHRIS:
Right,

[02:05:11] DEREK PARFIT:
um, now that applies in ordinary prisoners’ dilemmas. Um, the point about perceptible, imperceptible, and trivial harms cases is that many people think, you know, even if you take into account the effect on others, if it’s hardly making a perceptible difference on anyone, people incline to think, well, that can’t be wrong. Uh, now I actually think that on a proper reading of how you should assess the effect of your acts, that’s a mistake.

But that’s reading that most people are not going to find very plausible, and that’s the case in which you say, “Well, what if you had– could choose whether everyone rather than no one did it?” Then everyone gives the right answer because by definition, these are cases in which, um, each act benefits the agent but imposes greater costs on others. Um, so it would be better for everyone if no one did it than if everyone did it.

A-and moreover, I should add, this is not restricted to cases of self-interest. It’d be better for everyone’s children if no one did it rather than everyone did it. Um, parents’ dilemmas.

Uh, um. I mean, as Broad says, it’s the moral microscope. It, it, it blows it up in such a way that then people are going to see.

Um, and I think that’s what if everyone did that? Kant was, as it were, developing that thought. But it only works in some cases.

[02:06:41] HOST:
Uh, final question, yes.

[02:06:46] AUDIENCE MEMBER:
(laughter)

Great. Um, Now I under– let’s take on Kant’s idea that, uh, this universal will will be good for everyone. But how does he account for the skill that some have in pursuing rational truths?

And, uh, while most people will struggle to perceive rational truths? Is rationality impaired or aided by nature? Um, or is it by training the rational mind in, like, schools or churches?

Or is it, is it a free choice, um, to inquire, which we sometimes don’t make because of risking discovery that some fears, discovery of our inability, or the fear that we don’t actually have this rational will? Is it one of those, or does he, does he answer that at all?

[02:07:36] DEREK PARFIT:
Uh, I think Alan would be more able to answer that than me.

[02:07:43] ALAN:
You’re the one who’s on the spot.

[02:07:44] DEREK PARFIT:
No, but you’re the person– That was a question about Kant’s views. Yeah. And you understand those much better than I do.

[02:07:54] ALAN:
But I don’t think, I don’t think I wanna try to answer it, so…

[02:08:03] HOST:
Well, um, I think the time has come for me to go and vote. Yes. And, um, you’re all invited back tomorrow, same time here again to hear the third of Derek Parfit’s Tanner lectures, which was titled simply Contractualism. Thank you very much.

(applause and cheering)

[00:00:00] ANNOUNCER:
Okay. Welcome to the third of Derek Parfit’s three Tanner lectures. Today’s the day the rabbit gets pulled out of the hat. Um, you’re all invited back tomorrow at this time to

(laughter)

discuss how it happened. Um, Derek Parfit lecturing on contractualism.

[00:00:26] DEREK PARFIT:
Contractualists ask us to imagine that we are all trying to reach agreement on which moral principles we shall all accept. According to what we can call the rational agreement formula, we ought to act on the principles to whose acceptance it would be rational for everyone to agree. I should say that people choose the principles to whose acceptance they agree.

People choose rationally, most contractualists assume, if their choice would be likely to be best for themselves, and we can start by making that assumption. Though there are some principles whose acceptance would be likely to be best for everyone, there are others whose acceptance would be best only for certain people. What would be best for the rich, for example, would not be best for the poor.

It may seem that in such cases, there would be no principle whose choice would be rational for everyone in self-interested terms. But everyone would know that everyone would accept only the principles that everyone chose. So what each person ought rationally to choose would depend on what others were likely to choose.

There would be no point in our choosing the principles whose acceptance would be best for ourselves if these principles would not be chosen by everyone else. What we ought rationally to choose would also depend on the effects of our failing to reach agreement. Most contractualists tell us to suppose that if we fail to agree, no one would accept any moral principles, so no one would believe that any acts were wrong.

This no agreement world would be likely to be bad for everyone. That would give everyone strong reasons to choose to try to reach agreement, and it might be rational for everyone to choose not the principles whose acceptance would be best for themselves, but the principles which other people would be most likely to choose. That might be everyone’s best hope of avoiding the horrors of the no-agreement world.

Now, such reasoning is for some writers the essence of contractualism. On this view, we should regard morality as if it were a mutually advantageous bargain. When people’s interests conflict, it would be rational for everyone to agree to certain principles to resolve these conflicts, And by appealing to that fact, these writers claim, we can justify these principles in the actual world in which there has been no such agreement.

Now, to make the imagined agreement easier to achieve, we can suppose that there’d be discussions and a series of votes, but there would have to be some final vote. It must be true that if we failed to reach agreement in this last round, we’d have lost our chance and couldn’t try again. In earlier rounds, it would be rational for us to try to reach agreement on terms that favored ourselves.

Only in the decisive final vote would it be rational for us to make our full concessions to others. I mean, there’s got to be a real danger of the no agreement world, and that was created by having one final vote. Well, there is now a complication.

The no agreement world would be bad, less bad for certain people, such as those who control more resources or have greater abilities. In a world without morality, such people would be better able to fend for themselves. These people would have less need to reach agreement, and that would give them greater bargaining power.

They could declare that they would accept only principles that gave special advantages to them, and such warnings or threats might be credible since these people would be more prepared to run the risk of no agreement. In some cases, moreover, it would be better for some people if there was no agreement. One example is the question of how much of their resources the rich ought to transfer to the poor.

If there was no agreement on that question, so that no one accepted any principle about what the rich ought to give, that would be much the same as everyone’s believing that

(cough)

the rich were permitted to give nothing, and that would be fine with the rich. For these and similar reasons, those who had greater bargaining power could win agreement on principles that gave special advantages to them, since it would be rational for others to give in to their threats. Well, some writers accept these implications of the rational agreement formula.

These Hobbesian contractualists defend a minimal version of morality. On Gauthier’s view, for example, since morality presupposes mutual benefit, it would not be wrong for us to act in ways that injure or kill other people if it would have been no worse for us if these people had never existed. Gauthier doesn’t quite claim that, but that’s the implication.

Um, Kantian contractualists like Rawls reject these implications. As Rawls writes, “To each according to his threat advantage is not a principle of justice.” But Rawls’ version of contractualism is not, I believe, Kantian enough.

In considering Rawls’s view, we can start with his assumptions about rationality. Rawls accepts the deliberative theory according to which we ought rationally to do whatever would best achieve what we most want after informed deliberation. As I’ve said, of those who accept this theory, many believe that it coincides with the self-interest theory, according to which we ought rationally to do whatever would be best for ourselves.

These people mistakenly assumed that after informed deliberation, each of us would always care most about our own well-being. Rawls does not make that assumption. He considers cases in which justice requires us to act in ways that would be very bad for ourselves.

Even in such cases, Rawls claims it might be rational for us to do what justice requires. We would be acting rationally if we would be doing what, all things considered, we most wanted to do. In his words, if a person wants with deliberative rationality to act from the standpoint of justice above all else, it is rational for him so to act.

Since the deliberative theory is desire-based, however, Rawls cannot claim that it would be rational for everyone to act justly. And when he discusses people who would benefit from injustice, Rawls claims that if these people don’t care about morality, we couldn’t honestly recommend justice as a virtue to them since they would not have sufficient reasons to do what justice requires. On the desire-based theories, we can’t have reasons to want anything for its own sake.

If people don’t care about something and wouldn’t care even after informed deliberation, we cannot claim that they have reasons to care. As Rawls writes, “Knowing that people are rational,” we do not know the ends they will pursue, only that they will pursue them intelligently.” Uh, similarly, when he rejects the view that something is right when an ideally rational and impartial spectator would approve it, Rawls comments, “Since this definition makes no specific psychological assumptions about the impartial spectator, it yields no principles to account for his approvals.”

” Now, this comment assumes that we have no reasons to care about the well-being of others. If Rawls believed that we have such reasons, he wouldn’t claim that if we knew only that someone was ideally rational, we could draw no conclusions about what this person would approve. Rawls’s claim would instead be that since this person was ideally rational, he would approve what he had reasons to approve.

For example, he would approve of acts that relieved suffering or saved people’s lives. We can now turn to Rawls’ suggested account of morality, which he calls rightness as fairness. As a contractualist, Rawls appeals to the principles that it would be rational for everyone to choose.

On Rawls’ desire-based theory, what it would be rational for people to choose depends on what they would want. And since Rawls cannot predict what everyone would want, he adds a motivational assumption. He tells us to suppose that when we were choosing moral principles, everyone’s main aim would be to promote their own well-being.

On this simplifying assumption, the deliberative and self-interest theories coincide. If we cared most about our own well-being, it would be deliberatively rational for us to make the choices that we could expect to be best for ourselves. So, though Rawls rejects the self-interest theory, his motivational assumption allows him to appeal to claims about self-interested rationality.

Rawls revises the rational agreement formula by adding a veil of ignorance. According to Rawls’s formula, we ought to act on the principles that it would be rational for everyone to choose if we had to choose these principles without knowing anything about ourselves or our circumstances. And Rawls gives several reasons for this veil of ignorance.

If we had full information, Rawls claims, we would often be unable to reach agreement. And if we knew nothing about ourselves, that would make us impartial. No one would know the facts that would give some people greater bargaining power.

Nor could anyone choose principles that were biased in their own favor. Though we would be choosing principles for self-interested reasons, our ignorance of who we are would force us to take everyone’s well-being into account. Rawls believes that his contractualism provides a systematic alternative to all forms of utilitarianism.

And this belief is surprising. If we appeal to a combination of self-interested rationality and impartiality, it’s hard to avoid utilitarian conclusions. Utilitarianism is, roughly, self-interested rationality plus impartiality.

Rawls is aware of this problem. He compares two versions of his formula. On what we can call the equal chance formula, if we were behind Rawls’s veil of ignorance, we should assume that we had an equal chance of being in anyone’s position.

Rawls says that on this assumption, it would be rational for us to choose a principle whose acceptance would make the average level of well-being as high as possible. By choosing this utilitarian average principle, we would maximize our own expected level of well-being. But Rawls rejects this equal chance formula.

If we were behind the veil of ignorance, Rawls claims, we should not assume that we had an equal chance of being in anyone’s position. On his preferred no-knowledge formula, we would have no knowledge of the probabilities. That would make it rational for us, Rawls argues, to choose certain non-utilitarian principles.

For this argument to succeed, Rawls must defend his rejection of the equal chance formula. When describing his veil of ignorance, Rawls writes, “There seem to be no objective grounds for assuming that one has an equal chance of turning out to be anybody.” This remark treats the veil of ignorance as if it would be some actual state of affairs whose nature we would have to accept.

But Rawls is proposing a thought experiment whose details are up to him. He could tell us to suppose that we have an equal chance of being anyone. What is wrong with that assumption?

Rawls himself points out that since there are different contractualist formulas, he must explain why we should appeal to his formula. “This formula,” he writes, “must be the one that is philosophically most favored because it best expresses the conditions that are widely thought reasonable to impose on the choice of principles.” Could Rawls claim that compared with the equal chance formula, his no-knowledge formula better expresses these conditions?

The answer, I believe, is no. Rawls’s veil of ignorance is intended to ensure that in choosing principles, we would be impartial. To achieve this aim, Rawls need not tell us to suppose that we have no knowledge of the probabilities.

If we supposed we had an equal chance of being in anyone’s position, that would make us just as impartial. And since there’s no other difference between the equal chance and no knowledge formulas, these formulas are equally plausible. Remember next that, as Rawls claims, the equal chance formula leads naturally to the utilitarian average principle.

Since Rawls cannot justify his rejection of this formula, Rawls’ contractualism does not, as he believes, provide an argument against utilitarianism. As Rawls points out, there’s another ground on which we might justifiably reject some contractualist formula. We may be right to reject some formula, however plausible it seems, if this formula’s implications conflict with some of our strongest moral beliefs.

Since Rawls assumes that utilitarianism conflicts with such beliefs, he might claim that we can justifiably reject the equal chance formula on the ground that it leads to an unacceptable conclusion. If Rawls made this claim, however, his contractualism would still provide no argument against utilitarianism. Rawls would be appealing to our non-utilitarian beliefs to justify our rejecting the equal chance formula and appealing to his no-knowledge formula.

So he couldn’t also claim that by rejecting the equal chance formula and appealing to his formula, we could justify our non-utilitarian beliefs. If we defend some argument by appealing to certain beliefs, we can’t then defend these beliefs by appealing to this argument. Rawls might retreat to the claim that though the equal chance formula supports utilitarianism, his no knowledge formula supports acceptable non-utilitarian principles.

Now, if that were true, Rawls’ appeal to his formula would at least show that veil of ignorance contractualists don’t have to accept utilitarian conclusions. When he appeals to his formula, Rawls argues that if we had no knowledge of the probabilities, we ought rationally to assume the worst and try to make our worst possible outcome as good as possible. We ought therefore to choose the principles whose acceptance would make the worst-off people as well-off as possible.

Since this argument tells us to maximize the minimum level, we can call it the maximin argument. Now, this argument has been widely criticized. Even if it was sound, however, it wouldn’t support an acceptable moral view.

Suppose that we must choose how to use some scarce medical resources. In one of the two possible outcomes, Green would live to twenty-five and a thousand other people would live to eighty. And the other outcome, Green would live to twenty-six, and these other people would live to 30.

On the maximin argument, we ought to choose the second outcome, giving Green her extra year of life. That’s the wrong conclusion. We can plausibly give some priority to benefiting those who are worse off, and Green is worse off by dying younger.

But this priority should not be absolute. It would be wrong to give Green one extra year of life rather than giving fifty extra years to each of a thousand other people, people who without these years would die almost as young as Green. Now, Rawls accepts what I’ve just claimed.

Though he applies his maximin argument to the basic structure of society, Rawls agrees that when we consider other questions about distributive justice, this argument has implications that are unacceptably extreme. Rawls rejects utilitarian theories because they fail to provide an acceptable principle of distributive justice. But as Rawls admits, his version of contractualism fails to provide such a principle.

Well, we can now look briefly at some other moral questions, such as whether and when it would be right to break promises or tell lies or impose certain risks on others. On Rawls’s maximin argument, when we choose between different principles about such acts, we ought rationally to choose the principles whose acceptance would make the worst-off people as well off as possible. There are here three objections to this argument.

First, it’s hard to apply. In the case of promises, for example, whom should we regard as the worst-off people? Are they the people who would lose most from particular acts of breaking or keeping promises, or those who would benefit least from the practice of promising, or those who would be on the whole worst off all things considered?

Well, perhaps it’s the last, but even if we could answer that question, that this can’t be the right way to choose between different principles. If one of two forms of the practice of promising would give much greater benefits to most people, that is not, as the maximin argument implies, irrelevant. Third, this argument forces us to ignore most non-utilitarian considerations.

According to utilitarians, when we are choosing between acts or principles, it’s enough to know the size and number of the resulting benefits and burdens. Most of us believe that there are several other morally relevant considerations. These include questions about responsibility, desert, deception, coercion, fairness, gratitude, and autonomy.

On Rawls’s version of contractualism, such questions are irrelevant. Though Rawlsian moral reasoning differs from utilitarian reasoning, it differs only by subtraction. When Rawls describes how people would choose moral principles behind his veil of ignorance, he writes that they decide solely on the basis of what best seems calculated to further their interests so far as they can ascertain them.

Rawls merely denies these people most of the knowledge that self-interested calculations need. There’s no way in which non-utilitarian considerations could possibly enter in. When he first presents his theory, Rawls writes: “It’s perfectly possible that some form of the principle of utility would be adopted, and therefore that contract theory leads eventually to a deeper and more roundabout justification of utilitarianism.”

He also writes, ‘For the contract view, which is the traditional alternative to utilitarianism, such a conclusion would be a disaster.’ Rawls, I believe, could deny that his theory justifies utilitarianism, but his claim would have to be that though his theory leads to utilitarian conclusions, it’s not plausible enough to support these conclusions. To reach a more successful version of contractualism, we should turn to a different formula and a different view about rationality and reasons.

According to the rational agreement formula, we ought to act on the principles to whose acceptance it would be rational for everyone to agree. According to what I’ve called Kant’s contractualist formula, we ought to act on the principles whose universal acceptance everyone could rationally will. Now, these formulas both require unanimity, since they both appeal to the principles that everyone could rationally choose.

But unlike the rational agreement formula, Kant’s formula doesn’t use the idea of an agreement. According to the agreement formula, more fully stated, we ought to act on the principles that it would be rational for everyone to choose if each person supposed that everyone would accept all and only the principles that everyone chose. According to Kant’s formula, we ought to act on the principles that everyone could rationally choose if each person supposed that everyone would accept all and only the principles that she herself chose.

So, in applying the agreement formula, we conduct a single thought experiment in which we imagine that we are all trying to reach agreement on which principles everyone would accept. But in applying Kant’s formula, we conduct many thought experiments, one for each person. We imagine that each of us applies Kant’s moral belief formula, one version of his formula of universal law, by asking which principles she could rationally choose if she had the power to choose or to will which principles everyone would accept.

And Kant’s contractualist formula just appeals to the principles, if there are any, which in these separate thought experiments everyone could rationally choose. So in that ways, it’s quite unlike ordinary versions of contractualism. Kant’s formula, I believe, better achieves Rawls’s aims.

One of these aims is to eliminate inequalities in bargaining power. On Kant’s formula, what it would be rational for each of us to choose doesn’t depend on what other people would be likely to choose. Since there’s no need to reach agreement, there’s no scope for bargaining, so no one would have greater bargaining power.

Consider next one of Rawls’ reasons for rejecting utilitarianism. Justice, Rawls writes, does not allow that the sacrifices imposed on the few are outweighed by the larger sum of advantages enjoyed by the many. According to several Kantian contractualists, utilitarianism goes astray because of the way in which it adds together different people’s benefits and burdens.

Utilitarians believe that it would be right to impose great burdens on a few people whenever that would give them a greater sum of benefits to others. According to these contractualists, to protect people from having such great burdens imposed on them, we should deny that the numbers count and should appeal instead to the idea of a unanimous agreement. By requiring unanimity, we can give everyone a veto against being made to bear such burdens.

I call this contractualism’s protective aim. Now, the rational agreement formula, as we’ve seen, fails to achieve that aim. Precisely by requiring a unanimous agreement, this formula gives greater power, not to those who most need morality’s protection, but to those who least need such protection, because their greater control of resources or other advantages give them greater bargaining power.

Rawls’ formula also fails to achieve this protective aim. Though Rawls’ veil of ignorance eliminates bargaining power, it prevents anyone from knowing whether they are one of the people on whom utilitarian principles would impose such great burdens. And since Rawls appeals to the principles whose choice would be rational in self-interested terms, he cannot plausibly deny that we could rationally choose rule-utilitarian principles, running the risks of bearing some burdens for the sake of possible benefits.

Since Kant’s formula neither requires an agreement nor has a veil of ignorance, it better achieves contractualism’s protective aim. On this formula, we ought not to impose burdens on anyone unless our act is permitted by some principle that this person could rationally choose even when she knows that she’s one of the people who would bear such burdens. Kant’s formula has other advantages.

Though Rawls’s veil of ignorance ensures impartiality, it does that crudely like frontal lobotomy. The disagreements between different people are not resolved but suppressed. Since no one knows anything about themselves, Unanimity is guaranteed.

In the thought experiments to which Kant’s formula appeals, everyone knows how their interests conflict with the interests of others. Since unanimity isn’t guaranteed, it would be significant if unanimity could be achieved. Well, whether unanimity could be achieved depends on what we ought to believe about rationality and reasons.

If we ought to accept a desire-based theory or the self-interest theory, Kant’s formula couldn’t succeed. If each person supposed that she had the power to choose which principles we would all accept, there’d be no set of principles whose choice would be rational for everyone in self-interested terms, nor would there be some set of principles whose acceptance would best fulfill everyone’s informed desires. We ought, I believe, to reject all desire-based theories, and though the self-interest theory is of the right kind in being value-based, this theory is too narrow.

On the wide val-value-based view that I believe we should accept, we have strong reasons to care about our own wellbeing and in a temporally neutral way, but our own wellbeing is not the supremely rational ultimate aim. We can rationally care as much or more about some other things, such as justice and the wellbeing of others. If we combine Kant’s formula with this view about reasons, Does it succeed?

Ought we to act on the principles whose universal acceptance everyone could rationally will? Well, I cannot here answer that question, but there are two other prior questions. Are there such principles, ones that everyone could rationally choose?

And if they are, what do they imply? For Kant’s formula to succeed, what we can call its uniqueness condition must be sufficiently often met. It must be true that at least in most cases, there’s some relevant principle, and only one such principle, that everyone could rationally choose.

If there was either no such principle or everyone could rationally choose two or more seriously conflicting principles, Kant’s formula would have unacceptable implications, since it would either permit or condemn too many acts. It might not matter, though, if everyone could rationally choose any of several similar principles. Such principles might be different versions of some more general higher level principle, and the choice between the lower level ones could be made in some other way.

Well, this formula’s uniqueness condition is, I believe, often met. Consider first cases in which some quantity of unowned goods can be shared between different people. No one has any special claim to these goods, such as a claim based on their greater needs or their being worse off than others.

And however these goods are distributed, the total sum of benefits would be the same. Now, most of us believe that in such cases, everyone should get equal shares. Kant’s formula appeals to the principles that everyone could rationally choose if each person supposed that everyone would accept whatever principles she chose.

We can argue, hey, A, if everyone could rationally choose the principles that, in such cases, gives everyone equal shares. B, no one could rationally choose any principle that gave them less than equal shares. C, only the principle of equal shares gives no one less than equal shares.

So D, that’s the only principle that everyone could rationally choose. Now, if we accept the self-interest theory, we must deny that everyone could rationally choose the principle of equal shares. On this theory, everyone ought rationally to choose some principle that gave themselves more than equal shares.

And we must also reject A if we accept a desire-based theory. There are many people whose informed desires would not be best fulfilled by their choosing the principle of equal shares. But I believe that as A claims, everyone could rationally choose this principle.

We wouldn’t be rationally required to choose some principle that gave us more than equal shares. Um. According to B, no one could rationally choose any principle that gave them less than equal shares.

Well, we can notice a difference here between consenting to acts and choosing principles. We could, I think, sometimes rationally consent to being given a less than equal share so that some stranger would get a greater share. Such choices would be generous and fine.

But that doesn’t imply that we could rationally choose some principle that gave us less than equal shares. Such a principle would apply not only to us, but all the members of some group. One example would be a principle that gave smaller shares to women.

If we are women, we couldn’t rationally choose that principle. We’d have no reason to want all other women to get smaller shares, nor would that choice be generous and fine. Consider next act utilitarianism or AU, according to which acts are right just in case they maximize the sum of benefits.

In the cases we’re now considering, the sum of benefits would be the same, whatever their distribution, so AU implies that it doesn’t matter how we distribute these benefits. We could permissibly give some benefits, some people, no benefits at all. Well, even the greatest act utilitarian rejected that conclusion.

“In such cases,” Sidgwick writes, “we ought to supplement AU with an egalitarian principle.” If Sidgwick had applied Kant’s formula, he would have admitted that no one could rationally choose AU since no one could rationally choose any principles that permits them to be given less than equal shares. Well, note next that for this argument to succeed, that premise could be much weaker.

It would be enough to claim that some people couldn’t rationally choose to be given a principle that gave them less than equal shares. According to the argument’s remaining premise, only the principle of equal shares gives no one less than equal shares. That’s clearly true, so as the Prince’s argument implies, Kant’s formula implies that everyone should get such shares.

Now, in most cases, the sum of benefits depends in part on how goods are distributed between different people. I believe that in some of these cases, at least, Kant’s formula can be used to defend other non-utilitarian distributive principles. But I have no time to discuss those cases here.

Someone may again object when we ask these questions about what people could rationally choose, our answers depend on our moral beliefs. You claim that in your example, some people couldn’t rationally choose the act utilitarian principle. Act utilitarians would reject that claim, since they believe that everyone could rationally choose to bear any burden, however great, for the sake of a greater sum of benefits to others.

Since our answer to these questions depends on our moral beliefs, Kant’s formula achieves nothing. Well, this objection, as I’ve said, seems to me mistaken. When we combine Kant’s principles with our beliefs about rationality, we can reach conclusions which conflict with our moral beliefs, and we can’t just readjust our beliefs about rationality.

However, these two kinds of belief are, as this objection claims, closely related. And by considering these relations, we can reach some wider conclusions. Our beliefs about rationality and reasons may partly depend on our moral beliefs, since we have moral reasons for acting, and some of these reasons are provided by facts about what’s right or wrong.

