2005-2006 Lecture Series

[00:00:00] INTRODUCER:
Welcome everyone. It gives me great pleasure, uh, to welcome all of you to Berkeley campus this afternoon for the Tanner Lectures on Human Values. And just to give you background both on the lectures and on the lecturer, er, uh, this, as many of you may know, this distinguished lecture series is presented annually at each of nine universities, including UC Berkeley.

The other universities are, uh, Cambridge and Oxford, and here in the United States, Harvard, Michigan, appropriately, Princeton, Stanford, Utah, and Yale, which is not a bad set of institutions to be, uh, associated with. Uh, the series was founded in 1978 by the American scholar, industrialist, and philanthropist Obert Clark Tanner, who was also a member of the faculty of philosophy at the University of Utah. Uh, he was also an honorary fellow of the British Academy.

Tanner’s goal in establishing the lectures through the Tanner Philanthropies was to promote a, a search for better understanding of human behavior and human values. He hoped that the lectures would advance scholarly and scientific learning In the area of human values, uh, and contribute to the intellectual and moral life of humankind. And maybe everyone always feels this, but if I ever felt that there was a time in history where we need to reaffirm human values, uh, and the intellectual and moral life of humankind, I think this is, this is one of those.

Uh, human values is defined a-as broadly as possible. It’s in the context of these lecturers. Uh, and the lecturers may be chosen from any discipline.

Uh, the lec-lecturers are international and transcend national, religious, and ideological distinctions. Uh, as, uh, many of you may know, the lectures from all nine universities are published in an annual volu-volume. And in addition to that, Oxford University Press publishes a series of books based on the Berkeley Tanner Lectures.

And the second volume of this series is now in press. I don’t know if you, Sam, if you brought an example of the—

[00:02:35] SAM SCHEFFLER:
I didn’t. No. No, that’s okay. It’s actually in print. The third one’s in press.

[00:02:38] INTRODUCER:
Good. Oh, okay. Great.

Good. That— that’s right. The, the Tanner Lecturer is chosen for his or her, or her uncommon achievement, I like that, And outstanding values, uh, abilities, pardon me, in the field of human values.

Um, here at Berkeley, the Tanner Lecturer is appointed through a faculty committee, uh, of which the chancellor is chair, uh, I’m embarrassed to say only an honorary chair. Uh, and I wanna congratulate the people who have done the real work, which is, uh, my colleagues, uh, Professor Sam Scheffler, uh, Hannah Ginsborg, Martin Jay, Victoria Kahn, Christopher Kutz, and R.J. Wallace for their brilliant choice of this year’s lecturer. You’ll hear more about him shortly from Sam, but our Tanner lecturer this year, as you all know, is Professor Alan Gibbard, who’s the Richard B. Brandt Distinguished University Professor of Philosophy at the University of Michigan.

A renowned ethical philosopher, he has made vital contributions to the field of ethical theory, particularly to the study of metaethics. Professor Gibbard will give two lectures addressing the topic of thinking how to live together, uh, a subject that manifestly is very relevant to our times. Uh, his lecture is entitled, and I quote, “Insight, Consistency,” and Plans for Living.”

And tomorrow, he will deliver a second lecture on, and I quote, “Living Together: Economic and Moral Argument.” Both lectures will be followed by commentaries from prominent and distinguished colleagues. Today’s commentary will be given by Professor Michael E. B-Bratman, Durfee Professor in the School of Humanities and Sciences and Professor of Philosophy at Stanford.

Professor Bratman is an internationally respected philosopher of action. His work is credited with furthering our understanding of moral responsibility, temptation, and self-control. Wednesday’s commentators are Professor John Broome and Professor Frances Kamm.

Uh, Professor Broome is White’s Professor of Moral Philosophy at Corpus Christi College at Oxford, and Professor Broome’s scholarship combines both economic and philosophic expertise. Professor Kamm is Littauer Professor of Philosophy and Public Policy at the John F. Kennedy School of Government And the Department of Philosophy at Harvard. Uh, Professor Kamm is a leading moral philosopher whose work has included exploring moral problems related to life and death situations.

On Thursday, to close our three-day Tanner Lectures event, there will be a seminar discussion including Professor Gibbard and all of the commentators. Both Wednesday and Thursday’s events will take place in the same location from 4:00 PM to 6:30 PM. Uh, now, without further delay, let me call on my distinguished colleague and friend, Professor Sam Shefford– Scheffler, pardon me, who’s the class of nineteen forty-one World War II Memorial Professor of Philosophy and Law, uh, and vice chair of the Tanner Lectures Committee to, uh, introduce Professor Gibbard.

And, uh, Professor Scheffler will also moderate the discussion that follows. Sam.

[00:06:40] SAM SCHEFFLER:
It’s a privilege to introduce Alan Gibbard as the Berkeley Tanner Lecturer from two thousand and five, two thousand and six. Professor Gibbard was educated at Swarthmore, where he majored in mathematics and minored in physics and philosophy, and at Harvard, where he received his PhD in philosophy in 1971. After teaching for five years at the University of Chicago and three years at the University of Pittsburgh, he moved to the University of Michigan in 1977, and he has remained there ever since, despite the best efforts of most of the world’s leading research universities to lure him elsewhere.

His many honors and awards include Guggenheim, ACLS, NEH, and Rockefeller Fellowships, a visiting fellowship at All Souls College, Oxford, an election to the American Academy of Arts and Sciences, the Econometric Society, and the American Philosophical Society, an organization which, despite its name, has very few philosophers as members. He’s also a past president of the Central Division of the American Philosophical Association, a similar sounding but very different organization. Professor Gibbard, as the Chancellor suggested, is best known for his groundbreaking work in moral philosophy and especially in metaethics, but he’s a wide-ranging thinker who has done important work, including some quite technical work, in such diverse areas as decision theory, social choice theory, political philosophy, the philosophy of language, metaphysics, and the theory of conditionals.

He’s the author of two acclaimed books, Wise Choices, Apt Feelings: A Theory of Normative Judgment, published in nineteen ninety, and Thinking How to Live, published in two thousand and three. Both of those books are part of a larger intellectual project of trying to understand the place of moral and other normative concepts and judgments within a broadly naturalistic view of the world. What is it that human beings as natural creatures are doing when they speak, for example, of a certain act’s being right or wrong, or rational or irrational, or when they say that one ought or ought not to behave in a certain way?

Gibbard is not the first person to ask these questions, but the theory he develops to answer them, which draws extensively on work in psychology, anthropology, and evolutionary theory, is original and compelling. His view has something in common with earlier so-called non-cognitivist theories, like those of A.J. Ayer, Charles Stevenson, and Richard Hare in the last century, but it far surpasses earlier forms of non-cognitivism in its intellectual power, interdisciplinary range, and philosophical subtlety and sophistication. In addition, Gibbard’s prose has a spare and graceful style that makes reading his work a memorable experience.

One reviewer of his first book wrote, quote, “Every so often, though not often enough, a philosophical book is written that addresses a deep problem with profound insight, subtle argumentation, and captivating style.” Alan Gibbard’s Wise Choices, Apt Feelings is such a book. It is an important book.

It is a beautiful book.” End of quote. Alan Gibbard’s title at the University of Michigan, as the Chancellor said, is the Richard B. Brandt Distinguished University Professor of Philosophy.

It is a wonderfully appropriate title, and in more than one way. Richard Brandt was himself a distinguished moral philosopher who taught at Swarthmore while Gibbard was an undergraduate there, and who subsequently moved to Michigan, where Gibbard eventually became his colleague. Gibbard has credited Brandt with having been one of the most important influences on his development as a moral philosopher, and for that reason alone, it is fitting that he should now hold a chair named in Brandt’s honor.

But it’s also fitting because it serves as a public reminder of the great tradition of Michigan moral philosophy, which was upheld for many years by Dick Brandt and his colleagues, Bill Frankena and Charles Stevenson, and which has continued uninterrupted to the present day. The middle decades of the twentieth century were, on the whole, a bleak period for moral philosophy in the English-speaking world, where the combined influence of logical positivism and ordinary language philosophy served to marginalize the subject and make it seem curiously uninteresting and unimportant. But Michigan was always an exception to this trend, and the renowned group of moral philosophers who worked and flourished there helped to keep the subject alive throughout that period.

There has been a c-powerful cohort of moral philosophers at Michigan ever since. And if no member of the earlier group was more distinguished than Richard Brandt, no Michigan moral philosopher, past or present, is more distinguished or more accomplished than Alan Gibbard. Indeed, he has few equals anywhere.

Gibbard has written of Brandt, quote, “I have always hoped to emulate his insistence on clarity, his rigorous honesty, his refusal to be swept along by trends, and his determination to press to the heart of matters.” All of these traits are in fact possessed to the highest degree by Alan Gibbard, and it is very– it is a very great pleasure to introduce him as this year’s Tanner Lecturer.

(applause)

[00:11:58] ALAN GIBBARD:
Well, thank you very much. Uh, wheth-whether it’s, uh, all true or not, it’s, uh, very, uh, uh, gratifying to hear all that, and it’s very gratifying to be at, uh, at Berkeley. I first came to Berkeley for Peace Corps training in nineteen sixty-three, and, uh, so– and I’ve been very attached to Berkeley ever since.

Uh, I also had the, uh, honor of, uh, meeting the Tanners during the early days of Tanner Lectures at Michigan, which I enjoyed very much. Uh, so, uh, uh, uh, uh, uh, as was announced, the title of the, uh, lecture today is Insight, Consistency, and Plans for Living. Uh, if, uh, if you can’t hear me in the back, by the way, uh, let me know.

Uh, I’ll stand

(cough)

a little way away from the microphone some of the time. Uh, well, Jonathan Haidt, the social psychologist, entitles a fascinating article, The Emotional Dog and Its Rational Tail. His topic is moral judgment, and the emotional dog is what he calls intuition.

Mostly, he argues, we don’t arrive at our moral conclusions by reasoning. We jump to them with the emotional judgments, with effectively valenced intuitions, as he puts it. We will often be firmly convinced that our moral judgments rest on sound reasoning, and that unless others are driven by bias, they will appreciate the force of our arguments.

He calls this the wag the other dog’s tail illusion. In fact, though, in our moral reasoning, we are not so much like intuitive scientists following the considerations where they lead, but like intuitive lawyers reasoning to preordained conclusions. Reasoning is effective on occasions, he concedes, with, as he puts it, adequate time and processing capacity, a motivation to be accurate, no a priori judgment to defend and justify, and when no relatedness or coherence motivations are triggered.

Mostly, though, what reasoning does is to construct justifications of intuitive judgments causing the illusion of objective reasoning. All this chimes in with Hume’s dictum: reason is and ought only to be the slave of the passions. Haidt himself isn’t talking about how moral judgment ought to work.

He’s offering a psychological account of how moral judgment does work. Now, even philosophers who stress reasoning have often thought that reasoning must rest ultimately on intuition. Intuitions give us the starting points of reasoning, and they tell us what follows immediately from what.

Reasoning thus strings together a series of intuitions. Haidt’s thesis isn’t just that intuition is crucial to moral judgment, but that it isn’t this stringing together that mostly drives moral judgment. Reasoning he defines as going by conscious steps so that it is intentional, effortful, and controllable, and so that the reasoner is aware that it is going on.

What’s powerful in moral judgment, uh, Haidt argues, will be the single emotionally valenced intuition that reaches its conclusion all by itself. Moral judgment doesn’t have to be that way, for all Hume’s dictum tells us, but that, Haidt argues, is the way moral judgments mostly are. We can ask whether what Haidt means by intuition is what philosophers have traditionally meant.

The paradigm of intuition in the philosophical tradition has often been geometric intuition by which we know the axioms of geometry. These are emotionally cool, whereas the intuitions that drive most moral judgment, according to Haidt, have emotions attached. It’s an important question whether the intuitions that ground morality are always tied in with emotion, but that’s not a question I’ll be addressing.

Later on, I’ll distinguish senses of the term intuition, but I won’t restrict the term either to hot or to cool states of mind. Now, we philosophers aren’t expert psychologists. It’s how we ought to reason that we are especially charged with assessing.

Often we do reason, even on moral questions, and I’ll assume that sometimes we should. The philosopher’s job in particular is to reason, and if we ought never to reason on morals, then we moral philosophers may need to find another line of work. Today, though, I won’t be engaged in moral reasoning.

That’s for tomorrow. My questions today will be about moral reasoning. What is its subject matter?

I’ll ask. How ought we to reason? If reasoning strings together intuitions, why trust its intuitive starting points?

I’ll talk about these broad questions today and then tomorrow scrutinize a particular piece of moral reasoning, one that purports to get remarkably strong moral conclusions from plain and clear intuitions. Moral intuitions are puzzling. We seem to have moral knowledge.

Indeed, some moral truths seem so utterly clear as to be pointless to state. It’s wrong to torture people for fun. Other moral questions are agonizing to ponder.

Are there any conceivable circumstances in which we would be morally justified in torturing someone? If we have moral knowledge at all, it seems, this knowledge must rest in the end on powers to intuit moral truths. G.E. Moore a hundred years ago elaborated arguments that moral claims aren’t claims that could be brought within the purview of natural science.

Two people could agree on all the facts of empirical science and still disagree morally. They could disagree, say, on whether, as Henry Sidgwick thought, pleasure is the only thing worth wanting for its own sake. The fault of the one who is wrong needn’t rest on ignorance of the natural facts or failure to grasp the concepts involved or any failure of logic.

Natural facts and conceptual truths aren’t enough to entail answers to moral questions. If we’re to have any moral knowledge at all, then the gap must somehow be filled. What else could it be filled by but a power of intuition, a power to apprehend some basic moral truths, though not by the senses?

Not all philosophers agree that morality lies outside the scope of empirical science, but I’ll be offering a picture on which it does and proceed on the assumption that The picture is right. Moreover, I would argue that even claims about rationality and science aren’t entirely within the rash the subject matter of science. Science itself rests on intuitions about the justification of empirical conclusions.

If that’s right, then it may not only be morality that raises puzzles about intuition. In the case of morality in particular, a chief puzzle is, it’s hard to see how beings like us could have powers of moral intuition. We are parts of the natural world.

Crucial aspects of any moral truth, though, don’t lie in the natural world. When we look at ourselves as parts of the natural world, as Haidt does, we won’t find a responsiveness to anything non-natural. We won’t even find the purported facts we claim to intuit I’ll begin what I have to say by sketching a view of ourselves as part of nature.

Moral right and wrong form no part of this view. It’s part of this view, though, I’ll be claiming, that we would ask ourselves moral questions and come to conclusions about them. How things stand morally is not part of the natural world, but our study of these matters is.

Uh, later I’ll be qualifying that, but, uh, for the moment, let’s stick to it. Uh, beings who think and reason about what to do, I say, answer questions of ought, at least implicitly, when they settle on what to do. Beings with our own psychic makeup make specifically moral claims.

I’ll speculate how these activities look within a value-free scientific picture. Then after that, I’ll turn to the plight of beings like us who figure in the picture, beings who think about what to do and think about right and wrong. Our answers to the questions we address will rest on intuitions.

But I’ll be asking, if intuitions are the sorts of states that figure in Haidt’s picture, why place any stock in them? So I’ll speak of nature, oughts, and plans. Begin then with us as living organisms who are part of the world of living organisms.

The upshot of natural selection is that genes become amazingly good at, as it were, working together to use us to make copies of themselves. So how, a great puzzle runs, have metaphorically selfish genes come to make people who are literally not entirely selfish? The answer can only be long.

The answer can only be long and controversial, and I won’t address this particular story in these lectures. Rather, I’ll ask about the place of oughts in the story in this world of ises. The story proceeds in terms of the metaphorical interests of the genes, the things that promoted the multiplying, uh, of the genes as this human genotype was formed.

And second, the literal aims, beliefs, and feelings of humans and protohumans. Genes proliferate in part by forming recipes for organisms that keep track of the world around them, very much including the social world. Knowledge guides action.

But it guides action in a way that proliferates genes only if the actions have the right aims, the right propensities to use knowledge to guide action. Well, knowing where a lion is won’t promote one’s genes’ reproduction if one’s reaction is to try to pet it. The beings in this biological picture of us then face questions of how things are, but those aren’t the primary questions they face.

The, the primary questions are ones of what to do, what to aim for, and how. Most organisms, of course, can’t be interpreted in any full-bodied sense as addressing these questions and accepting or rejecting answers to them. Dogs chase squirrels and bark at intruders, and much of the time we, like the dog, just act out of habit or emotion.

We, though, are of an intricately social species with language, differ from other animals in at least two important ways. Uh, first, our social emotions are especially refined and elaborate. A substantial amount of the human neocortex seems to function in the workings of emotions, and emotions include impulses to action.

Many of our feelings are intensely social, as with guilt and resentment, with shame and disdain. Second, we are beings with language, and we make judgments that we express with language. Here then are two pieces of speculation about our species.

First, we’re adapted to specific kinds of emotional reactions to social situations. These reactions include moral emotions of resentment or outrage and of guilt, guided by judgments of fairness. Emotions bring characteristic tendencies to action, so that resentment, for instance, tends toward punitive action.

Emotions thus affect reproduction through the actions they prompt. And so natural selection will shape the psychic mechanisms of emotion. Human emotional proclivities evolved the way they did because of this.

With humans also, though, I speculate, there evolved a kind of language-infused governance of emotions. We discuss together and have linguistically encoded thoughts that work to control our feelings. On feeling a flash of resentment that you took a piece of cake that I had hoped to have, I can reason that you were as much within your rights to take it as I would have been, and so there is no cause for resentment.

At this thought, my resentment may subside. If it doesn’t, and I complain, expressing my resentment, the rest of you may set me straight. If my resentment doesn’t subside, the actions it prompts may, in my social circumstances, work in the long run to hurt my reproductive prospects.

Hence come selection pressures for a genetic propensity to control emotions in certain sorts of social circumstances. My resentment is unwarranted, I judge, when you finish the cake. So how does a concept like warrant work when I say it’s unwarranted?

For purposes of delineating how reasoning with such concepts can go, I suggest we think of judgments of warrant as something like plans. I plan, as it were, under what circumstances to resent people for things they do. This talk of plans for feeling sounds artificial, I admit, but when we judge that resentment would be unwarranted in my situation, this judgment acts much as would a plan for my situation not to resent you.

Literal plans are carried out by choice, to be su-sure, and we can’t choose what to feel. Feelings, though, do respond somewhat to judgments of warrant, as they might in the example. It’s thus somewhat as if we planned what to feel, even though choice doesn’t figure in the way that plan and action get, uh, uh, choice doesn’t figure in the, in this case, in the way that planning for action gets realized, uh, by guiding one’s choices.

Questions of moral right and wrong on this picture will be questions of what to do, but with a particular kind of emotional flavor. What is it to think an act morally wrong as opposed to just silly or imprudent? Roughly, I propose it is to think that the act warrants resentment on the part of others and guilt on the part of the person who did it.

Specifically, moral questions, if this is right, are questions of what moral sentiments to have toward things. At their narrowest, they’re questions of what to resent people for doing and what to feel guilty for doing. To guilt and resentment here, as Howard Nye has urged on me, we need to add a prospective feeling of guilt-tinged aversion toward acts we might contemplate doing.

This emotion is negatively valenced toward the act, and so to plan guilt-tinged aversion toward an act is to plan to be against one’s doing it in a way that has a particular emotional flavor. Uh, whether planning this aversion must always go with planning all things considered not to do the act is an important question that I won’t try to answer here. I’m contrasting thin oughts in general and moral oughts.

Narrowly moral questions of right and wrong I’m treating as at base questions of what moral sentiments we ought to have and act from. Questions in the broad class we call normative questions. These include questions of what a person ought to do, all things considered.

They include epistemological questions of what we ought to believe, and they include questions of how we ought to feel about things. These, I’m saying, are all planning questions, questions of what to do, to think, and to feel. The narrowly moral questions are questions of how to feel about things, what moral sentiments to have toward them.

So now I’ll speak about explaining oughts. A dictum that we draw from Hume is that you can’t derive an ought purely from an is, and G.E. Moore argued that oughts don’t form a part of the natural world that empirical science can study. The picture I’ve sketched has the upshot that Moore was right.

The scientific picture tells us why organisms like us would have questions whose answers can’t be a part of science. The point is that not only do we think about how things are, but also we act and feel. Our actions and feelings figure in a biological account, along with goings-on in the head that lead to actions and feelings about things.

Questions of what to do and why, and questions of how to feel about things and why won’t figure in the picture. Yet the picture has us addressing those questions. Suppose I settle on helping a man in need, even though I won’t get any advantage from it.

I think I ought to help him and that it would be wrong not to do so. And so I help him. My coming to these conclusions must be part of any full and adequate naturalistic biological story of me.

The story, though, won’t contain any fact that I’ve got my conclusions right or not. It doesn’t contain a fact that I ought to help or that it’s okay not to. It doesn’t contain a fact that it would be wrong not to help or that it wouldn’t be.

Questions of what I ought to do and what it would be wrong to do or not to do are questions of… are not questions amenable to science. They are, I’ve been saying, questions of whether to help and of how to feel about not helping. A scientific, a scientific picture then has us asking questions that don’t have scientific answers.