This dependence can also go the other way. Our moral beliefs may partly depend on our beliefs about rationality and reasons. That can be true in two main ways.

First, we may be contractualists who believe that the rightness of acts depends on which of the principles that everyone could rationally choose. Second, our moral beliefs may partly depend on our beliefs about what is non-morally good or bad. And these beliefs may in turn depend on our beliefs about reasons.

Well, that dependence is less obvious. Pain is bad, most of us believe, in the sense of being something that we have reason to avoid. But some great philosophers didn’t have that belief.

Hume, for example, doesn’t use good or bad in reason-involving senses. That’s why he claims that it can’t be contrary to reason to prefer your own acknowledged lesser good. Hume uses good and evil to mean pleasure and pain.

When Hume calls pain bad, he means that pain is pain. While Hume would have thought it trivial to claim that pain is bad, Kant believed this claim to be false. Kant writes, “Good or evil is strictly speaking applied to actions, not to the person’s state of feeling.”” Thus, one may laugh at the Stoic who in the most intense pains of gout cried out, “Pain, however much you torment me, I will still never admit that you are something evil, kakon, malum.”

Nonetheless, he was right. Kant misunderstands this Stoic, taking the Stoic to mean that his pain isn’t morally bad. Feelings categories can’t be morally bad.

But the Stoic was denying that his pain is bad in the sense of being something that he has reasons to avoid. And Kant, I think, ignores that kind of badness. I believe that as well as being bad for the person who is in pain, pain is impersonally bad.

On this view, if more people suffer, that would be worse, even though it may not be worse for any of these people. Pain is impersonally bad in the sense of being something that everyone has some reason to prevent or relieve if they can, not just the person in pain. Some philosophers assume that if some outcome would be bad, we have a prima facie duty to prevent this outcome if we can.

Because Ross believed that we have no such duty to prevent our own pain, he concluded that our own pain isn’t bad. Each person’s pain is bad, Ross claims, only for other people, because only they have a duty to prevent it. Well, Ross reached that strange conclusion because he ignored or rejected these reason-involving kinds of badness.

Even if we have no duty to prevent our own pain, this pain can be bad in the sense of being something that everyone, including us, has some reason to prevent. Such reasons we can call impartial in the sense that everyone has them, whatever their point of view. And we can call some outcome best in the impartial reason-involving sense if this outcome has features that give everyone the strongest impartial outcome-given reasons to prefer this outcome and to bring it about if they can.

Now, that definition doesn’t assume that all impartial reasons are outcome-given. We may have other impartial reasons, such as reasons provided by people’s rights. And such reasons might outweigh our outcome-given reasons.

Suppose that some poor person gives us some wallet bulging with banknotes that she’s just found. We may believe both that we would make the outcome best if we told this person to keep the wallet and that our reasons to produce this outcome are outweighed by our reasons to try to return the wallet to its rich owner. On this view, we may have the strongest impartial reasons both to try to find this owner and to hope that this attempt will fail so that we could then permissibly produce the best outcome by returning the, the wallet to its poor discoverer.

Well, the view just sketched is semi-consequentialist in the sense that it gives some m-moral weight to the goodness of outcomes. On views that are consequentialist, the rightness of acts depends entirely on certain facts about what is non-morally good or bad. I mean, that’s how I’m going to use the term.

According to act consequentialists, acts are right just in case they make things go best. According to rule consequentialists, acts are right just in case they are permitted by one of the principles whose acceptance would make things go best. Now, consequentialists do best, I believe, to use best in the impartial reason-involving sense.

We can call such people reason consequentialists. These consequentialists believe that, one, to answer questions, answer moral questions, we must answer questions about rationality and reasons. Contractualists also believe that.

But while con-consequentialists appeal to claims about which outcomes would be best, contractualists appeal to claims about what everyone could rationally choose. But when outcomes are best in the impartial reason-involving sense,

(coughs)

they may be the outcomes that from an impartial point of view everyone could rationally choose. So if consequentialists and contractualists make the same assumptions about rationality and reasons, their theories may coincide. Though consequentialists appeal to what is best and contractualists appeal to what everyone could rationally choose, these people may reach the same conclusions.

They may find that in Mill’s phrase, they’ve been climbing the same mountain on opposite sides. These points have been widely overlooked. But that’s not surprising.

Many people don’t use best in the impartial reason-involving sense. Some people make no use of the concept of a normative reason. Others accept desire-based theories about reasons or the self-interest theory.

On these theories, there are no outcomes that everyone has some reason to want, since there are no outcomes that would be good for everyone or that would do something to fulfill everyone’s informed desires. We should next distinguish between consequentialism and utilitarianism. Consequentialists believe that the rightness of acts depends only on facts about what would make things go best.

Utilitarians believe that the rightness of acts depends only on facts about what would benefit people most. Now, some utilitarians are consequentialists since they believe that things go best when they go in the way that would benefit people most. But there are also some utilitarians who make no appeal to the goodness of outcomes.

On one such view, we ought to do what would benefit people most because that is how we can best treat everyone with equal respect and concern. Just as some utilitarians aren’t consequentialists, some consequentialists aren’t utilitarians. These people accept that the rightness of acts depends only on facts about what would make things go best, but they reject the other two claims I read out.

They believe that you can get them from what they believe. They believe that it would often be best if things went in some way that did not benefit people most. And on the view that’s most relevant here, how well things go depends in part on how benefits and burdens are distributed between different people.

On this view, one of two outcomes could be better, though it would involve a smaller sum of benefits, because these benefits would be more equally distributed, or because more of these benefits would go to people who are worse off. In their beliefs about the goodness of outcomes, these consequentialists accept distributive principles. Rawls claims that if a moral theory includes such principles, it’s not consequentialist.

In his words, “The problem of distribution falls under the concept of right as one intuitively understands it, and so the theory lacks an independent definition of the good.” Well, as we’ve seen, that may not be true. Reason consequentialists use good in the impartial reason-involving sense.

When these people claim that the rightness of acts depends only on facts about what would make things go best, they mean that the rightness of acts depends only on facts about what everyone has most reason to want from an impartial point of view. These people may claim that the goodness of outcomes depends in part on the distribution of benefits and burdens, because they believe that everyone has such impartial reasons to want benefits to be more equally distributed or to want benefits to go to people who are worse off. And there’s no useful sense in which that view isn’t consequentialist.

Last lap coming up. Kant wasn’t a consequentialist. But we are considering not Kant’s moral beliefs, but the implications of his principles.

According to Kant’s consent principle, we ought to treat people only in ways to which they could rationally consent. Act consequentialists can argue everyone could rationally consent to being treated in any way that would make things go best. Therefore, the consent principle never implies that such acts are wrong.

Now, that argument’s premise, I believe, is true. I believe that in Lifeboat, for example, White could rationally consent to our leaving her to die so that in the time available, we could save the five. Well, if that’s true, that view strongly supports the claim that everyone could rationally consent to being treated in any way that would make things go best.

If we could rationally consent to being left to die when and because that is what makes things go best, that strongly suggests that we could rationally give consent to having lesser burdens imposed on us for that reason. Well, if that argument is sound, as I believe, Kant’s consent principle could be accepted by act consequentialists. Kant’s contractualist formula, however, provides one premise of a counterargument.

We can claim we ought to act on the principles whose universal acceptance everyone could rationally will. We couldn’t all rationally will that everyone accepts the act-consequentialist principle. Therefore, this isn’t the principle on which we ought to act.

And this argument, I believe, is also sound. Sidgwick reluctantly concluded that we couldn’t rationally want it to be true that everyone accepts act utilitarianism. Of Sidgwick’s reasons for reaching that conclusion, most would apply to act consequentialism.

If everyone believed that it was right to do whatever would make things go best, that would be unlikely to make things go best. Well-known reasons make excuses in your favor. Stealing is all right if I need it more than him, et cetera.

Things would be likely to go better if everyone had certain other moral beliefs. It’s wrong to steal. If we knew that to be true, we couldn’t rationally will that everyone accepts act consequentialism, so Kant’s formula would imply that we should reject this view.

With some help from Sidgwick. Return now to rule consequentialism. On this view, we ought to act on the principles whose acceptance would make things go best.

Now, that view is very different from act consequentialism, partly for the reasons Sidgwick gave, rule consequentialism supports principles which are much closer to most people’s moral beliefs. Rule consequentialists can appeal to Kant’s contractualist formula. They can argue— it’s my last argument— A, we ought to act on the principles whose universal acceptance everyone could rationally will.

B, everyone could rationally will whatever they would have most reason to will from an impartial point of view. C, what everyone would have most reason to will from this point of view is that everyone accepts the principles whose universal acceptance would make things go best. Therefore, these are the principles whose universal acceptance everyone could rationally will.

Therefore, these are the principles on which we ought to act. Now, if we accept Kant’s formula, which is premise A, accepting B and C would commit us to rule consequentialism. Some people would reject C because they believe that there’s no intelligible sense in which things might go better or worse.

Many philosophers keep saying that. But if these people have the concept of a normative reason, as some of them seem to do, they must understand the claim that some outcomes would be best in the impartial reason-involving sense. Well, these people might claim that there are no such outcomes since there are no impartial reasons.

And this must be claimed by those who accept either desire-based theories about reasons or the self-interest theory. I’m assuming, though, that we should reject these theories. We all have reasons to want some outcomes, such as those in which fewer people suffer or die young, and there are some principles whose acceptance would, in this impartial sense, make things go best or equal best.

When we apply Kant’s formula, we ask which principles everyone could rationally will or choose if each person supposed that everyone would accept whatever principles she herself chose. Now, if each person chose the principles whose acceptance would make things go best, that choice would itself make things go best. And according to premise C, that is the choice that from an impartial point of view, everyone would have most reason to make.

Now, we might challenge C by appealing to our non-consequentialist moral beliefs. We might claim that if people accepted the principles whose acceptance would make things go best, that would sometimes lead these people to act wrongly. But such claims are irrelevant here.

When we apply some contractualist formula, we shouldn’t appeal to our beliefs about the wrongness of acts. As Rawls says, we should bracket these beliefs. Otherwise, the formula would be trivial.

So as I say, according to contractualists, whether acts are wrong depends on which principles everyone could rationally choose, and that would be trivial if which principles everyone could rationally choose depended on which acts are wrong. So we should appeal to our moral beliefs only at a different stage when we are deciding whether to accept some version of contractualism. If it’s too deeply counterintuitive, we’ll reject it.

There are, I believe, no other grounds on which we could reject premise C. If each person chose that everyone accepts the principles whose acceptance would make things go best, this choice would itself make things go best in the impartial reason-involving sense. In other words, this is the choice that each person would have the strongest impartial outcome-given reasons to make.

Now, such reasons may be outweighed by other conflicting impartial reasons, since there may be cases in which we ought not to choose what would make things go best. But as we’ve just seen, no such conflicting reasons are relevant here. We can’t appeal to our beliefs about right and wrong.

Uh, nor would it be plausible to claim that the people who we are discussing in this imagined thought experiment, that it would be morally wrong for them to choose the principles whose general acceptance would make things go best. I mean, it’s not as if in doing that, they’d be breaking a promise or something of the kind. Uh, so C is restricted to the particular case of the choices that would be made in this thought experiment.

If no such conflicting reasons are relevant here, then I think all that are left are the impartial outcome-given reasons, and that’s enough to show that everyone could rationally… Sorry, that that’s what everyone would have most reason to choose from an impartial point of view. Now, according to the argument’s remaining premise, everyone could rationally will or choose whatever they have most reason to choose from this point of view.

Now, that claim isn’t consequentialist, since it doesn’t assume that all reasons are outcome-given, and I did have several pages defending that claim, but I’ve now realized that it’s enough for me to make a much weaker claim. I believe that in many cases where no other special considerations apply, uh, we could rationally choose what would make things go best, whatever the cost to ourselves. Uh, we could rationally give up our life to save two strangers or even one stranger.

But this argument doesn’t require an assumption at all as strong as that. Um, and one way of bringing that out, which is topical on this gloomy election post-mortem day, uh, is to make a point about referenda, uh, not as they actually, uh, occur, particularly in America, but how thinking more rationally they should occur. Let’s just compare four choices.

Um, if some rich person asks herself Whether she thinks she should or would give ten thousand dollars a year to the poorest people on– some of the poorest people, the people who earn less than two dollars a day. I think she could rationally give that. Some people may think, \”No, she doesn’t have sufficient reason.

Ten thousand is quite a lot of money.\” Um, but that isn’t what this argument needs. Suppose that that person could choose to give up ten thousand dollars a year, in a way that gave to the poorest people in the world ten thousand million dollars a year, that’s to say ten billion dollars a year. Well, then she would have sufficient reason to give up the ten thousand, even on a pretty weak version of this kind of view.

We merely need to suppose that her reasons to care about the well-being of the poorest people in the world aren’t more than a million times weaker than her reasons to care about herself. Similarly, to take a case where people are more likely to resist the initial claim about rationality, uh, we can have a case in which we’re reallocating resources in a way that affects people’s health and life expectancy. And, uh, suppose the question is: Could one person rationally choose a distribution that would reduce her life from 80 years to 70 so that one other person could live to 60 rather than 20.

Well, it’s a lot to give up ten years of life even to give 40 years to someone who’d die at 20. Some of us may think we can’t claim that everyone has sufficient reason to make that great a sacrifice. But suppose that by giving up ten years of life, we could enable ten million people to live to sixty instead of twenty.

Well, then I think we do have sufficient reason to give up ten years. Well, that may seem irrelevant. I mean, how could we have this ability?

Um, by giving up ten thousand dollars to give to the poorest people in the world, ten thousand million dollars. Easy. Have a referendum.

Just need to suppose that there are a million rich people, and the referendum is that everyone will be taxed, ten thousand dollars will go to the poorest people in the world. Well, if I vote, if I have the casting vote in this referendum, and I vote yes, it’s only gonna cost me ten thousand dollars, but it will make ten billion dollars go to the poorest people in the world. Now, that is in effect how this premise of this argument could go, um, Because we’re asking what everyone could rationally will if she were choosing the principles that everyone would accept, and we can assume that many people would act on them.

Well, if you choose the principles that everyone accepts, the principles that would on the whole make things go best, compared with some set that would make things go a lot worse, that’s very like the cases I’ve just done. Um, even if you know that these principles that would make things go best would impose some burdens on you, the benefits to others would be so enormous that I think you have sufficient reason to make that choice. Um, well, that’s all I need to say, I think, to defend the argument, so I just end with very brief conclusions.

According to Kant’s formula, we ought to act on the principles whose universal acceptance everyone could rationally will. As I’ve now argued, what everyone could rationally will is that everyone accepts the principles whose acceptance would make things go best. So on Kant’s formula, these are the principles on which we ought to act, and that’s what rule consequentialists believe.

Now, as well as providing an argument for rule consequentialism, Kant’s formula may give this view a stronger foundation. According to many rule consequentialists, what ultimately matters is how well things go. If that’s true, however, why is it wrong to act in ways that would make things go best when and because such acts aren’t permitted by certain principles?

That’s the standard objection. Well, rule consequentialism may instead be founded on Kantian contractualism. What’s fundamental here isn’t the belief that it ultimately matters most how well things go.

It’s the belief that we ought to act on the principles whose universal acceptance everyone could rationally will. Now, on this view, what everyone could rationally will is that everyone accepts the principles whose acceptance would make things go best. But because this view appeals to Kant’s formula, it may better resist the thought that it can’t be wrong to do what would make things go best.

Such acts may be wrong, not because they contravene the principles whose acceptance would make things go best, but because they contravene the principles whose acceptance everyone could rationally will. Well, I haven’t been defending rule consequentialism. My aim has been in part cartographical.

On one standard moral map, most of us are shown as holding pluralist views of the kind that Sidgwick calls common sense morality. This map also shows two main kinds of systematic theory. One kind of consequentialist, with utilitarian theories as the main examples.

The other kind of Kantian theories and various forms of neo-Kantian contractualism. Now, this map I’ve argued should be redrawn. Of the different ways of thinking about morality, it’s Kantian and contractualist theories that do most to support consequentialism.

And that isn’t surprising. Rawls writes that in adopting his contractualist formula, we have substituted for an ethical judgment, a judgment of rational prudence. Rawls’s version of contractualism makes some moral assumptions, as in the way in which by imposing a veil of ignorance, Rawls makes our reasoning impartial.

But in applying Rawls’s formula, we can’t appeal to our beliefs about which acts are wrong. So we should expect to reach utilitarian conclusions. As I’ve said, utilitarianism is roughly rational prudence plus impartiality.

Rawls’s aim was in part to provide an alternative to utilitarianism. By turning to Kant’s contractualist formula, I’ve argued we can achieve that aim. On Kant’s formula, we appeal to the principles whose choice would be rational, not in purely self-interested terms, but in response to every kind of reason.

These principles, we can argue, aren’t utilitarian. By appealing to Kant’s formula, I believe we can support various distributive principles, and this formula allows us to defend other non-utilitarian principles, such as principles about retributive justice or some kinds of perfectionism. Now, I would reject that, but the, the retributive justice, that it’s allowed for as a possible move.

Though Kant’s formula counts against utilitarianism, it supports consequentialism. And as before, that isn’t surprising. On this formula, we achieve impartiality by appealing to the principles whose universal acceptance everyone could rationally choose.

And in applying the formula, we can’t appeal to our beliefs about which acts are wrong. So we should expect to reach consequentialist conclusions. What everyone could rationally choose are the principles whose acceptance would make things go best.

Consequentialism is roughly wide value-based rationality plus impartiality. Well, if we believe that we ought to avoid consequentialist conclusions, the obvious course is to think about morality not in a Kantian or contractualist way, but in Sidgwick’s way. Now, that claim may seem surprising since Sidgwick is an act utilitarian.

But Sidgwick assumes that we must appeal to our beliefs about which acts are wrong. This appeal should be critical since we should try to clarify these beliefs, and we should consider all good arguments for and against them. Uh, the longest part of Sidgwick’s Methods of Ethics does just that.

But Sidgwick thought we can’t hope to answer moral questions by appealing only to claims about what everyone could rationally choose. We must appeal to our moral intuitions. There is, however, an alternative.

Scanlon claims that rather than appealing to the principles that everyone could rationally choose, we should appeal to the principles that no one could reasonably reject. And Scanlon’s contractualist formula appeals to claims about what’s reasonable in a moral sense, but that does not, as some suggest, make that formula circular or trivial. And of the different versions of contractualism, Only Scanlon’s, I believe, may be as good as the Kantian formula when combined with the view about rationality and reasons that Scanlon and I both accept.

(applause and cheering)

[01:08:02] ANNOUNCER:
Okay, we’ll now hear from the third of our three commentators, Professor Thomas Scanlon. Professor Scanlon is the Alford Professor of Natural Religion, Moral Philosophy, and Civil Polity at Harvard, where he’s taught since nineteen eighty-four after having, uh, taught at Princeton for the previous eighteen years. Professor Scanlon is one of the leading moral and political philosophers of our day.

His nineteen ninety-nine book, What We Owe to Each Other, is one of the landmarks of contemporary moral philosophy. He’s a former Tanner Lecturer himself and also a former MacArthur Fellow. And as we’ve just heard, according to Derek Parfit, he’s Kantian moral philosophy’s last best hope, so we’re particularly lucky to have him with us today.

[01:08:46] DEREK PARFIT:
Tim. That was a pretty good intro.

[01:08:50] ANNOUNCER:
Yes, I know.

[01:08:56] THOMAS SCANLON:
Thank you. There’s a great deal in this extremely rich and complex lecture that we might discuss. Uh, so in my comments, I’ll have to be selective.

I’ll resist the temptation to talk about the early uh, discussion of Rawls and some of the later cartological points that Parfit makes. Uh, and I’ll concentrate simply on the argument that’s offered at the very end of the lecture on page three of your handout to show that what Parfit calls Kant’s contractualist formula leads to, or at least does not exclude, rule consequentialism. What I’ll do really is just to walk you carefully through this part of the argument, since it seems to me that this is where we’re likely to find the to Gavagaiian moment or the occasion of lapine manifestation or whatever you want to call it.

Uh, so let’s– You, you can turn to page three of your lucky number programs there and, and look at A, B, and C. Looking first at A, you’ll recall that Parfit states Kant’s contractualist formula as follows: We ought to act on principles whose universal acceptance everyone could rationally will. This principle appears to be at least potentially permissive. That is, it leaves open the possibility that there is more than one set of principles whose universal acceptance everyone could rationally will.

In this case, the principle wouldn’t entail that we must act on one of these principles rather than the other. Acting on either would presumably be at least morally permissible. So what the argument Parfit is offering may show is the court– is that according to this formula, we are morally permitted to act on rule consequentialism, uh, but not necessarily that we are required to do so if there’s also some other, uh, principle that everyone could rationally will.

The second step in Parfit’s argument is a claim about what people could, in fact, rationally will. And here I’m going to concentrate on the argument as printed rather than, um, the more colorful, uh, and, and, uh, emotionally and morally gripping, uh, reformulation which Derek gave in his, uh, c-concluding remarks. I’ll return to that variant, uh, in, in a little ad lib, uh, a little later on, but I’ll follow my own text here and sticking with the, the, uh, original formulation.

So in the original formulation, B said, everyone could rationally will whatever they would have most reason to will from an impartial point of view.

[01:11:29] DEREK PARFIT:
I was keeping that formulation.

[01:11:31] THOMAS SCANLON:
You were, yes. Okay, yeah, right. Well, you’re just arguing for it by, by, by you.

Is that right? So you were giving a different argument for it than I… I’m sorry, it wasn’t a reformulation.

All right. Um, I’ll, I’ll return… I’ll say something about that argument a little later on.

My first observation here is that, as Allen Wood remarked in his comments on Parfit’s first lecture, uh, this is a rather, I mean to say, distinctly un-Kantian idea of what one could will. When Kant speaks of what one could or could not will, he has in mind what he takes to be the constraints on willing that are constitutive of rational agency itself. We couldn’t will something if in so willing we would violate these constraints.

Uh, he, he, he does not, that is, Kant does not, I think, take the question of whether one could rationally will a certain action to be one that we should answer, as Parfit does, by asking whether one could have sufficient reason for so acting, taking into account the balance of reasons on, on opposing, uh, on opposing sides. Weighing reasons in this way is not something that Kant engages in, to my– not very much anyway. And indeed, I think that many have been attracted to Kant’s view in part because it seemed to provide a way of answering moral questions without appeal to this kind of balancing of reasons.

It may be that, as I myself believe, this appeal is illusory because the weighing of conflicting reasons can’t be avoided in reaching moral conclusions. But in adopting a criterion of what one can will, which depends upon the balance of reasons in this way, what Parfit calls Kant’s contractualist formula departs from Kant in, I think, of an important way, and remains Kantian only in a looser, modern sense. This second step in Parfit’s argument claims that in order to show that a person could rationally will a given principle, it is sufficient to show that this principle is what that person has most reason to will from an impartial point of view.

[01:13:31] DEREK PARFIT:
So what is an impartial point of view? I take it that this point of view is defined, as Parfit says of an argument of Rawls’s in another context, by subtraction. It’s defined by beginning with all the reasons that a person has, and then dropping from consideration all the reasons that depend on how the object of willing is related to that person, him or herself.

This includes, for example, the reasons one has to favor an outcome because it would be better for oneself, the re– and the reasons one has for favoring an outcome because it would be better for one’s family or one’s friends or one’s favorite football team, uh, and so on. What someone would have most reason to will from an impartial point of view is what he or she would have most reason to will, leaving aside all such person-relative reasons. So if these person-relative reasons are left out, what is left in?

I’ll mention several categories. First, there are what might be called purely impersonal reasons. Suppose I decide to give money to the Berkeley Physics Department because I think that discovering fundamental truths about the nature of the universe is something worth doing, something we all have reason to support.

That would be an impersonal, purely impersonal reason. It has to do with simply thinking of this objective as being one that we have reason to achieve because of its inherent properties. Insofar as we don’t think that my giving my money away in this fashion would be foolish or mere caprice, we must agree that the investigation of such truths is something that one has at least some reason to support.

I also might support, on similar grounds, general education in science or advanced education in science for a variety of people, or education in music, because I think that these forms of education promote the development of talents in people that are intrinsically valuable, deep scientific understanding or musical understanding and skill.

[01:15:36] THOMAS SCANLON:
These would be, I think, what Parfit calls perfectionist reasons when he says we might have– because they have to do with producing valuable, developing valuable talents in people. Now, I might think that the reasons that I have to support research in physics would count equally in favor of supporting basic research in biochemistry, just another way of forwarding the goal of understanding the fundamental truths about, uh, the universe and the life within it. But I might decide that the fact that biochemical research not only would uncover these deep truths, but also might lead to the prevention of disease and to the alleviation of human suffering.

And I might think that this provides me with an additional reason for supporting that kind of research. So I might switch my gift from physics to biochemistry. I don’t know, the, the chancellor isn’t here today, but, uh, if he were here, I could do this, I suppose, if I had actually given any money.

(laughter)

Um. This would also be an impartial reason, right? The first impartial reasons had to do with simply the value of these accomplishments.

The second kind of impartial reasons would have to do with the value of alleviating suffering and, uh, combating disease. Would this second kind of impartial reason be a moral reason? It would certainly be a morally approvable reason.

You might think that a person who gave money for that on that ground was a good good chap for that reason. But as Parfit said in responding to John Searle’s question in the discussion after the first lecture, E-even if I gave my money to biochemistry for this reason, I might not be doing it because I took this to be morally required. I might just think that suffering is in itself something that anyone who can prevent has reason to prevent.

And I agree with Parfit that in this case, we shouldn’t say that my reason is a moral one unless all that we mean by this is that it’s one which is pro– which properly elicits moral approval. It’s not a response to morality, even though it elicits, I might say, moral approval. It need not be.

But this distinction between reasons that are, say, morally approvable, uh, but independent of morality and those that are genuinely moral in their basis may be more difficult to draw in some of the other cases to which Parfit refers. Parfit says that the reasons we have from an impartial point of view include considerations of distribution and also considerations of retributive justice. I take it that what he has in mind here is this.

If two social situations involve the same aggregate well-being, but in one case, this well-being is much more evenly distributed, much more equally distributed than in the other, this fact could provide us with reasons to prefer that situation, the more equal one. Would this be a moral reason? Well, it clearly would be a moral reason if it were just the reason that inequality is morally objectionable, so its alleviation to something morality requires us to, uh, uh, prefer.

It, it may be less clear than it was in the case of suffering and the elimination of disease what other form this reason could take than as a reason that has this kind of moral, uh, foundation. I see the force of this claim, but I think there’s also something to be said on the other side, which is, I take it, the side that Parfit has in mind in formulating his view. In a society marked by a high degree of inequality, some people may lack any degree of meaningful control over their lives.

They may also have the burden of being stigmatized in the eyes of, of their fellows, and hence in their own eyes, as inferior and unworthy. Like suffering, lack of control and stigmatization in the, the eyes of others, and hence in your own as unworthy and inferior, are, are things that people have reason to want to avoid. They’re, they’re bad for people, you might say.

And there are therefore also reasons that, there are also therefore things that other people have reasons to regret and to want to see alleviated. This would be so, I think, even if, although this is hard to imagine, uh, these, uh, forms of stigmatization or lack of control were brought about in ways that were not in themselves unjust through some kind of natural accident or something. So I think one might have an impartial, although not explicitly moral reason for preferring a more equal distribution to a less equal one.

I leave aside the question of whether the same might be said in regard to retributive justice. That is to say, in regard to the reasons somebody might have for preferring outcomes in which well-being is distributed in a way that’s more in accord with moral merit. Finally, the reasons that, in, I think in Parfit’s view, one has from an impartial point of view, include reasons whose basis does lie, unlike the ones I’ve just been outlining as not being essentially, not necessarily essentially moral.

It didn’t call but impartial reasons, in his view, also include reasons whose basis does lie in moral requirements. So, for example, the fact that an action would be wrong, would, uh, would, would be wrong or would violate someone’s rights can give us a reason not to do it. And we have reasons perhaps to prefer certain kinds of institutions to others because they’re just and don’t involve the violation of people’s rights.

These are impartial reasons, too. And we might have i-impartial reasons of this kind, perhaps to promote the welfare of our children. These now wouldn’t be partial reasons, personal reasons.

It’s not just that because they’re my children, I love them and care more about them than everybody else, but because I’m their parent, I have an obligation to take care of them. So I would have a morally based reason to take care of them, and I hope, uh, we might say, given the age I’ve now reached, that children, you know, have the same kind of, uh, impartial, uh, reason to take care of their parents as, as well. Okay, so we have many kinds of impartial reasons, uh, s– many of them non-moral, uh, but i-in the last analysis, some, uh, moral reasons as well.

The class of reasons that, according to Parfit, one has from an impartial point of view is thus very broad. And the claim that anyone could rationa-rationally will whatever he or she would have most reason to will. From an impartial point of view, it draws strength from this breadth, right?

It’s the fact that the reasons one has from an impartial point of view aren’t all of the reasons one has, but they include enough that if there’s a case on based on those reasons to do something, then y-you can conclude that the rest of the reasons aren’t going to out-outweigh it. Now, what are the reasons that might outweigh it? Well, the reasons that might outweigh it are mainly personal reasons, reasons of personal sacrifice.

Of course, though, any sacrifice on your part will already have been taken into account in, in impartially because it’s someone’s sacrifice, right? Uh, so Parfit’s idea is that the o– given that they’ve already been taken into account in the impartial reason, your personal reasons can’t, uh, make it irrational for you to do what you have most reason to do from an impartial, uh, point, point of view. Just let me say one thing, though, here about, about the revised or, or the additional argument that, that Parfit– Derek offered in the, in the talk.

Uh, what we’re, what we’re, what we’re tending toward here is a conclusion about rule consequentialism taken as a view that would differ in its consequences or might differ in its consequences from alternative forms of contractualism, let’s say. And it seems to me that the examples that, uh, Derek mentioned, where, you know, by making a certain sacrifice, we could create huge, bring huge benefits to people elsewhere, not just huge in the aggregate, but also huge to each one of them in that example. Um, it’s likely that other forms of contractualism would also require the same th-th-those, those contributions in that case.

Um, what we’re interested in here is a case for rule consequentialism, which might poss– might deviate, might require things of us that other forms, that other, other contractualist views wouldn’t require. So I think it’s crucial that we focus on the, the content of B, which says that— to go back to B here, I’m not just to ad-lib it, uh, everyone could rationally will, rationally will whatever they have most reason to will from an impartial point of view. But that–

So it seems to me that not every case of what you have reason to will from an impartial point of view will be one in which the advantages are so staggeringly great. In many cases, they will, right?

[01:24:18] DEREK PARFIT:
Yeah.

[01:24:18] THOMAS SCANLON:
But I, I don’t think you can explain in your reply, but I don’t, I don’t see that it follows That whenever one has–

[01:24:23] DEREK PARFIT:
that’s right.

[01:24:25] THOMAS SCANLON:
Um, in many cases, you know. So it seems to me the more controversial cases would be the cases where that wasn’t true. True, but no doubt we’ll get persuaded later that, uh,

[01:24:35] DEREK PARFIT:
That’s true in every case. So as I said, the case for B, uh, the idea that it’s at least sufficient, we have suff– at least sufficient reason, maybe not conclusive, but sufficient reason to do whatever we have reason to do from an impartial point of view, depends upon the breadth of that category. The effect of the next step in the argument, step C, is somewhat to narrow that breadth.

C says, what everyone would have most reason to will from an impartial point of view,

[01:25:07] THOMAS SCANLON:
is that everyone accepts the principles whose general acceptance would make everything go best. Well, when does everything go best? Everything goes best in the sense that Parfit has in mind here, if it goes in the way that we have what he calls the strongest impartial outcome-based reasons for wanting it to go.

(cough)

This is a consequentialist idea, but not a utilitarian one. Impartial outcome-based reasons include impartial reasons of all the kinds I’ve listed, except for those in my last category, except possibly for those, that is, the morally based ones. Uh, at least he certainly means to exclude some of the reasons in that last category, since that includes reasons that are not outcome-based, but, but depend on things like violations of rights and the wrongness of particular forms of behavior.

Um, I’m not certain whether he means to exclude all of them. I will, I will assume, uh, that that’s probably what’s going on, but let’s continue. It’s important to ask then why, why it is that some or all of the reasons in the third category, the morally based category, shouldn’t be taken into account in determining when things go best, right?

That’s, that’s what what C does is to narrow our focus from all the impartial, uh, reasons we’ve got to the outcome-based ones. Now, on a utilitarian conception, we’d have a ready answer. Because the considerations in that final category, the morally based ones, would be ruled out because they are not in themselves matters of individual well-being, and that’s all that counts according to utilitarianism.

Perhaps the consequentialist conception that Parfit has in mind, rule– might be said to rule them out because they’re not reasons for preferring outcomes. This wouldn’t rule out all the reasons in the last category. I think one might have explicitly moral reasons for preferring a state of affairs, let’s say, a state of affairs in which no one’s rights are violated to one in which rights violations occur.

But I don’t think that outcome-based moral reasons of this kind capture all the force of morality as we ordinarily understand it, and as I accept it. I believe we have reasons not to violate people’s rights which aren’t reducible to reasons to prevent the occurrence of rights violations. This is a familiar dividing point between consequentialist outlooks and deontological ones.

But I don’t see any grounds inherent in the argument being considered so far for adopting this consequentialist restriction. This is an argument for consequentialism. It’s not supposed to pick up any consequentialist assumptions along the way.

So there’s another reason for excluding the the reasons in my third category, I think. And this is relevant not to the idea of consequentialism, but to the supposed and assumed purposes of the argument we are considering and the purposes and aims of Kant’s contractualist formula in particular. The idea is that what we’re trying to do in deploying a principle like Kant’s contractualist formula is to establish a procedure for answering questions about the rightness and wrongness of actions.

So we should not appeal in carrying out that procedure to these very considerations. That is, we shouldn’t appeal to our ideas about which actions are right and which actions are wrong. We can understand this rationale along the same lines as Rawls’s rationale for excluding appeals to ideas of justice from the range of reasons that are to be considered by the parties to his original position in deciding which principles to accept.

That would defeat the purpose of the exercise, and the assumption here is that Kant’s contractualist formula is an exercise of the same kind. One might wonder why, then, we can allow people’s reasons for accepting or not accepting a principle to include ideas about which distributions are objectionably unequal, or about when people are or are not getting the punishment they deserve. And I take it that Parfit’s response would be that, as I argued above on his behalf, these reasons do not need to be at base moral reasons.

They may track morality, they may be morally approvable, but we might have reasons for caring about those things which wasn’t caring about them because they’re right or because they’re un– Otherwise, it would be unjust or something like that. So these are reasons that can be included within the ambit of consequentialism of the kind we’re discussing and can be appealed to without jeopardizing the aims of the construction. What this rationale seems to support, however, is not the exclusion of all reasons that are not outcome-based, but the exclusion of all morally based reasons.

Sorry, if not the exclusion of all reasons that are not outcome-based, but the exclusion of all morally based reasons, no matter whether they are outcome-based or not. The result would then be to exclude all non-consequentialist reasons, but also perhaps some reasons that a consequentialist, uh, might appeal to, thus yielding a slightly restricted form of conseq– of rule consequentialism. But that’s okay.

It’s rule consequentialism nonetheless. The reasons that remain will be allowed, not because considered as reasons they form some significant class, but because the objectives of the principle under consideration do not require their exclusion. The conclusion that Parfit arrives at in this way could, it seems to me, be generalized as follows.

As long as we understand what Parfit calls Kant’s contractualist formula in the way he proposes, a form of rule consequentialism will always be one possible outcome. The only way to avoid it will be to make the argument circular by bringing in as reasons, um, as reasons, cons– reasons people have for will– sorry. The only way we could exclude it would be by bri– by bringing in as reasons people have for willing or not willing the acceptance of a principle, the very non-consequentialist moral considerations that the argument seeks to support.

So this doesn’t mean that, that Kant’s contractualist formula has to lead to Kant– to consequentialism, because, as I said earlier, these principles are… These steps are permissive. They just say it, it’s not gonna be contrary to reason, so to speak, to choose the consequentialist path.

So where does this bias toward consequentialism come in? There seem to me to be two steps. The first is the formulation of the relevant thought experiment as one in which we ask, looking at things from each person’s point of view in turn, what principles he or she could will to be universally, could will to be universally accepted.

The second step is the appeal to impartiality, which may serve here as something of a dialectical Trojan horse. The restriction to impartial reasons seems at first of an entirely plausible and innocent move. After all, it doesn’t seem that one’s conclusions about what is morally right should depend upon how one personally would stand to benefit from reaching a particular conclusion rather than another.

But suppose for a moment that we ourselves will not be affected one way or the other by the acceptance or rejection of a certain principle. Since we’re never involved in the kind of situation to which such a principle would apply. The idea of contractual– sorry, oth– sorry, others, however, are affec– may be affected greatly by the adoption or non-adoption of this principle.

By its adop– its adoption might affect some of them in a very seriously negative way while bringing small benefits to a large number, uh, of others. The idea of contractualism, normally understood, is that we should ask in such a case whether the claims of the few who will be harmed negatively, affected negatively are adequately answered by citing, “You’ve got to accept that because of the benefits that are brought to other particular people.” We confront these claims one-on-one, and we ba– we ask, does, does the benefit that a person on the benefiting side get constitute an adequate answer, uh, to the person who is suffering the loss on the other side?

Surely, in answering this question, we should be impartial. That is to say, we should consider these claims on their merits. I’m now supposing that we ourselves are not affected, right?

We’re just choosing these principles only affect other people, but so and we should be impartial. We should consider these claims on their merits without bias, without asking whether we have any affection for one person on one side and whether we know the person on the other side, uh, and so on.

But does this mean that we must answer the question by asking what would overall result in the best outcome? Produce the greatest sum of benefits? Uh, of course, it would seem to me, you’re not gonna be surprised to hear, that it doesn’t.

Impartiality does not entail consequentialism. But Parfit’s claim is not that the question in this case must be answered in this way in order to be impartial. He claims only that it may be so answered.

Such an answer is not excluded. As long as a person’s sacrifice would result in a state of affairs that was impartially better, he or she would not be open to rational criticism for making that choice. They would have sufficient reason to make it.

But we are choosing principles here for other people, right? We’re ch– And can we impose on someone else in making that choice any sacrifice that he or she would not be crazy to make, would not be positively irrational to make?

This seems to me a chillingly permissive criterion. And I suppose that this is just Parfit’s point against Kant’s contractualist formula. Uh, I agree with that point, uh, although I do not find it very surprising.

Once we see the permissive character of the formula, once we see that it’s just asking what people could accept, it’s not so surprising that we think, well, people could accept. Yeah, they, they could accept that. But what level of sacrifice can we demand of people without their actual consent?

If, if this is too permissive, where does the– where does it become too much to ask to impose this for the greater good? This is clearly a moral question. And the second thing that’s worrying about Parfit’s argument is the suggestion that there’s no way of answering this question within a contractualist framework in a non-question-begging way.

By setting up contractualism in a way that, that contains some answer to that question, we’ll be bringing in the very kind of consideration, um, that the contractualist system was meant to answer, the c-contractualist procedure was meant to answer. This might seem self-defeating for contractualism. I think Parfit

and I both hope, or maybe even both believe, that it’s not. But whether it is or isn’t is, as Parfit says, a question for another occasion.

(applause and cheering)

[01:36:03] DEREK PARFIT:
Now let me start with a comment about, um, your response to the referendum. I didn’t make clear enough how that applied and also the role of that premise in the argument. Premise B doesn’t claim that what that everyone could always rationally choose whatever they have most reason to choose from an impartial point of view.

I don’t need that claim. Um, as it stands, premise B just says, in this thought experiment, people could rationally choose whatever they have reason to choose because their other reasons for not making that choice don’t apply in this thought experiment. And now the point about the, um, referendum, it isn’t that

(clears throat)

this version of contractualism, uh, can justify voting yes for a referendum whereby if I give up the ten thousand dollars, but so do the other million rich people, ten billion will go to the poor. Certainly, other versions of contractualism can explain that too. My point just was that since that premise is about what people could rationally choose in the thought experiment, the question that they are facing is very like the question that we’d be facing in casting the decisive vote in such a referendum.

That was my only point. It was that in making that choice, the person might know, since there’s no veil of ignorance, the person is rich, she will know that if she votes for the principles that would make things go best, those principles are going to be more redistributive than many other principles. They will impose great taxes on the rich to help the poor.

And so in voting for those principles, she’s voting to incur a cost. But the benefit that she’s producing is a million times greater than the cost in my small example. In this case, of course, it’s billions of times greater because it’s everyone, not just now, but in the future.

So it was entirely directed at that point in the argument. It isn’t the general claim that everyone could always rationally choose what, from an impartial point of view, they have most reason to choose. Now, I’ll just very quickly go through these, um, comments on the screen.

Um, I agree that the version of contractualism I’ve been presenting isn’t Kant’s because it appeals to a view about rationality and reasons that is very unlike Kant’s view. That’s why I spoke of Kant’s contractualist formula. I think the formula is very close to Kant.

The principles whose acceptance everyone could rationally will. And I think that that formula, as I say, it’s just the collective version of that form of the formula of universal law in which we appeal to the principles that each of us could rationally will to be universal, and we just say, which are the ones that everyone could will? And I think it’s much closer to Kant than, say, Rawls’s version of contractualism, um, partly because, as we’ve seen, Rawls is assuming something like the rational agreement formula.

He imposes the veil of ignorance to limit people’s bargaining powers. So that conception of contractualism is the one on which if you don’t have a veil of ignorance, you get to the decisive final vote, and you have to know whether you can believe people’s threats and warnings that they’re not going to vote for the principle you’ve said you’d vote for and so on. That is completely alien to Kant.

The pressure of the need to reach agreement, making people make concessions to one another, and whether they believe each other’s threats couldn’t be more un-Kantian. What is un-Kantian is the view about rationality and reasons. I agree with that.

Um As Scanlon says, when Kant speaks of what one could or could not will, he has in mind the constraints on willing that are constitutive of rational agency itself. We could not will something if so willing it would violate these constraints. Well, there are two reasons why in developing a moral view we might appeal only to the constraints on willing that are constitutive of rational agency.

Let me give one example of those. I mean, it would be, you can’t rationally will an end, but fail to will what you know to be the necessary and unobjectionable means to the end. That’s constitutive of rational will, means-ends rationality.

Well, of the reasons for building a moral theory on such slender premises, one would be that it would be a more striking, successful achievement. We could show that any rational agent on pain of inconsistency or ceasing even to be a rational agent must accept our conclusions. That argument would have enormous importance, and moral philosophers continually try to do that.

Gauthier, for example, claims that by appealing to instrumental rationality, you can show that everyone’s rationally required to obey his minimal morality and not to kill people from whose existence you benefit. Um, I think that project just can’t succeed, and I think Scanlon agrees. Um, and if people have been attracted to Kant’s view because they’ve seen that that would be a more successful argument,

[01:41:42] THOMAS SCANLON:
yes, I think it doesn’t work. Now, there’s another reason why Kant and later Kantians might appeal only to these constraints on rational agency. That may be the total of their conception of non-moral rationality, maybe a very thin conception.

Uh, and I think many Kantians do have a very thin conception. When Onora O’Neill rejects the rational consent principle,

[01:42:08] DEREK PARFIT:
she assumes that in asking whether someone could rationally consent to being treated in some way, uh, all we’d be appealing to is whether there’d be an internal inconsistency given that person’s desires. Uh, she does mention the alternative of appealing to a substantive account of rationality, appealing to claims about ends that are good, but she says that that’s unacceptable since it commits you to an unbelievable form of Platonism. Um, and I think Kant himself, uh, didn’t have a rich conception of reasons in between instrumental rationality and morality.

That’s what I said yesterday. Um, okay, next point. Well, what’s an impartial point of view?

The argument appeals to an impartial point of view. Um, Scanlon says this is defined by subtraction. That’s not inaccurate, but it’s a much too thin description, and that’s partly because it ignores the idea of an impartial point of view.

Um, now there are several ways of defining such a point of view. Most straightforwardly, we can ask what we could rationally choose or have reason to prefer, um, if we were not involved, if we were an impartial observer. More complicated but more accurate, I think, is the way in which, as, uh, Nagel particularly eloquently argues, throughout our life, not just in practical reasoning, science and everything, we have this constant ability to shift back and throw from our personal point of view and an impartial point of view.

And in epistemic reasoning, the impartial point of view is the objective point of view, and so on. Uh, I think it shortchanges the idea of what we could will from an impartial point of view to just say that’s reasons with some of them crossed out. Um, But that’s a minor point.

Now, I turn to the objection to the crucial steps in the argument. Um, Scanlon says that my class of impartial reasons is very broad, but they then seem to narrow the breadth. Um, I’m excluding an appeal to explicitly moral reasons, reasons that are given by the rightness of acts.

But he then says, it’s important to ask why some or all of these reasons shouldn’t count in determining when things go best. And that wasn’t my claim. Um, they can count as helping to determine what makes things go best.

We can agree that it’s worse if more people’s rights are violated. It’s just that, as Scanlon went on to say, in applying a contractualist formula, the only kinds of reasons that we can’t appeal to are ones that presuppose our beliefs about what’s right and wrong, and that’s going to presuppose– that will be true of our beliefs about what people’s rights are. Um, now, um, Scanlon then says, what this rationale seems to support, the one I’ve just given, is not the exclusion of all reasons that aren’t outcome-based, but the exclusion of all morally based reasons, whether outcome-based or not.

Now, I just, uh, I have no idea which kinds of reason Scanlon has in mind there, so that’s just a question.

(laughter)

Um, now, however, it’s quite important that the C is not narrowing down the range of admissible reasons in any way other than this minimal way to make the formula avoid being circular and trivial. And many non-conscious– non-consequentialist reasons remain fully relevant. That’s true of all the personal reasons provided by our own well-being or that of our children, those we love, and so on.

Um, in asking which of the principles that everyone could rationally choose in the thought experiment, there’s no exclusion of their personal reasons and the reasons provided by the well-being of those they love. Um, So you say, surely in answering this question, we should be impartial, but does that mean that we must answer it by asking what would be overall the best outcome? It, of course, seems to me that it doesn’t.

Impartiality doesn’t entail consequentialism. Again, I think that misunderstands the argument. I know it may look like a rabbit out of the hat getting exercise.

I don’t think it is. Um, my claim isn’t that impartiality entails consequentialism. It’s that impartiality plus contractualism, which tells you to exclude appeal to your beliefs about the wrongness of acts, may entail some form of rule consequentialism.

And so, as Scanlon rightly went on to say, um, this argument can be taken not as an argument for rule consequentialism, but as an argument against contractualism. Uh, I’m happy to have it taken that way. Now, finally, the main substantive question at the end.

Take the cases of the great burdens that we could impose on a few people for the sake of the greater sum of benefits to others. Now, when we ask whether Kant’s formula supports a principle which allows us to impose those great burdens, We don’t answer that by asking whether that principle’s acceptance would result overall in the best outcome. Kant’s formula imposes a limit on the burdens that people can be permissibly made to bear for the sake of benefits to others, And it’s nothing to do with the goodness of the outcomes.

It must be true that these people could rationally choose to bear these burdens for the sake of these benefits. If they couldn’t, then we can’t impose it on them. Um…

And then finally, this is the last thing I shall comment on. Parfit’s claim is not that the question must be answered in this way to be impartial. He claims only that it may be so answered.

As long as a person’s sacrifice would result in a state of affairs that was impartially better,

[01:49:02] THOMAS SCANLON:
he or she would not be open to rational criticism in making it. But can we impose on someone else any sacrifice that he or she would not be crazy to make? That seems to me a chillingly permissive criterion, and I suppose that’s Parfit’s point against Kant’s contractualist formula.

I agree with the point, although I don’t find it very surprising. For some reason, I’m reminded of this moving scene in the film Spartacus, when Kirk Douglas and someone else, was it Tony Curtis? I don’t know, are fighting and trying to kill each other,

[01:49:36] DEREK PARFIT:
but it’s because they love each other, and they know the survivor will be cruelly put to death, so they’re trying to kill the other for the sake of the other. Um, now, we’ve got something like this slightly going on here.

(laughter)

because I’m claiming we may need Scanlon’s version of contractualism because of these faults in Kant’s, which seems to me the other best version. And Scanlon is defending Kant’s version, and I’m attacking his defense

(laughter)

because that will make his version more important. Um, Okay, comments on those last remarks. He says, \”I agree with that point, though I don’t find it surprising.\”

Well, I agree that when we ask which kinds of principles Kant’s contractualist formula would support when combined with what Scanlon and I believe to be the right view about reasons, it isn’t surprising that the formula supports rule-consequentialist conclusions. If some argument is both sound and pretty short, we shouldn’t be surprised that when we consider this argument’s premises, that it’s going to support its conclusions.