The picture shows, too, why these considerations, uh, why these questions aren’t luxuries, but must be central questions to us. And from the scientific picture comes an account of what these questions are. They are questions of what to do and how to feel about things people do or might do.

If these are the questions, we don’t need to worry that they concern queer goings-on that form no part of the fabric of the universe, as J. L. Mackie put it. They are intelligible questions, and they are questions of first importance. Uh, I’ve been contrasting questions of empirical science and questions of what to do and how to feel.

I should note, though, that this may not be getting matters quite right. Perhaps the two-way split that I’ve made really ought to be a three-way split. First, as I’ve been saying, there’s the empirical picture of us as parts of the natural world, shaped as a species, as it were, by natural selection, and shaped as individuals in society by complex social dynamics, a complex human ecology.

The empirical sciences of psychology, sociology, and anthropology, and the like all contribute to this. Next, though, there’s a part I haven’t singled out, interpretation. We understand some of these natural goings-on as beliefs, assertions, plans, and the like, with which we can agree or disagree.

The ought part then comes third in the list as we seek answers to the questions we can be interpreted as asking. So we have three areas of inquiry: psychosocial science, interpretation, and normative inquiry. When I speak of a person as thinking that she ought to help, and when I say that this amounts to deciding to help, I’m interpreting certain goings-on in her as the having of these thoughts.

As a first approximation then, I’m saying ought thoughts are plans. Thinking what I ought to do is thinking what to do, but the dictum needs refining. Uh, thinking what to do can go in two stages.

In the first stage, I form my valences or preferences. In the second stage, if there’s more than one thing I equally and most prefer from among my alternatives, I pick one. Not out of preference, but out of the necessity to choose if I’m not to be like Buridan’s ass who starved to death, uh, midway between two equally attractive bales of hay.

Uh, My strictly normative thinking is a matter of the first stage. We would call this part concluding what’s, Or, or let’s call this part, uh, concluding what’s okay to do and what isn’t. When it’s okay to do something and not okay not to, then I ought to do it.

Thinking what I ought to do then is not all of thinking what to do. Rather, it’s the part that matters, the valence stage. This ties in with worries about the right direction of explanation.

I started out explaining ought beliefs as plans, but this, even if it’s right, doesn’t explain normative belief in general. It doesn’t explain belief in ties for what it would be best to do, the belief that more than one alternative would be okay, the belief that something is rationally okay to do then has to be explained in some other way. And once we have this explanation, it’s easy to explain the concept ought.

That a person ought to do a thing just means that it’s okay to do it and not okay not to do it, as I said. Since we can’t explain an okay in terms of plans, perhaps we’re forced to become normative realists. We start, according to the normist realist, uh, establishing that being okay to do is a property we can know some acts to have, and then go on from there to explain the concept ought and what plans consist in.

So we begin with the property. Uh, that’s the objection. I’ve tried to explain the concept ought in terms of plans, but the explanation, it turns out, according to the objection, could only run the other direction.

Well, I answer that we can explain both both these concepts, uh, okay and ought, in terms of something we do on the way to planning, uh, forming valences. The explanation is oblique. I don’t offer straight definitions of the terms.

Okay and ought in terms of planning. Rather, I say what believing an act to be okay consists in. To believe it okay is to rule out preferring any alternative.

It’s thus to rule out a kind of valence. Normative judgments, we can say, consist in valences and restrictions or rulings out, uh, of valences. This, I’m claiming, explains the philosophically puzzling notions of what one ought to do and what it’s okay to do.

It explains the to-be-doneness that John Mackie thought to be no part of the fabric of the universe. It explains how G.E. Moore and other non-naturalists could argue so convincingly that ethical thought deals with non-natural properties. Many philosophers think that the right direction of explanation is the opposite.

An answer to the question of how to live, they would say, is just a belief as to what we ought to do and what it’s at least okay to do. Now, of course, anyone who says this has the burden of explaining what ought and okay mean. If they can’t, or if their answer involves strange and incredible things like non-natural properties, I then say that my ex– my direction of explanation is better.

I start my explanation with something intelligible, with decision and the valences and restrictions that get a person to the final stage where, if need be, he goes from indifference to picking something. So I now speak about intuitions again. I return now to the subject I started out with, to moral intuition.

I’m treating moral inquiry as inquiry into how to live and how to feel, how to engage people and their actions emotionally. Often, though, moral in– moral inquiry is conducted by consulting moral intuitions. And indeed, Sidgwick, Ross, and others have argued that moral reasoning couldn’t get off the ground without moral intuitions.

This alleged power of moral intuition Mackie attacked as incredible as a purported mode of knowledge that is unlike any other we know. How could we be in any position to intuit moral truths or normative truths in general? No answer is apparent in bi– in the biological picture I sketched.

The non-natural facts are absent from the picture, and so are any powers to get at non-natural truths by intuition. Interpreting the natural goings-on as thoughts and judgments doesn’t change this. If moral knowledge must depend on intuition, we seem driven to moral skepticism.

Intuitions would give knowledge, perhaps, if we had a kind of inner eye that peers into the non-natural layout of moral facts, but that’s not a picture to take seriously. Another stance we can take toward intuition is just not to worry. We rely on, we rely on intuition, after all, for mathematical knowledge.

And so why should morality be more constrained in the ways we can know it? Now, the question of how we have mathematical knowledge is difficult. Still, at least for arithmetic and geometry, mathematics is part and parcel of empirical knowledge.

The knowledge we get by counting, measuring, and the like. Our abilities to get numbers right, are aspects of our abilities to get right such empirical matters as the number of pebbles in a basket. If our abilities to get morality right were like this, there wouldn’t be the same puzzle about them.

There would be difficult philosophical explaining to do, as with our knowledge of arithmetic and geometry. But there would be no sheer mystery as to why evolved beings like us would have powers of veridical insight into the realm of morality. Another possibility would be that intuitions matter because the moral question just is what our moral convictions would be in reflective equilibrium when we’ve given adequate heed to everything that would affect our considered moral beliefs.

Moral intuitions would matter then as starting points for reaching reflective equilibrium. I’ll be claiming, though, that moral claims aren’t claims in interpreted psychology. The question of what we would think if such and such conditions obtained is mostly an empirical one, along with the further question of how to interpret the state we would then be in.

I have been saying that the moral question isn’t what we would think in such and such conditions, but what to do and how to feel about things we do or might do. These questions aren’t answered by interpreted psychology alone. Now, it might seem that I have escaped the problem of relying on intuitions.

If normative thoughts are plans or valenced restrictions on plans, then to come to normative conclusions, we just have to plan. This, however, doesn’t free us from intuitions. As we plan, we’ll weigh considerations for and against actions.

Take even a case of non-moral planning, uh, say, thinking whether to go to the store. In favor, I might weigh the consideration that there I can get rich chocolate cookies. I can go on to evaluate whether to weigh that consideration in favor.

I settle what to weigh into my decision and how, and form a string of considerations that support other considerations. At some point, though, the string comes to an end. Perhaps I weigh the fact that I’d enjoy a thing in favor of doing it, and on no further ground.

Uh, and perhaps I weigh the chance that I’d suffer, uh, if I, uh, ate chocolate cookies, I’d suffer from the health effects, uh, against a plan to get chocolate cookies. Weighing enjoyment in favor and suffering against, on no further ground, amounts to having an intuition about why to do things. Intuitions, then, apply to planning and not just to thinking how things stand.

If I keep challenging my thoughts about what to do and why, I end up grounding my planning in intuition. I accept then that normative thinking rests on intuition. This seems to raise the same question again.

And why think that we can intuit why to do things? Like questions go for thinking how to feel and why. Why think we can intuit a why and why not to feel certain ways about things?

But thinking of ought judgments as plans leads to an answer. I intuit, we said, that the chance that I’d suffer if I did a thing is reason not to do it. But to say that I have this intuition is just another way of saying that I confidently weigh the chance of suffering against doing a thing and on no further ground, even if I ask myself why.

This use of the term intuition, uh, is in an empirical non-normative sense, uh, to use it as Haidt does, as a certain kind of state of mind that is open to empirical study. We could instead use the term, though, in a normative sense. An intuition, we could say, is a state of mind of accepting something, uh, not on the, uh, not on the basis of further reasoning, even upon challenge.

Uh, and that it’s one that we ought to place some trust in. To think something an intuition in this sense is to plan to rely on it. I’ll call intuitions in the non-normative sense, in which they figure in psychology, a de facto intuitions.

Uh, these are judgments made confidently on no further ground, with no felt need for further grounds, even upon challenge. Intuitions in the normative sense, I’ll call intuitions de jure. These are de facto intuitions to rely on.

It’s a normative claim, then, that de facto intuitions are genuine intuitions, and it’s a claim we need, uh, I’m arguing, for coherent planning. So I now talk about ideal conditions. I’ve been stressing the distinction between non-normative psychological questions of how we do form moral judgments and normative questions of how we ought to.

What we will plan under what conditions is a psychological question, whereas normative questions are planning questions of what to do. The two are closely tied to each other, though. We can ask the planning question of when to trust our own planning.

We can ask under what conditions to trust our planning most. That amounts to asking what conditions are ideal for planning. Ideal conditions, we might conclude, involve such things as full information, vividly taken in and contemplated, and an alert, engaged, and dispassionate frame of mind.

If we come to a precise view about what those conditions are, we can then ask the psychological question of what in those conditions we would plan. I face a moral dilemma, suppose. I’ll give as an example a simple and far-fetched dilemma that I’ll talk more about tomorrow.

A father stands on the bank of a river where two canoes have capsized with children in them. His own daughter was in the far canoe, and he can rescue her. Alternatively, though, he could rescue two children of strangers who are nearer to him.

So he can’t do both. What ought he to do? The first is an ought question.

Uh, now we can ask another kind of question. How would we answer this first question in ideal conditions for judgment? If we get an answer to the second psychological question, we’ll come to an answer to the first.

Suppose I conclude under ideal conditions for judgment, I judge that he ought to rescue his daughter, even though that means only rescuing one child when he could have rescued two. Relying on myself as I judge in ideal conditions, I can now say he ought to rescue his daughter instead of the other two children. It’s not that the moral conclusion is entailed by a finding in, uh, interpreted psychology.

Rather, what’s going on is this: What we call conditions for judg-judgment, uh, well, what– When we call conditions for judgment ideal, we mean that those, uh, that judgments in those conditions are ones to trust. To accept this is to plan to trust such judgments.

So I accept the claim. In ideal conditions, I would judge that the man ought to rescue his daughter. So equivalently, I accept the second thing there.

Uh, the judgment that he ought to rescue his daughter is one to trust. Well, to accept this is to plan to trust this judgment, the judgment that the man ought to rescue his daughter. Uh, to trust the judgment means being disposed to emulate it in one’s own judgment.

So following through on the plan, I make the judgment, the man ought to rescue his daughter. If then we could settle under what conditions to trust our normative judgments, then we could come to normative conclusions on the basis of interpreted empirical findings. From the empirical finding that in those conditions for contemplation, I judge that the man ought to rescue his daughter, I could reason to judging that he ought to rescue his daughter and voice my state of mind by saying, “He ought to rescue his daughter.”

This isn’t deriving an ought from a psychological is alone, for there’s an intervening normative premise. The premise amounts to this: that what I’d find, uh, right or wrong in those particular circumstances is, uh, right or, or wrong. That what I would then think ought to be done, ought to be done Possibly then we could find a systematic way to move from psychological findings to moral conclusions.

In many kinds of cases, after all, a transition from is to ought is entirely obvious and uncontroversial. If you’re driving late to work and a child will be killed unless you stop, then you ought to stop. How to weigh a child’s life against arriving promptly at work is something we’ve settled beyond need for further review.

If the conditions under which to trust one’s normative judgment were similarly unproblematic, then the problematic parts of ethics would be reduced to questions of interpreted psychology. The move from is to ought still wouldn’t be one of entailment, but it might be systematic and trustworthy. We aren’t at that point yet, though, and if we did get there, it would still be important to distinguish ought questions from psychological questions to keep track of what we had achieved and what our basis was for accepting the odd conclusions we accepted.

So I’ll talk about co-coherence and inconsistency. Plans, I claim, require moral intuitions, but I need to make this claim more precisely. At the moment, I can find it clear that the fact that I’d enjoy something weighs in favor of doing it.

I can then rely on this as a premise without relying on the further psychological premise that I find this obvious. No thoughts about intuition enter into my thinking, and I haven’t skipped over any steps that would be needed to make my thinking explicit and fully cogent. Over time, though, I can plan what to do only if, at least implicitly, I do place some stock in my earlier conclusions without rethinking them.

I trust my earlier conclusions, and I can’t be justified in doing this unless the fact that I earlier found something obvious on no further ground is at least some reason to accept it. Planning requires thinking that the is of interpreted psychology that I implicitly accept an ought and would accept it explicitly if challenged on no, no further ground supports accepting the ought. I must not only have de facto intuitions, but I must trust them.

I must, I must treat them as intuitions de jure. I don’t mean, though, that de facto intuitions are to be trusted entirely. Seeming intuitions can clash, and indeed, seeming intuitions about what to do can clash severely.

The trust to put in them can only be defeasible. Even if moral claims didn’t mean what I say they do, and even if the visual model held good for intuitions of moral right and wrong, we’d have to test intuitions against each other and revise them in light of conflicts. Philosophical work on normative ethics, much of it consists in engaging in this refinement of intuitions.

But there’s no end of controversy as to where the weight of corrected intuition falls. I’m suggesting that we might get further by conceiving our questions as ones of what to do and how to feel about things and why. This won’t end our dependence on intuitions, but we could see if the intuitions we now rely on are more tractable.

Much of what I’ll be doing in the next lecture will go over ground familiar to moral philosophers, and we’ll have to hope that the resulting treatment makes contact with ordinary moral thought, or there would be little reason to trust what I was doing.

[01:01:10] AUDIENCE MEMBER:
Mm-hmm.

[01:01:11] ALAN GIBBARD:
A lot of what I’ll be saying tomorrow stems from decision theory and from arguments that decision theorists have made. We can think of decision theory as a systematic development of intuitions about what to do and why. Let me begin with a famously unintuitive result in decision theory due to Richard Zeckhauser.

So you’re forced to play Russian roulette, but you could buy your way out. Um, what is the most you would be willing to pay, the question is, to remove the bullet, reducing your chance of shooting yourself from one and six to zero? Or that’s the first question.

Uh, Once you answer it, we ask a more complex question. You’re instead, it turns out, forced to play a worse version of Russian roulette with four bullets in the six chambers. What’s the most you would pay, the question now is, to remove one of those four bullets?

And in particular, is it more or less than before? Most people answer less. But you should pay more, an argument from orthodox decision theory goes.

This problem is equivalent after all to a two-stage problem as follows. In the first stage, you’re forced to play with three bullets and no chance to buy yourself out. In the second stage, if you survive, you’re forced to play with two bullets, but you can pay to remove both.

Well, the amount to pay in the second case then is anything you would pay to remove both of two bullets if they were the only two bullets. Surely more than to remove one sole bullet. This case and others like it have been staples of debate on the foundations of decision theory, and ways out of this conflict of intuitions have been proposed.

Uh, the first thing to note, though, is that the intuitions and conflict are strong. In removing two bullets, uh, well, is removing two bullets worth more than removing one

(clears throat)

if you thereby empty the pistol? Uh, surely. Uh, does anything in, in these choices matter, uh, but the chance of surviving and how poor you will be if you do?

Well, not much. Uh, those seem to be the predominant considerations. It doesn’t matter then whether you must play the four-bullet game or the two-stage game since they involve choice among the same chances of death.

Does it matter if you choose at the start of the two-stage game what to play if you survive the first stage, or decide once it turns out that you’ve survived the first stage? Uh, clearly not. Orthodox decision theory goes against the, uh, intuition for this case, but any alternative orthodoxy will violate one of these strong intuitions, too.

I myself am convinced that what we gain in intuitiveness when we depart from the orthodox views and decision theory in cases like this are less than what we lose. That would be a long argument, though, and I can’t expect you to ex-accept this conclusion on my say-so. So, What I hope to convince you of is that some of our strong intuitions will have to go whatever we hypothetically decide to do in the Zeckhauser case.

In tomorrow’s lecture in particular, I’ll proceed as if the conclusions of orthodox decision theory, um, are right, but you should note that part of the argument remains to be discharged, and it is controversial among the experts whether it can be. So let me speak on why coherence in action matters. The constraints of decision theory are ones of consistency in action or something close to it.

Whether they are full-fledged matters of consistency is a tricky question, and so I’ll use the word coherence. Uh, why though does coherence and plans for action matter, especially when they’re plans for wild contingencies that we will never face, like being forced to play any of several versions of Russian roulette? With questions of fact, the problem with inconsistency is that when we– when a set of beliefs is inconsistent, at least one of the beliefs is false.

I’m suggesting that we think of ought questions in the first instance as planning questions. Answers to them may, in the end, count as true or false, but we don’t start our treatment with talk of truth or falsehood and help ourselves to these notions in our initial theorizing. With incoherent plans, I accept, the oughts we accept in having those plans can’t all be true, but that isn’t at the root of what’s wrong.

As a first approximation, I can say incoherent plans can’t all be carried out. If I plan to be here today and also plan to be on top of Mount Kenya, believing that I can’t be both places on the same day, my beliefs and plans are inconsistent. Either then, my belief that I can’t be in both places is false, or one of my plans I won’t carry out, no matter what choices I make.

Some of the plans I’ll be talking about tomorrow, though, are wild contingency plans that I’ll never be in a position to carry out anyway. I might talk about such wild plans as plans for what to prefer in the con-contingency of, say, being Brutus on the Ides of March. I won’t talk about that, but I could.

Uh, and some of the states of mind that can be coherent or not with others won’t be simple plans, but constraints on plans or beliefs. I plan to pay more, forced to play Russian roulette, to empty the pistol of two bullets than of one. The problem with inconsistent plans is that there’s no way they can be realized in a complete contingency plan for living.

For each full contingency plan one might have, something in the set will rule it out. Or more completely, we’ll have to talk about an inconsistent, uh, combination of beliefs, plans, and constraints. If a set of these is inconsistent, there’s no combination of a full contingency plan for living and a full way the world might be that fits, And judgments get their content from what they’re consistent with and whatnot.

So to preview, uh, today I’ve contrasted biological thinking about us and the normative thinking that the biological picture has us engaging in. A rich enough biological picture, I think, explains why a highly social linguistic species like ours would engage in normative thinking and discussion, and in moral thinking and discussion in particular. I also talked about intuitions.

We couldn’t coherently proceed with normative thinking without giving some trust to some of our de facto intuitions regarding them as intuitions de jure. Indeed, I would claim that this applies to thinking of all kinds, but I haven’t gone into that today. At the same time, some of our strong intuitions are inconsistent with each other, and so our trusted de facto intuitions, to be coherent, must be guarded.

In tomorrow’s lecture, I’ll take this, uh, very high-level, normative thinking about intuitions and reasoning and turn it to morality. Our lives are social, and a large part of thinking what to do and how to feel is thinking how to live with other people. We address these questions partly each by ourselves and partly together in discussion.

I’ll keep my eye on moral thinking as thinking how to live with each other and on the question of how to regard our moral intuitions. Tomorrow, though, I’ll be engaging in the kind of thinking I’ve been describing. In particular, I’ll be, I, I’ll pursue a kind of moral argument that may be very powerful, but that raises difficult questions.

This is an argument that owes the most to Berkeley’s late John Harsanyi. It leads to conclusions that clash with strong moral intuitions, and I’ll be trying to think through the force of those intuitions. Tomorrow, then, instead of just describing moral thinking as thinking how to live with each other, I’ll engage in moral thinking in a reflective and highly theoretical way.

Thank you.

(applause)

[01:12:12] SAM SCHEFFLER:
Uh, we’re going to have some comments from, uh, Michael Bratman, but before we do that, let’s, uh, take a short break, just five minutes or so, and then we’ll resume. As the Chancellor said, we’re fortunate to have three wonderful commentators who’ve joined us for this week’s events: Michael Bratman, John Broome, and Frances Kamm. Professor Bratman will make some comments now, uh, and Professors Broome and Kamm will comment on tomorrow’s lecture, and then all three commentators will offer additional observations at the Thursday seminar.

Michael Bratman, today’s commentator, is the U. G. and Abbie Birch Durfee Professor in the School of Humanities and Sciences and Professor of Philosophy at Stanford, where he’s taught since receiving his PhD from the Rockefeller University in 1974. Professor Bratman is one of the world’s most important and influential philosophers of action. He’s written widely on issues of intention, agency, and practical reason.

His 1987 book, Intention, Plans, and Practical Reason, is a contemporary classic, and he’s also the author of two volumes of collected essays, Faces of Intention and Structures of Agency. We’re very fortunate to have Professor Bratman with us this week. Please join me in welcoming him.