If there’s a surprise, it comes at an earlier stage. I think what may be surprising is that a consequentialist formula, which I believe is closer to Kant’s own views than any other, and one that Kant came very close to suggesting, that that, when combined with a view about rationality and reasons which isn’t itself consequentialist, should nonetheless lead to rule consequentialist conclusions. That’s what’s surprising.

As, uh, Rawls says, contractualism is the alternative, and that’s its often its ja-justification. Similar claims applies to Rawls’s view. It isn’t surprising.

No, I’ll leave that out. I’ll, I’ll go back to Kant and Scanlon. Um, well, how strong, uh, is the objection that appeals to these cases?

Now, I think Scanlon’s form of the objection seems overstated or misleadingly stated. He asks, can we impose on someone else any sacrifice that he or she wouldn’t be crazy to make? Well, when I say that someone could rationally make some sacrifice, as when White could sacrifice her life to save the five, or you could sacrifice your life to save Black’s life, I’m not claiming that if White and you made those choices, you wouldn’t be being crazy.

I’m claiming that those choices would be fully rational since White and you would have sufficient reason to make them. Um, note next that by using the word sacrifice, that may be suggesting that these great burdens would be imposed as a means of benefiting others, as in the case of Bridge, in which we’d save the five by means of killing White.

But we need also to consider Lifeboat, Tunnel, and Earthquake. We impose a great burden on White if in Lifeboat we save the five rather than White, and if in Tunnel we redirect the train so that it kills White rather than the five. Now Scanlon believes that we’re morally permitted to kill White as a side effect of saving the five and to save the five rather than White.

And it counts in favor of this version of contractualism that it permits or requires those acts. Um, if we turn to Bridge, this version of contractualism again permits us to impose this burden on White, and here we may find the formula unacceptable because we’d be killing White as a means, not as a side effect. We may believe White has a right not to be killed as a means.

Now, we might call that the deontological objection to this version of contractualism. And, uh, I think this is a very understandable opposition. On a contractualist view, which appeals to what we could rationally choose in some way that makes the results impartial, we shall obviously take into account benefits and burdens that different principles or acts would impose on people or give them, and some other things as well.

But it’s much less clear that we should take into account the facts to which the deontological constraints appeal, which are facts about the causal relations between benefits and burdens. Now, as I say, if White could rationally save the five by killing herself as a side effect, she has no reason to regard that as better than saving the five by killing herself as a means. So I think there is a very obvious tension between the contractualist view and deontological constraints.

Um, now I’ll just say very briefly, if we ask what a contractualist formula would imply with respect to those two cases and other such cases, and also I think what Scanlon’s contractualism implies, I think what it supports is the following double principle, which is that doctors must never kill their patients either as a means or as a side effect of saving more people’s lives. But in non-medical emergencies, with trains out of control or avalanches or fires or whatever, everyone should do whatever would save most lives. Um, now it’s easy to see why, uh, doctors should not be permitted to kill their patients, and that doesn’t draw mean side effect.

Just as much reason to exclude that. Um, well, deontological objections are not Scanlon’s main concern. He’s more concerned with distributive objections, the single person who’d bear the great burden.

I just end with a comment about his own proposals. He says the idea of contractualism is that we should ask whether the claims of the few who will be harmed negatively are adequately answered by citing those of the many. By claims here, Scanlon means moral claims.

Uh, Allen Wood protested yesterday against appealing to the distinction between the rational and the reasonable. Well, I accept Wood’s objection to Rawls’s use of this distinction, but reasonable can be used in a specifically moral sense, and that’s what I take Scanlon to be doing. And, um, a choice might fail to be rational without being unreasonable, and vice versa.

I think they’re different. Um, a choice is unreasonable in the moral sense if the chooser is not giving enough weight or the right kind of weight, uh, to other people’s well-being or moral claims. Now, in the cases that we’re now concerned with, great burdens on a few people for the sake of benefits to others, the objection wouldn’t be that you aren’t giving enough weight to their well-being.

You might be giving just as much weight as to everyone else’s. It’s that you’re not giving enough weight to their claim not to have this great burden imposed on them. Um, and that brings me to what I take to be the most important difference between Scanlonian contractualism and the neo-Kantian contractualism that I’ve been describing.

Combines Kant’s formula with a non-Kantian view on reasons. Um, since the neo-Kantian version appeals to claims about what everyone could rationally will or choose on non-moral grounds, it’s more ambitious than Scanlon’s version. So, since Scanlon appeals to what’s reasonable in the moral sense, the premises of his arguments, as he says, are somewhat closer to their conclusions, and more must be said to show how this version of contractualism isn’t circular or trivial, uh.

Uh, Scanlonian moral reasoning sometimes allows us to appeal to our moral intuitions. We may be able to support and explain conclusions about which acts are wrong in an illuminating and non-circular way. But that distinction isn’t as sharp, however, since there are borderline cases.

And I think it may be least sharp in the kind of cases we’re now considering, those in which by imposing great burdens on a few, we could give small benefits to many others. Um, and these are the cases that may count most strongly against the Kantian and in favor of Scanlon. Uh, And that could be expressed as follows.

Even if the few could rationally will that everyone accept a principle that permitted such burdens to be imposed on them, our acceptance of this principle and our rejection of a principle that protects them from that might be unreasonable in the moral sense because we’d be giving too little weight to their moral claims not to be treated like that. Now, that may seem too close to arguing that we’re being unreasonable because it would be wrong to impose these burdens on the few. Um, it would be pointless to say that people have moral claims to be treated in some way if and because such treatment is required by some unrejectable principle.

If we also say that this principle is unrejectable because these people have moral claims not to be treated in that way. Well, I think the circularity objection isn’t decisive. I think that most of it might show is that when we consider certain cases, Scanlon’s conjecture’s formula isn’t really doing independent work, since it’s allowing us to appeal in this somewhat indirect way to our intuitions about the wrongness of acts.

There may be many other kinds of case, however, and many other questions to which that objection doesn’t apply. In these other cases, I think the formula does independent work, helping to explain and justify our beliefs and providing arguments against others. But that’s a subject for another day.

(applause)

[01:59:47] SPEAKER 4:
Candidate for that,

(laughter)

because one thing that should be striking about, uh, Kant, Kant’s presentation of the formulas is that he presents the formula of autonomy often in something like the form: Act on those maxims, maxims that could make themselves or could, that you rationally will to be laws, right?

[02:00:15] DEREK PARFIT:
Yeah. Uh- Yeah

[02:00:16] SPEAKER 4:
universal laws. Uh, and that’s much stronger, uh, requirement than you get in the formula of universal law, where all you are asking is whether it’s, whether, whether it’s conceivable, uh, considering only grounds of thinkability and conflicting volitions, right?

[02:00:34] DEREK PARFIT:
That’s right.

[02:00:35] SPEAKER 4:
So what we, what we, what we really seem to want o-out of Kant is some kind of criterion. Well,

(coughs)

which maxims are they that make themselves, yeah, into universal laws?

[02:00:47] THOMAS SCANLON:
And, um,

(coughs)

and I, I think that, uh, Kant’s contractual formula, as you present it, is a, an interesting candidate for such a criterion. Um, I’m not as deterred as Scanlon is, or maybe even as you are, by the thought that Kant doesn’t accept your view about impartial reason- Yeah, giving senses of good, since the way I read what Kant says has has to say about, uh, rational beings as ends in themselves and about our own perfection and the happiness of others as ends that at the same time are duties, I don’t think he’s at all distant from your view that there are these re- partial-

Yeah, reason-giving goods. Yeah. Um,

[02:01:36] DEREK PARFIT:
But there are really three considerations in Kant that make me not sure, uh, whether, uh, Kant’s contractual formula, as you presented, is a reasonable version of the formula of humanity or the formula of autonomy. Um, and let me try briefly just to say what I think those are. Uh, first, Kant’s theory of ethics in the broad sense, Sitten, is divided into two parts, right and ethics, Ethik, right?

Yeah. And, and there’s a, there’s a, a controversy among Kant people, how you, how you understand this. That is, is the formula of autonomy or formula of universal law, or whatever you take the universal formula of the, of the moral law, does it govern both realms Or does it govern only ethics and the principle of right, which is that an action is right if it or its maxim can be, uh, can is consistent with everyone’s external freedom according to universal law, right?

Is that a separate and distinct principle governing the area of right? I’m of the view that it’s a separate and distinct principle. Yeah.

All right. Suppose I’m right about that. Kant also holds that among the duties belonging to ethics is the duty to respect right.

[02:03:05] THOMAS SCANLON:
Yeah.

[02:03:05] DEREK PARFIT:
All right. So, A-and Wright provides what looked to me like very non-consequentialist constraints on what your duties are. So that would be one source of worry about whether a contractualist, whether Kant’s view could be contractualist and or, or, or consequentialist rather, in the way you’re trying to develop it.

A second worry I have is that when Kant turns to ethics, that is the, the ethics part, Uh, he doesn’t seem to be a consequentialist, but he does seem to be a teleologist. That is, all the reasoning about, uh, what your duties are, whether they’re strict duties or wide duties, is based on this idea that we have these duties of virtue, the, the, uh, our own perfection and the happiness of others. Um, but, but his way of thinking about these things isn’t consequentialist, I don’t think.

And it’s, um, and it certainly deliberately eschews certain kinds of principles that, you know, what would go best on the whole and so on, that kind of thinking. Because the idea is I, it, it’s up to me to choose which bits of my perfection and the happiness of others I make my end. I can–

I don’t– I’m not responsible in any way to choosing the maximum.

[02:04:32] THOMAS SCANLON:
Yeah.

[02:04:33] DEREK PARFIT:
And, uh, I, uh, and I, you know, A-and it doesn’t matter whether, as a result of my setting those ends, anything like the maximum happen, you know, happens. And I’m not—I don’t get more merit for pursuing more of your good. I get more merit for sacrificing more of my good to pursue your good or for putting forth more effort or something like that, but I don’t get more merit for pursuing more of your good.

So that also, that’s another sort of general area in which there’s more I could say about that, but maybe I, I should move on.

[02:05:07] THOMAS SCANLON:
The other thing, uh, is the relation of the au-formula of autonomy to the formula of the realm of ends. And there I s– the way I read the formula of the realm of ends, I am unsure what…

I mean, you don’t mention a certain factor that comes in very strongly in the realm of ends that you could, perhaps you could me-uh, mention. And that is the formula of the realm of ends, in effect, tells us to unify our ends with the ends of other people. Uh, it, it isn’t focused solely on how to balance conflicting ends.

It more, it, it might say in your terms that it’s something impartially good that people modify their ends in such a way that their ends coincide. And so… A-and in fact, Kant thinks there are two kinds of social relations that we have an ethical, not a juridical duty to enter into.

One is friendship, and the other is what he calls the ethical community, and he means by that something like the church. Yeah. And in both cases, what is decisive about those relations is the way in which they lead to your, uh, one way, one term he uses when he’s talking about friendship, is the swallowing up of your ends in a collective end, right?

Kant is, uh, is very communitarian or communist or whatever you wanna say about this. That is- Yeah, he thi- the, the formula of the realm of ends involves Ultimately, you’re not holding to your own good as a, as a, as a, as an end, uh, all by itself.

You, you ought to choose ends in such a way… So the laws that would, would follow a realm of ends would, would, uh, would involve something like trying to make people’s ends converge and be swallowed up into a larger end. Right.

And-

[02:06:58] ANNOUNCER:
Alan, maybe-

[02:06:59] THOMAS SCANLON:
Yeah.

[02:06:59] ANNOUNCER:
you’d like to pipe in.

[02:07:00] ALAN GIBBARD:
Yeah. Maybe let, uh-

[02:07:00] THOMAS SCANLON:
That’s- I’m, I’m finished. I’m finished. I’m sorry, I-

[02:07:03] DEREK PARFIT:
No, that’s okay. No, that, that’s very helpful and relevant. I mean, of course I know Kant wasn’t a consequentialist, and I’m very much thinking only of the Kant of the groundwork because most of the formulas that I’m talking about don’t recur in the Methods of Morals.

Um, really, all that I’m claiming, um, is that the formula, as I interpret it, is very close to Kant, and I think it involves really only two steps. One is, uh, the move that in the formula of universal law, could I rationally will that it be universal? On one version, could I will that everyone accept the maxim or that everyone act on it?

But I think it’s clear in Kant there’s also, could I rationally will that everyone believe that to be permissible? I’ve got at least five quotations in which he says no. Could I will that everyone believe it to be permissible?

Well, that can be restated as could I will that everyone accepts a principle permitting it? And I think he assumes that everyone would reach the same answer to that. And I think, well, actually, that’s not so because of the impartiality objection.

They wouldn’t all reach the same answer. But I’m then I’m saying, but if you have the right assumptions about rationality and reasons, you can get the intersection because there are some principles that everyone could rationally choose. But my only point is the formula itself, and why I think the reference to the realm of ends is relevant, is that that introduces the idea of every rational being legislating, legislating the moral law.

And it may be misleading to call it Kant’s contractualist formula because, of course, it’s worlds away from contractualism as ordinarily conceived, in which there’s debate and mutual interchange and threats and warnings and, you know, calculations to reach agreement. Worlds away from that.

It’s just the ones that each rational being could will to be universal. Now, are there some principles which is true? That’s really all that I’m claiming.

Um, so I, I mean, if you think that is a reading of those elements in Kant, that’s really all I’m claiming. Now, what makes Kant so great, greatest moral philosopher, is that he’s so enormously thought-provoking and stimulating. That’s what my drab hero Sidgwick said about Kant, that he found him so enormously thought-provoking and stimulating.

I think Sidgwick missed a lot in Kant. Sidgwick very badly misunderstands Kant’s categorical formula. He takes it just to be if an act is right for me, it must be right for anyone else in similar circumstances.

That’s a big misunderstanding. Um, so what I’m doing, I’m just drawing on one of the elements, Ken, but I’m sure it’s there. Um.

[02:09:49] ANNOUNCER:
Okay, I think we probably have time for one short question if someone has. Jay?

[02:09:54] JAY:
Uh, yeah, a short question about the, um, discussion of Rawls’ version of the conversation.

[02:09:59] DEREK PARFIT:
Yes, I’m sorry, that seemed a bit dismissive as I read it, but, uh, yeah.

[02:10:03] JAY:
Uh, well, um, you know, I mean, I’ve always, as many people have always been pu-been puzzled about this exclusion in Rawls’s, uh, original position of information about, um, you know, the, the probability that you’d occupy each of the positions, um, in the, in the world that you’re, you’re choosing principles for. Um, but let me have a very, um, no doubt, inadequate stab at trying to defend, uh, defend exclusion of that kind of information, um, from the Rawlsian, uh, choice situation. I, you know-

[02:10:33] DEREK PARFIT:
Well, it’s not so much information, it’s why it, it’s objectionable to tell everyone to suppose that they have an equal chance.

[02:10:41] JAY:
Right. Yeah.

[02:10:41] DEREK PARFIT:
What’s wrong with that version? That’s the idea.

[02:10:44] JAY:
Right. Yeah. Um, okay.

The, the idea would be there, um, there are situations where, uh, your, your, your chances of occupying each of the positions that, um, you know, outcome positions that you’re choosing, uh, that, that your choice, uh, will determine is, um, is that’s important information to take into account. It’s crucial to the rationality of your decision and such. You’re in the game show, you’ve won twenty thousand dollars, there’s a car behind door number five, you know, and, uh, and you’re trying to decide whether to, you know, to bet this or there, there’s a car behind one of these doors and you’re trying to decide whether to, uh, you know, to bet the money on, on the chance that the door you pick will have the car behind it and such.

But in the, the kind of situation that Rawls is imagining, you’re choo- you’re choosing principles to, to govern, uh, the whole of your life, which could radically affect your life prospects, right? And the idea might be that, uh, under that, those kinds of situations, the, the assumption that you’re going to have an equal chance of occupying each of the possible positions will, in some sense, uh, downplay or be the wrong response to the, the significance of the badness of certain potential outcomes.

[02:11:54] DEREK PARFIT:
I agree. I agree. I mean, following, uh, Nagel and Tim, um, reducing the question whether it’s justified to impose these great burdens on a few people to whether it would have been rational for them in self-interested terms to choose principles that might permit that, uh, is just not the right way to do it.

And, um, the objection to telling them they have an equal chance is that then they’re not going to take sufficiently seriously what happens to the people on whom these burdens are imposed. And as many people have said, uh, Rawls’s rejection of the equal chance really is that it’s leading to the wrong conclusion. But I don’t think that affects my claim that that means that his appeal to the veil of ignorance conception can’t support his non-utilitarian conclusions, because his only reason for rejecting the version of it that supports utilitarianism is that it leads to the wrong conclusions.

So–

[02:12:49] JAY:
well, that, that seems potentially just a little bit too quick. It might not be, the objection might not be simply that this is the r– You know, the… it would be the, uh, the wrong outcome that would be, uh, or the wrong set of principles that would be, um, selected on the assumption of, um, you know, equal probability.

Uh, but, but somehow that the decision procedure has to be designed in a way that, uh, takes… that, that expresses in some sense the significance of a difference between, uh, individual lives in some way.

[02:13:19] DEREK PARFIT:
Yes. That might explain to you why, um- And this isn’t the way to do it. That, that isn’t the way to do it.

[02:13:23] JAY:
Yeah. Yeah, yeah, I agree.

[02:13:25] DEREK PARFIT:
Okay. Um, I mean, but I was, as I was poi-pointing out, there are other objections. One is that, um, as Rawls himself admits, um, his argument doesn’t support an acceptable distributive principle in morality. No, I don’t think that’s what you’re saying.

[02:13:39] JAY:
Of course, yeah.

[02:13:39] DEREK PARFIT:
Uh, but then most of all, that I think it excludes all non-utilitarian considerations because it’s just what would best promote their own self-interest. Um.

[02:13:52] ANNOUNCER:
I’m afraid we’re out of time, but we have, uh, tomorrow’s, uh, seminar, which there’ll be more opportunity for discussion. In the meantime, please thank Derek Parfit and all the commentators for their comments.

(applause)

[00:00:01] (laughter)

[00:00:02] MODERATOR:
Attentive readers among you may have noticed that the, uh, the, uh, uh, moderator today is supposed to be Robert Post, and that, uh, I am not Robert Post. Robert unfortunately had to be out of town today, so I’m going to be pinch-hitting. Let me just say a word about the format.

Um, each of the three commentators will be given up to twenty minutes or so to make some um, additional comments as he or s- she sees fit about any aspects of these lectures. Uh, after each of those comments, Derek will, um, make a brief reply. And after that, um,

(cough)

uh, uh, we will, uh, invite questions from the audience. And following that, um, we’ll we hope to have some general discussion. At the very end, Derek be given, uh, some time to make a final– some final remarks about anything that’s come up that he wishes to, uh, expand on.

This is supposed to be a seminar and a discussion, and we’ll try to, uh, be at least partly faithful to the spirit of that idea. So, uh, without further ado, uh, we should begin. Uh, you’ve already, uh– I’ve already introduced all three commentators, so I won’t do it again.

I did just want to say a word of thanks to the three commentators. Uh, for three busy academics to take a week out of their schedules in the middle of the term, uh, and go to another institution to do something like this is extremely difficult and burdensome. And of course, it’s a tribute to the interest and importance of Derek Parfit’s work that they’re willing to do it, and a great stroke of good fortune for us.

So, uh, thank you to, uh, Alan, Susan, and Tim very much for helping to make, uh, to make this week so rewarding. And the first comment, we’re gonna go in the order, uh, the same order that has that prevailed during the week, so Alan will begin.

[00:02:12] TIM SCANLON:
In my remarks today, I will take up some of the examples Parfit used in his opening lecture. I admit that my remarks go off at a sort of forty-five degree angle from Parfit’s lectures themselves. Let it be a measure of my general agreement with Parfit’s lectures that in order to find something to controvert, I have to wander off in this way.

But examples of this kind and their use in moral philosophy has long been a pet peeve of mine. I’m not sure myself just why, and so I’m going to try to say. I’m very uncertain that what I’m going to be saying today is really true or defensible, but it’s something I’ve wanted to discuss for a long time.

In May of 2001, the Tanner lecturer at Stanford was Dorothy Allison, author of the novel, Bastard out of Carolina. Allison didn’t talk much about moral philosophy as such, but she did discuss a lifeboat problem that she had heard from a philosopher. Her reaction was to reject the problem, to refuse to answer it at all, on the ground that she refuses in principle to choose between one life and five lives.

Even to pose the question in those terms, she said, is already immoral. The real moral issue raised by such examples, she thought, is why provision had not been made for more or larger lifeboats. To many philosophers, your remarks will no doubt seem naive, wrong-headed, and unreasonable, but they seem to me the most sensible and right-minded reaction to such problems that I’ve ever heard from anyone.

I’m going to refer to these kinds of examples not as lifeboat problems, but as trolley problems. None of Parfit’s examples are actually about trolleys, though two of them are about trains. They’re all examples where the main point is that you must choose between saving more people from death and saving fewer.

Since a human death is in general bad, given this information alone, it’s natural to think that the option that involves fewer deaths is to be preferred to the one that involves more deaths. The examples gain their poignancy from the fact that this apparently obvious point suddenly begins to seem questionable or even counterintuitive when the fewer deaths are caused in the wrong way. The problem posed is always what we think we should do when faced with such stark choices, and how what we think about this is supposed to bear on various moral principles that philosophers have proposed or might propose.

To take trolley problems in the spirit in which they’re intended, you’re supposed to think that at least some of the issues they pose might be difficult to decide if you had to face them in real life. You are supposed to think that how we decide these difficult issues must have some important implications for the fundamental principles of moral philosophy. But I don’t think either of these things is true.

Like Allison, I think the right way to deal with trolley problems is to refuse in principle to answer them, or rather, as I would modify it, to refuse in principle to take them in the spirit in which they’re intended. I think that to take in, uh, in the way philosophers want you to might not be merely immoral, but perhaps even worse, it might be bad moral philosophy. I don’t think our so-called intuitions about any examples are worth much to philosophy unless they are honest and critical moral reactions to the kinds of problems that might actually arise in our lives, hence problems to which we could reasonably think that our moral education and experience might serve as some sort of reliable guide.

But trolley problems, as philosophers pose them, are virtually never of that kind. They casually assume we are certain about things that we could never be certain about in real life, and they cavalierly omit facts about the context that we would know about in real life, facts on which what we ought to do would vitally depend. Virtually all trolley problems are such that just witnessing such a situation in the real world would produce feelings of horror and anguish in us, no matter what we did or didn’t do.

If we place most of these problems in a more realistic context, that feature would remain, But it would be– It would not– But it would usually not be difficult to know what the agent should do.

The decision would very likely not turn on any of the factors to which the trolley problem philosophers try to draw our attention. Suppose a moral philosopher gave you the following example: A group of white people are stranded on one rock, and a group of black people are stranded on another. Before the rising tide covers both rocks, we could use a lifeboat to save either the white people or the black people.

Which should we save? Since the philosopher has told you nothing about how many people are in each group or even anything else about them except their skin color, I would hope that you would resist giving any answer at all to the philosopher’s question. If you did have the intuition that you should save the group of white people or even the group whose skin color is the same as your own, then I would hope that you would resist answering on the basis of that intuition, and then you would even dislike yourself for having had that intuition at all.

Certainly, you should not think that agreement with such an intuition ought to serve as a test that basic moral principles ought to pass. What is most objectionable here is the conversational implicature of the philosopher’s question itself. The question implies, namely, that you have been given enough information to answer the question as posed, or at least enough to have some intuition worth reflecting on about, about what the answer should be.

In the case of the example I’ve just described, that implicature is morally offensive all by itself in a very obvious way. Most trolley problems differ from that example, in that in them we have been given information about the situation that is at least prima facie morally relevant. The number of people on each rock is at least not obviously and offensively irrelevant to our decision in a way that their skin color is.

But it may still be true that in trolley problems, we have typically not been given enough information or the right information to have intuitions that are worth much to moral philosophy. In real life, people go to a lot of trouble to arrange things so that no one will ever be placed in the position that, for example, the bystander in the train examples is placed. There are sound moral reasons why this is so, reasons closely connected to Dorothy Allison’s sound observation that it’s already immoral to ask anyone to decide between one person’s life and five people’s lives.

Her point, I believe, is that even if some choice inevitably has the consequence that either one will die or that five will die, it is still immoral to look at the choice only in that way. In light of this point, I want briefly to discuss each of Parfit’s three examples, making, unlike the philosophers who posed the problem, some more realistic assumptions about things that we would know and about things we would not know. Assumptions about facts whose omission or inclusion is responsible for the philosophical deception I think the examples perpetrated on us.

Lifeboat. When faced with a situation like lifeboat, there is only one morally defensible policy. We must seek to rescue all six people as quickly and efficiently as possible.

Perhaps it is true that following this policy, we should set about rescuing the five before we rescue White, but only because in that way we will go farther, faster, and with greater certainty toward achieving our goal, which is rescuing all six people. If for some people, for any reason, we thought we could go farther, faster, and with greater certainty toward the goal of saving all six by rescuing White first, say, because White’s rock is right on our way to the rock with the other five on it, then we should do that. It is also highly relevant here that in the real world, if both rocks are in imminent danger of being swept under the water, then you could not know for certain that you must choose between saving White and saving the five.

In the real world, if you set out to save all six and took the best means to this end, then there would always be some chance that you would succeed in saving all six. And if both rocks were about to go under, there would also probably be a significant chance that no matter what you did, all six would drown. If a philosopher simply stipulates that we are certain that we can save all and only the inhabitants of exactly one rock, then we should be clear that he’s posing a problem sufficiently different from otherwise similar problems we might face in real life, that any intuitions we have in response to the philosopher’s problem should already be suspect.

What is certainly clear about a situation such as lifeboat is our intuition that if any of the six drown, the result is tragic, probably a ground for grief and traumatic memories for years afterwards. We should regard ourselves as having failed significantly no matter what we did, even if our failure was inevitable and not our fault. Another vivid intuition is that we need to call to account whoever is to blame for the fact that there were not enough lifeboats, find out why this happened, and take steps to minimize the chances of its ever happening again.

If you are a decent and thoughtful person, these intuitions would be at least as strong as any intuition you might have about what you should actually do about White and the five. Yet trolley problem philosophers focus attention short-sightedly on what we should do in the immediate situation, and they would tend to deride us as philosophically dense if we even bothered to express the intuitions that are clearer and more natural. The fact that those intuitions are irrelevant to what interests them ought already to make us distrust both their moral and their philosophical judgment.

Tunnel. Trains and trolley cars are either the responsibility of public agencies or private companies that ought to be, and usually are, carefully regulated by the state with a view to ensuring public safety and avoiding loss of life. I take the following five points to be reasonable guesses about the kinds of rules and policies that would result from their responsible regulation.

One, every precaution should be taken to prevent runaway trains or trolleys. Where trains or trolleys run out of control, a heavy burden of responsibility is borne by those in charge of the rail system and by those who are responsible for regulating it. When any such example is mentioned, the first thing a right-thinking person will be reminded of is the unconscionable privatization of the British railways under the Thatcher regime, which is directly to blame for several derailments and many deaths.

Two, people should not be permitted to be on tracks where they might be endangered by runaway trains or trolleys. There would ideally be provisions for physically preventing anyone from being there. If either the five or White in the tunnel have nevertheless succeeded in disobeying the rules and entered such dangerous areas, they should be regarded as doing so entirely at their own risk.

Three, mere bystanders, too, ought to be physically prevented from getting at the switching points of a train or trolley. They have no business touching in such equipment under any circumstances or for any reason. Four, mere bystanders who meddle with the switches are morally to blame for any harm they cause and should be liable to criminal prosecution.

If that meddling causes the death of anyone, whether it caused the death of one or the death of five, then they should be held criminally responsible for any death or injury they may have caused. Five, if the person near the switching points is not a mere bystander, but an employee of the railway system whose job it is to deal with emergencies, then that person must be given strict rules how to deal with cases like Tunnel. What the employee ought to do is then strictly determined by the rules.

His primary aim would be to ensure that no one will ever be injured or killed in such situations. It would be a ground for criminal prosecution of the railway, railway’s management if the employee were given no guidance except to consult their private moral intuitions or only very general principles like kill as few people as possible, or the duty not to kill takes precedence over the duty to save lives. Such abstract moral principles might play a big role in the justification of the instructions the employee was given, but it would be a hysteron proteron to regard reflections on the employee’s situation as a good way of justifying those principles.

Now let’s apply these points to the tunnel. As mere bystanders, we have no business touching the switching points for any reason. It is relevant that in the real world, we could not be sure h– we n-know how to operate the mechanism properly.

For all we know, our attempt to save the five might result in wrecking the runaway train and killing dozens of people on board. Further, in the real world, if we see five people in one tunnel and one person in another tunnel, we have no way of knowing whether farther down the track from the one there are not also many more people we would also be killing by switching the points. In the real world, for all the bystander could know, the five people might be interlopers present on the track illegally and entirely at their own risk, perhaps with some criminal intent. While– while White is an employee of the railway who is there on the job, of course, with every guarantee that it will be a safe place to work.

Of course, if I were the bystander who correctly did nothing, I might very well second-guess myself in my nightmares for years afterwards, torturing myself with the thought that there might have been something I could have done to save the five. This might be a natural human reaction to the horrible scene I had witnessed, but my feelings of guilt and self-reproach, though perhaps understandable, would be irrational. Far worse would be the truly monstrous state of mind of the bystander who switched the points, killing White but saving the five, and who then thought that he had been treated unjustly when he was sentenced to prison for manslaughter.

Perhaps the five interlopers would show up at his parole hearing to argue on his behalf, but the parole board would be unwise to listen to them. I would also recommend that he not spend his time in prison reading journal articles about trolley problems. It would not be good for his moral rehabilitation.

Bridge. Many of the same points apply here as apply to tunnel, except that here the wrongdoing of the bystander who acts to save the five is obviously far graver. For here, the bystander surely must suppose that White, in walking on the bridge over the train, is walking in a place where people have a perfect right to walk, and to regard themselves as free from the risk of harm from the deeds either of railway employees or meddling bystanders.

The five, however, can be presumed to have entered a forbidden zone at their own risk. To kill White to save the five would be, in this case, not merely manslaughter, but murder. The bystander does have the consolation, as he sits in his cell contemplating the long, dismal life before him, that some of the great phil– greatest philosophical minds in the world, holding professorships of moral philosophy at prestigious insti- universities think it worthwhile to reflect on the moral intuitions that put him where he is.

I hope I will be forgiven for wishing I could deprive him of his last one solace. If cases like Tunnel or Bridge occurred in the real world, there would or should be a big public outcry against the railway and perhaps against the regulating agencies of the government. The question whether one died or five died would be of much less importance to the protesters than the fact that a runaway train had caused death.

If it were further to come to light that the choice of who died had been at the mercy of some bystander acting on their private moral intuitions, this would be a further ground for public outrage, and it would or should make little difference whether the bystander had chosen the death of one or the death of five. Examples like lifeboat, the tunnel, and the bridge seem theory-driven to the extent that they appear to assume the basic subject matter of normative, normative ethics consists solely in reckoning up the goodness or badness of states of affairs, also taking into account the various causal relations human actions may have to those states of affairs. Trolley problems are often little more than vehicles for representing abstract principles deciding these matters.

And I suppose the examples are natural ones to reflect on for anyone who has made that assumption. But those of us who do not accept it should not be expected to take the examples at face value. It has occurred to me that perhaps my animus against trolley problems comes from my rejection of consequentialism.

But part of my thought is that it is a bad consequence when someone has to make these kinds of decisions on these grounds. And I would think it a good feature of some moral principles that they would lead to the prevention of situations where people find it, uh, find themselves asking and these questions and trying to answer them on these grounds. On the other hand, trolley problems appear to some people, Or, or one reason trolley problems appeal to some people is that at times it seems to them that the only honest way to confront many social policy decisions is to see them as frank trade-offs between the deepest interests of different people.

Surely it is true of many social policy decisions that if they are made one way, then these people will be hurt, and if they are made the other way, then those other people will be hurt. In some moods, I think evil moods, or under the influence of some moral theories, bad theories, it may seem that the only honest way to look at any moral problem is simply to see it starkly in these terms. I do agree that there are indeed desperate situations in real life, in conditions of war or anarchy, pestilence, famine, or natural disaster, where it can look as if the only way to think rationally about them is simply to consider coldly and grimly the numbers of people, the amount of benefit and harm, and the kind of actions available to you that will produce the ben-benefit and harm.

(breathing)

So I’m not saying that trolley problems can’t ever help us to think about what we ought to do in these extreme situations. It’s rather that such situations really are much rarer than trolley problems may lead us to believe, and that we don’t necessarily have such clear intuitions about how to deal with these extreme situations as reflections on trolley problems may lead you to think we have. In real life, situations of this kind, in war for instance, what happens is that we have been deprived of humanizing social institutions like those that pro– should provide enough lifeboats or prevent runaway trains and keep interlopers off the tracks and away from switches that make it possible to look at the world in better ways or impossible to look at the world as trolley problems might lead you to.

We think of war as a morally unacceptable condition, in part because in war it can sometimes seem rational to look at the world in monstrous ways. Our first task as moral beings, I think, is to view things in better ways than this, and even to change the world so far as we can to bring it about that there are other ways of viewing it rationally. But if you take situations in life that are not as barbarous as war, or even take human life as a whole and choose to view it as trolley problems lead you to, then that amounts to a voluntary choice on your part to turn that situation, or even human life as a whole, into something barbarous like war.

Some people may think these, m- these sentiments must be motivated by something like a principled belief in the sanctity of human life. I’ve thought about this a good deal, and I think this interpretation of my objections would be wrong.

For one thing, I don’t even believe in the inherent sanctity of human life. I think that a lot of humbug and pernicious superstition are involved in the popular moral use of this idea. For another, I think I would have much the same objections to trolley problems even if the examples involved lesser goods and evils than life and death.

I think my objections have more to do with something like the Kantian ideal of a realm of ends. The rough idea is that we should not think about moral problems in terms of trade-offs between competing human ends, but should try to understand the answer to every problem as one that treats all people as ends and leaves out no human ends except those that exclude themselves from the harmonious system or realm of rational ends. I don’t think this line of thinking is really opposed to the line that Parfit has presented in his lectures.

If Kant’s contractualist formula involves trying to create a realm of ends, then everyone has reason to accept the principle that we should want no one to be in the position of the bystander, uh, or life… or the lifeboat operator, or to have to think about situations in the way these people seem forced to think about them in the trolley problems. If this principle is also more important and more basic than any principle about what a person in that situation should do, then I think that shows that the trolley problems involve misleading implications about what kind of problems and ways of thinking are most basic and essential for moral philosophy.

(mm-hmm)

Thank you.

[00:22:44] MODERATOR:
Derek, can I reply?

[00:22:46] DEREK PARFIT:
Yes. Well, I’ll just say, uh, I agree that it makes a difference what kind of examples we’re appealing to, and I think the simplest distinction is three kinds of imagined space. Um, there are some cases that involve deeply implausible assumptions.

There are others that, no, could happen, but it’s rather rare. And then there are cases that are very common. Now, I think the cases you’ve been discussing, none of them are in the first category, but I think that tunnel and bridge are pretty rare.

So is Kant’s example, should you lie to the would-be murderer? And some utilitarians say it’s no objection to my view that it implies that it would be right to kill one person as a means by saying we needn’t consider that kind of case. Similarly, we could say it’s no objection to Kant’s view that you should never lie, that we needn’t consider whether it’s a would-be murderer who asks you where the victim is.

We could just say, ignore cases that don’t often occur. Now, but lifeboat, I think, is a case that occurs all the time. I think we’re in lifeboat all the time.

But there’s an assumption that I think we always bring to bear on these cases. You assume there’s no other relevant difference. You can save one person or five, but there’s no other relevant difference.

Now, perhaps you can’t be certain that the chances of saving one are about as good as the chances of saving five. And I think cases in which we have a decision how many people to save are extremely common. There are basically four kinds of answers.

Some people think you should always try and save the larger number. Others, some Kantians think, well, you certainly ought to save at least one person, but you don’t have to maximize and try and save more if you can. Some people think you don’t have any obligation to save these people’s lives, even if you fairly easily could.

That’s most people’s view. And then some people think it’s immoral even to ask the question. Now, I think when you said that you were rewarding us for a rather strenuous week.

But, uh, you have very strong intuitions about all of these examples. And I’m somewhat puzzled why-

[00:24:54] TIM SCANLON:
Mm-hmm.

[00:24:55] DEREK PARFIT:
-you think that we shouldn’t discuss them. But we-we’ll come back to that.

[00:25:01] MODERATOR:
Okay. Um, we will defer further discussion of that puzzle for later. Susan.

[00:25:09] SUSAN:
Thank you. Uh, can you hear me? I don’t know if this is on or… Can you hear me?

[00:25:14] MODERATOR:
It’s supposed to be on. Move in a little closer.

[00:25:16] SUSAN:
Okay. Uh, first of all, before I get distracted and forget, um, since Sam started by thanking us, I would like, on behalf of us, to thank you, the Philosophy Department, the Tanner Committee, the Tanner Fund, Ellen Gobler, if she’s here, uh, and Derek Parfit for an incredibly stimulating week. Uh, I’m sure it’s been all of our pleasures, certainly mine, so.

Uh, I’ve discovered about myself that whenever I’m giving lectures in, in class or at home to my children or anywhere else, I always need to make three points. And, um, so for no other reason, I have three comments to offer, uh, right now. Uh, the first you’ll be happy to hear is very small.

Ah, it’s just this, uh, it’s natural given the content of these lectures, and perhaps even more natural given, uh, the philosophical dispositions of my other distinguished commentators, um, that the discussion has been among the various versions of contractualism, uh, Kantianism and consequentialism. You know, which one’s right, which one’s the best, how do they relate to each other, and so on. But I, I think it is useful at some point to step back and remember that that doesn’t exhaust, uh, the options of moral thought, that there are, uh, excellent philosophers whose, um, response to both Kantian slash contractualism and consequentialism is a curse on both your houses.

Um, or less emotively, uh, to say, look, there are some assumptions made by theorists in both these camps that, uh, they have in common and that there might be some reason to question. And, um, maybe I wasn’t holding up my end of, uh, um, bargain of your inviting me here by not br- mentioning that before. Um, the– there’s only one aspect of that that I wanted specifically to bring out.

Um, and that is that what Kantianism and contractualism in probably any form, but certainly in the Kantian forms and consequentialism have in common is the, uh, maybe not assumption, but at least the operating, uh, principle that, that we’re looking for one supreme moral law. That there’s supposed to be one fundamental criterion of rightness and wrongness, and that-that’s what we’re looking for. Um, that certainly in the written version of, of the lectures, uh, gets mentioned.

You know, this– we’re looking for the supreme moral law. This can’t be it, this… Maybe this is it.

Um, and there’s at least one way in which I think that, that has substantive consequences, which I will bring out in a minute because it connects to the second point I wanna make. Uh, the second point, uh, basically takes off from and concerns what went on in the first lecture, so I hope, uh, you all remember that. Uh, it began, uh, with, uh, Derek’s mentioning Kant’s formula of humanity, formula of ends in themselves that says, “Act in such a way that you treat humanity, whether in your own person or in the person of another, always at the same time as an end and never simply as a means.”

Now, Parfit began by focusing on the idea that you shouldn’t treat someone as a mere means. Uh, that, that aspect of the formula is what he started with, and he understood this to be equivalent to the idea that, uh, in treating someone, um, in using someone for some purposes, you don’t treat him as, as a mere means if you restrict the way you’re willing to treat him, um, according, you know, by some morally significant criterion, presumably some morally significant criterion having to do with their value or their good. Um, now I didn’t go back and look at the exact wording.

Um, that was the way I, I remembered it, and I’m not sure whether your use of morally significant there actually, uh, fits with your later use of moral versus non-moral grounds here because it wasn’t, uh, as I understood it, uh, to treat someone as not a mere means, just meant that there was… it wasn’t– didn’t mean that you restricted yourself on grounds of what was morally wrong, but on some… Because this was supposed to be a way of figuring out what was morally wrong.

However that may be, um, I wanted to say something about about that condition. That is, that interpretation of what it is to not treat someone as a mere means. Uh, it seems to me a perfectly reasonable way to use the words to ha-have it mean what you do, but that strikes me as a very broad and very

, uh, weak criterion Um, so much so that I suspect that only a psychopath, um, would actually be someone who doesn’t, uh,

(coughing)

who doesn’t obey the principle of not treating people as a mere means. I mean, even your ordinary gangster, you know, doesn’t kill someone unless they have some reason to. Um, uh, and mere inconvenience doesn’t usually constitute a reason.

Of course, there’s always the issue of whether they get caught. I mean, it would be hard to tell perhaps. But I think the, uh, there is a very strong, uh, human reaction not to treat people as mere means in that very, uh, broad sense of, well, I won’t just do anything, uh, to them in order to meet my aims as efficiently and conveniently as possible.

Um, as, uh, as Derek himself had mentioned, you know, the, uh, animal researcher doesn’t treat her animals as a mere means if, um, you know, there are some ways in which s-she won’t perform experiments because they cause certain kinds of undue pain or, uh, mutilation. Uh, the Chinese robbers didn’t treat your mother as a mere means because they only took half of her, um, her belongings. Uh, actually, when I started thinking about it, it, it seemed to me, except for kind of used tissues and, you know, a-actual garbage, there’s almost nothing that I treat as a mere means.

I mean, I don’t, I don’t think I treat my houseplants as a mere means. I, I– you know, even though, uh, I often just would like to be rid of them, I can’t just- I just can’t throw them out.

Like, you know, they’re living things. I try to bring them in.

(breath)

And, um, you know, as, as some of you, uh, know, I’ve just recently moved house. I have these huge boxes and bags of stuff that I shipped from storage, and I have to go through it all and try, you know, try to sort it out and, you know, cut down. You know, I can’t- There are things that, that really are literally mere means.

I mean, they are tools, but if they’re in perfectly good condition, I can’t just throw them out or smash them to bits. I, you know, I have to give them to Goodwill or do some- All right?

So it seems to me that I’m not even treating mere means as mere means. And that is, I think, just a way of showing that that as a condition is really very weak. Um,

(laughter)

but, uh, Derek then goes on to take, uh, what he takes to be a, an interpretation of what Kant has in mind about how we treat persons, and that’s his principle of rational consent. Don’t treat someone in a way in which

(clears throat)

to– they could not rationally consent. Now, that is a more significant proposal. I don’t think when I, I place that restriction on my houseplants, uh, on my treatment of houseplants and so on.

Um, but it’s very far from what I had at least previously understood Kant to be interested in, uh, um, or saying with the formula of ends in themselves. Um, I don’t really, uh, I wanna try to defend my interpretation of Kant, but what I had previously understood at least as Kant’s, uh, concern is one that, whether it was Kant’s concern or not is, I think, a recognizable moral concern. Um, which is, as I say, I think very far from the constraint that you don’t do things to them or with them to which they could not rationally consent.

Uh, what I had understood, um, well, it was closer to what Allen Wood said on, uh, in his reply to that lecture. Uh, you know, when Kant says treat humanity as an end, uh, oddly enough, what he means by humanity is rational nature, which is, um, it is very odd because that doesn’t seem to be in any way equivalent to humanity. But it was– the idea was to, to treat humans qua rational beings as ends, and in particular to respect them as rational beings, to respect them as… at least this is what I thought, and what in any case seems to me a worthwhile moral aspiration, to respect people as themselves possessors and centers of reason, including practical reason, as beings who give the law to themselves, who make their own value judgments and can make their own choices about what to do and what to allow to be done to them.

And so as I understood it, the idea that we should treat people as ends in themselves, at least in its fullest, um, realization, would involve allowing them, as far as possible, to express and pursue their own values, to exercise their own reason, insofar as that’s compatible with other rational beings being allowed to exercise theirs and pursue their ends. Um, but that understanding of what’s going on, um, in treating people as ends in themselves would necessarily urge us to restrict ourselves according to what they actually consent to, um, or at least what they– what we have reason to think they would consent to if we were able to check with them, rather than restricting s-themselves as Parfit’s condition of rational consent says to what they could rationally consent to in, in the sense merely that there is sufficient reason. So if they did consent to that, it wouldn’t be, uh, irrational.

Um, it’s interesting that– I mean, the story of the Chinese robbers can actually illustrate the difference between Parfit’s condition of rational consent and this other condition that, of actual consent, in that the robbers, uh, not only restricted themselves by saying, “I’ll only steal half of your mother’s stuff,” but she– they asked her, did, did she want the engagement ring or did you want to keep the wedding ring? Um, That is, they asked for her actual consent, her actual choice.

And so to that extent, they respected as a

[00:36:57] TIM SCANLON:
Her?

[00:36:58] SUSAN:
center of practical reason, whereas it would have been compatible with the principle, with Parfit’s principle of rational consent to look at the two rings to say, “Well, this one’s more valuable.

[00:37:08] TIM SCANLON:
Okay.

[00:37:09] SUSAN:
You know, it’s worth more money or something, so we’ll leave you that one.” Right? I mean, that would be certainly, um, something they could

[00:37:16] TIM SCANLON:
Okay.

[00:37:17] SUSAN:
say she could rationally, you know, agree to. But she might not, and the fact that they asked her actually showed that they were, at least to that degree— um, honoring and it– you know, the principle that I regarded as more, more Kantian or in any case, um, a better thing. It was better that they asked her and let her decide which ring to keep.

Uh, to take another example, uh, let us suppose that opera is a good thing, um, that it would therefore be, you know, good for you to go to operas, to learn about operas, to cultivate an interest in operas, and so on. Um, now, if that’s right, as I understand Parfit’s way of thinking, then you could, you could rationally consent to going to an opera since it would be good for you if you did. Um, and that therefore, if I were just to force you to go to the opera, I would be meeting–

I mean, other things being equal, uh, I would be meeting the requirement of rational consent. I would be doing something to you that you could rationally consent to. Um, but I would not be respecting your rational nature in the way in which I understood that, um, if I just unilaterally decide that you have to go and watch opera whether you want to or not.

Uh, if my kids had heard, you know, had thought of this, this would have been their reply when I made them go to the opera. But, um, in any case, it was all right when they were my kids. I mean, they’re still my kids, but they’re older now.

Um, but most of us think there is something wrong with this, with just making people do things because they could rationally consent if they don’t, and if we know that they wouldn’t, given the choice. Uh, now, Parfit did consider the idea that we restrict ourselves, uh, not to what people could rationally consent to, but to what they actually did, um, actually consent to or would actually consent to. Um, or even to the subset of things that they would rationally actually consent to.

Uh-

[00:39:33] TIM SCANLON:
Well, well, yeah,

[00:39:34] SUSAN:
that is, it has to be- any rational, I mean, that is, leave out the irrational things that they want, right? Um, he referred to that as the veto principle or, or some-something with veto in the title, and points out that that’s too strong.

(mm-hmm)

Um, For White, for example, might not actually consent to being passed over so that you can save the five others, uh, even in the lifeboat case, um, much less than any of the others. And, and of course, the others might not consent to our passing over them to save White. So if, I mean, it’s unacceptable, obviously, to, uh, you know, to accept the veto principle as a condition of wrongness that would, you know, that would lead to unacceptable conclusions.

I agree that it leads to an unacceptable conclusion that, you know, there are times when you cannot, uh, you know, get somebody’s actual consent, uh, or,

(Thank you,)

or act only, or restrict yourself according to what they would actually consent to. Sometimes you have, you know, you just have to, uh, bypass that, and that therefore, I agree with Parfit that the veto principle is too strong as a criterion of wrongness, as an absolute criterion. You can– You know, this is a, a condition of something morally permissible.

Um, but this is how my first point actually connects to the second point. That is, um, the fact that it, uh, that it’s too strong as an absolute requirement of moral permissibility in Parfit’s case is, well, that’s too strong, so let’s go on to find some other different principle that might not be too strong. Which if you’re searching exclusively for the principle that will, that will settle, all the, uh, questions is an appropriate thing to do.

But if you’re looking to understand the moral world by saying, well, you know, if you’re just looking to figure out, well, what things are morally important considerations? What aspirations should we have? What kinds of things should we morally aim for?

What’s a wrong-making feature? Then to discard what seems to me to be a very plausible, um, moral desideratum, namely that, that you get people’s actual consent or something sufficiently close, uh, to that, uh, because it’s too strong for a final criterion is to then lose something, uh, you know, very important. And so it makes a, a very big, uh, it’s a very big loss to the moral picture, I think.

(breathing)

Um, I-I– One last re-remark in connection with this, Um, it’s often seemed, uh, that a difference between consequentialist thinking on the one hand and K- and Kantian outlooks, um, and contractualist outlooks on the other, is that consequentialists tend to make or, uh, don’t have a problem making unilateral decisions about how to deal with the world. Um, and there’s this, you know, it sometimes seems like a certain kind of arrogance that they’re going to figure out what’s best for everyone or the world as a whole. Whereas a Kantian or contractualist perspective is gonna take more seriously and as, uh, you know, um, as as a real restriction on what they can do, what other people think, want, and consent to it.