(audience applauding and cheering)

[01:13:34] MICHAEL BRATMAN:
Thanks, Sam, for the introduction. It’s a pleasure to be here. Allan Gibbard’s overall title for this fascinating pair of lectures is Thinking How to Live Together, and I think a good place to start is the ambiguity of in this title.

We can talk about thinking, perhaps thinking conducted by a single individual, that has as its content how we are to live together, and we can talk about thinking together, a kind of language-infused shared activity about how to live. Gibbard evidently has both ideas in mind. One of my concerns is with how these ideas interact.

Begin with the main outlines of the views presented this afternoon. We face questions of what to do and why. That we face these questions is a fact about the kinds of beings in the natural world that we are.

However, a full and na-adequate naturalistic biological story will not, Gibbard avers, contain any fact that our answers to these questions of what to do and why are right or not. Here, Gibbard agrees with G.E. Moore. Gibbard also issues appeal to non-natural facts about what we ought to do, facts that are not part of the naturalistic story.

Here, Gibbard disagrees with Moore. Nevertheless, as Gibbard sees it, our answers to these practical questions do constitute a kind of judgment. Our judgments about what to do are, to a first approximation, plans about what to do.

And judgments about what to do are the basic case of judgments about what one ought to do. Ought thoughts are plans. To make a practical normative judgment, a judgment about what someone ought to do or has a reason to do is at bottom to plan.

This may sound jarring. Gibbard’s talk of plans is not just talk of plan-like contents. A recipe is a plan-like content, a content one might think about without planning to act on.

Talk of plans here is talk rather of attitudes of planning to. Such planning attitudes aim at changing the world, at making the world fit them. But judgments try to fit the world.

In this sense, judgments and plans have different directions of fit. So how can plans be judgments? Well, Gibbard would acknowledge some such distinction, at least initially, between ordinary empirical beliefs about the natural world and attitudes of planning to act.

(coughs)

(inhales)

The view is that nevertheless, an important subset of these planning attitudes behaves in a sufficiently belief-like way to count as judgment. Given the temporal extension of our thought and action, an agent needs to some extent to trust her earlier plans in her later planning, just as she needs to trust her earlier beliefs. Both planning and belief involve a kind of cross-temporal self-trust.

And just as beliefs are subject to constraints of consistency and coherence, so are one’s planning attitudes. These constraints are to some extent specified in formal decision theory, and as the Zeckhauser Russian roulette example is intended to show they can lead to striking results. Further, just as one’s beliefs are subject to pressure towards agglomeration, towards putting them together into an overall view of the world, the various planning attitudes of an agent are subject to pressure towards constituting at least part of a complete contingency plan for living.

We can ask which possibilities are and which are not consistent with one’s overall plans, just as we can ask which possibilities are and which are not consistent with one’s set of beliefs. And a basic tenet of interpretation is that, quote, judgments get their content from what they are consistent with and what not. All this is offered in support of the idea that we can see plans as sufficiently belief-like to count as judgments.

I’ll come back to this. First, let’s explore the various forms of planning that Gibbard identifies with various kinds of normative judgment. What’s moral judgment?

On the theory, moral ought judgments are, to a first approximation, plans about what moral sentiments to have, in particular, plans about what to resent people for doing and what to feel guilty for doing. Where these emotions of resentment and guilt play a central role in our social world. What about judgments not directly of what one ought to do, but of what there is reason to do?

Well, we not only plan what to do, we plan what is to count and in what ways in our deliberation about what to do. We have plans about weights and deliberation. Where, to avoid circularity in the account of the idea of a reason, it seems we need to understand deliberation as a mental activity that can be characterized in a basic case without appeal to that idea of a reason.

Judgments about reasons are plans about what to weigh in deliberation, so understood. This gives us a way of understanding the idea of an intuition, understanding it without appeal to some special power to apprehend something non-natural. Some plans about what to weigh are basic, in the sense that on reflection, we don’t support them by appeal to yet a further plan about weights.

Such a basic plan about weights amounts to an intuition about why to, about why to do things. In that sense, we ground our normative thinking, that is our relevant planning, in intuition. Such an intuition is a normative judgment.

It’s not just a tendency to give weight to some consideration, avoidance of suffering, say. It involves a plan to give such weight, and in that way, a plan to rely on one’s tendency to give weight. Such intuitions are not just de facto intuitions, they’re de jure, and we can express their content as something like: such and such is a reason to act in certain ways.

We also have plans about when and under, under what conditions to trust our planning, and so when to trust our ought judgments. We might plan to trust our planning under, say, conditions of full information vividly presented. Such plans for trusting planning are judgments about what conditions for planning are ideal.

Given such a judgment about what are ideal conditions, that is a plan for trusting planning in those conditions, I can ask, if I were in those ideal conditions, what would I judge I ought to do in certain circumstances? What, for example, would I, in ideal conditions, judge I ought to do if I were one of the fathers in the example Gibbard sketches? If I determine that I would, in those ideal conditions, judge that I ought to save my own child, then if I follow through on my plan to trust judgments made in those ideal conditions, I arrive at the judgment that I ought to save my own child, where that judgment involves a plan so to act in this scenario.

So I can, in a way, reason to an ought conclusion from a factual statement about what I would judge in certain ideal conditions, but only because in the background is a plan to trust judgments made in those ideal conditions. So my ought judgment’s not reducible to a naturalistic claim in interpreted psychology. As Gibbard notes in, in his, uh, two thousand and three book, Thinking How to Live, This leaves open the possibility that I think that what I would judge in ideal conditions differs from what you would judge in ideal conditions.

As I see it, in such a case, there’s no judgment we would reach in ideal conditions. We’re at an impasse. This planning theory of normative judgment requires a qualification.

We sometimes decide between options, neither of which we think strictly superior to the other. Perhaps I think two routes from Palo Alto to Berkeley are equally desirable. Maybe I should say equally, uh-

Well, you know what I mean. Yeah. Okay.

Or perhaps, as in Sartre’s famous case, I must choose between staying home with my ill mother and fighting for the Free French. And I think the conflicting considerations are incommensurable, and so neither option is superior to the other. In each case, we can suppose I reach a decision, so there’s one option I plan to perform.

But I don’t think I ought to do specifically what I plan to do. What I think is only that I ought to perform one or the other of the options. So planning to act does not ensure thinking I ought so to act.

This leads to a complexity in Gibbard’s theory. To think and act okay, he says, is to rule out preferring any alternative. To think one ought to act is to think it’s okay

So to act and o– not okay not to. To rule out a preference is, I think, the view is, it– to rule out a preference is to plan not to prefer. Where to prefer, preference here is strong preference.

To prefer X over Y involves not preferring Y over X. Thinking it okay to do something is to plan not to prefer its alternatives to it itself. In our first example then, I plan not to prefer route one over route two, and I also plan not to prefer route two over route one. I think each route okay.

Thinking I ought to do s– to A is to plan not to prefer A’s alternatives to A’s it- to A itself, and to plan to prefer A to its alternatives. And I take it Gibbard’s assuming that if I plan to prefer A to its alternatives, I thereby plan to A. So this new story of thinking I ought as a complex plan to prefer matches the original story of thinking I ought as a plan to act.

But the new story aims to work better for cases of ties and thinking only that an option is okay. That anyway, is the theory as I understand it. Let me now briefly raise some questions.

Some touch on issues Gibbard’s discussed further in his, uh, in both of his books, but mo-most recently the two thousand and three book, and my aim is to create a context in which he can tell us more. First, what should we say about cases of thinking one ought to A, and yet, out of weakness of will, planning instead to do something else? Perhaps I think I ought to limit myself to one glass of wine.

I have no commitments about tonight. Um, yet I give into temptation and drink a second glass. It seems that, sad to say, my thought that I ought to stick with one glass fails to involve my planning to stick with one glass.

That’s my problem. Second, we turn to cases of ties and thinking okay. Suppose the boy in Sartre’s case finds himself with a slight felt preference for mom, but he still thinks the options involve incommensurable considerations, and that neither option is uniquely what he ought to do.

He thinks it’s okay to perform either option, but he doesn’t plan not to prefer mom over the Free French. Indeed, he continues slightly to prefer mom. In light of these two kinds of problems, perhaps we should put more theoretical weight on plans for weighing rather than plans to act.

On the theory, remember, plans for weighing constitute judgments about reasons. When I think I ought to stop with one drink, my plans for weighing favor one drink over two, and I know this. Perhaps a theory like Gibbard’s can say that my thought that I ought to stop with one drink expresses this known structural feature of my plans for weighing, that they weigh in favor of one drink over two.

And it can do that even if I plan to have a second drink. That plan is a plan for action, not a plan for weighing. Again, the boy in Sartre’s case has plans for weighing that leave unsettled his decision between mom and the free French.

And he knows this. Perhaps a theory like Gibbard’s can say that his thought that either is okay expresses this known structural feature of his plans for weighing. And this structural feature is compatible with his preference for mom, since a preference need not be a plan for weighing.

Return now to the basic idea of identifying normative judgments with plans. If you judge one ought to A in circumstance C, and I judge one ought not to A in circum–

(coughs)

in that circumstance, we’re not merely differing, as we might if we have different tastes for flavors of ice cream. We’re disagreeing. But if you plan to A in C, whereas I plan to refrain from A in C, we are, it may seem, not disagreeing, but just differing what we each plan to do if we’re in C. Well, the features of plans Gibbard has highlighted today as supporting the idea that plans are judgment-like, which we recall were pressures for consistency, coherence, agglomeration, self-trust over time, assignment of content by appeal to what’s consistent with one’s plans.

Well, those pressures also explain why you and I may be disagreeing, not just differing, in plan. Well, so far these features are individualistic. They explain perhaps why when I change my plans, I’m in a certain case disagreeing, not merely differing with my earlier self, since I’m kind of going up ag-against my self-trust over time.

But it seems that we do not yet have the resources to explain why you and I may be disagreeing when I think we ought to do A and C, and you think we ought not. It’s here, I take it, and this returns to where I started in these comments. It’s here, I take it, that Gibbard would appeal not simply to my thinking about how to live and to your thinking about how to live, but to our jointly thinking together about how to live.

When I change my own plans over time, this can count as disagreement with my earlier self, in part because, given the temporal extension of my thought and action, I need to see my earlier plans as having a defeasible claim on me now. In contrast, it’s certainly possible for me not to treat your ought thoughts as making any such claim on my ought thoughts. But insofar as we’re engaged in thinking together about how to live, I will see your thoughts as making a claim on my thinking, a claim that’s analogous to the claim of the thoughts of my past self on my present thinking.

There is then pressure to keep track of the extent of agreement between us, just as there’s pressure on me to keep track of the extent of agreement between me now and me earlier. So our differences in plan can constitute disagreements in plan, which is the conclusion we wanted to reach. And I take it Gibbard would say that there are selection pressures in favor of dispositions on the part of a highly social linguistic species like ours to engage in such shared thinking about how to live.

The idea then is, as I understand it, is that the s– it is the sociality of normative thinking rather than independent facts, natural or non-natural, about oughts and reasons of the sort appealed to by the normative realist. It’s the sociality of normative thinking that explains the nature of interpersonal normative disagreement as disagreement in plan. So let me close with two brief questions about this.

First, suppose you think you are, in the sense I pointed to earlier, at an impasse with another person, and for that reason, you opt out of shared thinking with that person about how to live. Can Gibbard make room for your thought that nevertheless, in your normative judgments, you are disagreeing, not merely differing with that other person? You’re not part of this shared thinking, but you still think of yourself as disagreeing.

The second question I have concerns just how to think about this shared thinking, and I’ll, I’ll finish with this. Consider shared thinking in particular about reasons for action. In such shared thinking, as it is understood within Gibbard’s theory, are the participants trying to converge on a coordinated shared standpoint of reasons?

Or is each simply trying to arrive at the best set of plans for weighing, in part by being responsive to the views of others, and so by keeping track of the extent of agreement, but not in the pursuit of agreement? In short, in such joint thinking about how to live, as it functions in Gibbard’s theory, do we track agreement in a sense that involves aiming at agreement, or do we only keep track of the extent of agreement?

[01:30:00] ALAN GIBBARD:
Okay.

(applause)

[01:30:11] SAM SCHEFFLER:
Alan, do you wanna say something? You can either speak from here or there as you wish.

[01:30:21] ALAN GIBBARD:
Well, I want to thank Michael Bratman for, uh, a wonderful set of comments which, uh, are, are marvelously clear and, and accurate. Uh, actually, when I was, uh, beginning to develop this theory, I was at the Behavioral Science Center at Stanford, and, uh, Michael and I had weekly lunches, and he, uh, pressed me on various aspects of the theory. And a lot of what I, uh, ended up saying was very responsive to his, uh, pressing me.

Uh, in a way, I’ve, uh, you might say I’ve drawn away from some of the positions that I, uh, was or some of the, uh, answers I was trying to give under the, under his pressure, and, uh, we’ll have to, uh, ask why. Uh, also the, uh, uh, thinking about planning owes, uh, an immense amount to Michael, although Michael, uh, makes many distinctions about planning that, uh, he, he may feel that I’m riding roughshod over. Uh, so, uh, central questions about my approach, uh, trying to understand, uh, normative judgments, ought judgments as, uh, as, uh, in a way of like, uh, plans or plans for valences and, uh, restrictions on plans, uh, weakness of will.

So I plan, uh– So I think I ought not to have the, uh, second glass of wine, but I do have it. If I do have it, then I’m, uh, not, we could say, planning at that very moment, uh, not to have it because, uh, a feature of, uh, what my plan is that if, uh, if my plan really says to do such and such, uh, avoid the se– don’t take the second glass uh, at this very moment, uh, then that’s what I do.

Uh, well, it’s always seemed to me that it isn’t terribly clear what’s going on in situations like that. Plato thought that I always do what I think best, but that I, uh, under the pressure of temptation for the second glass, my, uh, The, uh, judgment that, uh, drinking– that it’s best not to drink the second glass of wine, uh, uh, uh, wavers, um, seem to me that’s, uh, not a bad description of what’s happening in cases like that. So the question isn’t whether I, uh, think as a stable matter, uh, that I, uh, ought to or think over the long term that I ought not to take the second glass of wine.

All it takes is, uh, all it takes is, uh, losing my conviction that I ought not to at the very moment of, uh, taking the second glass of wine. But also my, uh, my, uh, state of mind in that situation, we, we might say isn’t, uh, one of being very clear in what I plan. So I have a general plan, uh, never to take a second glass of wine and not to take a sec-second glass of wine and, uh, in particular, uh uh, this evening, uh, and then, uh, at a certain point, I take it, uh, do I, uh, plan not to?

In, in a way, I’m at a uh, I’m– I’m like– it’s not, uh, likely that I have a uh, a clear sort of change of plan. I, uh, I get into a state where it isn’t very clear whether I’m, uh, planning not to have this glass, but somehow having it anyway or, uh, having it. Um, now in the– now under partly, uh, under pressure from, uh, Michael’s cross-examinations, and my, in my first book, I, um, I took the view that, well, there are two sort of control systems here.

They, uh, uh… There’s one control system that’s a control system from accepting norms. I go on accepting the norm that says, “Don’t have more than one glass of wine,” “but I have it anyway.”

Uh, now I, uh, now other systems of control, my responsiveness to social pressures, my appetite, uh, take over. So my normative judgments are the judgments of one subsystem. But we can ask, uh, maybe I should go back to thinking that, but we can ask, uh, Uh, well, do we have such, uh, distinct subsystems?

Uh, it seems more of what’s happening is that I… I’m just not, uh– I’m just of more than one mind in my planning and one part of the planning I’m in– and my long-term planning doesn’t prevail at this point.

Uh, what we should say I think I ought to do at that particular moment, uh, just isn’t all that clear. Although, uh, a minute before, it’s clear that I think I ought not to, and then later that night, it’s clear that I think I ought not to. So that’s, uh, so I, I, I guess I, uh, think that since, uh, so-called weakness of will is a kind of wishy-washy state, uh, maybe it requires a wishy-washy description.

Uh, now, for, uh, for, uh, for the Sartre case, uh, again, under, uh, in my first book, under, uh, pressure from conversations with Michael, I had, uh, an account of what, uh, thinking that, uh, rationality doesn’t determine whether I should, uh, tend to my mo- tend my mother or join the resistance. Uh, I had a price- precise account of what, uh, that state was, uh, consisted of, uh, accepting certain higher order norms. Uh, it seemed to me, though, that there’s a kind of, uh, uh, there’s a problem with the view that, uh, rationality doesn’t settle the matter.

Uh, Then suppose I, uh, suppose I so conclude, but I decide to join the resistance. I, uh, I get on the train to where the rendezvous is. Uh, uh, after all, it’s something that is rationally permissible to be doing.

I arrive there, and, uh, it’s still rationally permissible to tend my mother. When I get to the rendezvous, uh, I say, “Well, it’s rationally permissible to tend my mother,” so I get on the, uh, train and go back to where my mother lives, and then I think, uh, but it’s still rationally permissible to join the resistance, so I get on the train and go to where the resistance is. Uh, are we to think that this is a, a perfectly–

that I, I’m at every moment, uh, perfectly rational? Well, that seems like a, an incoherent way of, uh, proceeding over time. I spend all my time on the train, my mother is neglected, the resistance is joined.

Uh, uh, yet it seems to be allowed by the view that, uh, there’s, uh, nothing rationally compelling about either course of action. Uh, am I not doing something rationally permissible at every point? So, uh, in my first book, I tried to allow for this.

Now I’m not sure I should be trying to allow for this. Ah. Hi.

As Michael, uh, sketched, I, uh, I argue that, uh, it’s often important in our planning to put our heads together, uh, thinking together on how to live. Now, there are two sorts of spirits in which we can think together on how to live. One is trying to come to a joint plan, uh, Uh, that’s what’s going on, I think, in Charles Stevenson’s talk about, uh, disagreement and, and attitude.

Now, I’m talking about, uh, something, uh, maybe less common for us to put our heads together about, but thinking about, uh, the situation of a particular person and asking a ki– a hypothetical question. Uh, if one is in that person’s circumstances, uh, what to do, what to, what to prefer. Um, now that’s a different sort of question because we might decide that in the case where, uh, uh, I can gain some slight advantage at, uh, at your great expense, and you can gain some slight advantage at my great expense, the thing for me to do is, uh, look out for number one.

The thing for you to do is to look out for number one, even though we will both be, uh, uh, both be left in the lurch that way. Uh, so it’s coherent to think about what to do in your case and, uh, to think together about what to do in your case and come to a conclusion that isn’t at all in accordance with what we would, uh, decide if we were thinking about, uh, thinking together about a, a joint plan of, of action. Uh, so what I’m talking about is, uh, is the latter.

I hope that the latter will coincide with the former, that, uh, the question of what to do in my position and what to do in your position comes out as, uh, congruent with, uh, what to want us both to do. Uh, but I don’t think that that’s, uh, guaranteed by the, the logic of the, the normative concepts. Um, so in the, uh, in the dormitory bull session, the plan to, uh, always do what’s got the most in it for you, uh, may be one that, uh, we adopt, uh, thinking together about everybody’s circumstance.

But when it comes to the question of, uh, who brings the beer, there’s a lot of pressure toward, uh, toward joint planning, and a lot of what goes on in life is the interplay between joint planning and individual planning of that sort. Uh, so I thank Michael very much again for, uh, this very clarifying commentary.

(applause)

[01:42:17] SAM SCHEFFLER:
We have time for some questions if, uh, if people would like. Yes, Jay Wallace.

[01:42:23] JAY WALLACE:
Uh, yes.

[01:42:24] SAM SCHEFFLER:
Oh, I, I think, I think Ellen wants you to come stand by the microphone, if you would. Is that right, Ellen? Is that what you’re gesturing about?

[01:42:30] ELLEN:
Yes.

[01:42:32] JAY WALLACE:
Well, I, I formed a plan of not attaching much weight to well, to Ellen’s preferences, but victory.

(laughter)

But this is weakness of will, uh,

(laughter)

uh, ju-just a, a, a few questions that are more in the way of requests for clarification about the theory. I wanted to start with moral judgments, um, and, and, uh, their, the relation between planning and, and moral sentiments like resentment and guilt. Um, I guess I f-

I still feel hung up, um, on the problem that it, it, it seems infelicitous in general to talk about planning to feel emotions, um, and to the extent that seems an infelici- just insofar as the emotions don’t seem subject to direct volitional control. Insofar as that’s infelicitous, the, the whole, uh, way of unpacking, uh, these judgments of moral wrongness in terms of plans to feel resentment or guilt, um, has, has an um, you know, an air of implausibility about it to my, to my way of, um, of hearing things. It just– I,

I can’t, um, myself reconstruct a situation where I think I’ve actually formed a plan to feel emotions in different kinds of circumstances. It doesn’t seem like the sort of thing we plan to do. And so, uh, an account of moral emotions that’s, uh, of, of moral judgments that’s cashed out in these terms seems, seems peculiar to me.

I mean, uh, I, I think you adverted to a situation of this kind, but, but, you know, if I, if I’m planning to do something while thinking that it’s wrong, um, you know, it’s like, it doesn’t seem right to think of that as a case where I’m, I’m planning to do something, and I have a kind of codicil to the plan, plan where I, I, I’m, I’m planning to feel guilt when I do it. Um, just doesn’t seem like the right way to think about the, the situation. And, and in general, I think I’m still hung up on the, on the fact that the, that the emotions, insofar as they’re not really subject to direct volitional control, don’t seem like proper objects of planning agency in the first place.

Uh, another related question to this part of the theory, and, and maybe I’ll stop there for, for now, is about the emotion side. You’re talking that we’re, we’re, we’re– on the account, we’re, we’re, um, making plans for emotions like guilt and resentment, uh, and that’s supposed to explain, I guess, moral judgments of, of, of, of, wrongness and, and so on. But, but to my– at my kind of naive way of thinking, resentment and guilt seem to presuppose judgments of wrongness in the first place.

You don’t– The difference between resentment, for instance, and other kinds of disappointment or anger you might feel about something as precise is, is connected to the idea that that particular emotion, um, presupposes a kind of, uh, I would have thought, uh, normative judgment. I, I– that seems to me the phenomenologically kind of plausible way to, to think about resentment.

But if that’s true, then it looks like, um, you know, we can’t analyze these judgments, uh, judgments in terms of the emotions. The emotions themselves presuppose a kind of, um, account of, of, of, uh, the moral, moral judgments that, that, um, that isn’t available, uh, given the way you’re trying to yourself unpack these judgments. So could you say a bit more about those issues?

[01:45:45] ALAN GIBBARD:
Um, well, on, on the, uh, talk of plans for feelings, I, I, uh, I agree that, uh, with the, the ordinary usage of plan is, uh, confined to things that could be carried out by, by choice. Now and like, uh, suppose I don’t know, uh, uh, I’m puzzled about how to feel about something that somebody did that was disadvantageous to me. Uh, if I come to an answer, which I might express by saying, “Well, it, uh, doesn’t make sense to resent him for that,” uh, then I’m in a kind of mental state that, uh, uh, in a way, uh, operates somewhat as, uh, plans do for action.

Uh, uh, there’s some tendency to carry out one’s plans for action, but, uh, you don’t invariably carry out your plans for action. There’s some tendency to not to feel resentment when you are in the mental state of thinking it doesn’t make sense to, uh, feel resentment for that, but, uh, you may find yourself resenting it anyway. Uh, so there’s a, uh, a kind of analogy between, uh, judgments of, uh, how to feel about something that seems like nat- uh, natural language to me and, uh, and, uh, how one does feel that’s, uh, that’s analogous to the connection between, uh, uh, judgments of what to do in a situation and what one does do.

So I think that it’s only, uh, that degree of analogy with plans that I, uh, really need for this. Now, uh, certainly if we can’t understand, uh, resentment, uh, apart from, uh, understanding, uh, moral right and wrong, uh, prior to that because, uh, resentment involves, uh, a judgment that something is morally right or wrong, uh, then, uh, um, then this, this won’t work. Um, we’ve got to remember, of course, that explaining what, uh, moral right and wrong, uh, what judgments of moral right and wrong have as their content has been, Ah, a baffling problem.

Philosophically, it’s given rise to some, ah, theories that, uh, people wouldn’t take very seriously, and the ones that, uh, or that it’s hard to take seriously and, and the ones that, ah, the ones that we can take seriously seem to have problems with them. So, uh, if I’m going to res– uh, if in order to resent something, I have to judge that it’s wrong, presumably it’s, uh, going to be in a sense that I, uh, don’t have a proper explanation, uh, of. But still, that’s possible that, uh, uh, the person who resents has to judge something’s wrong, even though the, even though, uh, philosophers can’t come to a good account of what it is to, uh, to do that.

Um, but I think, I think we should understand, uh, resentment as a, um, as an em-emotional reaction that’s, uh, continuous with the, uh, emotional reactions of, uh, of, of beasts with the, uh, the, the kind… the anger of a dog barking or something like that. Uh, so what’s the difference between, uh, uh, resentment or outrage and, uh, and anger? Well, I think resentment or outrage are, uh, sort of species of anger, uh, and, uh, some, uh, uh, and I think that our, uh, our moral concepts would work if we, uh, say that, uh, uh, the question of whether something is reprehensible really is the question of whether to be, uh, angry about it, uh, or in a somewhat cooler state and the, with the, uh, emotional flavor of anger, uh, from a, uh, presumably we’d have to say from a, uh, detached, impartial point of view if one’s engaged in it, uh, uh, emotionally.

And, uh, likewise, I would want to explain guilt as a, uh, in the case of guilt, presumably it’s not a, uh, not a feeling that’s, uh, we know to be continuous with, uh, feelings in the, uh, in the beasts, although, uh, I guess some, some dog owners might attribute guilt to dogs. Uh, I think that’s probably a, a pretty human emotion, but I think we can, uh, understand it as a kind of a natural psychic kind that has, uh, certain sorts of, uh, uh, tenden- tendencies to be aroused in very– in certain ways, tendencies to, um, uh, produce, uh, appeasing actions, uh, uh, demonstrations of how bad one feels, uh, about what one has done. Uh, so, uh, I think the way to explain emotions is not to start with some, uh, some cognitive content that nobody can explain and say that it’s the, the feeling that something, uh, has that, uh, uh, is that way, but to, uh, give a kind of, uh, uh, a symptomology– I mean, give a… say what the symptoms are and say that it’s the, uh, the state in, maybe in human beings or maybe in human beings in a certain sorts of cultural circumstances that, uh, has that pattern.

[01:52:44] MODERATOR:
Uh, Niko Kolodny.

[01:52:54] NIKO KOLODNY:
Um, this is, I think, just a point of clari– a question, uh, for, uh, clarification. I wasn’t exactly sure why on your theory as it now stands, um, there’s, um, necessarily pressure to… I mean, why you have to say if, um, I believe all things considered that I ought to X, and I’m completely wholehearted and clear about that, why it necessarily fo- why there, why it should follow that, uh, um, I decide to X.

I mean, it seems that now as the theory stands, if I believe that I ought to X, then I believe it’s okay to X, and I believe it’s not okay not to X. And that consists in my not, uh, my ruling out preferring any alternative to X and my ruling out ref– Uh, I guess, failing to prefer any alternative, um, to not X. Um, but that’s as far as we get, and unless we hear more about what it is to plan to prefer something, it isn’t clear why it then follows that, um, you in fact plan to X.

So you might be right that there really aren’t cases of, uh, a wholehearted weakness of will, but it wasn’t clear to me why, uh, your theory pushed you in– as it now sta-stands, pushed you in, in that direction. Um.

[01:54:23] ALAN GIBBARD:
Well, well, one way to put it would be if I, uh, rule out not doing something that, uh, uh, that amounts to doing it. I don’t really rule out omitting it. But you rule it…

uh, at, at this very moment if I, uh, uh, unless I actually set out to do it, except that, uh, I may be of more than one mind about that and, uh, and it may be the timorous part of me or the, uh, or the careless part of me or something that, uh, that seizes control at the moment.

[01:55:03] NIKO KOLODNY:
Can I say something?

[01:55:04] ALAN GIBBARD:
Yeah, sure. Sure.

(laughter)

[01:55:10] NIKO KOLODNY:
I, I could see why, I could see why ruling something, I could see why ruling something… If I ruled something out, why ru-ruled out not doing something, then I’m committed to, I mean—

[01:55:20] ALAN GIBBARD:
Mm-hmm.

[01:55:21] NIKO KOLODNY:
Then I have decided to do it. But why, if I rule out preferring any alternative to something? I mean, that’s the, that’s the step, the step from judgments about planning to prefer to, um, uh, sort of facts about planning to prefer to facts about, um, planning to do.

That, that was the thing that I was confused about.

[01:55:41] ALAN GIBBARD:
Well, I, I, I’m not sure whether I said something inconsistent with this, but, uh, but to, uh, think I ought to do it is to plan to prefer it to all the alternatives. Uh, so the, uh, if I think that something is, uh, let’s see. But the question is why thinking it’s not okay, uh, amounts to that.

Well, uh, uh, Well, let’s see. Think-thinking it’s okay rules out rejecting it from preference. Uh, Th-thinking that it’s not okay not to do it, uh, consists in, um, um, uh, ruling out, ruling out not, uh, uh, let’s see.

Sorry, I’m losing that train of thought. Michael, did you want to?

[01:56:53] MICHAEL BRATMAN:
Could I just jump in for one second? ’cause I, I… Actually, this came up really qui– uh, something like this came up in my comments, and I said I thought Alan was supposing, and you could see why he would suppose. Is this okay? Can I– is this working?

[01:57:06] ALAN GIBBARD:
Yeah, that’s-

[01:57:06] MICHAEL BRATMAN:
That, um, that if you plan to prefer to do A, that involves planning to A. And so that’s how you get from the more complicated story that’s introduced to deal with ties and so on to the original story. And so I, I thought that’s really the s- direct answer.

Now, maybe they shouldn’t be assuming that, but that seemed to be… And so I’d just like to know, is, is that right? I mean-

[01:57:34] ALAN GIBBARD:
Yeah. Yes. Yeah.

And, and I think that’s part and parcel of the structure of the theory. Can I just say one more thing? Yeah.

That is, there tends to be in the theory a very close. Well, you can see it when he talks about planned consistency as the standard… well, this is from the second lecture, but you got the idea.

(laughter)

The plan consistency in terms of, um, the kinds of consistency coherence constraints that decision theorists worry about. ’cause the cons- the consistency coherence constraints decision theorists worry about are basically on preferences, right? But the way Alan’s thinking about plans is there’s really a kind of translation from planning to do something and, and preferring to do it.

So when you have… So, so that’s sort of they’re not being pulled apart, whereas what you were suggesting was maybe we could pull them apart more. But okay, yeah.

[01:58:20] SAM SCHEFFLER:
Uh, Frances Kamm.

[01:58:26] FRANCES KAMM:
Maybe I shouldn’t ask a question now because I’m supposed to ask them on Friday.

[01:58:30] SAM SCHEFFLER:
It’s okay. Thursday?

[01:58:33] FRANCES KAMM:
I was just wondering, Alan, how you deal with supererogatory conduct. Namely, I think that… I mean, in your sense of ought, uh, I’m not sure.

I certainly don’t think that I ought to do, give up my life to save, you know, five other people. But it may come to the, to it that I actually prefer to do that than what I think, uh, maybe there are some things that I think I ought to do. Maybe they’re m-minimal morality or something, call the police to save them or something of that sort.

But how do you deal with cases where you prefer to do something and you give moral weight to it because you think it’s morally better, but it doesn’t actually seem to fall into the class of what I think I ought to do, and yet it is something that I prefer to what I think I ought to do. I’m not sure. I–

It’s just a question. I was wondering how that fits in.

[01:59:24] ALAN GIBBARD:
Uh-huh. Well, I’ve, I’ve been talking about morality in a, in an extremely narrow sense here that has to do just with, uh, with right and wrong, what one morally ought to do. Oh, and I would think to, um…

And that, uh, uh, leaves out part of what’s morality in a pretty narrow sense, namely the kind of positive end of, uh, the moral approbation end. And so, uh, so there are feelings of moral approbation. Brentano talks about, uh, uh, good is what, uh, being what loved for is correct and, uh, various moralists talk about, uh, uh, love as a kind of response to, uh, what we now call moral supererogation.

Uh, and, uh, love is probably not quite the right word, but, uh, moral approbation seems, seems right. So, uh, so we have the question of, uh, uh, uh, what we ought to feel moral approbation for and what we ought to feel, uh, self-approbation for, these positive feelings. So, uh, uh, something is supererogatory if, uh, others ought to feel approbation for the, uh, person who does it, and, uh, the person himself, uh, ought to feel, uh, self-approbation for it.

Uh, I think those are feelings that we recognize, and then I would cash out the, the, the ought by saying that to think that, uh, one ought to feel approbation of for the persons doing it is to, uh, uh, plan, or not exactly plan, but do this thing that’s analogous to planning to, to feel that way.

[02:01:25] FRANCES KAMM:
I, I, I, I’m not sure. The, the point I want to make is that sometimes people do the things that are supererogatory.

[02:01:32] ALAN GIBBARD:
Mm-hmm.

[02:01:34] FRANCES KAMM:
Though they think they– it’s not the case that they ought to do them.

[02:01:36] ALAN GIBBARD:
Mm-hmm.

[02:01:37] FRANCES KAMM:
And in the course of doing that, they actually fail to do what they would have thought is, is the thing they ought to have done.

[02:01:44] ALAN GIBBARD:
Mm-hmm.

[02:01:45] FRANCES KAMM:
So that is to say– But there’s a different sense of–

[02:01:47] ALAN GIBBARD:
So, so, so the thing they thought they ought morally to have done or the thing they ought– thought they ought rationally to have done?

[02:01:54] FRANCES KAMM:
No, that they thought that… In the sense, in the more narrow sense in which I would say that something is something I ought to have done in the sense that the supererogatory is not something I ought to have done.

[02:02:03] ALAN GIBBARD:
Yeah. It’s still permissible to do something. Yeah. Right. Right? Right. So, so it’s permissible, but not, uh, not something you ought to do.

[02:02:09] FRANCES KAMM:
It’s not something I ought to do. And in doing what this supererogatory action, I could fail to do the, what we would ordinarily call the obligatory action, but there would be nothing wrong in doing that.

[02:02:21] ALAN GIBBARD:
Um- Uh, well, wait, uh, wait. If, uh, if an action is obligatory, then there is something wrong with not doing it. Uh, right.

[02:02:28] SAM SCHEFFLER:
Wanna give an example, Frances?

(laughter)

[02:02:31] FRANCES KAMM:
Examples. I forget my own examples. But, you know, um, you, uh, you have a duty to, um, you know, to meet someone, uh, and it’s very important to them that you meet them.

[02:02:44] ALAN GIBBARD:
Mm-hmm.

[02:02:45] FRANCES KAMM:
So that they would be within their rights-

(laughter)

not to waive your, you know, according them what you owe them to meet them. But on your way to the meeting, you know, you, you see someone who’s drowning.

[02:02:55] ALAN GIBBARD:
Mm-hmm.

[02:02:55] FRANCES KAMM:
Now, he happens to be drowning in shark-infested waters. You asked for an example.

[02:03:00] ALAN GIBBARD:
I mean- Yes.

(laughter)

Uh-huh.

[02:03:02] FRANCES KAMM:
You asked for it.

[02:03:02] SAM SCHEFFLER:
That’s why I asked for an example.

[02:03:05] ALAN GIBBARD:
You must be masochistic. You want me here to begin with, but you asked for it. So, you know-

[02:03:11] FRANCES KAMM:
(gasp)

mm-hmm. -it’s not obligatory for me to jump into shark-infested waters, but I do it, right?

[02:03:17] ALAN GIBBARD:
Mm-hmm.

[02:03:18] FRANCES KAMM:
It’s perfectly all right to do it, even though I’m not carrying out my obligation to that other person.

[02:03:24] ALAN GIBBARD:
Uh-huh. Okay?

[02:03:25] FRANCES KAMM:
And it’s not because it’s a case of conflicting obligations. I– it was supererogatory, and yet it seems-

[02:03:30] ALAN GIBBARD:
Mm-hmm.

[02:03:31] FRANCES KAMM:
-perfectly right that if I want to do that, though I don’t have to do it, that I should do it. Oh, that I should do it even though I’m not doing what-

[02:03:38] ALAN GIBBARD:
Uh-huh.

[02:03:38] FRANCES KAMM:
-you know, I have an obligation-

[02:03:40] ALAN GIBBARD:
Yeah.

[02:03:40] FRANCES KAMM:
-to do. That’s the sort of case I was imagining.

[02:03:42] ALAN GIBBARD:
But it’s not the case that only an obligation can, uh, can, uh, can, uh, neutralize another obligation because, uh, I mean, you have ha-Hare’s example of, uh, he’s promised to take his children to the zoo, but, uh, a visiting philosopher comes, and it would be nice to show the visiting philosopher the, the colleges. Uh, he isn’t obligated to do that, but, uh, but that, uh, means that the, uh, that the pro tanto obligation to strong, pretty strong pro tanto obligation to take his children to the zoo, uh, doesn’t turn out to be a, an obligation all told. So I, I guess I would have thought it’s morally permissible to, uh, to, uh, not to brave the shark-infested waters and, uh, keep the appointment.

Uh, it’s also morally permissible to, uh, brave the shark-infested waters, and it’s also morally admirable to do the latter, but not morally admirable to keep the appointment.

[02:04:45] FRANCES KAMM:
Okay. Maybe… It’s fine with me.

(laughter)

I was just wondering how you would deal with cases- where there-

[02:04:51] ALAN GIBBARD:
Yes. Yeah.

[02:04:51] SAM SCHEFFLER:
I think we have time for one more question. Chris Kutz?

[02:05:03] CHRIS KUTZ:
Thanks. Hi. So the– I was, um, interested in the stress you wanted to place on the biology of all this and the- evolutionary biology of it in particular.

And two questions. One is I think mostly for clarification. I wasn’t sure I saw what work the emphasis early on in the lecture on evolution was doing.

I understand w- how the naturalistic commitment makes the story come out as it does, but I didn’t hear or see deployed much of the- mm-hmm, much of the evolutionary stuff, and I’m wondering if maybe that’s coming up or if that’s in the background.

What constraints is that placing on the kind of argument you’re giving? That’s the first question.

[02:05:38] ALAN GIBBARD:
Well, part, uh, most of my thought on this is, uh, things that I didn’t really lay out today. And, uh, so, uh, and I guess I wasn’t making it very clear what, what that was. I, I think that, uh, there are evolutionary pressures for, uh, discussion of the, uh, uh, although mostly for the kind of discussion that, uh, that’s involved in coming to a, a joint plan as, uh, as Michael, uh, uh, distinguished it.

So, uh, are there evolutionary pressures for, uh, uh, coming to plans for what an individual, uh, should do? Well, I think there are evolutionary pressures for, for, for planning, certainly. Uh, there are evolutionary pressures for once conversation is available for getting the, uh, for an ability to get what help in your planning you can get from your, uh, from your fellows.

So I think that’s the kind of, uh, thing that I would appeal to there.

[02:06:55] CHRIS KUTZ:
So is that adding something to, um, an alternative premise that, uh, planning around these things helps individuals secure their interests?

[02:07:04] ALAN GIBBARD:
Uh, well, the, uh, the evolutionary story would be that, uh, that in the old days, it helped the genes secure their metaphorical interests. Now, uh, the genes’ metaphorical interests and our literal interests, uh, uh, do tend to coincide, not, uh, not fully, but, uh, but, uh, to, to a fair amount since, um, so, um, so there’s of course, uh, uh, given these tendencies that, uh, in the old days, uh, helped the genes multiply, uh, in those days and in these days too, uh, help us, uh, achieve things that are good for us.

[02:07:53] CHRIS KUTZ:
Um, so do I have time for a second related question?

[02:07:56] ALAN GIBBARD:
Sure. Yeah.

[02:07:57] CHRIS KUTZ:
Second question then was, um, if we are stressing a biological account, I was wondering, this came up a little bit in your colloquy with Jay, um, about biologically more primitive kinds of emotions. I can understand using the language of planning with these, um, sort of more reflective moral emotions, resentment-

[02:08:12] ALAN GIBBARD:
mm-hmm and the like.

[02:08:13] CHRIS KUTZ:
But what about, um, emotions that, like disgust-

[02:08:16] ALAN GIBBARD:
Yeah or,

[02:08:17] CHRIS KUTZ:
um, sort of sheer unreflective rage that one might think- Mm-hmm would place, play a role in a kind of biological history of- Mm these moral emotions. They seem harder to attach to both the notion of judgment- Mm and to the notion of plan.

[02:08:31] ALAN GIBBARD:
Right. Well, these are things that aren’t really under the kind of, uh, normative control in our psychology that, uh, that moral emotions are. Disgust is very interesting because, uh, it’s primarily, uh, an emotion directed toward, uh, things that are, uh, toward contamination, things that will make you, uh, make you sick.

Uh, but, uh, then it gets, uh, moralized and, and we find that, uh, a lot of the, uh, a lot of moral rhetoric is built around disgust. Uh, so the– if you look in the, uh, OED under pollution, it seems that a lot of the earlier, the early citations for it, uh, use the metaphor of a pigsty, but it’s applied to, uh, moral behavior. So I, I think, uh, when we are disgusted by, uh, sleazy politicians or something like that, that, that kind of, uh, secondary disgust, uh, is under some, uh, nor– normative governance, uh, uh, in the, in the human psyche.

And, uh, I, I can– we could discuss whether that’s the, uh, the sort of thing to be disgusted by. But, uh, whether maggots are the sort of thing to be disgusted by isn’t, uh, isn’t a live question, uh, because, uh, our, uh, our answers to that aren’t going to affect our disgust.

[02:10:07] SAM SCHEFFLER:
Okay. I think we’re out of time for today, but I hope you’ll all come back tomorrow for the second lecture at the same time, same place. And in the meantime, thank Alan and Michael for a great talk.