And, um, so in that respect, the difference between what– restricting yourself according to the principle that says what they could rationally consent to, and restricting yourself according to a principle, um, restricting yourself as much as is possible according to a principle of what they actually consent to is, um, uh, tracks this difference. If I’ve gone too long, I will… I, I have uh, another point, but I won’t…

I can just forget it.

[00:43:27] MODERATOR:
Another minute or two, perhaps?

[00:43:29] SUSAN:
Yeah, okay. I can do… I can make a…

The last point is very sketchy, and I could actually just make it very short. Um, and it has to do with the theory of reasons that Parfit wants to supplement, uh, his contractualist formula with. Right, um, the idea is that if we’re gonna restrict ourselves to what everyone could rationally consent to, then of course it’s important to know what they could rationally consent to, which, uh, is to say, what is it that you think there is sufficient reason to approve of or, or, uh, prefer and so on.

Um, I mean, the shortest way to make my comment is just to say, uh, it would be very nice to hear a lot more of that because there hasn’t been a chance other than saying it’s value-based rather than desire-based. Um, we haven’t heard very much more of it. Um, I guess the, the one worry I had that was– that I would specifically like addressed is that, um, even though in comments you had said at one point, uh, it, it struck you that a lot of moral philosophy, um, seemed to think there were only two kinds of reasons, self-interested reasons, um, or desire-based reasons and moral reasons, and that was it, and you thought there are lots more, and I certainly agree with that claim.

Um, there was something in the way you talked that suggested that you thought, well, there are, there are only two kinds of reasons, or um, self-interested reasons and impartial reasons. Um, and that’s it. And I, I guess the, the one little clue that made me nervous about this was that in the, the one case that I recall in which you felt pretty confident, uh, something might be irrational was the case of someone, uh, consenting to the idea that we save Black’s leg rather than his life.

And I thought, well, why… You know, unless the only two options are do something that’s in your self-interest or do something that’s best from an impartial point of view, Why would that be irrational?

[00:45:38] DEREK PARFIT:
Well, a large number of very good points here. I don’t have a very good memory, and that’s the first I’ve heard them. I’ll try and go backwards just over the main ones very quickly, and then we’ll go on.

Um, you need to assume other things are relevant and equal. Um, if Black, Uh, is 90 and you are a ballet dancer age 20, I think Black could kill himself to save your leg because he could plausibly think what he had to lose is not very much in saving you from a great burden. But I think if you don’t assume there are these other morally relevant differences, it would be in s– one way admirable if I sacrificed my life to save your leg.

But I don’t think you have a sufficient reason to do it, and I think people who tried to persuade you not to do it would be doing so for that reason. Um, with respect to the rational consent principle, I never intended it to be put forward as the supreme principle of morality. I suggested that you might have three principles.

Uh, one says, when your acts would affect only one person, then you might say he should have a veto. It’s up to him to decide what you do. I’d have a slight rider, unless his decision is very irrational.

Two, there may be some things that you should never do to people without their actual consent, even when it affects other people. Those would be the veto. But then the question is with respect to all other matters.

Now, um, I was working from the phrase that the man whom I deceive cannot possibly consent to my treatment of him. And of course, Korsgaard, O’Neill, and others take that to support the veto principle that you should never treat anyone in any way to which they don’t actually consent or they would not actually consent. You agree that’s much too strong, and I was simply trying to see whether there is some plausible principle that you could add, having given the veto to certain restricted cases and said, “Well, you have to choose it, you’re the only person affected.”

Can we do anything else in terms of predicting what people could rationally consent to? But it wasn’t meant to be a single principle. I mean, the supreme principle.

And the case of you being forced to go to the opera sounds rather like my case of rape. I don’t think you could rationally consent in advance to being forced to go to an opera if at the time you didn’t consent. So I think that’s okay.

That’s not an objection to the principle.

(laughter)

Um, uh, turning then to merely as a means. Um, I said, uh, significantly and relevantly affected because if there are certain trivial things that you wouldn’t do to someone, even if it would serve your purposes, then you’re very close to treating the person merely as a means. Uh, so it is a matter of degree.

Um, I think very often people do treat each other merely as a means, treat other people merely as a means, and I think that’s certainly wrong. Um, but now it raises a wider question, which is: Should we use simple terms and phrases in anything other than their ordinary sense? I’m assuming that when Kant claims you shouldn’t treat someone merely as a means, he means that, And I think my account is the only thing that it could mean.

Now, some Kantians say, “No, no, he’s using it in a special sense.” For example, some of them say, if you see someone in great distress and you ignore them, you just walk on, you’re treating them merely as a means. Now, that’s just false.

You’re not treating them as a means at all, let alone merely as a means. The thing to say there is you’re treating them as a mere thing, as if their well-being doesn’t matter. Similarly, Korsgaard and O’Neill say that if you treat someone in some way to which they don’t actually consent, or you use any force to stop them acting in some way, you’re treating them merely as a means.

Suppose you and I are in a desperate situation from which only one of us can escape alive. I stop you from sacrificing your life to save me so that I can sacrifice mine to save you. Of course, Korsgaard and O’Neill, I’m treating you merely as a means.

Well, I don’t admire G.E. Moore nearly as much as other people do, but I do like his literal-mindedness. And if we say what Kant says, for example, you should never lie.

(coughing)

But Kant doesn’t mean never by never, he means often, sort of.

[00:50:21] AUDIENCE MEMBER:
So then you shouldn’t say this.

(laughter)

[00:50:23] DEREK PARFIT:
He says we should never lie.

(laughter)

And merely as a means has a perfectly clear sense.

[00:50:27] AUDIENCE MEMBER:
That’s what happens.

[00:50:29] DEREK PARFIT:
Uh, then the map. Well, I quite agree. I agree.

Uh, I mean, I would broaden the map perhaps by saying most of us hold ordinary common sense pluralist morality. Then you might say, well, there are four kinds of systematic theory, not just two. There’s consequentialist utilitarianism.

Then there’s Kant, then there’s contractualism, and then there’s virtue ethics. Now, I’m inclined to think that those four systematic theories actually are all going up the same mountain, and we’ll find they reach the same place at the top. Uh, but they then– they may then disagree very strongly with ordinary common sense morality.

And I’m in some ways deeply suspicious of these systematic theories. And the main respect I am is this, and this cuts across the map. It’s the question of the method.

Sidgwick was an act utilitarian, but he thought the method of moral thinking is to examine your deepest moral convictions and submit them to the greatest scrutiny. Whereas Kant and contractualism have a method according to which you shouldn’t appeal to your moral intuitions, and I think that’s deeply questionable. Why shouldn’t you appeal to your moral intuitions?

And so I wasn’t intending to scorn these systematic theories. I was trying to discuss what their most striking features were, and people say that’s the wrong way to do it. I just think that would be wrong.

I, I treat that with tremendous respect.

[00:51:57] SUSAN:
Um.

[00:52:00] MODERATOR:
Thank you. Tim?

[00:52:02] ALAN:
Well, here we are back in the realm of moral theory with me, but I’ll try to describe, uh, moral theory in a way that incorporates some of the pluralism that, uh, Susan and Derek has spoken out for. So this tries to be, although it may not seem that, it may not seem that way to you at first, uh, but I hope it’ll look more like moral theory with a human face.

[00:52:24] STAFF MEMBER:
Can you take the microphone?

[00:52:25] ALAN:
Oh, sorry. Yeah. Yeah. Okay. Is that better?

[00:52:30] TIM SCANLON:
K-can– Am I audible now? Maybe I’ll move it over here.

(microphone shuffling)

Good. Thank you. Moral theory is generally understood, I think, or at least by theorists, as, as having two sorts of aims.

Uh, one aim is to clarify the content of morality as we ordinarily understand it. And on the other hand, it aims to explain why these requirements, as we ordinarily understand them, should have the distinctive importance that they seem to have. Morality seems to identify a set of principles, requirements, or standards of behavior that are distinctively significant.

We should not obey them, disobey them, violate them trivially. So why not? What’s so important about them?

So I’ll refer to these two tasks as attempts to answer the question of moral content and to answer the question of normative ground. That is why, why care about these things? Now, to answer the latter question of normative ground, we need to have some general characterization of, of the morality that we’re talking about, right?

So you have to characterize the, the subject in some general way. We have to describe the kind of claims that these moral, various moral requirements

(clears throat)

all express, or you might say characterize what it is that they’re all about.

(clears throat)

And then we need to explain why claims of that kind or claims about that subject matter, whatever it is, if correct, um, have s- the distinctive kind of importance that they claim, uh, to have for us. So, following Kant, we might call a general characterization of what morality is about or what kind of principles there are, uh, a statement of the fundamental principle of morality. It’s the general omnibus description of what kind of claims we’re making when we’re, uh, when we’re making moral claims.

Now, there’s a problem here, and this is my first, uh, gesture in the direction of pluralism. Uh, I think it’s more than a gesture. There, there’s a problem here, I think, about what morality in– that’s the object of this, uh, supposed investigation, uh, should be taken to include.

Because I think the term morality is used by different people to cover really quite different things, and in fact is used probably by each of us, uh, to, to apply to requirements and standards which don’t all have the same kind of normative ground, as they aren’t all important to us for the same, for the same kind of reason. I mean, looking first at the interpersonal case, Uh, some people, uh, I think people who at least write newspaper editorials, um, think that sexual conduct is a central concern of morality. If anything has to do with sex and, uh, it’s, it’s a moral issue.

Others, most people who write in philosophy journals, uh, think that that’s a rather peripheral question if it’s involved, uh, if— If, if it really is an issue, uh, at all. And I don’t think this represents, uh, on the part of these two, uh, parties, a fundamental disagreement about the same subject matter. As they’re talking about the same thing, they disagree about what its content is.

I think, I think it’s pl- more plausible to understand this as two groups of people who are using the term morality in somewhat different ways. They have different ideas about what, what, what the nature of the importance of moral requirements is, what, what they’re, what they’re about, uh, rather than disagreeing about the truth about the same subject that they’re, that they’re talking about. And, and I think there are many other cases of, of this kind.

I won’t, I won’t go through a lot of them, but I, I think that, as you say, the term morality is used in lots, uh, lots of different ways to refer to standards that have many different kinds of, um, claims of importance. So I think that one of the things that theoretical reflection about morality can do, and that is the investigation of this first question of what is the normative ground of morality? Why is it something we should take seriously?

I, I think that kind of investigation can make an important contribution simply by bringing to light differences of this kind. We reflect on what we, uh, think morally, and we discover and we say, “Why should that matter?” And it turns out that we’re, we’re actually not talking about just one thing, but a lot of different things of different, different kinds of significance.

Um, and of course, I don’t think that all these different things are equally worthy of a-adoption and adherence. I’m, I’m not a relativist in that sense. But still, I think that one, one important kind of clarity we should seek is, is to identify different conceptions of morality and try to investigate the different reasons that people might have or that we may think we have, um, to give one or another of these, uh, some distinctive kind of force in deciding how to live.

Now, I started off saying there were two questions that moral theory investigated. Uh, one is the question of what the normative grounds of morality or the various different parts of morality and kinds of morality are, and the other was the question of moral content. Um, now in, in attempting to answer each of these questions, we may face, uh, a, a problem of circularity or a threat of triviality in the answer we give.

In explaining why we should take some particular set of moral requirements as authoritative, we can’t simply say, “Well, they’re authoritative because it would be morally wrong not to take them seriously.” That just sort of repeats the same, the same issue again at the higher, higher level. But on the other hand, uh, an answer to the question of normative ground, why morality is something we should take seriously, can’t be entirely unrelated to the idea of morality that we’re, that we’re talking about.

Uh, it’s got to deliver the kind of significance, uh, that morality seems to have, whatever that kind of significance, uh, is. Second, as to the question of the content of morality, I think it’s natural to suppose that an adequate specification of a fundamental principle of morality, a way of characterizing it for purposes of asking the question why care about it, should also, I think certainly this is a theoretical ambition, should also pri-provide some kind of basis for deciding what the content of morality is. There ought to be a connection between the characterization that explains why I care about it and what we’re gonna– the conclusions we’re gonna be led to if we start taking that, uh, start taking that, um, seriously.

Now, in an extreme form, we might think that this characterization of, of, that we give of morality for purposes of asking the question, of answering the question of why we should care about it, should itself provide a basis from which we can decide all the questions of right or wrong without any further appeal to our moral, uh, intuitions about what’s right and what’s wrong. And I actually think that’s an unrealistic, uh, claim. I don’t–

This is my second kind of pluralism. I, I don’t think, uh, that’s very, that’s, uh, that’s very plausible, uh, either. Um, but one of the, one of the requirements if we’re gonna try to do this, um, is, is, is, is that, you know, one, one of the reasons for thinking we’re gonna have to look to our ideas about the content of morality in order to get from any principle to the conclusions, we can’t just get… go directly, is that the, the intervening steps, the considerations that we appeal to, uh, are going to have to be ones that strike us not simply as following from the principle, uh, the, the fundamental principle we’ve stated.

They also have to be ones that seem more or less significant. So two examples might illustrate this point which have come up in our discussion. In Parfit’s second lecture, he pointed out that insofar as, as Kant’s fundamental principle, uh, his, his, as he, in various forms, he was understanding it, is taken to determine the content of morality via the notion of a contradiction.

So, suppose we’re thinking here of the universal law form of the categorical imperative, insofar as the universal law form is supposed to determine the, the, the answers to questions of right and wrong, because we’re gonna ask whether, uh, uh, the maxim of that action would be one that could be conceived, uh, without contradiction in, in, in various ways Parfit considered. And one of the things that Parfit brought out in that, in that lecture was that although there may be some cases where this might, would, might give the right answer, it doesn’t seem to give the right answer for the right reason. The fact that your maxim might be contradictory doesn’t seem to be importantly connected with, with what would actually make it, make it wrong.

So it seems like we, we don’t want a principle, so to speak, that leads to conclusions, uh, about what’s right and wrong without going via intervening steps that have independent, it seemed to have indep– seemed to us independently to have moral significance. Similarly, picking up a point from the third lecture, the fact, if it were a fact, that it would be rational for someone in Rawls’ original position to reason as if he or she had an equal chance of being in any social position wouldn’t seem to me to show that social institutions that maximize average utility are morally justified, no matter how that utility is distributed. Rather, what it would show is that a thought experiment that simply involves people gambling, uh, is, isn’t very morally relevant.

So you, you, you don’t, you don’t, you, you don’t want a theory, again, a theoretical structure that just leads to conclusions. You want a theory that leads to conclusions through intervening steps that match what seems independently to be morally, um, morally relevant. Now, Kant’s various specifications of the fundamental principle of morality, um, have attracted enormous attention over the years and have attracted our attention in this discussion.

Um, in part because they have the merit of seeming to deal particularly well with the first of the problems I mentioned, that is the circularity problem as it faces an explanation of, uh, of the, um, normative significance of morality. Why sh- why we should care about it. Um, he locates the normative authority of morality in a view we must take of ourselves insofar as we think of ourselves as agents at all.

He thinks these various forms of the principle are ones that we’ve got to, we’ve got to accept for that, for that reason. Um, and insofar as this is the ground that he finds, the– insofar as he finds this is the reason why we should take, uh, morality, uh, so seriously, Um, he’s found a foundation for it, um, which is free of any charge of begging the question simply by repeating a moral idea, by saying, “Well, you gotta do it because it would be morally wrong,” right? You do, you gotta do it because otherwise you won’t be acting.

So when you’re thinking of yourself as acting, you have to think of yourself as acting, uh, according to this, uh, formula. This is clearest perhaps in the argument Kant offers for the universal law form, for example, in the third section of the groundwork, but I think it applies as well to the various independent or seemingly independent arguments he offers for his formula of, of humanity. So on the one hand, he seems to have provided, uh, his answer to the question of the ground of, of, of these fundamental principles in a way that seems, doesn’t seem to be open to a charge that, that it’s not morally that– sorry, that it’s, that it’s morally question-begging because there’s already moral content built in.

But on the other hand, the formulae themselves, particularly the formula of humanity and the formula of the realm of ends, but also the formula of universal law insofar as it’s read in a way that connects it with these other formulae, are formulae which seem to it– themselves express ideas that, ideas about what it is to respect other people and what it is to be to form part of a moral community with other people, which have very deep moral resonance. So these principles are– have, have this great place in the history of the subject, I think, because Kant both was giving what seemed like a non-question-begging answer to why we should care about them, but the things that he was identifying as having this non-question-begging answer were things that seemed morally relevant. They seemed, they seemed to tie up with our deep moral feelings, and this seems a wonderful achievement, uh, uh, on Kant’s, uh, part.

Marrying the abstract idea of rational agency with substantively appealing moral ideas in this way was a brilliant, a brilliant move. Uh, unfortunately, it seems to me it doesn’t work. Uh, his attempt to show that insofar as we think of ourselves as agents at all, we must accept these moral ideals as fundamental constraints, uh, on our practical reasoning is brilliant, daring, a monument of philosophical ambition and ingenuity, and entirely unconvincing.

But still, it’s, it’s a great thing because it, it, it seems to do both of these things. That is, to derive from non-moral starting points something that seems to have great moral content. Now, one response to this failure would be to cleave to the appealing ideas of community that are expressed in Kant’s formulae, and to take them as themselves offering an explanation of, of morality’s normative basis.

So this is to sail closer to the rock of circularity. Say, why should you care about morality? Well, because it’s only in thinking in the terms that morality describes that you will be treating others as ends in themselves or tr– or, or, or behaving to them in the way, uh, uh, that’s part of a, of a realm of ends.

And those things are just im- appealing on their own f- on, on the face of it. And I think modern contractualism, in the form that I espouse, goes in that direction. It, it tries to, it tries to give, it, it gives up a little bit by way of our worry about circularity, and tries to identify at least part of what we think of as morality as being important for us because it’s a way of having the right kind of, of relations with, with other people.

Now, you can debate whether that kind of contractualism is really Kantian, having given up, uh, this fundamental aspiration of Kant’s, but I’ve already t- we already talked about that last time. So some of what I’ve said so far might be summed up as follows, in terms of a variety of answers that we might give to the question of normative ground.

In particular, to the answer we might give to the question, why should we take what other people have reason to want or what they might have reason to agree to, uh, into account in our own decisions about what to do? I’m going to consider four possible answers to this question. First, we might have reason to give what other people have reason to want or what they have reason to accept weight in our own decisions simply because things go better for these other people.

Things go better in the world when these other people’s lives go better. That is, we have reason to promote states of affairs in which their interests are better fulfilled. Their interests are, as Parfit might say, intrinsically reading, reason-providing for us.

(clears throat)

So this you might think of as a purely teleological answer to the question of normative ground. We, we want to take their interests seriously because we have reason to make the world go better, and that makes the world go better. This brings up a question that emerged in yesterday’s discussion.

I had suggested in my comments after the lecture yesterday that what Parfit calls the impartial point of view was defined in his lecture merely by subtraction. F- that is, we define the impartial view by starting with all the reasons that a person has and removing those reasons that are personally based, that the re– the reasons a person has only because of the way he or she is affected, or only because of the way his or her children are affected, or only because of the way his or her football team is affected, or whatever it might be. Parfit replied, as I understood him, that taking the impartial view also involved adding reasons that wouldn’t, that one wouldn’t have simply from one’s own personal point of view.

At least this , I took him to be making this suggestion, and he cited here Thomas Nagel’s description of the move from a personal or subjective to an impersonal or objective point of view, where we look at things, setting aside our own thing. And then on, on that view, it seems that when we move from the personal point of view to the objective one, new reasons come into view, namely the reasons that, that we have in virtue of other people’s pains or other people’s aspirations that we didn’t have when we were looking at things merely from a personal point of view.

Now, perhaps I simply misunderstood Parfit on this point, um, but I was led to my remarks about, uh, arriving at the impartial point of view by subtraction, um, by r- by my reliance on his remarks that, for example, someone else’s pain or, or pain in general is not only something that the person who is in pain has reason to avoid, um, it’s something that others have reason to avoid or prevent if they can, uh, as well. And I took this to be a, a perfectly general claim he was making about the reasons we have, not a special claim about the reasons we have if we take a certain special, uh, point of view. If we were to begin from a desire-based or a self-interest-based conception of the reasons people have, then obviously some move, such as a move to an impersonal, impersonal outlook, would be required in order for us to come to have reasons to prevent other people’s pain.

But Parfit was rejecting, uh, self-interest-based and desire-based conceptions of what reasons we had and starting off with a much broader, um, a broader conception. And it seems to me that Nagel, Nagel’s idea of the personal point of view is a rather narrow conception of what reasons we have, and that’s why the– for him, the move to the impartial point of view adds reasons. Whereas it wasn’t clear to me, although I’m– I don’t think in, you know, I’m not here interested in fighting for one side or other of this, but it seems to me that given the broader conception of reasons that Parfit started with, um, uh, reasons to prevent other people’s pain are already included, so to speak, in the basic package of reasons that we get if we just sign up.

You don’t have to buy extra channels

(laughter)

That’s it. You don’t have to move to the impartial supplement, to get the impartial supplement in order to get those reasons that come with the package. Okay, so that was a bit of a digression or reference the last time, following on my discussion of one way in which we might have reason to take into account in our decisions about what to do, what other people have reasons to, namely because what their reasons just are also reasons for us, right?

No contractualism there. That’s just straight objective to objectivity of reasons. So let me– but now let me consider three other answers.

We might have reason to take into account what other people have reason to want or what they’ve had reason to agree to, because we will be better able to pursue our own aims if we secure the cooperation of others. But no system of cooperation would be stable unless other people have a reason to accept it. And this means that any such system of cooperation must require us to take other people’s interests into account to some extent in deciding what to do.

This, crudely put, is David Gauthier’s strategy at which Parfit mentioned. It might be seen as attempting to found morality on self-interest. Although I think this is a little bit of a misnomer, since all that’s assumed is that we have reason to accept what will advance our aims, and it isn’t necessarily true.

It needn’t be assumed that these aims are purely, uh, self-interested. Okay, that’s the first quasi-contractualist, uh, reason. We need to, we need to take other people’s reasons in, reasons into account because we need their cooperation, and we need it in a stable form, and we’ve got to buy them off in order to get them to go along.

Second possibility, um, we might have reason to be concerned with what others have reason to want, because if we don’t think in this way, we will be failing to understand correctly our own status as rational agents. If we understand what it is for us to be rational agents and for our reasons to count for us in the way that they do, we, we have to see that, that applies to other people a-as well as it does to us. This is Scanlon’s position as I understand it.

Third, we may be said to have reason to be concerned with the justifiability of our action to others because otherwise our relations with them will be based, as Rawls says in his early writings, simply on force and circumstance. They will be asked to accept the way we treat them because they have no choice other than to accept it. And we won’t, we will be asking them to live with us on those terms.

This is an unattractive way to live with them, you might say. We want to live with them on better terms. We want to live with them on terms of equality or terms that, that acknowledge their, their standing.

So this is contractualism as I would understand it, and I think as Rawls describes it in his earlier articles and in A Theory of Justice. Now, any of the last three alternatives I’ve described, that is the Gauthier version, the Kant version, and my, my, my last Rawls, Rawlsian or Scanlonian, uh, contractualism might be called contractualism. Although the label seems to me not so, uh, appropriate, uh, for– not so clearly appropriate for Kant’s alternative because it doesn’t start from an idea of the value of others.

Rather it deduce, it deduces that from the requirements of practical rationality. The cartographical conclusion that Parfit reached at the end of his third lecture might be put as follows, in terms of the two questions that I’ve been considering. A moral theory might be called consequentialist because of the answers that it gives to the question of normative ground.

That is, namely, it gave the first answer, that we should, we should take into account the interests of others because those interests count toward making the world better. Um, or sec-second, a theory might be called consequentialist because of the answer to the question, answer it gives to the question of moral content. Namely, it says the way we ought to decide what to do is to ask which consequences, uh, would be better.

And it might seem that a view that is consequentialist in the first of these senses, that is, it takes promotion of the best of consequences to be what morality is all about and what gives it its rationale, point, and claim on our attention, would provide the best foundation for a consequentialist account of the con-content of moral requirements. But in fact, I take this to be the claim, Parfit’s cartographical claim at the end of the third lecture. But in fact, a contractualist answer to the question of normative authority need not exclude a consequentialist answer to– a consequentialist answer to the question of moral content.