(applause)

[00:00:01] INTRODUCER:
Okay, welcome to the second of the two Tanner Lectures by Professor Alan Gibbard of the University of Michigan. The title of his lecture today is Living Together: Economic and Moral Argument. Professor Gibbard.

[00:00:26] ALAN GIBBARD:
I, I appreciate the brave people who returned. Uh, So the title has been announced, and, uh, we’re beings who think about how to live. We each live with others, and we think how to live with each other.

Sometimes a person will think about such things by h-herself, and sometimes we think and discuss together. These are truisms, but I argued in the first lecture that the truisms are rich in consequences. They explain, if I’m right, the philosophically puzzling area of thought we call normative, thought that somehow involves oughts.

I guess we’re not ready for this yet. Uh, I want today to, uh, ask whether such a self-understanding could have any bearing on questions of right and wrong, of good and bad. Yesterday, I talked about moral concepts without using them.

I did metaethics, not normative ethics, not the work of thinking what is right and what is wrong and why. My metaethics leaves room for any coherent answer whatever to normative questions of what’s right and what’s wrong to do, and a wide range of possible answers are coherent. But I want to explore whether the metaethical picture I sketched contributes at all to making some answers to normative questions more plausible than others.

In doing so, I’ll have to pass lightly over controversies familiar in the literature, giving quick and insufficient arguments on issues that have been extensively and subtly debated. My– Uh, so I’ll talk about a social contract and the strains of commitment. Uh, my late colleague, William Frankena, said, “Morality is made for man and not man for morality.”

And his saying is widely quoted.

(chuckles)

He told me that he regretted ever saying this, but I don’t see that he had anything to regret. If morality should matter to us, if we should adhere to morals, st- demands, even at great sacrifice, then morality shouldn’t be arbitrary. Concern for morality should be out of concern for something that makes morality of value.

And how could that be of even being of value for people? Uh, I don’t mean to rule out other sentient beings, but in this lecture, I’ll st-stick to people. Most philosophers, I think, will agree with Frankena’s saying, but we fall into contention when it comes to drawing out the implications.

Moral inquiry and philosophy come, often comes in either of two broad styles. One is pragmatic, thinking what’s in morality for us, for us human beings, and asking what version of morality best serves us. The other broad style is intuitionist, in one important sense of that term.

Consult our moral intuitions, revise them as need be to achieve consistency, and embrace what emerges. The point isn’t that these two styles of moral inquiry need entirely be at odds with each other. Henry Sidgwick, over a century ago, advocated a morality made for humanity and insisted that there could be no moral knowledge without intuition.

And I end up agreeing with Sidgwick on this. Still, the two approaches are initially quite different in spirit. If morality is for humanity, then we might expect utilitarianism to be right.

Moral rules, we might expect, will tell us to act for the benefit of humanity. The right act will be the one with the greatest total benefit to people. Utilitarianism, though, notoriously conflicts with strong moral intuitions.

As a simple example, I’ll start with the case from yesterday of children drowning. I’ll then broach a line of argument that appeals to other intuitions and seems to lead back to a utilitarian answer. The case illustrates a much broader systematic argument for utilitarianism, one that draws on decision theory and was most notably advanced by Berkeley’s own John Harsanyi

Well before he came to Berkeley. Aspects of the argument have been widely debated, and my aim is to use the debate to explore how moral inquiry might proceed if it consists in doing the sort of thing I claim in thinking how to live together. The case is due to Diane Jeske and Richard Fumerton.

So two canoes of children capsize in rapids, and a man on the bank can rescue some, but not all of the children. So close to him are two children, and he could rescue both. Further away is his own daughter, and alternatively, he could rescue her, but not the other two.

Utilitarianism seems to say that, faced with this grim choice, he should rescue two children rather than his own. Many people have the strong intuition that the father is morally permitted, perhaps even required by the duties of parenthood, to rescue his daughter, even though he must then let two children drown instead of one. This example is contrived in the style of many philosophical examples.

The hope is, though, that such examples can let us examine considerations in isolation that get too complex to handle clearly in the kinds of morally fraught situations we’re apt to encounter. I’ll now introduce the style of argument I’ll be exploring. Imagine now that the situation is a little more complex.

There are two fathers on the two riverbanks by the rapids, and two canoes are swamped, each with two children. For each father, his own children are in the further canoe. Each could save either the two children closest to him or one, but not both, of his own children in the far canoe.

The rule to give preference to one’s own children, if each father follows it, means that each father loses a child. The rule to save as many children as possible, regardless of whose they are, means if each father follows it, that no father loses a child. Perhaps in this freak circumstance, the fathers could quickly reach an agreement that each would rescue the other’s children.

They would then each have a contractual obligation to act as utilitarianism would command. And for this case, the contrast between intuition and utilitarianism might disappear. In its prescriptions for this particular case, a social contract would thus coincide with utilitarianism.

Let’s return, though, to the first story, uh, with one father whose child was in the far swamped canoe, uh. Uh, so suppose that in advance the two fathers contemplate this contingency. Uh, one of them will be on the shore with one of the his two children in the swamped in the far canoe.

Oh, um, and the other is two children in the near canoe. Oh, so he could rescue both. Alternatively, uh, but one of his chil- own children is in the far canoe, and he can rescue her, but if he rescues her, he can rescue nobody else.

And we’ll suppose that they have equal probabilities of being in either position, and whatever agreement they is made will be kept. So what sorts of contracts might they come to to cover this eventuality? Well, uh, one contract would be the, uh, the loyal one, we might say.

Uh, I’ll call it L to save one’s own child. Each save one’s own child, uh, one of the two children, because that’s the most each father could save or, or the father could save of his own children. Or we could have the maximizing, uh, a contract to save the most children possible.

So the upshots, say, they are shown on the right. In the case of the, uh, of the loyal, the contract for loyalty to one’s children. In case, uh, you’re the lucky father on the shore, uh, you don’t lose any children.

In case you’re the unlucky father, uh, who isn’t there, both your children are lost. In, uh, so in the case of the, uh, in the case of the maximizing contract, if you’re this time, the lucky father is the one who isn’t there and then doesn’t lose any children, but the unlucky one is the one who is there and contractually bound to rescue the other’s children and then loses, uh, one child of his own. So, uh, in the lucky case, uh, one doesn’t lose, uh, any children.

In the unlucky case, uh, the unlucky father loses two children, uh, if the contract is to be loyal to one’s children, uh, and loses only one child if the contract is to rescue as many children as possible. So utilitarianism and contractarianism in this case give the same answer. Uh, save the most possible.

So in short, so long as whatever agreement they reach will be kept, then the first agreement in the unlucky case means losing both one’s children, whereas the second, in the unlucky case means losing, losing only one child. Each is a terrible loss, but losing both children is even worse than losing one, and the two cases are equally likely, we’ve supposed. For each father, then the second agreement, the maximizing the number saved, uh, offers better prospects.

Again then, for the case in question, the two approaches give the same answer. Utilitarianism says to rescue as many children as one can, and so does the social contract that people would make if they knew that the social contract would be kept. Well, this kind of argument generalizes.

John Harsanyi in the nineteen fifties proved two famous theorems that apply, theorems which I think should be more famous than they are among students of moral philosophy. The import and the limitations of his theorems have been widely debated, and I’ll be exploring how some aspects of the discussion might go if moral inquiry is the sort of thing that I think it is, uh, planning how to live together. First though, let’s explore further the case of the children and the swamped canoes.

The two fathers have agreed what to do in the contingency, and now one of them finds himself in the dilemma on the riverbank. He has agreed to save the two children that aren’t his, but still, of course, he’s strongly moved to save his own child. What motive might he have to keep the agreement and let his own child drown?

His motive might be one of fair reciprocity. “He would have done the same thing for my two children if our positions had been reversed,” he could say to himself. Still, he faces the question of whether to reciprocate.

Possibly fair reciprocity will insufficiently motivate him, and he’ll fail to reciprocate in his desperate situation what the other father would have done for him and his children. A further question arises, too, then: Would the other father have been sufficiently motivated? If the other would have reneged had their positions been reversed, then the father on the bank loses his rationale from fair reciprocity.

Here then, the upshot of contractarian thinking deviates from that of utilitarian thinking. Suppose for now that I’m right that if the two could make an agreement with full assurance that the agreement would be kept, they would agree on the arrangement that utilitarianism prescribes. In this way, utilitarianism can stem not only from motives of benevolence, but from motives of fair reciprocity.

That’s only, though, if the motivations of fair reciprocity are never overwhelmed by other motives, and the parties have full assurance of this. A contractarian that heeds the limits of motives of fair reciprocity will be quite different. What would we have agreed on under the constraint that the motivations we would have to keep the agreement would be sufficiently strong if everybody knew that the agreement would be kept?

Well, that will depend on a psychological question. How strong are the motives of fair reciprocity? How strong can we trust them to be under very various relevant circ– conditions?

We can see now why there might be a contractarian excuse for rescuing one’s own child in preference to two others. If we’d been able to make any agreement whatsoever and make it effective, we would have agreed to rescue as many children as possible, no matter whose. But we can’t produce such strong motives, and under the constraints of how strong motives of fair reciprocity can be, we wouldn’t have made such an agreement only to expect it not to be kept.

Well, I’m touching on what John Rawls called the strains of commitment. In most of the rest of this lecture, I’ll ignore them and explore other questions. I’ll consider contractarian arguments that assume full assurance of full compliance, severe though this limitation is.

Any full explanation exploration of contractarian arguments, uh, utilitarianism, and moral intuitions, though, would have to pay great heed to the strains of commitment. I’ll now talk about the separateness of persons. One way of arriving at utilitarianism is to say that morality consists of benevolence, of impartial concern for all involved, including oneself.

Uh, Rawls responded that impartial benevolence is a weak motive and that a far stronger motive is fair reciprocity. Tim Scanlon puts the motive differently, roughly as a concern to live with others on a basis that no one could reasonably reject. The canoe case suggested a way in which all these might coincide, at least for the case of full compliance.

Fair reciprocity consists in abiding by a practice, if it’s the practice we would have agreed to before we knew who would be in what position. To such a practice, no one could reasonably object. The question we ask in moral inquiry, I’ve been saying, isn’t the psychological one of what motives we do have and how strongly, but the question of what motives to have.

It’s a planning question, a question of how to live with each other. Nothing in metaethics that I’ve laid out dictates an answer. Still, the ideals of fair reciprocity and of living with others on a basis they could not reasonably reject seem good candidates for what to want in one’s dealings with others.

These aims are vague, but I propose to think together with people who might be brought to share these aims and try to work towards specifications worthy of pursuit. Morality, it’s often said, is grounded in respect for persons, and utilitarianism fails in that it can prescribe actions that violate people’s rights and fail to respect them. I can’t, of course, go over the history of systematic attempts to ground morality and respect and get non-utilitarian conclusions.

But my own reading of this history is that these attempts have not had any great success, and our brief discussion of the canoe case, uh, illustrates why coherent non-utilitarian theories are so elusive. The vague aims of fair reciprocity and of dealing with others in a way that no one could reasonably reject do seem good places to start in working out what aims to have, what aims in the vicinity of aims that having respect might consist in having. What we would have agreed on seems highly relevant to what respect demands.

Before I scrutinize contractarian arguments further, though, I’ll say a few things about why I don’t think respect leads us straightforwardly in directions that oppose utilitarianism. Well, Rawls charged that utilitarianism ignores the separateness of persons. A person is not to be sacrificed for the sake of another.

One person’s gain doesn’t compensate for another’s loss. Well, thinking in terms of gains and losses or sacrifice, though, requires a base point for comparison. And so we’ll need some rationale for heeding one possible base point as opposed to others.

Uh, so suppose we have persons Ida and Jay, and states A and B, with Ida better off at state A and Jay better off at state B. Let’s give numbers to how well off they are, as I’ve got up here. Um… Okay.

So Ida’s gain in going from s-state B to, or to state A, we might say, doesn’t compensate for J’s loss. Uh, Ida goes from five to nine, an improvement, but J goes from three to one, uh, a loss. Uh, J has only one life to lead.

We can’t sacrifice him for Ida’s benefit. But if we frame matters differently, we get the opposite conclusion. In going from state A to state B, Ida loses, going from nine to five.

Uh, Jay gains, to be sure, but he and Ida, we can say, are separate persons, and Ida can’t be sacrificed for Jay. To choose between the two seeming upshots of the separateness of persons, we must choose between state A and state B as a base from which gains and losses are measured, its sacrifice is attributed. Rawls seems to choose the state with the worst-off person, state A in this case.

Rawls objected that we can’t sacrifice the well-off to be, to benefit the badly off. Uh, Nozick and some others who appeal to Kant say that we choose as the base state for comparison, the state in which people are entitled to what they would have. Uh, Rawls replied that when the basic structure of society is at issue, we’re asking what entitlements to institute.

Intuitions that invoke ownership and other entitlements are very strong, and they may well be wired into the human psychic makeup. They’re very sensitive, though, to framing effects. Even a person’s self-regarding choices are affected by attributing ownership.

So consider endowment effects. Uh, we give a person a coffee mug and ask him if he’ll trade it for a chocolate bar, and he says no. It seems he prefers having a mug to having the chocolate bar.

But if we’d given him the chocolate bar instead, he would have refused to trade it for the mug. It seems he would then prefer having the mug to having the chocolate bar. Or sorry, I’ll— it seems he would prefer having the chocolate bar to having the mug.

Uh, the only difference is which of the two objects he frames as already his. How can we find some basis for attributing entitlements then that’s independent of a pragmatic test, independent of evaluating the consequences of a system of entitlements by a standard that doesn’t assume the importance of the entitlements in advance? Well, Nozick tried, but he left the basis of what he was saying unexplained and seemed to appeal to the pragmatic advantages of systems of property.

Well, I conclude that we can’t talk of gains and losses and sacrifices until we identify some base point for the comparisons. It’s true enough that we’re separate persons, but nothing about what we may permissibly do follows from that by itself. Our strong intuitions do latch on to some base point or other, but not in any consistent way.

Perhaps we could establish some base point as morally relevant. One way to do so, though, would be to be the way I’ll be exploring. Ask what we would have agreed to from behind a veil of ignorance, what we would have agreed to treat as morally relevant.

So I’ll talk about Harsanyi’s theorems. Well, the point of morality, I’m taking it, is to live with each other on a basis that none of us could reasonably reject. No one has a reasonable objection if the system we live by is what we would have agreed to in fair conditions.

And one way to make conditions fair is a veil of ignorance. We saw in the canoe case that this may yield utilitarian prescriptions. Harsanyi argued that this upshot generalizes.

His argument starts with the coherence of plans for action. Decision theorists have shown that if a way of ranking actions satisfies certain conditions, then it’s as if the person chose by maximizing an expected value. It’s as if the person formed degrees of credence in the relevant eventualities, attributed levels of value to the various possible outcomes, and then took the alternative that held out the greatest expectation of value, reckoned with those degrees of credence and levels of value.

Whether the axioms needed for this are really compelling has been a matter of great contention, but the argument for the standard conditions, as I’ll call them, goes much like the argument I gave yesterday for Zeckhauser’s Russian roulette example. Peter Hammond systematizes such an argument, though his proof, I’m afraid, is couched in a fearsome mathematical apparatus. By the standard conditions, I’ll mean any of various sets of conditions that have been shown to yield the result.

And coherent plans, I’ll assume, are plans that satisfy these conditions. It’s highly contentious whether this is so, as I indicated yesterday, but I’ll be exploring what we should think if it is so. So Harsanyi proved two theorems.

Uh, the first conveyed something like Rawls’s original position, where behind a veil of ignorance, no one knows what position he’ll occupy in society. Think of valid moral rules as the rules one would choose, assuming an equal chance of being anyone. Assume one’s preferences are coherent in that they satisfy the standard conditions.

Then Harsanyi proved one will prefer the rules that would yield the greatest total utility. Here, by individual utility, I mean the scale that represents one’s preferences, given that one will turn out to be that person. Well, the second theorem is this.

Suppose that prospective individual benefit is coherent, and so is desirability from the moral point of view. Suppose also that morality is for humanity, at least in the following sense. If one prospect is better than a second for each individual, it’s the better prospect ethically.

Uh, this is called the prospective Pareto condition. Then the desirability from a moral point of view, he proved, is a weighted sum of individual benefit. The only way ethical evaluation could satisfy these conditions and depart from utilitarianism is by weighing one person’s benefit more than another’s.

Economists represent the theorem in graphical form. So we have a set here of all the, uh, all the possible combinations of, uh, Ida’s benefit and Jay’s benefit, uh, Ida’s shown horizontally, Jay’s shown vertically, that are, uh, achievable by social policy. Think of these as prospects that are achievable.

Well, suppose this is the, uh, presumably they agree on some point on this frontier. Uh, they agree on some point, uh, where you couldn’t make both of them better off at once. If they agreed here, you could, uh, improve things for both of them, so why would they agree on this rather than this?

Okay, suppose the point is this one, then the argument is just, well, um, that’s taking things as far as they can go in a certain direction, and that amounts to say that that one’s maximizing some weighted, uh, uh, sum of the benefit of one and the benefit of the other. Okay. So the challenge to anyone who wants to get a non-utilitarian morality out of thought on a social contract is how to evade the force of Harsanyi’s two theorems.

If you’re going to be a non-utilitarian, you’ll, uh, you will adopt moral rules that none of us would have chosen for his own sake, unless he knew of some special way that he and not others stood to benefit. And any evaluation of the prospects that various different moral orders bring must either violate some demand of rationality or weigh one person’s utility above another’s, or rank some prospect best, even though another prospectively benefits everyone more. Now, Harsanyi’s theorem has been much discussed, if not sufficiently.

The quick, careless statements of the theorems that I have given would require close scrutiny, and important parts of the needed scrutiny are in print. Uh, John Broome is possibly the, uh, world’s leading expert on the ethical aspects of this. Uh, what I can hope to do is just to select a few issues that are relevant to these debates, uh, framing the theorems as parts of moral inquiry as I’ve been picturing it.

So back to talking about a person’s good. Uh, what is it? A person’s good or benefit?

In the tradition that Harsanyi worked in, a person’s good is a matter of her preferences. We gauge the strength of her preferences by the preferences she would have among gambles. If you would risk a one in a million chance of being killed by a car, but no more, to cross the street to buy a chocolate bar, then the benefit to you of a chocolate bar counts as one millionth of the harm to you of being killed.

This notion of benefit has at least three chief problems. One seems tractable enough. Philosophers hasten to add that the preferences must be considered and informed.

A second is more difficult. The preferences of any decent person won’t just be for that person’s own benefit. The person will care about others.

He will care about fairness. He’ll care perhaps about living with others on a basis that no one could reasonably reject. A person’s benefit is at best one component of his considered informed preferences.

So what component is that? Well, the third problem interacts with the second. What people prefer for themselves differs from person to person.

Some differences aren’t fundamental. I dislike asparagus, and, uh, my wife Beth loves it, so I prefer not to eat it, and Beth prefers to eat it. Basically, though, we both want to enjoy our food, and we’re different in what we enjoy.

Other examples, though, might be different. When I taught school in Ghana, people told me that one thing they lived for was a big funeral. Well, that puzzled me.

But as I thought about it, I realized that a big funeral indicates how one is loved and respected in the community. And to be loved, respected, and missed did seem to be, to be things a person could intelligibly put great stock in. Still, once we distinguish carefully what goes on in one’s life and what happens after, people maybe differ in whether they care for their, for their own sakes, uh, how they’re regarded after death.

When I stand behind a veil of ignorance and choose a social ethos to institute, I contemplate that I may turn out to be you with your basic preferences, and I may turn out to be me. with my basic preferences. You and I may differ even in the basic preferences we would have if our preferences were fully considered and informed.

Rawls stressed this and proposed instead of, uh, of looking to self-regarding preferences, that we look to what he called an index of primary social goods. Primary goods he de-defined as things one wants, whatever else one wants. Saying all this, however, leaves it as part of his project to find some basis for this index.

Rawls thought the problem was mitigated by his theory, since he only had to deal in broad categories like income and opportunities, and only had to identify whether the worst off were better off under one arrangement than under another, since he was trying to make the worst off best off. In fact, though, his theory ends up placing great demands on this index. Uh, what turns out to matter in his theory, uh, at least as I read it among various things he says, is the life prospects of those in the worst starting positions in life.

Uh, well, to tote up on certain prospects, we need more than an ordering from best outcomes to worst. We need to know what differences are big and what differences are small. Uh, we need what measurement theorists call a cardinal scale, and we need to be able to compare people on that scale.

I’m asking about the basis we might have for these comparisons. As for Scanlon, uh, he never engages directly the kind of perspective argument that lies at the center of Rawls’s original position and Harsanyi’s theorems, as far as I know. Uh, I’ve followed Harsanyi and Rawls in saying how we can dismiss some objections as unreasonable.

So you do something and I object. You reply, “That’s the established way we do things,” and that’s what you would have wanted before you knew which of us you would be and how in particular it would affect you. This seems to show my objection unreasonable, and from crediting such dismissals, uh, Harsanyi draws powerful consequences, or as I’m, uh, putting Harsanyi’s theorems.

Uh, we now see, though, that he must place great demands on the notion of individual benefit Scanlon offers an extended critique of the notion of welfare or a person’s good, and this, I take it, constitutes his response to Harsanyi. His critique doesn’t rule out the possibility of dismissing objections as unreasonable on prospective grounds in the kind of way I’ve been discussing. It does, though, place an advocate of such tests, uh, place on, on an advocate of such tests the burden of saying what notion of benefit could play this role.

Scanlon allows that a notion of a person’s good can be tailored to play a role in moral theory. Clearly, though, from the way he himself develops his test, he doesn’t think the test of what we would have wanted from behind a veil of ignorance plays anything like the broad and systematic role in a coherent moral theory that Harsanyi thought it played. Scanlon’s critique of the concept of a person’s good is a serious one, and I’ll be particularly concerned to respond to it.

I’ll be arguing that we can’t derive the needed notion of individual benefit directly from the preferences that people have or even the preferences they would have in ideal conditions. Instead, forming a conception of benefit, of the intrinsic reward of a life, as I sometimes have called it, is part of ethical thinking, part of thinking how to live among other people. That fits a pattern of what Scanlon himself concludes.

But if the retort I’ve imagined always deflates a claim that an objection is reasonable, then thought of prospective benefit will have a much larger role in coherent ethical thinking than Scanlon gives it. So I’ll talk about preferences for being another. To illustrate and explore the problem, uh, let’s return to the simple toy case of Ida and Jay.

I’ll suppose first that we understand the notion of a person’s good. So suppose again that how well off Ida and Jay would be in states A and B, uh, go as follows, and this is the same diagram as before with, uh… So in, uh, structure A, Ida would have nine and Jay would have one.

In structure B, Ida would have five and Jay would have three. Uh, we can also give the expected, uh, value of, uh, having, uh, equal chances of being Ida and Jay. Uh, five in the case of an equal chance of being, uh, of nine and one, and four in the case of an equal chance of five and three.

These numbers are chosen so that the, um, uh, index of a prospect is, uh, just how prospectively well off someone is who, uh, who has those outcomes. Okay. Well, Harsanyi’s argument favors Structure A.

Uh, before they both knew who they would be, uh, both would prefer structure A with expected level five to structure B with expected level four. J comes out badly under structure A, but A is the going structure, and he has no reasonable objection to it. Well, I turn now to the objection raised by Gauthier, Suzumura and others.

Uh, what do the numbers in my simple table represent? They represent evaluations from behind a veil of ignorance of people who are choosing basic social structure only for their own prospective benefit. We’re asking what notion of benefit feeds into the moral force of the rejoinder.

Uh, of the rejoinder, that’s the structure you would have chosen. Well, Ida, from behind the veil of ignorance, contemplates two possible, two possibilities: that she will be Ida and that she will be Jay. Now, Ida in the world, let’s suppose, wants a big funeral, and at state A, she gets it, whereas in state B, she doesn’t.

Uh, does she have this preference behind the veil of ignorance? Well, suppose she does, but Jay doesn’t. So Jay understands that in the case, in case he turns out to be Ida, his actual concerns very much include having a big funeral.

But being Jay, he doesn’t intrinsically care about having a big funeral. He’s indifferent between being Ida and having a big funeral, and being Ida and unexpectedly be-being buried after death without ceremony. Being Ida, he understands, includes wanting a big funeral.

But as he is behind the veil, he’s indifferent between wanting a big funeral and getting it, and wanting a big funeral and not getting it. “If I get it,” he figures, “I won’t be around to enjoy it,” and if I don’t, I’ll never miss it.” Well, once we distinguish Ida’s preference for being Ida in A from Jay’s preference for being Ida in A, we get a more complex table like the one up here.

Um, so we have to distinguish Ida’s preferences for being in state A or state B, and in one case, being Ida, being Jay, uh, from Jay’s preferences for, say, being… Well, in particular here, uh, Ida’s preferences for being Ida in state A with a big funeral, uh, she wants that a lot, and Jay’s preferences for being Ida in A with a big funeral, which Jay is, uh, indifferent to, uh, or the case where he turns out to be Ida. So Ida’s evaluation of being Ida under structure A includes four units for having a big funeral and five for other aspects of how things go for her.

And Jay’s evaluation of being Ida under structure A includes only the five units. From behind the veil, he places no value on actually having the funeral. Uh, well, Ida now can’t refute Jay’s objection by saying that A is the state he would have chosen if he hadn’t known who he would be.

He would have chosen state B, which gives him an expected utility of four as, uh, well, five. That should, uh, that should be and, uh, as opposed to three. Uh, suppose, though, the structure B has been instituted, and Ida, not getting a big funeral, objects.

Or since she’s not around to object, suppose someone objects on her behalf. Well, Jay can’t deflate the objection by saying that B is the structure she would have chosen if she hadn’t known who she would turn out to be. Uh, so for each state, there’s an objection that can’t be shown unreasonable, at least in this way.

Unless we could find some other way to show one of the objections unreasonable, we’re damned whichever state we institute. Well, that fits in with Scanlon’s critique. Uh, there’s no one coherent notion, he says, that will do the jobs that welfare or a person’s good has been asked to do in much ethical thinking.

Roughly, first determining, uh, first what a person will choose insofar as others aren’t affected, what others concerned to benefit him will choose to promote, and what counts as the person’s good for purposes of moral thinking. We’re asking about the third, what counts as a person’s good for the purposes of moral thinking, uh, and indeed, as Scanlon told us, not finding a way to read off a person’s good, uh, from her preferences. One appealing way out would be to let Ida be the judge of her own good.

The problem remains, though, as Rawls insisted. From behind the veil of ignorance in his system, we’re choosing among alternative basic structures of society. What people will want at base might be highly affected by the kind of society we choose to have been nurtured in.

Ida might have been indifferent to a big funeral if she’d grown up in a different sort of society, in an alternative social order that is open to choice from behind the veil of ignorance. Well, Rawls, as I say, was responding partly to this kind of problem when he set up his index of primary social goods, but I, I claim offered no adequate, uh, defensible rationale for its solution. Uh, so I’ll be asking what– whether such a rationale can be provided.

So back to the question of, uh, person’s good. Uh, let’s call the retort I’ve been discussing the “you’d have agreed” retort. It does seem to have moral force, and we’re left with the question of what kind of force to accord it.

Objections to a social order often are on the grounds that it unfairly disad-disadvantages someone. Such an objection itself appeals to a notion of advantage or benefit. And so if the retort is incoherent because it requires a notion of a person’s good, then so was the original objection.

Let me get back then to the more basic perspective with which I started. We need to think what to prefer in our dealings with others and why. When I think why to prefer something, there’s a challenge I could put to myself.

If I find myself drawn to a preference, I can ask, for whose sake? Frankena is saying morality is made for man and not man for morality, uh, tells us to be suspicious of preferences for which we can’t find this answer. Or sorry, for which we can’t answer this question.

How, we’ve been asking, does this bear on thinking how to live with others? How does it bear if an ideal of fair reciprocity is one to pursue, as what living with others fairly on a basis they could not reasonably reject consists in? In thinking how to live with others, I start with incoherent impulses and preferences.

An ideal answer to how to live would be formally coherent, among other things, and that requires working my raw incoherent preferences into something more coherent. There are, though, many formally coherent plans for how to live, most of them with little to recommend them. If, though, the ideals that I’ve been exploring are ones to develop and make coherent, then our answers to how to live should take a form we might call humanistic, to live for the sake of people, oneself very much included.

That, of course, leaves the question of whether alternative coherent plans for living might equally deserve the name humanistic. But for now, I’ll appropriate the name and try to work out this ideal. What is it to put one’s preferences in a form that is not only formally coherent, but humanistic too?

Humanistic preferences vaguely are preferences for the sake of people. One way you might try interpreting this is by a number of loose axioms. So first, uh, the axiom will be, uh, prefer most to live with others on a basis that no one could reasonably reject on one’s own behalf.

So this is an imperative that I direct to myself. Prefer most to live with others on a basis that no one could reasonably reject on her own behalf. Uh, this is a plan or a fragment of a plan, but with crucial terms left unexplained.

We must specify what it is to reject a basis for living with each other reasonably and on one’s own behalf, And hence the next axiom. A rejection of a social arrangement on one’s own behalf is unreasonable if, absent information about which person one would turn out to be, one would have chosen that arrangement for one’s own sake. Okay.

So this leaves to be explained, uh, choosing a thing for, for one’s own sake. So we could try s-saying, uh, one chooses what one most prefers. That is, one.

And so in particular, one chooses for one’s own sake what one most prefers for one’s own sake. And moreover, for the case of being person J in social arrangement X, what one prefers for one’s own sake is what is to J’s good in X. So this leaves another uninterpreted term, the term good in the phrase person J’s good in state X.

So what then constitutes a person’s good? We’re still left with that question. Now, though, we could say that a person’s good is whatever plays the role of a person’s good in this string of four conditions.

The question of what constitutes a person’s good is part of the question of what to prefer in life. The string of four conditions is a formal constraint on how to live with others. The constraint is to live with others on some specification or other of the ideal of fair reciprocity, which specification is a further planning question of a further question of how to live with others.

If our preferences for how to live, as we struggle to make them coherent, do take this form, then we can go on to argue in Harsanyi’s way that behind the veil of ignorance, one would choose the social arrangement that in prospect maximizes the sum of people’s goods. Preferring to live on that basis, one prefers to do one’s part in a utilitarian order, provided that everyone else can be fully expected to do so. None of this settles whether to have preferences that take humanistic form, and so nothing Harsanyi showed demonstrates that the right answer to how to live includes this dictum to reciprocate in utilitarian arrangements.

Other ways of living are coherent, and one of those might be the way to live with others, for all Harsanyi and I have shown. I hope, though, that I’ve given some glimmering of what speaks for such preferences as a way of filling out the wish to live with each other on a basis that no one could reasonably reject. And Harsanyi did show that any alternative to preferences for living that give some valuation to the role of a person’s good, uh, must have a disturbing feature.

Either the preferences that constitute adhering to that alternative are formally incoherent, or for every possible way of specifying a person’s good, they include preferences for a societal structure that’s prospectively worse for everyone. I come to a view of what a person’s good is, then, if and when I come to have preferences that take this humanistic form. We come to a joint view and discussion of what a person’s good is if we all come to have preferences that take this form, and for each of us, the same valuations play The, the role in these, uh, uh…

The same valuations play the role that these conditions assign to a person’s good. A few comments on this approach. I’ll comment somewhere this leaves things, and then I’ll raise some questions that we can discuss further tomorrow.

First, if fair reciprocity takes a utilitarian form in the way I’ve pictured, that only answers a sharply delimited moral question, namely what to do with full compliance, uh, for the scheme we w- would have wanted. It doesn’t say anything about what to do if others can’t be expected fully to comply with such a kingdom of ends. Second, the preference to live on a basis of mutual respect may not be overriding.

And coming to a view on how to live, I might find the contractual ideal strongly appealing, but still plan to opt out in extreme circumstances. What contracts we might make if we know I’m bound to be like this, I haven’t explored. Third, I’m distinguishing a person’s good from her preferences, and surely I’m right to do so.

But still, there must be some close tie between the two. Unless I find that people normally do care deeply about their own good, something’s gone wrong with the conception of a person’s good that I’ve constructed. That a person will care about her own good, though, is not something that stems from the sheer logic of the concept of a person’s good.

The question of what’s to a person’s good, it, I’ve been saying, boils down to a planning question. It’s a question of what, if anything, to respond to as a matter of fair reciprocity, of what to treat as fair conditions for having decided the basic ethos to govern us, and supposing, supposing that if the going system is one we would have chosen in those conditions, then its rules are the ones to follow and to feel as just. Being the way we are, the things we’ll accord the status will be things that we do find important.

Often indeed, the more we know about what it’s like to be a person, the more we’ll find that the things that the person feels important, indeed are important. We’ll find that there are things to want for the case of being that person and things to want to have weight in deciding social arrangements. Still, I’ve insisted that the two can come apart.

I can conclude that something I put great stock in isn’t really at base a thing to want. Uh, I can find myself caring about whether my parents would have approved of what I’m doing and yet conclude that doing what they would have approved of is not something to want for its own sake. I can have the same thoughts about something another person puts great stock in.

If I may be mistaken, others may be, too. Well, suppose, though, what a person wants isn’t for that person’s own good or for anyone else’s, for that matter. Still, doesn’t respect for a person mean giving the person scope to pursue the goals that the person adopts?

Isn’t respect then basically different from heeding a person’s good? There are weighty reasons to allow people scope to pursue their freely chosen ends. They’re far more likely to be sensitive to what’s to their own good than anyone else is.

And the social structure that most realizes the good of people will stress effective freedom. Moreover, aspects of autonomy, obscure though that notion is, must surely be part of what to want and what to have weighed on one’s own behalf in social arrangements. Even if, even if a person is mistaken on what to pursue in life, an important part of what to want for the case of being that person is scope to think through what’s important and pursue what she concludes to be important.

Concern for a person’s good includes wanting her to have scope even to make her own mistakes. Do the moral requirements of respect for persons go beyond what these broadly utilitarian considerations support? As I’ve characterized a person’s good, it’s hard to see how they could, at least so far as morality is a matter of living together on terms that no one could reasonably reject.

A person’s good just means the magnitude to reciprocate in terms of whatever that is. If the things to heed out of respect aren’t a matter of a person’s good, then by this definition, they aren’t things to enter at the basic level into fair reciprocity. My colleague Stephen Darwall offers a brilliant and illuminating alternative way of defining a person’s good, a way that merits careful study.

I’ll sketch it quickly in my own terms as I understand it. So a person’s good, Darwall says, is what to want given and insofar as one is to care about the person. Care, he treats as a characteristic human attitude, as a natural kind of the human psyche.

Darwall thinks that one can raise even the question of whether to care about oneself. If the answer is yes, then by definition, one’s own good is what to want out of the care to have for oneself. Darwall himself thinks that the morality of care is not all of morality, that there’s also the morality of respect.

What to want if and insofar as one is to respect oneself may thus be different from what to want insofar as one is to care about oneself. And the morality of respect may deal in things, of even the good of people. Now, in a way, Darwall’s concept of a person’s good fits nicely into a Harsanyi-Rawls sort of original position.

We can now specify that each party from behind the veil of ignorance is to care about herself exclusively. But if a person’s good, as Darwall defines it, is not the only thing to reciprocate in terms of, then my own way of defining a person’s good is not equivalent to Darwall’s, since my own way makes such a divergence impossible. It’s not to the good of people in my sense– or sorry, it’s to the good of people in my sense, not Darwall’s, that the Harsanyi argument applies as drawing out the implications of fair reciprocity.

It’s my concept that requirements of coherence in thinking how to live with others, uh, drive us to. Well, none of what I have said, I should say finally, rules out coherent ultimate disagreement on what a person’s good consists in. Well, how could it?

Uh, questions of how to live are not settled by requirements of formal coherence alone. There remain then further questions. Among people whose conclusions are coherent and humanistic in form, but who still disagree, how can they live together on a basis of mutual respect?

And what if their conclusions aren’t even humanistic in form? Uh, of course, we have to ask, uh, what if people aren’t even coherent in their evaluation of the evidence and their resulting views on how to live? We face questions of what kind of imperfect political consensus to promote for the world as it is and as we might help to make it.

Uh, on the need to address these questions, utilitarians and non-utilitarians will agree. So questions of justice and of right or wrong, I started out, started out by saying, Bigger questions of how to live with others and how to feel about one’s own actions and the actions of others. I’ve been tracing a path for our planning how to live with each other, to take a path from Scanlon to Harsanyi.

I hope that I’ve conveyed a little of what recommends the path. Anti-utilitarian intuitions are easy to elicit, but even our strong intuitions are incoherent. And a coherent view on how to live will run into strong contrary intuitions at some point.

Harsanyi’s theorems don’t rule out all opposing moral views as incoherent, but they do indicate a kind of incoherence that attempts to, uh, at a close match to moral intuitions will run into. If an intuitive moral restriction isn’t one that we would have chosen for the sake of humanity, We need to ask, uh, why lend credence to it? If our thinking how to live with others takes the form I’ve been advocating, then our answers won’t fall prey to this objection.

Thank you.

(applause)

[01:08:25] INTRODUCER:
We’re again going to take a very short break, just a couple of minutes, and then we’ll resume with comments from Frances Kamm and John Broome. Okay, um, I think we should resume. As I said, we have two commentators today.

The first commentator is Professor Frances Kamm. Professor Kamm is the Littauer Professor of Philosophy and Public Policy at Harvard University, where she’s on the faculty of the Kennedy School of Government and the Department of Philosophy. Prior to moving to Harvard in 2003, she taught for many years at NYU, where her primary appointment was in the Philosophy Department, and she also held appointments in the schools of Medicine and Law.

Professor Kamm is one of the leading moral philosophers of our day. Her extensive and ingenious writings address central theoretical questions in ethics, as well as important issues of practical concern. She’s the author of Creation and Abortion and a monumental two-volume work, Morality, Mortality.

She’s received numerous fellowships and awards and sits on various editorial boards. It’s an honor to have Professor Kamm with us for these events.

(applause)

[01:09:39] FRANCES KAMM:
I keep my watch here. I just hope time doesn’t go faster on the West Coast, does it? Can you hear me?

Yes. Okay. So, um, but I’ve gotta get this down so I can see my…

I’ll hold it here. Okay. Uh, first of all, let me thank the department or whoever it was who invited me and, uh, the Tanner Committee.

And, uh, it’s an honor to be here, and it’s an honor to, uh, be with my great colleagues. Allan Gibbard, uh, warns us that we must be prepared for conflicts between our intuitive judgments about cases and what on reflection we consider reasonable. The judgment of what is reasonable will also involve intuitive judgments, he thinks, though not about cases necessarily.

I suspect I have been asked to comment, though I’m not a professor of metaethics, um, because, uh, I rely so heavily in my own work on intuitive judgments about cases. At one point, Gibbard describes what might be thought to be the method of those like myself. Try to get a consistent set of intuitive judgments and just go with that as the correct morality.

But I think this leaves out at least the need to explain, uh, what intuitions that seem to be in conflict or inconsistent are, and whether they really are in conflict and inconsistent. And also to find out what deeper principles underlie the case-based intuitions and judgments that we have, for I do think that the plausibility of the, uh, principles, such underlying principles, is necessary to justify ultimately our intuitive judgments. I should also note that Thomas Scanlon, uh, ‘s version of contractualism, to which, uh, Alan Gibbard refers, uh, very much aims to explain and justify the just the sort of case-based intuitions that Gibbard thinks actually conflict amongst themselves and with contractualism.

Now, with that as an introduction, uh, let me ask some questions. Uh, I have some more questions I hope I will be able to get to on Friday, but I’ll, I’ll try to do what I can today. Sorry, Thursday.

Don’t want to confuse you. Uh, Gibbard says that morality is for persons. It consists of humanistic preferences, which are preferences for the good of people.

The minimal sense of this, he says, is that, uh, if a state of affairs is better for someone and no worse for anyone, then it is simply morally better. This is what he calls the prospective Pareto condition. A more robust interpretation is that morality seeks the maximal good for people, which is the maximal sum of individual utilities, and thus the maximal expected utility for each individual behind a veil of ignorance.

I think there are reasons to doubt, first, that morality is only for persons or sentient beings, and that the parts of it that are for persons take the form of arrangements that produce highest expected utility for individuals. Scanlon himself claims that there is a part of morality this, uh, that is concerned with impersonal values. For example, it would be morally wrong to pave over the Grand Canyon or destroy great works of art, even if this was not bad for people.

Admittedly, uh, Scanlon thinks that there will be, in his system, contractualist system, uh, only indirect ways to give weight to those values, but nevertheless, they are part of morality overall. Shelly Kagan and Larry Temkin have argued that the Pareto condition is just wrong. It can be morally worse if someone gets what is good for him, even if his getting it is not worse for anyone else.

Consider an evil person who has utility twenty and a good person who has utility ten. It could simply be morally better if people have only as much good as they deserve in absolute terms, and also better if they, if their relative utility is in proportion to their desert. If this is so, it could be morally worse if the evil person is better off than the good person, even if this, his being so does not worsen the condition of the good person in any way.

It would be better if his utility were less than the good person’s, at least if this is not below the point of his absolute desert, even though this loss of utility to him does not improve the condition of the better-off person at all. Some proponents of equality as an impersonal value have also argued that it could be morally worse if some are better off, even if this does not make anyone else worse off. They call for producing equality even by leveling down when this is the only way to achieve equality.

That is to say, making some worse off, even though this improves no one else because it results in equality. I actually think this view to be plausible must be connected to the desert view that I described f– previously. A final example emphasizes the possibility that morality is not always about what is good for people, but rather about expressing or representing the good of people, or their worth, as it is sometimes called.