We start off with contractualism of the Kantian variety that he describes. That can perfectly well lead to consequentialism as an answer about what, what to do. Um, and, uh, this is what, this is what he argues, uh, is, uh, is, is, is likely, uh, to happen.

This is what, this is what he argues, uh, is, uh, is, is, is, is, is likely, uh, to happen in that particular version of con-contractualism. Now, it remains true that a teleological answer to the question of normative authority might make consequentialism about moral content more inescapable. It’s hard to see once you start off with the idea of promoting the best state of affairs is what Rawls is about, You’re never gonna get any other answer to its content.

But, um, and it may be if you start off with a contractual starting point about the rationale for morality, it, it, it may not necessarily lead to a consequentialist, uh, conclusion. But I think Parfit’s idea is that if it does so, it does– it provides consequentialism with a more secure and less controversial, uh, perhaps, uh, starting point. So the, the cartological– cartographical question had to do with the idea that things could be called contractualist or consequentialist because of their answer to the question of why care about morality or because of their answer to the question, what’s the content of morality?

And the two don’t necessarily march together. Now let me conclude here. I think I’m doing all right.

Uh, a couple minutes. All right. Uh, let me just state, say quickly my other, my other, um… I have two pages.

My other, um, uh, pluralist point about the question of moral content. How and to what degree can a characterization of the nature of morality provide us with a basis for an answer to the question of, of its, of its content? If we, if we, if we characterize a fundamental principle for the purposes of explaining why we care about morality, should that also tell us of exactly what we ought to do?

Now, it– I think not. It may be reasonable to expect there to be a single fundamental principle of morality in the sense of a single characterization of what that particular kind of morality is about, which explains its normative authority. But how reasonable is it to expect there to be a single principle from which all of the content of morality could be derived without intervening appeals to moral judgment?

Utilitarianism famously offers such a principle, and some may regard this as one of its attractions. But, but this monistic character is, from another point of view that I would take, a source of its great implausibility. That’s not what all of…

You can’t get all the answers out of that one, out of that one way of thinking. Now, Parfit’s version of consequentialism, because it allows a much wider range of reasons to count toward determining the best states of affairs, these aren’t all about well-being, yields a doctrine that’s much more able to accom-accommodate the apparent diversity of moral phenomena. It still offers a fundamental principle, but applying that principle now requires a great deal of normative judgment about the relative importance of all these diverse kinds, uh, of, of, of reasons, not just a comparative judgment of what will produce the most happiness.

So consider this example. In my comments yesterday, I suggested that in some cases, the best state of affairs might be produced by following a principle that required great sacrifices by a few people for the sake of very small gains to a great many others. And I said that one of the few– that if one of the few would have sufficient reason to make such a sacrifice, but also sufficient reason to decline to do so, it seemed questionable whether this sacrifice is one that we could say was morally required.

But it might be replied that I was here being unfair to Parfit, uh, or unfair to this kind of consequentialism, since I failed to take into account the diversity of considerations that this kind of consequentialism includes. In particular, the state of affairs resulting from a principle of the kind I mentioned, even if it would involve a greater sum of well-being, might be objectionable on distribu-distributive grounds. It was unfair, or was it, was it, it produced a bad distribution to take a lot from a few and spread the benefits, uh, over, uh, over many.

This reply seems to me to have force, and I think I should have taken it account, into account better in what I said yesterday. Distributive considerations may indeed rule out the most threatening principles of the kind I was referring to. But to decide whether a principle would fail to yield the best state of affairs because of the inequalities it would involve, even though it produced a greater sum of well-being, we would need to decide how object– how objectionable are these particular inequalities, and we would need to balance that against the gains in aggregate well-being.

This brings us back to the question which we considered yesterday of whether the reasons we have to avoid an inequality are or are not at base moral reasons. I suggested yesterday that there are several grounds for objecting to inequality that do not seem to me to be essentially moral. For example, the disadvantage of losing one’s liberty and the costs of stigmatization.

But distributive reasons of the kind one would need to appeal to in order to resist a principle of the kind I’ve just been discussing, a principle that imposed large costs on some in order to bring small benefits to others, might not be of either of these types. It might not be either loss of liberty or, or a stigmatization, but simply inequality and benefit that would be the, that would be the, um, objection. And it isn’t, isn’t clear to me that that, that objection wouldn’t basically be a moral, uh, an objection of moral character.

So this would raise the question then of how much difference there would be from the point of view of the charge of circularity between consequentialism of the kind that brought in distributive considerations of this kind and asked us to bene- weigh these moral objections to inequality against other benefits. How much difference would there be between that kind of consequentialism and a kind of c-contractualism that involved a me– an admittedly moral idea of whether it was reasonable for someone to reject or wasn’t reasonable for someone to reject the demand that he or she shoulder shoulder this burden? From the point of view of underlying moral content, it didn’t seem to me that there was much to choose between the two.

[01:19:10] DEREK PARFIT:
Over here. Uh, thanks very much. Um, I…

There’s rather little I have to disagree with there. Perhaps I should say that… No, I wouldn’t say that.

First, I think you were perhaps understated Kant’s account of the reason to care about morality. When you say, “If we don’t do so,” we’ll be failing to understand correctly our own status as rational agents.” I don’t think you’d think that it’s your failure to understand your status that gives you the reason.

That just seems much too slender.

[01:19:49] TIM SCANLON:
Um, I feel-

[01:19:51] DEREK PARFIT:
Now, you may-

[01:19:52] TIM SCANLON:
If you fail to recognize it as a reason, you will be showing that you haven’t taken on board.

[01:19:55] DEREK PARFIT:
Yes, but, but that, you should. But that isn’t why you should do it, because if you don’t act rightly, you’ll be failing to understand your own status. That just seems to me-

[01:20:04] TIM SCANLON:
No, no. So choose one.

[01:20:06] DEREK PARFIT:
Now, you made some excellent points about the impartial point of view, and, um, I’m inclined wholly to agree with the claim that even from your personal point of view, you have reasons to care about all these things that we have reasons to care about. So I think what I might say here is this, that what happens when you move from an impartial point of view, that’s a question of the strength of these various reasons. I think on the view that Tom Nagel and I hold, and I think Sidgwick held, I have some reason to care about everyone’s well-being, but I have much stronger reason to care about my own and the well-being of those I love.

And so, when Sidgwick thought the extreme version of this,

(background chatter)

from a personal point of view, you have reason to do what’s best for you or for those you love, but from an impartial point of view, you have reason to do what’s best, all things considered. That’s a question of, of the strength of the reasons. From the impartial point of view, each person’s well-being matters equally.

Mine matters no more than anyone else’s. Now, as Sam very well argued in his first book, you can’t live entirely from the impartial point of view. You have to recognize that we must and can justifiably care much more about what happens to us than those we don’t.

So is that an acceptable amendment, that the reasons of the second, whether the point of view is personal or not, is a matter of their strength?

[01:21:42] TIM SCANLON:
Well, I would say that the change in strength is achieved by subtraction. That is, you subtract the personal reasons you have to care about, about your reason, and what you’re left is their impartial strength. So that I don’t, I don’t see a difference between that and my point about subtraction.

[01:21:56] DEREK PARFIT:
Um, well, you subtract the strength by saying I’m to ignore the fact that it’s my well-being, that I’m the person who will die. Well, the natural way to express that, that you’re ignoring that fact is you’re thinking of it from an impartial point of view.

[01:22:18] TIM SCANLON:
Well, I was defining the… I suggest that you define the impartial point of view by starting with all the reasons you’ve got and, and factoring those out. Now, that’s what– what I’d offered was this definition of what you meant by the impartial point of view.

This is, I think, it’s perhaps becoming-

[01:22:31] DEREK PARFIT:
Yeah. Okay.

[01:22:32] TIM SCANLON:
Academic in a not altogether favorable sense.

[01:22:34] DEREK PARFIT:
Okay. Now, uh, the main questions are questions at the end about, um, consequentialism and contractualism. Uh, there’s a huge difference which I alluded to but didn’t say much about between principles that apply directly to acts and principles that apply only indirectly because they apply to the disposition to act in certain ways or the principles that we apply to act or the rules.

Now, um, act consequentialism is monistic. Uh, there’s just this one principle that tells you how every act is to be assessed. Rule consequentialism isn’t monistic, and it has a structure actually that’s pretty similar to contractualism and to some forms of Kant’s view.

Um, In each case, you ought to act on the principles whose universal acceptance everyone could rationally will, or on the principles that no one could reasonably reject, or on the principles whose general acceptance would on the whole make things go best. So there’s no difference with respect to the degree of moni-mononissity there.

[01:23:46] TIM SCANLON:
And indeed, as

[01:23:47] DEREK PARFIT:
you said and I said, a lot of these can come together. You could plausibly think that you ought to act on the principles whose general acceptance would make things go better, because those are the principles that everyone could rationally will to be universal, and those are the principles that no one could reasonably reject. Um, so those three views are, as far as I think, exactly the same with respect to the monistic pluralistic.

They have one single higher level criterion for which are the relevant principles, but then the principles can be as complicated as common sense morality.

[01:24:25] TIM SCANLON:
Now, you said some other things that we could discuss, like for example, perhaps I’ll just discuss one, but then I think we might, uh, open it more widely. Um, when you talk about the distributive reasons, I think there are many things going on, because at one point you said, uh, one might resist the principle of the kind I’ve been considering that imposed large costs on some in order to bring very small gains to a number of others. Well, I think the question whether very trivial benefits and burdens can outweigh major ones on a few people isn’t itself a distributive question, and there’s an independent reason for thinking that some gains are too trivial to count when other things are at stake.

Uh, but that’s a nitpicky point of detail. So, Sam, uh-

[01:25:15] SAMUEL SCHEFFLER:
Okay. Yeah. Okay.

[01:25:19] TIM SCANLON:
Um- Oh, no, I didn’t want to end with… just, just one, one question to you.

Um, I’d like to ask you whether you thought that the main argument that I end lecture three with and that you mainly discussed, uh, that’s the argument for the view that Kant’s contractualist formula leads to rule consequentialism. Uh, whether that argument is first valid, and then whether the second and third premises are true. The first premise is Kant’s contractualist formula.

My claim is that leads to rule consequentialism.

[01:25:57] DEREK PARFIT:
Uh, and if you thought it’s valid and those two premises are true, so that it’s relevantly sound, um, then I’d be pleased

(laughter)

.

[01:26:04] TIM SCANLON:
But I, I conceded and I agreed that I, which what I thought your argument showed, not that it led to, uh, rule consequentialism, but that it didn’t exclude it. That’s one of the things that it could lead to, since it’s a permissive. Yeah.

I was reading it as a permissive, uh, uh, principle. So yes. Yeah.

In that sense, I agree, yes.

[01:26:22] DEREK PARFIT:
To save five other people. Now, that’s wholly compatible with thinking that you have many other personal reasons, reasons to sacrifice your child’s life, and save your child, reason to help the projects you’re committed to, and so on. So I think at that point, the fact that I was taking simple cases, my life or the lives of five, I didn’t think affected the argument.

Now, with respect to premise B, in a way, I’d quite like to turn to that because it bears on, uh, Tim’s point that you think the argument at least permits, uh, the rule-consequentialist conclusions rather than requiring it. Um, firstly, B may look like the claim that we could always rationally choose whatever we would have most reason to choose from an impartial point of view. And what I meant to be saying, firstly, is that I’m inclined to think that’s so, but it’s not required by that argument because that premise is about what someone could rationally choose in the thought experiment to which this contractualist formula appeals.

She’s supposing that she can choose which of the principles that everyone accepts, and therefore that most people act on. Okay. Um, and in that context, um, the point of my referendum examples are that as long as she thinks she has some reason to care about the well-being of each other person, even if it’s only one millionth as much as her reasons to care about her own well-being, it’s fairly plausible to think that she ought to choose the principles that would, on the whole, make things go best.

And that’s because it’s like the case in which, instead of my giving ten thousand to the poorest people in the world, my giving up ten thousand, which is only one tenth of my income, let’s say, could bring it about that a hundred million benefits went to the poorest people. So that’s the point. It’s that in that argument, if you imagine that you’re affecting the whole of human history, well then if you compare the principles that would make things go best and principles that are significantly different because they’d make things go quite a lot worse, well then my claim just was, even if your reasons to care about the wellbeing of others were extremely small, the scale is going to-

[01:28:57] TIM SCANLON:
But doesn’t it, I mean, but- but how can you feel– I mean, why should we feel confident of that conclusion without attending at all to the variety of other kinds of reasons that you’re officially committed to allowing on the wide view?

[01:29:12] DEREK PARFIT:
Oh, I can include, I can include all that.

[01:29:13] TIM SCANLON:
That is, I mean, suppose her life is bound up with the life of a certain community or tradition that makes sense to her of her aspirations and so on, and that although principles that– and principles that would make things go best in your special sense, um, would have, um, wouldn’t make things go well for that community or that tradition, say. Um, is it unimaginable that there could be someone for whom it wouldn’t be– wouldn’t actually be rational to agree to those best principle– principles that are best in that very special sense? Or at any rate, don’t we need to think a little bit about the force of the actual other reasons that might be in play?

Do you wanna get it all out of these questions about size of contributions and so on?

[01:29:59] DEREK PARFIT:
Well, look, I, I agree that we do. But if you wait, I’m making two points. One is, and I think this would be quite widely accepted by non-academics, but perhaps I’m wrong. I think many non-academics, uh, I mean, it’s probably correct to say that a quarter of non-academics are rational egoists.

[01:30:22] TIM SCANLON:
Right.

[01:30:23] DEREK PARFIT:
Um, but I think many non-

[01:30:25] TIM SCANLON:
No, no, academic.

[01:30:26] DEREK PARFIT:
-non-academics would think that you could rationally sacrifice your life to save the lives of five others. You have sufficient reason to do that. Now, that is an enormous sacrifice. And if they think you could rationally give up your life to save five other people,

[01:30:47] TIM SCANLON:
Okay.

[01:30:48] DEREK PARFIT:
And what we’re now asking, you could rationally choose principles that would be somewhat worse for you and those you love, perhaps. But that isn’t to save five other people. It’s to make the whole history of mankind go enormously better.

[01:31:04] TIM SCANLON:
Yeah, but you’re putting it all in terms of what would be best for you and what would be a sacrifice for you. I mean, maybe what you’re doing—

[01:31:09] DEREK PARFIT:
No, no, no. I’m including you and your children and those you love.

[01:31:12] TIM SCANLON:
In a certain way you are. You’re including their welfare. But I mean, suppose that as I see it—

[01:31:16] DEREK PARFIT:
Mm-hmm.

[01:31:17] TIM SCANLON:
Um, what it would be doing would be consigning the, my community to a dismal future or acquiescing in their sort of dying out or the fading away of a tradition with which, uh, my life is deeply bound up. Um, so I might say I’d be perfectly happy to sacrifice my life to save millions of people, but it doesn’t follow that I’m willing to let my tradition or my community die out for the sake of what’s best in this very special sense that you have defined. I mean, well, I need to— maybe that’s right, and maybe that’s wrong.

I need— but we somehow want to get to the conclusion without even looking at any of these other reasons.

[01:31:50] DEREK PARFIT:
No, no, no, no. Look, look, I’m prepared to look at it, but if somebody says, my reason to care that my community continues is actually much greater than my reason to care about my own life and the lives of my children and those I love, then I think that person is very likely to be making a huge mistake. Um, there was some, uh, fellows of male-only Oxford colleges who cared enormously about preserving male-only colleges.

Um, and one of them made this eloquent speech that, you know, we go to great lengths to preserve Victorian architecture, even if we think that it’s ugly, because we think it’s important to preserve these distinctive units. So why not male-only colleges? Well, I think that has some weight, but I don’t think it has more weight than the lives of himself, his children, and those he loves.

Um, so I’m not meaning to exclude those, but, uh, I’d be surprised if bringing in the survival of my community is going to make enough difference unless Unless you’re introducing moral beliefs of a different kind, something like that. Clearly, if my community is, is the monastic community in the order which is the one,

[01:33:05] TIM SCANLON:
just be bringing in other values. I mean, it’s supposed to be a value-based view. Yeah. And if, and if the, if the view is that, um, it would, um, it would undermine values that I care deeply about.

[01:33:20] DEREK PARFIT:
Yeah, yeah. I mean, you might get an example that, um, someone who cares greatly about the natural wilderness, something like that. Now, you might say, I mean, it would make the whole history of humanity go enormously better, but, you know, the, the natural wildernesses would be interfered with, and I think that outweighs.

I don’t want to regard it with, um-

[01:33:45] TIM SCANLON:
The question is whether someone’s life couldn’t be devoted to a cause or a tradition or a value in such a way that for them it wouldn’t be rational to consent to the principles that would make things go best in your sense.

[01:33:58] DEREK PARFIT:
Well, I think you’re there appealing to a version of the desire based on view. You’re appealing to the strength of that person’s commitment to this project.

[01:34:09] TIM SCANLON:
Why does it have to be to the strength of their desire as opposed to the role of the value in structuring their sense of the, of their own life and its significance? I don’t see why it has to be strength of desire. I mean, if you really take the value-based view seriously, then it seems to me you really have to look at the role that these different values can play in people’s lives

[01:34:28] DEREK PARFIT:
Right, but, but-

[01:34:29] TIM SCANLON:
and the way they can give rise to reasoning.

[01:34:30] DEREK PARFIT:
Right. But the point is it’s got to be a value that they can plausibly give greater weight to than their own death at the age of twenty. I mean, if somebody could– if, if a twenty-year-old could rationally give half our life to save five people, let alone a million people, then how is this project that gives meaning to this person?

[01:34:55] TIM SCANLON:
Well, some twenty-year-olds may. I, I, I mean, well, okay.

[01:34:59] DEREK PARFIT:
Well, I don’t think there are many twenty-year-olds who don’t have sufficient reason, but okay.

[01:35:05] TIM SCANLON:
Well, that’s somebody else. Yeah. I have a question about, uh, a-about picking up on this pluralism stuff, and maybe it’s just a request to hear more about what you think is at stake in, um, in engaging with questions about the normative structure

(cough)

of rightness and wrongness. Yeah. But, um, but you’ve got the, um, the wide value-based theory of reasons, which I’m sympathetic to in some sense, which involves lots of what one in a, you know, uh, earlier idiom might have referred to as fine-grained intuitioni- intuitionistic judgments about plur– you know, a diverse range of values and, uh, hard to quantify, uh, attempts to weigh these different values and think about what would be appropriate responses to them and all of that.

A lot of which subsumes some of the, um, traditional territory or coincides in various ways with some of the traditional territory of moral reflection in some ways. Um, one, one natural thought is that if you’re, if you’re prepared to go so far, why not just c-be pluralistic, as it were, all the way down and just say there’s the landscape of, uh, the, the tremendously diverse landscape of reasons and sometimes we talk about things as being right or wrong or morally right or wrong, but that’s just a summary way of talking about the balance of reasons and the actions that would be, you know, appropriate, uh, responses to this, this, um, incredibly multifarious, i-interesting, um, uh, ramified landscape of reasons. Um, we– And, and, you know, as Tim was saying, r-classifying responses as moral responses, uh, it’s not something that we have a r– that, you know, that forms a natural kind anyway in response, in, in terms of, um, describing a set of human responses or for that matter, a set of reasons.

So may, you know, maybe the– in the broadest sense, uh, in some sense, uh, questions of moral rightness or wrongness or questions of how appropriately to respond to the complicated landscape of reasons in situations where, you know, human well-being is at stake or something like that. In other cases, it might just be a, uh, you know, a question of how appropriately to respond to values. But, um, if you were to think about it that way, there would be no particular importance that would attach to giving a theoretical account of some subset of reasons that are the reasons, uh, constituted by, uh, moral rightness or wrongness itself.

These considerations would be, as it were, transparent from the standpoint of practical reflection, uh, and would, um, you know, the real reasons would be the, the concrete, um,

(coughs)

considerations that, uh, you would be correctly responding to when you’re properly weighing the diverse range of considerations and such. Uh, you know, um, one response to that might be that you need to give an– as Tim was saying, you need to give a unified account of the morality of right and wrong in some sense to make sense of, uh, the, the claim– the claims of morality on us, um, you know, uh, to be a s- a set of reasons that we ought to respond to in some sense. Yeah.

It wasn’t clear to me that that was an important concern of yours, at least in these lectures. And it seemed to me in some sense that, uh, uh, you know, the, the, the reason-giving considerations were all being done, uh, provided by the background, uh, wide value-based theory. Uh, but if you do think that it’s important because of its, uh, to give a unified account of the normative structure of morality, in part to illuminate the normative significance of a certain subset of reasons, uh, I’d, I’d like to hear more about, you know, how you think that story would go.

[01:38:31] DEREK PARFIT:
Well, that, that really does ask the central questions of this entire branch of philosophy. Uh, I… I am inclined to think that it would be a very bad mistake to replace our thinking about right and wrong as our thinking about what we have most reason to do, all things considered.

I think right and wrong are highly distinctive and separate and should be looked at, um, on its own, on its own terms. Although in the end, we need to return to the questions, what we have most reason to do, all things considered. Now, I think there’s been great oversimplification with respect to the question about wrongness.

I mean, to put it in its very simplest, I think the position is this. Um, some people believe that the concept of moral wrongness, as Tim and I believe about the concept of the reason, cannot be explained in other terms. Uh, you just mustn’t do it.

[01:39:35] TIM SCANLON:
Okay.

[01:39:37] DEREK PARFIT:
There are also various senses we can give to the word wrong, which give us distinctive ways of understanding morality. I think one that’s very attractive, which I take from Scanlon’s book, um, an act is wrong in what I call the moral reason-involving sense if it’s an act that, this is quoting from Scanlon, that violates standards of conduct that we all have strong reason to accept and follow, and an act that gives the agent reasons to feel guilty and gives others reasons for indignation and resentment. Now, there’s also a justifiabilist sense.

An act is wrong if it couldn’t be justified to others. There are many reasons involving analyzable senses of wrong, and I think actually when you look at people, they’re not talking about what’s wrong in the same sense. That doesn’t mean that they can’t be disagreeing, but it means that when you look at the disagreement, you have to take into account the different senses in which they’re using it.

Um, uh, so I mean, I’ve also had the thought– I was very struck in Scanlon’s original article that you can think that the well-being of others, in a way, matters as much as yours, and that you have reason from your point of view to relieve the suffering of others. But the thought that it would be wrong not to seems sufficiently distinct from the thought that as an impar– you know, from an impartial point of view, you have reason to do it.

Um, and I think, uh, Scanlon’s quite right to think that that is a distinctive thought, and one of the most compelling ways of bringing out what’s distinctive to it is by appealing to justifiability to others, which is closely connected to reasons to feel guilty and reasons for others to feel indignation and resentment. These ideas are all rather similar, and I certainly think it’s a great mistake to try and ram all these thoughts into one package. What do I have most reason to do, all things considered?

On the other hand, I do think that that’s the last question. That the people who say no, and Kant would have said this, but then he would have gone on, except by the answer. The last question is, you mustn’t do it.

Okay, you, it’s just wrong. You mustn’t do it. Now, the question is,

[01:41:56] TIM SCANLON:
well, does the fact that I mustn’t do it give me a reason not to do it? Now, if I said, oh no, it doesn’t give you any reason not to do it, it’s just that you mustn’t do it, well, that would be disaster. Um, now there’s a similar question.

It does give me a reason to do it, but is it what I have most reason to do? And he says, well, no, actually, you have most reason to act wrongly. I think that’s the ultimate question, but I don’t think we should bypass moral thinking to get to the question about what I have most reason to do.

I think we should let morality be fully displayed, work out the implications, and then return to the ultimate question.

[01:42:37] SPEAKER 1:
Very, but that still leaves somewhat obscure to me then your particular enterprise here, um, because you, um, display the ways in which Kant’s various formulations don’t capture what’s right and what’s wrong. And you, you have the, the rule consequentialism, which, well, what’s obscure is to what extent you are actually trying to do the thing that Tim was talking about, establishing the content of morality, finding something that will tell you what’s right and what’s wrong, and presumably that’s the role of your moral consequentialism.

So why, this connects also with Jay’s question, Why, um, why don’t we see right through those apparently moral considerations that you say we should keep alive, uh, to the, um, consequentialist, the rule consequentialist equivalents that, as it seems to me, your enterprise is supposed to be establishing? Or is that not your enterprise?

[01:43:51] DEREK PARFIT:
Well, no, my enterprise, um, was in a way more theoretical. I mean, I said on my picture of the moral map, you have common sense pluralistic morality, then you have these systematic theories. Now, I’ve just been struck by the fact that as far as I can see, there’s a kind of strange mismatch because Sidgwick, the act-utilitarian, thought about morality in the way that best supports common sense pluralist morality, deontological restrictions, you know.