So suppose it was true that beings that have a higher status, that beings have a higher status, if it is true of them that they have some sort of special authority over themselves and may not be used without their consent when it is seriously against their interests at the time that they would be used, even if they would be used to help many others, each of whom would be as badly off if not helped as those who would be used would be if they were used. For example, the standard illustration is if we may not kill one person to stop ten others from being killed. Now, some may interpret having this higher status as itself a good for a being that has it, even if it reduces the chances of his survival overall.

I think it may just be good that there are beings with a higher status without its necessarily being good for anyone. And I mean this in a morally serious sense of good. Now, let me move on from this question of, uh, whether the, uh, conditions of Pareto are absolutely required.

And let me raise some questions about the view that morality, uh, being for people involves its giving each individual his highest expected utility, and that this is what each would agree to behind a veil of ignorance when, uh, deciding, uh, for his own good. I will reflect on Gibbard’s case involving the two parents, um, with endangered children in considering these matters. Gibbard ultimately presents the parents’ case, as I’ll call it, by focusing on the potential loss to each of the two parents as what is at stake here.

A and B are their names. And he says it is worse for parents to lose two children rather than one. He ultimately imagines the case, I believe I understand it correctly, though I might be a subject to error here.

He ultimately imagines the case so that parent A has one child at home. I mean, he says, I think, parent A has two children on page three, but there’s one who’s in a canoe in danger. Parent B’s only two children need saving in another endangered canoe, but only A can be a savior, has the power to affect the situation.

Given this scenario, one possible state of the world would involve A saving his own endangered child, his– sorry, his one endangered child, thus winding up with two children, and B winding up with no children, having lost two. Another possible state of the world would involve A saving B’s children, in which case A winds up with one child and B with two. The second state of the world involves higher overall utility and also higher expected utility for each parent behind a veil of ignorance.

But it’s also– But it also involves the worst off parent being better off than the worst off parent in the alternative state of the world. As in that state, one parent winds up with no children, whereas in the state where A saves these two children, um, as if I’ve got it right, A still has another child somewhere else.

This leaves it open that it is the desire to pick an arrangement in which the worst off person is as well off as he can be, rather than the desire f- to maximize, uh, ex-ante expected utility that would or that could underlie an agreement or an arrangement that results in parent A having to save these children. Now, move on to another case. Suppose that parent B has four children, two of whom are in an endangered canoe.

The others are okay. Parent A has one child, and, uh, she is– that child is in another endangered canoe. In one possible state of the world, B winds up with four children and A with none when A saves B’s two children.

In another state of the world, B winds up with two children and A with one child because A saves his own child, two of B’s die, and B has two at home. Ex ante expected utility of each parent is highest if there is an agreement to bring about the first possible state of the world, as there will then be four live children, and we are supposing that A and B have an equal chance, you know, one has an equal chance of being parent A or parent B. But the worst off parent is better off in the second possible state of the world, for then he has one child instead of none. Suppose that all that were relevant is how things go for A and B.

Then it seems to me that behind a veil of ignorance, it is a state with lower expected ex ante utility, in which the worst off pos– the, w– the position in which the worst off position that anybody finds himself is better off. Is the better off position, namely then has one, one child rather than none. And I think that this would be preferable if we take seriously the good of each person as a separate person.

We should be concerned with the eventual good of each person, not just their expected utility behind the veil of ignorance. Of course, the sort of maximin reasoning, as it’s called, uh, that I’ve just been employing now, you know, getting the state of the worse-off parent as high as possible, could still result in the requirement that B, the person who has four children, if he was the only possible person who could save anybody, should save A’s one child rather than his own two children in the endangered canoe, given that B has two children at home. Okay.

This is one reason I do not think that the maximin solution is ultimately correct either. Ah, but what do I know? I’ve only got my intuitions.

Anyway, next, um, suppose that parent A had only one child and she is in the endangered canoe, and parent B has only one child, who is in another canoe going down along with an orphan. No sharks, but there’s an orphan here. One possible state of the world would involve A saving his one child, and B winding up with no children.

Another possible state of the world, A winds up with no children because he saves B’s one child. If we just focus on the parents’ good, it might not seem to matter much which scenario is chosen behind the veil of ignorance. But of course, the loss to the children themselves should also matter.

And given that they are the only ones whose– who stand to lose their lives, it might be thought that This is what matters most. And of course, an agreement to save two children rather than one maximizes each chi-child’s ex ante expected utility without violating a concern for making the worst-off position as good as it can be. These considerations suggest that A should save the two children and let his own child die.

But these considerations also imply a lot more. They imply that A should kill his child if this would save the other two children. Similarly, all that supports the view that an agreement would be chosen that results in an ag– that A, an, a, an agreement would be chosen that results in an arrangement that gives A a requirement to let his child die to save B’s two children, would also seem to support the view that A should kill his child if it would help save B’s two children.

It is notoriously hard to see why the difference between killing and letting die should matter morally in a utilitarian system. And considerations that support the view that B should abandon his two children when he has the power to do the saving in order to save A’s only child, given that B has two at home, would also imply that he should kill his two children in the canoe if this is necessary to save A’s only child. For it is also hard to see why the distinction between killing and letting die should matter in a system concerned with maximizing the position of the worst off.

Now, Gibbard, of course, offers a reason why none of these arrangements might be sustainable, namely the strains of commitment. This involves the idea that one cares more for one’s own child than for engaging in a fair reciprocal relations with others in supporting an arrangement that is best for everybody. Before considering the idea of the strains of commitment and what work it can do, let us consider, however, another approach to explaining the unsustainability of the specific requirements that we have been considering.

Some might argue that a utilitarian perspective does not in fact require A to either let die or kill his child to save B’s children. And this is the case where, you know, B has only two children, the original case that Allan Gibbard began with. This is because a parent who could do this, someone might say, could not also have the sort of strong love for his child that in most other circumstances and overall makes for better lives for parents and their children.

So those who contract for the sake of humanistic concerns and maximal utility would want the stronger form of love to exist, and not the requirement to abandon or kill one’s child with which that stronger love might be incompatible. Those who emphasize impersonal values might put this differently by saying that, um, it’s simply they might, they might simply think that a greater value, the possibility of this super strong love entering the world, could be worth even overall losses of life and wellbeing by comparison to a world in which only the shallower form of love exists that permits for a parent to do that, uh, killing or letting die. But notice that such a more, either more intrinsically valuable type of love or even a super strong love that, on the whole, maximizes utility, would also lead to strains of commitment in the following sort of case, and this is where I consider what work strains of commitment can do.

A’s child is drowning. The only way to save her is for A to kill B’s child or children. A might suffer strains of commitment with respect to a rule that prohibits this sort of killing.

Yet this does not in any way diminish our sense, and here I’m just talking about our intuitive sense, that A is morally required not to kill B’s child or children. Just watch its own sadly d- child die. That has to be the fate.

The fact that strains of commitment do not undermine the existence of a requirement in this case, or lead us to think, suggests that it is the content of the supposed requirement that A let his child die in order to save B’s children in Alan Gibbard’s case. It’s the content that makes this not actually be a requirement, rather than that there is such a requirement, that, that rather than that there is no such requirement due to strains of commitment or due to the intrinsic or consequential good of having super strong love, crazy glue love, or we might call it, right? There is another reason why I think strains of commitment do not account for why there is no requirement on A to save B’s children and let his own die.

It is always possible to coerce people into carrying out their duties, and this might maximize utility. For example, suppose A knew that it would be legally permitted and required for the state to put his daughter to death if he failed to let her die and save B’s two children. Then there would be no good reason for him not to carry out his supposed duty and save B’s children.

The killing of a child who was impermissibly saved, as horrible as you think that might be, could be a system that deters other parents from avoiding their utilitarian duties, and so be justified overall on utilitarian grounds.

[01:27:55] ALAN GIBBARD:
Okay.

[01:27:57] FRANCES KAMM:
So what I want to emphasize here is that I haven’t give any, given any positive prescriptions or a positive account to justify various sorts of requirements, um, though I have suggested how impersonal values might play a role. I’ve tried to, uh, focus on, uh, some questions I have about the arguments for expected utility rather than, for example, maximin. And I’ve tried to raise some questions about the reasons that people give for thinking that, uh, we do not have, uh, certain sorts of requirements that I find inadequate as reasons for why they don’t exist.

Thank you.

(applause)

[01:28:45] INTRODUCER:
Our next commentator is, uh, Professor John Broome. Since 2000, Professor Broome has been the White’s Professor of Moral Philosophy in the University of Oxford and a Fellow of Corpus Christi College, Oxford. Prior to that, he had held academic posts at the universities of St. Andrews, Bristol, and London.

[01:29:02] SAM:
He is, by training, both an economist and a philosopher, and in his research, he has used the insights of each discipline to help illuminate the other to the very great benefit of both. He is also a leading contemporary theorist of normativity and the author of five important books: The Microeconomics of Capitalism, Weighing Goods, Counting the Cost of Global Warming, Ethics Out of Economics, and most recently, Weighing Lives. Professor Broome is a fellow of the British Academy and also a fellow of the Royal Society of Edinburgh.

We’re delighted to have him with us this week, and please join me in welcoming Professor Broome.

(applause)

[01:29:44] JOHN BROOME:
I’m gonna be an amateur technician for a minute. Well, thank, thank you, Sam, for those kind remarks, and, uh, thanks to all of you for this invitation to come and speak here. It was a, a, a great honor.

Um, I like this lecture, uh, a lot, and I find myself agreeing with Alan, um, on, uh, a, a great many points. Um, I’m, uh, going to use this opportunity of commenting today to follow up just one of his more important points. In fact, what I’m going to do is, uh, add some extra support to it, add a different argument for, uh, one of Alan’s claims, namely that, um, Harsanyi was wrong to think that we should try and build a notion of goodness out of individual preferences, and instead that we have to concentrate on the good of, uh, individuals.

Um, I’m going to add an argument in support of that, but, uh, I’m also going to suggest that, um, uh, Alan slightly misled us at a couple of places along the path from Scanlon to Harsanyi, which, uh, he took us along. So, um, I’m broadly in support, though I’ve got a couple of, um, uh, disagreements on the way. It’s fortunate that I’m broadly in support because as you can see, I needed Alan to, uh, put these, um, uh, what came as transparencies onto the screen here.

Um, and I hope that I’m now going to be able to, um, use the technology, uh, following his, his lead. Um, as, uh, Alan said, Harsanyi produced two, uh, theorems, um, both intended to give support to utilitarianism. Uh, the first was about the choices that a person would make if she was placed in, uh, behind a, a veil of ignorance in a, uh, an original position and was asked to make choices between, uh, alternative social arrangements.

The second theorem is very different, and that’s the one that I’ve got written, uh, up here, and this is the one that I’m going to concentrate on. This theorem has nothing to do with an original position, uh, or a, a veil of ignorance. It’s not in the contractarian, uh, tradition at all.

It certainly has nothing to do with putting oneself in somebody else’s shoes in evaluating, uh, how well Their life, uh, is going. The first theorem that Alan mentioned certainly does involve that, putting oneself in someone else’s shoes, but this one, uh, uh, doesn’t. And insofar as this gives any support to utilitarianism, it’s support that doesn’t come from the direction of contractarianism.

The basis of it, uh, is different. In fact, we could say its basis is, uh, an intuitive one. What this theorem does is offer some, uh, assumptions at the beginning, which are supposed to come from intuition, and then by inference, it derives a conclusion, uh, from them.

What does it mean? I’ve got it written up there. Uh, let me go through it, st- um, read through it.

It’s got some assumptions, first of all, and the first one is that each person has preferences that are coherent, uh, Alan used that term, and what it means is that they conform to the exp– uh, axioms of expected, uh, utility theory. The second assumption is that social preferences are coherent. That’s to say, they too conform to those same axioms.

Now, we’re obviously going to have to worry if we want to interpret this theorem in detail, what Harsanyi means by social preferences, and we could spend a lot of time talking about that. I’m not going to. I’m just going to, uh, do the same as Alan does on this and take social preferences, uh, to mean simply, as he puts it, uh, ethical betterness.

To say one thing is socially preferred to another, I’m going to take to mean simply that it’s ethically, uh, better, uh, than the other. And the third assumption is that these social preferences satisfy the Pareto condition, which Alan mentioned, and I spelt out in a bit more detail there. Uh, they say, uh, it, it con-contains two clauses.

First, if everybody’s indifferent between two prospects, then the social pros– the social preferences are indifferent. And the second, uh, clause is that if one person prefers the first of two prospects, uh, to the second, and no person prefers the second to the first, then socially, social preferences prefer the first, uh, to the second. So those are the assumptions of the theorem.

The conclusion I’ve got written down there, um, and I’m just going to read it out because the interpretation is debatable. The conclusion is that social preferences can be represented by an expectational utility function that’s a sum of expectational utility functions that represent the preferences of the people. Now, that’s hard to understand and hard to interpret, and actually, I’m not going to spend time on that because for a reason you’ll see in a moment, we don’t need to worry about the conclusion.

We’re going to… We have, uh… There’s going to be a reason to ignore the conclusion, uh, as you might say.

But the only thing I’ll say about it is that it looks sort of utilitarian, and that’s what Harsanyi was hoping to get out of it, that there was some sort of a utilitarian, uh, uh, conclusion would emerge. That, uh, third condition there, the condition that social preferences satisfy the Pareto condition, that’s where Harsanyi formalizes the idea that, um, ethical betterness, goodness we might say, is to be founded on the preferences of, uh, the individuals. That condition says overall goodness depends on, uh, people’s preferences, and that’s something that Alan is opposed to, and it’s something that I’m opposed to too, but my reas-reason is different from Alan’s.

You’ll remember Alan’s arguments against it. His principal argument, uh, was one that, um, uh, involves talking about, uh, Ida and Jay and their preferences behind a veil of ignorance for being in each other’s, uh, positions. But remember, preferences of those sorts for being in the other person’s shoes play no part in this theorem at all.

They do in Harsanyi, Harsanyi’s other theorem, but they don’t, uh, in this one. So those sorts of considerations are not going to be applicable to this, uh, this theorem, and I’m going to offer a different sort of consideration that is, uh, uh, uh, applicable. It arises from a different theorem, which I hope I can produce.

Yes. Here’s another theorem. Um, the three assumptions on that theorem are exactly the same as the three assumptions of Harsanyi’s theorem that I just, uh, stated.

I just copied them out. They’re the same three assumptions. The conclusion is that everybody agrees about the probability of every event.

And that’s a falsehood. It’s not true that everybody agrees about the probability of every event. For example, some people think that the world is going to warm by about four degrees.

It’s likely, it’s probable that the world will warm by about four degrees in the next century, and other people think that’s not likely at all. And some people think, uh, people attach different probabilities to, uh, the Democrats winning the next presidential election. People assign lots of different, uh, very often assign different probabilities to the same event.

People do not agree, uh, about the probabilities of all events. And moreover, that doesn’t mean that any of them are necessarily irrational. There are certainly some circumstances in which we might expect that rational people would agree about the probabilities, uh, uh, of events.

For example, playing roulette in the casino, the roulette table is set up as a machine for delivering particular events with what might be called objective probabilities, and perhaps all rational people ought to agree on those objective, uh, probabilities. So at the roulette table, maybe we can expect that all rational people would, uh, agree about probabilities. But in life, away from the roulette table, that isn’t so.

The sort of events that we come up against in life are ones where rational people might very readily assign different probabilities, uh, to them. And that means that Harsanyi’s theorem simply doesn’t apply to life because the three assumptions of the theorem applies, uh, imply, uh, a conclusion that in life is false. It’s not true that everybody agrees about the probability of every event.

Even if we assume that everyone was rational, that would not be true. So one of Harsanyi’s assumptions that his theorem– in his theorem must actually be a false assumption. Harsanyi was implicitly thinking of life as like a roulette wheel.

He assumed that all probabilities are objective and agreed, uh, by everyone, but they aren’t. So we can’t learn anything about life from Harsanyi’s theorem, that second theorem of his, as it stands. There’s one point in, uh, there’s one point in Alan’s lecture where he, he threatened us with, um, uh, Harsanyi’s theorem if we were not sufficiently utilitarian, because if we were not, he pointed out, we would violate one of the assumptions of the theorem.

But it’s true, we will if we’re not sufficiently utilitarian, violated. But that’s not a threat because we can’t help violating it. One of those assumptions just has to be false.

And given that this theorem depends on assumptions and one of them has to be false, that means we just can’t learn anything about life from that theorem as it is. Now, uh, I- I was going to, um, to try and explain to you the basis of this theorem, but I think, um, I won’t try and give you a general explanation of why it’s, why this theorem applies. Instead, what I’ll do is go straight to a discussion of one of Alan’s examples which illustrates the theorem.

It’s not gonna provide a, a s- a, a proof of it, but it’s gonna give you, uh, an illustration of why when people disagree about probabilities, things go wrong, and things don’t happen in, uh, the ways, the sorts of ways that Alan’s at one point was implicitly, uh, assuming. So I’m going to use Alan’s father’s example, the final, uh, version of Alan’s father’s example, just to illustrate the difficulties that arise when people have different assessments of, uh, probabilities. Uh, so here is the, uh, example formulated in a table that’s slightly different from, uh, uh, Alan’s.

Um, the example is, is, uh, his final version, as I said, when, um, there are two fathers, uh, each with, uh, two children, and either of two things, um, might happen. Um, it might happen that Father A finds himself on the shore with Father B’s two chil-children in the nearest canoe and one of his own children in the further canoe. Or it might happen that Father B finds himself on shore with, uh, two of Fathers A’s children in the nearest c-canoe and one in the furthest canoe.

And the question at issue is what rule, what policy would the fathers, um, should they adopt in, uh, determining which of the children to save? So there are two policies here, the policy of saving your own and the policy of saving the most. Now, Alan assumed, in his formulation of it, that, um, the, there are, uh, objective probabilities of a half each for these two events to happen.

So there’s a probability of a half that Father A might find himself on the shore, probability of a half that Father B, uh, might. And he points out that if the fathers only care about the number of their own children, uh, who will be saved, and they agree that the probabilities are each a half, then both of them will prefer this policy of saving the most. Because if they pick the one of saving their own, then, uh, each father will either, uh, look at fa– look just at A. Father A in the policy of ha-saving their own will either have two children or ni-no children surviving.

Under the policy of saving the most, Father A will either have one child or, uh, two children surviving. And having either two or one is better than having either two or zero, so both would favor the policy of saving, uh, the most, which is the utilitarian solution. Now, this is obviously a highly contrived example, but it’s more highly contrived if you assume that what determines which of the two states will take place, that the two events, uh, will take place, is some sort of a… something like a roulette wheel, a coin toss, with an objective probability of a half.

If this was an example from life, there wouldn’t be any objective probability of which position these people were going to be in. Um, and it’s quite possible that they would assign different probabilities to the different events. So let’s suppose that happens, and let’s suppose, uh, that Father A actually thinks what’s likely to happen is that it’s going to be he who finds himself on the shore and f- uh, in a position with two of Father B’s, um, uh, children near him and, uh, um, his own further, uh, further away.

So he thinks that this possibility, A being on shore, is the likely one. Uh, and this one, B being on shore, is the unlikely one. Now, and suppose he thinks that’s much more likely, then he’s going to prefer the policy of saving your own, ’cause in the thing that he thinks is likely to happen, he ends up with two children saved, whereas in the case which he thinks is likely to happen under this policy of saving the most, he ends up with only one child saved.

So if he thinks that, uh, it’s likely that he’s going to be on shore, then he’s going to favor the policy of saving their own. And if, on the other hand, Father B has the opposite judgment about probability and thinks he’s likely to be unsure, then he’s going to end up with– he’s going to favor the policy of saving their own too. So both fathers in this case will prefer the policy of saving your own.

Now that means that if they were to agree, to get together to agree what would happen, they would agree on the policy of saving your own, which is not the utilitarian policy in this case. Uh, this– that’s the one that only saves two children as opposed to saving three children. So we cannot assume even that, uh, um, when the decision is made by agreement between the fathers, that we will end up with the utilitarian solution as, uh, Alan assumed, um, when people make different judgments about probabilities.

But that was, uh, bes- Uh, beside the point, ’cause it’s not contractarianism that I’m, I’m talking about. I’m talking about Harsanyi’s other theorem, the one that has no contractarian, uh, uh, basis.

What conclusion does this, uh, can we draw from this? Well, the Pareto condition, which is implicit in the theorem, says that save your own is the better policy, because that’s the policy that both fathers prefer. Um, that’s a very hard thing to believe, uh, in this case, it’s saying save your own is a better policy, because the effect of that is that two children survive as opposed to three under the policy, uh, of saving, uh, the most.

And in fact, I c-could, um, uh, explain that, um, it would actually be against the coherence of the betterness relation nor the social preference relation to accept the conclusion that the policy of saving your own, uh, is the best. Um, so that would illu- that illustrates how the Pareto condition gets into conflict with the coherence condition. Coherence in this case would be in favor of saving the most, the Pareto condition would be in favor of saving your own.

The two things are in conflict, and that illustrates why the, why the theorem fails. Now, I think that that tells us that we have to give up the Pareto condition. The Pareto condition takes us wrong in a case like this because it takes us towards a policy, uh, of saving the road.

It has to go. Um, so what can we do, uh, instead? What can we put in its, uh, uh, place?