It’s just you mustn’t kill someone as a means. Then we have Kant’s principles and Kantian contractualism, And I was just struck by the fact that, but as far as I can see, that most strongly supports consequentialism of a rule-consequentialist kind. I wasn’t, as I said at the end, arguing for rule consequentialism.

Um, I was trying to see where I thought these different methods of thinking about morality lead. So in that sense, it was cartographical or theoretical. Um, does that leave you with a puzzle?

[01:44:58] TIM SCANLON:
Yes. Where are you on the map? Uh, that is where, what– Where now do you put yourself on the map as a result of this enterprise? I mean, that’s my… That’s what’s obscure to me. Um, now you, um, you’re saying you’re not on it. You’re just pointing out, holding up the map for the rest of us.

[01:45:17] DEREK PARFIT:
Well, when you say, where am I on it? I can use the word wrong in several different senses, and I’ll tell you what I think to be wrong in this sense or that sense and so on. And then if you ask me questions about what I think I have most reason to do, I can start giving you answers to that.

But I don’t think I should come out and say, “Right,” what I am is an ideal rule consequentialist of the following kind.” I mean, I-

[01:45:43] TIM SCANLON:
Uh-huh.

[01:45:44] DEREK PARFIT:
I think that’s-

[01:45:44] TIM SCANLON:
Why not?

[01:45:47] DEREK PARFIT:
Because I think the truth is more complicated than that. Uh, I mean, my position is somewhat more like, I think, that of a physicist who knows that there are various competing theories about how, you know, how you explain this phenomenon. He, he alters his view from time to time about the plausibility of this and how well it’s working out.

But, um, I think that’s how we ought to regard these things. Um, now, that may seem a bit detached and clinical, but it’s, as it were, the most striking difference with Sidgwick as compared to all of his predecessors. For the first time, he was viewing it in what many people would regard as this academic, dry way, and that’s part of the reason why it suddenly became very boring.

Whereas Kant is gripping, okay?

[01:46:31] AUDIENCE MEMBER:
Oh,

[01:46:32] DEREK PARFIT:
but I, I think it’s– I think this subject is in very early days, and I think we need to do a great deal of just kind of clearing up what the views are, what the concepts are, how they support what they’re related to one another. And to say, “Oh, but where do you think we ought to end up?

What do you think the ultimate truth is?” It’s too soon. Too soon for me.

[01:46:55] TIM SCANLON:
But you said, uh, you, I’ve heard you say a number of different things of– One is this idea that it’s too s- in, it, that might count as answers to Barry. One is that it’s too soon to say. Another is that the truth, you think the truth is more complicated in ways that I took it-

[01:47:10] DEREK PARFIT:
Yeah.

[01:47:11] TIM SCANLON:
You could in principle articulate. And the third was that you Re- You had some reservations about systematic theories at all, and some attraction to sort of-

[01:47:19] DEREK PARFIT:
Well- -common sense pluralism.

[01:47:20] TIM SCANLON:
And I wonder-

[01:47:21] DEREK PARFIT:
Yeah.

[01:47:22] TIM SCANLON:
Right, right. which of those three things lies behind your reticence?

[01:47:25] DEREK PARFIT:
Well, here’s one reason. Um, I’ve said there’s this big distinction between, uh, two level theories like contractualism, rule consequentialism, and single principle. But in fact, that’s a huge oversimplification because there are enormous differences within those categories.

I mean, in the case of rule consequentialism, there’s an enormous difference between the one that appeals to ideal rules and the one that appeals to how you should assess actual rules that, Uh, we can’t judge whether a theory of this kind is going to work until we’ve looked at different versions of it, because there are very strong objections to this version and that addition, we need to look and see. So I just think it’s hasty to say, um, I mean, I don’t know enough about the ways in which these different views can best be developed for me to be confident in advance which I should end up with. Now, with respect to the systematic theories, here I think the position is this.

Um,

[01:48:28] TIM SCANLON:
The question is whether we ought to think about morality and try to answer moral questions in a way that excludes as irrelevant our moral convictions, our moral intuitions. And what’s striking about Kant’s various formulas and all contractualist

(clears throat)

theories known to me is that they take that form. Uh, in deciding what you ought to do, your moral beliefs never come in, okay? Now, most contractualists will allow that if the implications of this contractualist theory for what you ought to do conflict too strongly with your moral beliefs, then it’s reasonable for you to reject contractualism.

But in applying the theory, you don’t appeal to your moral beliefs. And there’s a baby question: why in our moral thinking should we not appeal to our moral beliefs? And one defense would be the skeptical defense.

Um, Scanlon said in his first article that contractualism rests on a limited form of skepticism. Well, some contractualists are more outspoken. Gauthier rejects all moral intuitions as just irrelevant because he thinks they’re all mistaken.

There’s nothing for them to be about. People believe there are independent moral truths, but there aren’t. The whole thing is a mistake.

And then, well, can we, out of the ruins of traditional moral thinking, erect a kind of substitute by appealing to means-ends instrumental rationality? He thinks we can. Um, but if a contractualist accepts, I mean, to put it in Tim’s terms, that there are conceptions of what’s right and wrong, uh, which aren’t themselves contractualist, they don’t appeal to the idea of what no one can reasonably object, and if there are independent truths about what’s right and wrong in those senses, then we need to be told why, in our moral thinking we should bracket all of that.

And that’s my main simplest question. Unless we think our moral intuitions are deeply untrustworthy, why shouldn’t we appeal to them in our moral thinking? And that’s what many of the systematic theories tell us to do.

[01:50:50] AUDIENCE MEMBER:
Uh, Tim, did you wanna-

[01:50:52] TIM SCANLON:
I had two, two thoughts about that. One in, uh, just in relation to Barry’s question. Uh, perhaps your position in relation to Barry’s question might be put as follows: Barry, you took to be asking you, “Derek, what do you think really is right and wrong?”

Yeah. I mean, which things are right and wrong according to you? What view do you take on that?

[01:51:11] DEREK PARFIT:
Yeah.

[01:51:11] TIM SCANLON:
And, and your response was, “I think that right and wrong are used in a number of different senses, can be clarified and so on, but they’re, but they, they’re, they’re distinct, and the answer to q- to, to right– what’s right and wrong may depend on which of those one is using, uh- I wouldn’t- I don’t think, quite the same way.

That you don’t think there’s a single answer to the question of which one of those is the correct account of right and wrong because they’re talking about some degree different, different things. Um, a-a-although you think there is the question, what– which one of those in which circumstances do I have most reason to take seriously? Yeah.

And, and that’s not a question about what’s the right answer about right and wrong. It’s rather a question about what I have most reason to do. Is that, is that a fair description of your point?

[01:51:55] DEREK PARFIT:
It’s not, it’s not just, I mean, I think there are different senses of wrong. Some of them I think it’s not worth worrying with, but there are several I think that give you a plausible thing.

And then if you’re using right and wrong in these different senses, then there are different kinds of theory, and there are very different ways of developing these theories. And I just think– I do think at the end, we’ve got to try and decide which are the most important ways in which an act can be wrong, right? How strong are the reasons that are given to us by that?

And within each of these, we have to decide, well, should we go for an ideal, uh, or should we go for something that brings in, you know, relativity to the community? I really do mean, I think it’s extremely early days, and I’m struck by the fact that, I mean, I used once to be interested in chess, and, you know, there are some questions in chess like, you know, the Sicilian variation. There’s a three-volume work on the Sicilian variation.

Far more attention is being given to that than to some of the most fundamental questions about normativity, reason, and morality. Um, and I don’t think I can predict which way it’s gonna go. Um, but I think a lot of people think, “Oh, goodness,” people have been thinking about morality for, you know, three thousand years, so surely everything that there is to say has been said and thought.

Absolutely not.

[01:53:18] MODERATOR:
Um.

[01:53:23] AUDIENCE MEMBER:
Let me ask this question.

[01:53:24] MODERATOR:
John?

[01:53:26] JOHN:
Yeah, um, Susan Wolf asked you the question whether you thought all reasons were either, uh, impartial or personal, and you talked about her example, but I didn’t quite hear what your answer to that question was, as in yes or no.

[01:53:39] DEREK PARFIT:
Uh, right. Well, I just did.

[01:53:41] JOHN:
I want to follow this up, but I don’t—

[01:53:42] TIM SCANLON:
Oh, sorry. Why, why don’t you—

[01:53:44] JOHN:
Yeah. Well— Well, was it yes or no for the former? I mean, was the answer yes?

[01:53:49] DEREK PARFIT:
What is the alternative? There are two related questions. There’s the distinction between agent-relative reasons and agent-neutral impersonal reasons.

Right. Uh, and then there’s the distinction between the personal point of view and the impartial point of view, and those are different. And, uh, they’re both complicated, and one of the complexities is that some moral reasons, which I would want to cast as impersonal, uh, are nonetheless agent-relative.

I mean, they don’t say, “I should minimize the incidence of deceiving, you know, stealing, tragic– I mustn’t do it, I shouldn’t do it.” That’s an agent-relative restriction, but it’s impersonal in some of my senses.

So, um, they don’t fall neatly into either of the two categories. There are clear cases, but then there are much more complicated ones which are much harder to classify.

[01:54:46] TIM SCANLON:
Yeah. I mean, the kind of cases I was wo-wondering about what you can say about were cases where, uh, suppose I’m considering, uh, my department. Yeah,

but I’m not picking myself out as special in that group. Yet I’m picking the group out by being my department. Um, impartial or personal, it doesn’t really seem that that classification, um, is apt for cases like that.

[01:55:14] DEREK PARFIT:
Well, uh, it would become an impersonal but agent-relative reason if you thought that everyone had the obligation to have the special relation with those with whom they work, their institution, and so on. It wouldn’t be in that way impersonal. If it’s true of you that, you know, you care greatly about these things, not your own well-being, it’s, you know, this institution to which you belong, but you didn’t think that that was a requirement on anyone in such institutions.

Um, that’s only one way of drawing that distinction. But I, I, I mean, I think part of the reason why Sidgwick despaired, uh, thought that the cosmos of duty was reduced to a chaos and so on, um, is that his conception of the personal and impersonal was much too simple. And if all you’ve got is what’s best for me and what’s best for everyone impartially considered, it can seem very difficult to bring them together and so on.

Once you throw in a much wider range of reasons that we have, uh, you may be able to sort of break the ice. Uh, Sidgwick, I think, just got into a rut which he just couldn’t, couldn’t get out.

[01:56:38] TIM SCANLON:
Alan, Susan, do you wanna say anything at this point before we start to run out of time and res– either in response to what Derrick said to you or, um, just to- no?

[01:56:47] ALAN:
No, I might wanna make a remark, uh, about a couple of things that he said. Um, I, I’m not sure quite what he meant by saying that the lifeboat case is extremely common. But, uh, I mean, I think it is extremely common, and I said this, that there are decisions where if you do it one way, you will hurt these people, and if you do it the other way, you’ll hurt these other people.

Those are very common. But that there are cases in which you have to think about it in those terms, or you really ought to think about it in those terms and the terms that you do in the lifeboat example, I’m not so sure there are— Those are common.

[01:57:30] DEREK PARFIT:
Well, um, if we think that it makes an enormous difference whether the people whose lives you could save are right in front of you, or makes an enormous difference whether or not other people could also save their lives, um, then there may be reason for thinking that the fact that each of us could easily save many people’s lives, uh, it doesn’t make this like lifeboat, ’cause in lifeboat they’re on rocks which are within sight or something like that. So there would be ways of saying, um, that this is not a good model for the wider picture. On the other hand, you were talking about war, um, and so on.

It’s extremely common for people running a health service, whether it be the medical people or whether it be the Minister of Health and the government, to have to make decisions, uh, that’s about whether it matters how many lives they save and what the relative importance are. So that’s the sense. Those are the two senses in which I thought that it was very common.

Um, but I agree, uh, life, uh, tunnel and bridge are not very common. And I wondered what you think about the claim that we can just ignore those cases, that it isn’t a good objection to utilitarianism, that it might tell you to kill one person as a means because that case is so unusual, and similarly, it isn’t a good objection to Kant’s view about lying that, you know, that would imply you should lie even to a would-be murderer, ’cause you can say, “Well, you know, that’s a, that’s one of those cases we then.”

[01:59:10] TIM SCANLON:
That’s the direction I wanted to take it, though it is sort of. Uh, that is, um, I don’t know that your intuitions about these very… It’s not just that they’re uncommon, but that they’re artificial. Uh-

[01:59:24] DEREK PARFIT:
Well, you had, you had very strong intuitions about all of them.

[01:59:28] TIM SCANLON:
Well, uh, that was the other point I wanted to address. I have strong intuitions when I add a bunch of assumptions that I don’t think you’re supposed to add. Now, maybe you are supposed to add them, but-

[01:59:37] DEREK PARFIT:
Well, let’s, let’s take a hypothetical case. A person says, uh, there are people on these two rocks, and, um, some of them are Black, others are white. Blacks on this rock, white on that.

Now, I quite agree with you. We should say, well, the fact that some are Black, some are white, that’s irrelevant. All I know is that there are some people on these rocks and some on others.

Some people would say, “That’s all I need to know. I should give everyone an equal chance.” Other people would say, “Well, what are the numbers?”

And that’s a straightforward question. Does it matter how many people you can save, or should you give everyone an equal chance? Now, as I say, in interpreting these cases, you’ve got to assume that there are no other relevant differences.

If you can save the lives of, right, five ninety-year-olds or one twenty-year-old, perhaps you should save the twenty-year-old. Um, so you just have to assume whenever people produce simply described cases that there are no other moral development differences.

[02:00:34] TIM SCANLON:
Yeah. The problem– then what I would… see, then, then what I would say is that I’d no longer have any clear intuition. No, I may have intuitions, but I don’t trust them.

[02:00:43] DEREK PARFIT:
I wouldn’t say that. So in other words, your view is that–

[02:00:45] TIM SCANLON:
I think they’d be bad data for moral philosophy.

[02:00:47] DEREK PARFIT:
Okay. But then, I mean, do you think that a case in which you could either save five, the lives of five 20-year-olds or one 20-year-old, and the chances of saving the five are the same as the chances of saving the one, and there’s no hope of saving all six. I mean, I think you could easily know there’s no hope of saving all six.

Um, then I think you might think, “Well, I should save five rather than one.” Or you might think, “I should toss the coin.” But I don’t think you’re missing any relevant information.

That’s a fully described case. I mean, the-

[02:01:22] TIM SCANLON:
No, I, I think you are missing, Rob, because I think you oughtn’t to make decisions on that kind of basis, and moreover, that if human life goes as we should want it to, you wouldn’t have to. That’s the line I want to take.

[02:01:40] DEREK PARFIT:
Well, that’s like saying, you should ignore the question whether you should resist aggression by force, because if human life goes as it wants it to, nobody will, you know, commit aggression. I mean, we have to respond to the actual world. Um, And what’s wrong with trying to decide whether you ought to save the one or the five?

[02:02:05] TIM SCANLON:
Well, there may be a, a, a real disagreement here about how necessary to the human condition it is that we think about things in the way that trolley problems lead us to. And you may think that it is necessary, and I may doubt that.

[02:02:22] DEREK PARFIT:
Well, Kant thought that it was necessary.

[02:02:24] TIM SCANLON:
Well, he may have.

[02:02:26] SUSAN WOLF:
Um.

[02:02:28] DEREK PARFIT:
Well, I mean, then, then you’re saying that, that, that there are many cases, uh, in which we just needn’t ask what we ought to do?

[02:02:38] TIM SCANLON:
No, well, uh, there are many examples where whether we have strong intuitions or not, we shouldn’t necessarily trust those intuitions as, as, uh, good data for morality. That’s, that’s the kind of conclusion I would draw.

[02:02:55] DEREK PARFIT:
Okay. Well then, I think you should withdraw some of the firm intuitions that you have here. Yeah. Because, you know, you very firmly said, um, that you shouldn’t switch the trolley from the five to the one.

[02:03:12] TIM SCANLON:
Well, I was making a bunch of assumptions. I was assuming, for instance, that if the trolleys were well-regulated, that would be an illegal act.

[02:03:20] DEREK PARFIT:
Well, I, I must say, what you said there-

[02:03:22] TIM SCANLON:
And, and I, you know, I made a bunch of other… Now, so if you don’t make any of those assumptions, then-

[02:03:26] DEREK PARFIT:
Yes, but-

[02:03:27] TIM SCANLON:
Okay. Then I don’t have the strong intuition.

[02:03:28] DEREK PARFIT:
Yeah, well, you did. You said, “You have no business touching such equipment for any reason,” and if you’re an employee, you should strictly follow the rules.” Now, I think that is what produced the difference between Germany and Italy in the Second World War.

Many Germans accepted that kind of view. We should follow the rules, and we shouldn’t interfere for any reason. And as Eichmann, you know, famously said to Straus, he said, “I’ve read the second critique.

I know about the categorical imperative.” No, I’m not claiming that Kant led- But his view that you have no business interfering for any reason, and you should follow the rules, I think that’s an extremely dangerous, uh, view for people to hold.

And Italians-

[02:04:15] TIM SCANLON:
Oh, I give you-

[02:04:16] DEREK PARFIT:
Their humanity break through, and they behave much better.

[02:04:18] TIM SCANLON:
You wouldn’t know that you wouldn’t wreck the train, that you wouldn’t know that there were people down, uh, farther down on the track. If you remove a lot of those assumptions-

[02:04:25] DEREK PARFIT:
Well, but, well, but that-

[02:04:26] TIM SCANLON:
You know, you can doctor these cases so that it– what I said isn’t any longer reasonable. Well, and then I don’t think it’s reasonable either.

[02:04:34] DEREK PARFIT:
Well, but that’s like, I mean, what you said about that case is like Kant said when, you know, the murderer asks where his victim is. For all you know, if you tell him where the victim is, that may save the victim’s life.

(sighs)

Well… We have to go on the balance of probabilities, and very often, although you’re not certain, you will know that it’s much more likely that this will save them or that this will lead to their death. Um.

[02:05:06] TIM SCANLON:
Susan, did you want to weigh in maybe?

[02:05:09] SUSAN WOLF:
Uh, well, I think rather than reintroducing another subject, I should stop. But I would make one comment about the, uh, value of trolley problems and so on.

(laughter)

I, um, even though I don’t go in for, uh, doing moral philosophy that way, I, um, it’s actually my impression that almost anything that moral philosophers have created, um, happens in the world in some, um… So I was actually surprised that Derek conceded so much, uh, that, oh, well, not, not only lifeboat. Uh, it seems to me they all happen.

And if I recall, the trolley problems were introduced as a way of dealing with abortion issues. I mean, real live issues about what doctors may and may not do. And, and so I guess I think though it’s very healthy to recognize the ways in which it’s, um, you know,

[02:06:07] DEREK PARFIT:
Well, it can be-

[02:06:08] SUSAN WOLF:
if we get highly fictionalized, it’s also not so unconnected with real problems, and, um, it may be a lack of imagination to think that, that that’s so stuck in that-

[02:06:20] TIM SCANLON:
Well, Bernard Williams, who’s not particularly a friend of transportation ethics, made the point in, in, in Critique of Utilitarianism, and that one of the reasons for resorting to hypothetical examples is that it’s an important resource in thinking about actual cases. Uncommon cases occur actually all the time, and one of the things we do when we confront them is to try to think what we’d say if the cases were a little different in a certain relevant respect. So that partly it seems the danger of the line of, um, criticism that you’re suggesting is that it makes sort of cut off conversations that we need to be able to engage in if we’re gonna say anything about the cases we do encounter

[02:06:59] ALAN:
Well, I snuck in remarks that said essentially that. That there are, there are… This– They certainly do give you, uh, resources for thinking about extreme cases, right? Right. I didn’t deny that.

[02:07:14] TIM SCANLON:
Well, I mean, the point is they may give you resources for thinking about actual cases, the actual cases you are trying-

[02:07:19] ALAN:
Well, I know. But I considered these cases are extreme. Yeah.

[02:07:24] TIM SCANLON:
Okay.

[02:07:24] DEREK PARFIT:
I didn’t think it’s very often true that you can only save some people by means of killing others. I think those cases are pretty rare. I don’t think it’s a good defense of utilitarianism to say that they’re rare, but I think they are rare anyway.

But I think saving one or many, and that’s just an instance of saving one or many from burdens, needn’t be life. Uh, those are extraordinarily common, I think. That’s why I wanted to distinguish so sharply between that and as a means or as a side effect.

[02:07:56] ALAN:
Well, I didn’t deny that these cases are, uh, that cases in which this is true are common. What I’m denying is that there is, there need be no other way of thinking about this.

[02:08:08] DEREK PARFIT:
Well, the most striking thing you said is that it’s immoral even to ask whether you should save the one or the five. Now that– I mean, you might say two things. It’s immoral to ask that question if you’re not confronted with that actual dilemma.

You shouldn’t waste your time thinking about that. The other is, even if you are in that position, it’s immoral to ask, what should I do? Now, why is it immoral to ask that question?

[02:08:35] ALAN:
Yeah. No, I didn’t mean to imply it in the mo– in the extreme case. I did quote… I, I, I think– So what’s immoral– Dorothy Allison was saying that, and I quoted her with some approval.

[02:08:46] DEREK PARFIT:
The point is that if you don’t–

[02:08:47] ALAN:
But why is it immoral to ask? What I said was, if you don’t have to think that way, it is probably immoral to m– to, to think that way voluntarily. That’s what I would say.

[02:09:01] DEREK PARFIT:
So it’s immoral to ask the question, does it matter how many people your acts would benefit or burden? Is that an immoral question to ask?

[02:09:11] ALAN:
Not necessarily, no.

[02:09:12] DEREK PARFIT:
But that is the question.

[02:09:13] ALAN:
No. I mean,

[02:09:14] TIM SCANLON:
But that, that you sh– But that should be the, the… The trolley problems force you… I mean, they give you only this information, really, and you’re not supposed to add a bunch of other information.

And so they suggest that you’re dealing with a case in which, uh, that’s the only relevant information. Well, but look, can I say that? Yeah.

Well, but what they’re concerned- those kinds of cases are very rare, where they’re the only relevant information.

[02:09:42] DEREK PARFIT:
No, no, no. What they’re doing is they’re saying it isn’t the only relevant information, how many people you are benefiting or burdening, but the question is, is that one of the relevant considerations? Does it matter whether you’re saving one person’s life or a hundred people’s lives?

Now, some people think, no, you should give everyone an equal chance.

[02:10:04] TIM SCANLON:
Um, but I don’t think that if you say, well, suppose that other things are relevantly equal, the only difference is that you’re either curing one person’s blindness or ten people’s blindness, um, that it’s immoral to ask whether the numbers count. You know, when you say all other things are relevantly equal, So I, I find it very hard to have intuitions in which– about cases in which I’m trying to take that stipulation seriously.

[02:10:38] DEREK PARFIT:
Well, I mean, and one easy way of doing it in which you, you don’t know whether one of these ten people is young Adolf Hitler or something like that, uh, or is a mass murderer or whatever. All you know is there are ten thirty-year-olds over here, and they look as if they’re in their twenties, and there’s one over there. Um, if you don’t know the further information, then all you know is you can save one life or ten, and, uh, it would matter if, if the ten were age ninety and the one was twenty, which is why I’m saying, well, other things equal, they’re about the same age.

Um.

[02:11:18] MODERATOR:
Um. The hour is getting late. I think yesterday, uh, in posing one of his questions, I think, uh, Jay Wallace, uh, gave us an example of quiz show in which you have to pick what’s behind door number one, two, or three.

Um, we have a real-life example. Uh, behind this door, um, there is actually a reception. Um, and your only choice is whether you would like to attend it.

But I thought that before, uh, to, to give you a little time to think about that, uh, we might first ask Derek, Derek, if he’d like to have a last word or give us any last thoughts in response to, uh, any of the issues that have been raised or any issues that haven’t been raised. Any final words? You don’t have to have them.

[02:12:06] DEREK PARFIT:
I don’t think so. I mean, I just want to repeat early days that, I mean, I– In preparing these lectures, I was actually rather staggered to realize that the questions that I was asking and the views that I was considering are ones that weren’t even asked by the philosophers I was writing about, Hume, Kant, Sir David Ross, and Rawls, because they didn’t even understand them in the same sense.

Now, if we haven’t been even understanding one another, it’s very early days. So no, I don’t think I have more to say.

[02:12:48] MODERATOR:
Well, in that case, please join me in thanking Derek and Pete.

(applause)