Well, I think what we need to do is just what Alan says. Give up the idea that overall goodness is going to be determined by the preferences of individuals because in examples like that, that gets us the wrong, uh, conclusion. The preferences of individuals take us wrong in a case like this.

Instead, we have to recognize that overall goodness has got to be founded on the good of the individuals, not on the, uh, preferences, uh, of the individuals. Um, and we can actually simply reinterpret Harsanyi’s theorem in those terms. Stop talking about preferences, start talking about goodness instead, and we end up with this, which is a reinterpreted version of Harsanyi’s, uh, theorem.

I just crossed out preference and put in good instead. We now have three, uh, new assumptions that each person’s good is coherent, that general good is coherent. Nothing to do with preferences now, but, uh, good.

By general good, I mean overall goodness that we were taking to be equivalent to social preference, but I’ve now made that explicit. But, um, I’m now not talking about individual preferences, I’m talking about individual goodness. I’m assuming that the structure of each person’s goodness conforms to expected utility theory.

That’s what the first assumption says. Then I assume what corresponds in this reinterpretation to the Pareto principle, uh, namely that good satisfies what I call the principle of personal good, but it’s the same as the Pareto principle, just reinterpreted instead of good, in terms of good rather than in terms of preferences. And we can now simply apply the same mathematics and get the corresponding conclusion.

The same theorem will give us, um, uh, uh, what’s written down there as the conclusion. I’ll read it out. General good can be represented by an expectational utility function that’s a sum of expectational utility functions that represent the good of the people.

Now, that, that conclusion is actually one we should pay attention to, because that one can be derived from assumptions that are not necessarily false or not inevitably false, as, uh, Harsanyi’s own theorem, uh, was. This one is a conclusion we need to pay attention to, but unfortunately, there isn’t time, uh, to do that. But let me say once again, I think you can see that it sort of looks utilitarian.

It’s not utilitarian. It certainly is quite a lot weaker than what utilitarianism actually says, but it may give some sort of support to utilitarianism. And it avoids the difficulty that I, um, uh, described to you that Harsanyi’s theorem encounters.

Because the problem of Harsanyi’s theorem w-was one to do with disagreement amongst probabilities, and that doesn’t arise in this case. Why doesn’t it arise? It’s because when you assess another person’s good, you’re going to have to do it on the basis of your own probabilities, not her probabilities.

Her probabilities determine her preferences, but your judgment about what’s good for y-her has got to depend on your probabilities, not her, hers. So for example, suppose you s-see somebody s-heading into a dangerous situation or what you think is a dangerous situation. You would judge that that’s bad for her on the basis of your judgment that she’s standing into danger.

And the– even if her probabilities are quite different, and she thinks whatever she’s doing is perfectly safe, that won’t stop you from judging that she’s doing something that’s bad for her. She no doubt prefers to do it on the basis of her probabilities, but your judgments of good for her will depend on yours. So that’s means that when you are stating this theorem– When I’m stating this theorem, the good that I’m talking about is assessed in terms of my own probabilities, and those will remain the same all the way through, uh, the statement of the theorem.

There will be consistent probabilities when the theorem is stated in my voice, and there will be consistent probabilities when it’s stated in your voice too. Uh, so there will be– there won’t be the problem, um, that arises. So I think this gives a different argument, different from Alan’s, but for the same conclusion f- for as Allen’s.

Harsanyi’s second theorem, when you look at it carefully, can only be understood properly if we f- change our attention away from people’s preferences and concentrate on the good of people. Thank you.

(applause)

[01:52:06] ALAN GIBBARD:
Okay. Well, well it’s a great privilege to, uh, to have commentaries by, uh, two such marvelous philosophers. And, uh, apparently we all have to do a lot of thinking between now and tomorrow.

Uh, but let me say a few things about, uh, each of these, uh, uh, wonderful commentaries of the position I took on intuitions, of course, is quite, uh, quite ambivalent. I was saying that, uh, on my kind of view, we’re dependent on intuitions also. So, and so the question is going to be how, uh, how, uh, sort of case-based are the intuitions?

What, uh, what rationale, uh, will we be able to discover for them? Are we, uh, do things come out when we try to make them coherent the way Sidgwick thought they did with, uh, some very broad, uh, intuitions or, uh, that are, uh, far– whose, uh, connection to, uh, everyday moral thinking is quite indirect, or do we get intuitions that are mo- much more closely connected with, uh, everyday intuitions? I was urging that we understand, uh, intuitions as bearing on this planning question of how to live with, uh, others, uh, assuming that we should live on with others on a, uh, that an important goal is to live with, to have is to live with others on a basis of mutual respect.

But that doesn’t in itself rule out, uh, intuitionism in a style quite different from, uh, from Sidgwick. Uh, now let me just make a f-few broad comments, uh, uh, about the way, uh, systematic considerations bear on, uh, intuitions. If we’re, uh, in the case of equality, uh, I see lots of reasons to want equality, uh, the kinds of social relations it makes possible, uh, the, uh, the the dangers of dominance from when people are unequal.

Uh, and certainly I think that we have, uh, uh, a Jonathan Haidt sort of, uh, emotional intuition about, uh, equality. Uh, these, uh, come, come with intuitions about, uh, hierarchy as, as well. Uh, so Alan Fiske thinks that universally people uh, organize social relations in, in four different ways, uh, uh, and those include, uh, market probably fourth and primitive conditions, uh, uh, communally, uh, hierarchically, and egalitarian.

And so I think there are strong emotional, emotional intuitions attached to both, uh, equality and, uh, hierarchy. But, uh, in the case of equality, uh, there’s a question of why to, why to want it, and, uh, why to want it. I… one reason to want it is, uh, that for non-moral reasons, that’s, uh, what we’d want, Uh, that would be the thing to want for oneself in conditions where, uh, I didn’t know whether I was going to be on top or on the, on the bottom, and that leads to the, uh, usual utilitarian style, uh, arguments for equality.

Uh, if we’re talking about a greater equality than we would want for non-moral reasons, not knowing, uh, who we would benefit by, and if those are the, the wants to have for ourselves in those circumstances, that, uh, I’m puzzled about why to want that much equality. Uh, I’m also, uh, quite drawn to, uh, to expressing things in our actions, but that leaves a, a puzzle of why am I expressing, uh, one thing when I kill someone and expressing something different when I fail to save someone when I could is uh, is this a kind of expression that, uh, values, uh, crocodile tears which are very expressive, but, uh,

(coughs)

but don’t very much bear on action? Uh, well, uh, seems to be, again, one question to ask is, well, suppose, uh, suppose we ask the question what to prefer for a social order, uh, not knowing in particular how we and those dear, uh, dear to us are going to be affected, um- Uh, well, it seems to me there are lots of reasons to want, uh, strong prohibitions on killing, uh, and, uh, lots of reasons to, uh, be worried about, uh, equally strong prohibitions on, uh, failing to, to rescue. Duties not to kill are, are more focused.

They establish a, a bright line. Uh, so do we really think that, uh, these prohibitions on, uh, this greater stress on not killing than, uh, than on rescuing is, uh, something that, uh, uh, that has nothing to do with what to want for non-moral reasons, uh, prospectively? It seemed to be that the, uh, that, uh, there are strong reasons to want such a distinction prospectively, and, uh, that that’s the place to look for, uh, for, uh, a, a view that, uh, that, uh, reconciles our intuitions about, uh, uh, how to want to live with other people, uh, in society.

Um… So, um, okay, well, uh, let me leave other things for, uh, tomorrow from this, uh, marvelous commentary. As, uh, John said, I, uh, there’s– there are large areas of agreement, uh, but that–

But he says I misled people, and, uh, I guess I, I guess I did. Uh, I talked about theorems, and there are two ways of thinking about theorems in applied mathematics. Uh, one is as interpreted theorems.

Now, uh, I talked about the Pareto condition. Pareto was, uh, as I understand it, adamant on using, uh, using preferences and finding, uh, ordinal notions. Uh, that’s not entering into my treatment, so I was thinking about the, the formal structure of the Pareto condition, and we both agree that the formal structure can be given an, an interpretation in terms of preferences and in terms of a person’s good, uh, and we both agree that the interpretation in terms of a person’s good is the one that’s, uh, morally relevant.

Uh, so, uh, Uh, uh, now the, the second Harsanyi theorem, uh, John says, uh, has nothing to do with, uh, with putting oneself in each other’s shoes. Uh, he says it has nothing to do with, uh, contractarianism, and, uh, I think I was using it as a, as a different kind of, of argument from the contractarian one, but, uh, we both agree, uh, Francis doesn’t, that it’s an argument with a, a great deal of force. Uh, but it has something to do with putting oneself in other people’s shoes on some interpretations and not on other interpretations.

Uh, John, uh, at least here, is, uh, helping himself to the, the notion of a person’s good. I’ve been worried about Scanlon’s, uh, critique, and so, uh, trying not to help myself to the notion of a person’s good. Also, I want the notion of a person’s good to come out of realizing that, uh, that moral thinking is, uh, thinking how to live with each other and why.

Uh, and so I want the concept of a person’s good to, uh, come out of such thought. So we can evaluate a, a state by, uh, uh, uh, the prospective good of the people involved, interpreting the prospective good in terms of, uh, doing the thought experiment of, uh, uh, as Hare would put it, uh, knowing that one is forthwith to be that person, and then, uh, thinking, uh, what to want, what counted for, uh, purposes of reciprocity. Uh, under that interpretation, uh, the, uh, the theorem does have something to do with put-putting oneself in other people’s shoes.

Um, uh… Okay. So, um, so we end up in the, in the same way.

Anyway, uh, I try to get there by reinterpreting, uh, Harsanyi’s, uh, interpret– by changing from Harsanyi’s interpretation of the, uh, Harsanyi theorems in the, uh, Pareto tradition. Uh, I think at one point I slipped back into talking about preferences, but, uh, but that was a mistake. Uh, we both agree that it requires this, uh, reinterpretation.

Uh, I’m, uh, suggesting that we can understand the reinterpretation in terms of, uh, thought experiments of, uh, uh, supposing that one’s to be the person in question. Uh, thank you.

(applause)

[02:03:20] SAM:
We’re just about out of time, so I think that rather than try to squeeze in a question or so today, I think we’ll just, uh, ask you to hold your questions for tomorrow. Tomorrow’s format will be that there will be, uh, additional comments made by each of the three commentators, uh, and then there’ll be discussion amongst the panel, and there’ll be, uh, ample opportunity for people to ask questions from the audience as well. So please come back tomorrow at, uh, same time, four o’clock, four ten, uh, for the third and final installment of this Tanner series.

Thank you very much.

(applause)

[00:00:00] (laughter)

[00:00:01] MODERATOR:
All right. Uh, we-welcome to the third of our, uh, uh, uh, Tanner Lecture events this week. Uh, today’s format will, will involve a, a kind of seminar and discussion.

Uh, the three commentators will be presenting further remarks about, uh, the two lectures that Alan Gibbard has, um, has given us this week. Uh, we’ll follow that with a break, and then there will be a response to the, the three new sets of comments from, uh, Alan Gibbard, uh, which will then um, segue into an open discussion. Uh, and I want to alert you in advance to the fact that after the discussion is over around 6:30, we will, uh, there will be a reception with some refreshments that will magically appear behind this, uh, this wooden wall.

Wall, the wall will, will open. And, um, so, so that’s something to, to think about and keep in mind. Um, uh, we’ll be starting with, uh, with Michael Bratman.

[00:00:55] MICHAEL BRATMAN:
Okay, I should come up there, I suggest.

[00:00:57] MODERATOR:
Uh, I, I think that’s fine. Is that comfortable for you?

[00:00:59] MICHAEL BRATMAN:
It’s fine. So let me thank the Tanner Committee for again, for the privilege of participating in this wonderful event. Central to Allan Gibbard’s planning theory of normative thought is the idea that plans are subject to constraints of consistency and coherence that to some extent parallel such constraints on beliefs.

This idea is an important element in the argument that planning to act or to prefer can be a kind of judgment. In particular, a normative judgment. I agree with Gibbard that attitudes of planning are subject to constraints of consistency and coherence, but I want to examine Gibbard’s story about why this is so.

Gibbard’s thinking about plan consistency and coherence is influenced by another feature of his view, namely that the plans we need to appeal to in understanding normative thinking include plans for contingencies one knows one will never face.

(coughing)

If I judge that Caesar ought not to have crossed the Rubicon, then I in some sense plan not to cross were I Caesar. But I know I will not be Caesar. This lends special force to Gibbard’s question, quote: “Why does coherence in plans for action matter, especially when they are plans for wild contingencies that we will never face?”

Gibbard notes that a problem with inconsistent beliefs is that they cannot all be true. But he does not think we can simply extend this to incoherent plans. He does, in the end, think we can say of plan-laden judgments that they are true or false.

This is the quasi-realism of his 2003 book. But this talk of true plan-laden judgment must come at the end of the story when we have earned our right to it. We are now, however, at the beginning of the story, and we need a different kind of answer to the question why plans are to be coherent.

A word about terminology. Gibbard does talk of consistency of plans and of plans with beliefs, but his preferred terminology in talking about plans is incoherence. So with an exception noted below, I’ll follow him there.

Gibbard notes that, quote, “Incoherent plans can’t all be carried out.” But Gibbard does not think this gets to the bottom of things. As noted, it’s part of his theory that many of the plans that constitute normative judgment are wild contingency plans that will never be in a position to carry out anyway.

So he suggests we need to ask, why should it matter if these wild contingency plans are ones that one will never be in a position to carry out? I think Gibbard’s right to reject an easy extension from an account of belief consistency and coherence by appeal to truth to an account of plan coherence. And I think he’s right to wonder why the coherence of wild contingency plans should matter, since it’s not clear why it would matter if all of them could not be carried out.

The issue we face is how to proceed once this is granted. First, though, let me say a bit more about how Gibbard is thinking about constraints of coherence on plans. As he explains in his second lecture, he’s thinking primarily of what he calls the standard conditions appealed to in decision theory as together ensuring that a person’s choices between options can be seen as maximizing an expected value.

These standard conditions are conditions on preferences concerning options. One such condition is that one not strictly prefer A to B and also strictly prefer B to A. Another condition is one is transitivity of preference. Yet another condition is a completeness condition.

As between two options, one is to have some preference or other. And these preferences are conceived as ranging as well over lotteries as well as over simple options. Gibbard supposes that the relevant constraints of coherence on plans are in effect constraints on preferences, constraints of the sort captured in these standard conditions.

He’s also assuming, I suppose, that to plan to A is to prefer to, is to prefer A to its alternatives, and he’s assuming, and this seems problematic to me, that to prefer A to its alternatives is to plan to A. I’ll briefly come back to this at the end. What about plans for weights, which figured in his picture as I– as we talked about at the first lecture?

Gibbard seems to be thinking in this context about coherence and consistency. He seems to be thinking of such plans for weights as implicit in a complete contingency plan over actions. Coherence applies to plans for weights indirectly, is the view, I think, by way of coherence constraints on preferences over options.

I myself would prefer a model of our practical thinking that sees plans for weights as elements in their own right, but I’ll mostly put this aside here. So Gibbard’s appeal to coherence of plans is in a way better seen as a story about coherence of preference. Still, it remains true that one central kind of incoherence of plans that’s his focus in mind is inconsistency of all the different things you plan to do with your beliefs about the world.

And it– and it– this is the kind of inconsistency I’ll focus on below. So what is Gibbard’s answer to his question about why coherence in plans for action matters? This is what he says, and I th– I put a lot of weight on this passage, so I’m just going to read it to you.

I apologize if it’s– easy to miss this, I think, in the lectures, but but I believe I’ve got them right. Okay, but we can find out. Okay.

The problem… here’s the quote. “The problem with inconsistent plans is that there’s no way they can be realized in a complete contingency plan for living.” For each full contingency plan one might have, something in the set will rule it out.

Or more completely, we’d have to talk about an inconsistent combination of beliefs, plans, and constraints. If a set of these is inconsistent, there’s no combination of a full contingency plan for living and a full way that the world might be that fits, and judgments get their content from what they are consistent with and what not. So that’s his answer to the question.

So his idea, I take it, is this: We assign content to the agent’s overall set of beliefs and plans by asking what overall possibilities are consistent with those plans and beliefs. So inconsistency of plans given beliefs, even while contingency plans that are, in the theory associated with normative thinking about scenarios will never be in, like that of Caesar, will tend to baffle the holistic assignment of content, and so baffle the treatment of those plans as judgments with the precise content. Call this a content theory of why planned coherence matters.

We can express this in terms of the idea of interpretation, which Alan introduces in the first lecture. Coherence of planning attitudes, at least coherence for the most part, that’s actually an expression from his 1990 book. Um, at least coherence for the most part is needed in interpretation to ascribe content.

If you seem to plan to A And you seem to plan to B and you seem to believe that A is contingently incompatible with B, well then, there’s significant interpretive pressure against saying, against saying, that you really do plan to A, really do plan to B while believing they’re not co-possible. Our interpretation of you tends to be to that extent thwarted.

Perhaps one can be interpretable while violating a strict constraint of coherence, but such violations must be the exception. To be interpretable, one must conform at least for the most part. So the constraints of coherence on plans and preferences are an aspect of the second of the three areas of inquiry Gibbard identifies in the first lecture, namely interpretation, in which we, as he says, understand some of these natural goings-on as beliefs, assertions, plans, and the like with which we can agree or disagree.

And this area of inquiry is, according to Gibbard, to be distinguished from normative inquiry. So constraints of coherence on plans and beliefs enter once we do interpreted psychology. We can then try to say that certain aspects of that interpreted psychology, the planning aspects on the view, constitute the ought judgments essential to normative inquiry.

In normative inquiry, we can go on to consider whether one ought to be coherent in plan. But the judgment that one ought to be coherent in plan, if that’s what we do judge, is not what is at the the root of why coherence in plan matters, according to Gibbard, as I understand it. According to the theory, what is at the root is the connection betwe-between coherence and interpretability.

And this seems a virtue of the theory. If we were instead to say that coherence and plan matters because we ought to be coherent, where this is an ought judgment within normative inquiry, it seems we might be threatened with a circularity. To explain what that ought thought consists in, we need to appeal, according to the theory, to plans as judgments.

But to explain the nature of those planned judgments, we would need that ought thought. We avoid this circularity if, to explain why plans are judgments, we appeal not to the thought that we ought to have coherent plans, but rather to the thought that coherence, for the most part, is constitutive of planning in the sense that it’s a constraint on interpretability. I do have a worry.

A minimally realistic psychology will include, in addition to attitudes of planning to act, something like non-instrumental desire. Suppose then that I desire, non-instrumentally, though I won’t keep saying that, to be fabulously wealthy, and also desire to lead the life of a scholar. In the world as I know it, however, these desires can’t both be fulfilled.

That certainly does appro– does pose a problem for me. But it doesn’t show that I’m failing to satisfy a relevant constraint on desires. Such known contingent conflicts of the desire seem an inevitable feature of our lives.

What we reasonably aspire to is not eliminating these conflicts, but negotiating a life in light of them. This contrasts with desiring to be wealthy and desiring to be poor. Desiring things that are of necessity in conflict does seem to violate a relevant constraint on desire.

But that’s different from desiring things that are known only to be contingently incompatible. Now consider planning. Suppose I plan to be fabulously wealthy and also plan to lead the life of a scholar despite my belief that these are contingently incompatible.

Here I agree with Gibbard that there is a violation of a coherence constraint on planning attitudes. Since it’s here the talk of consistency seems most natural, let’s call this a constraint of strong consistency. That is, consistency of plans both among themselves and with beliefs.

But now we need to ask, why are planning attitudes subject to this strong constraint in a way in which desires are not? To answer that question, we cannot just note that incoherent plans can’t all be carried out. After all, my desires to be wealthy and to be a scholar can’t all be carried out in the world as I know it.

But that’s just the human condition. The problem is to say what’s special about planning attitudes that explains this difference between planning and desiring. And my worry is that the content theory, as so far developed, does not by itself answer the question of why in particular planning attitudes, in contrast, say, with desires, are subject to the constraints of strong consistency.

After all, we’re able, in interpretation, to assign contents to desires. And we can agree or disagree with what others or our past selves desire. None of this seems to require that desires are subject to a constraint of strong consistency.

So we need to know why, in contrast, inconsistency of plans with beliefs baffles interpretation and content. Perhaps the idea that plans are judgments does depend on this appeal to strong consistency, though when you kind of see it this way, if you, if you’re one of those people who thinks of desires as judgments, you, you probably want to stop there, but we’re not going to go there. So perhaps the idea that plans are judgments does depend on this appeal to strong consistency, but at this point in the argument, we’re looking for reasons to agree that plans are judgments, so we can’t appeal to the id– to that idea to defend the role of strong consistency in plan content.

I think then that we need at least to supplement the content theory. And the natural strategy here is to say more about the roles planning attitudes play in our lives. As I see it, and I think Alan would agree, the basic roles of plans and planning are coordinating roles, both within the life of the agent and socially.

Plans of action coordinate our actions over time and socially in the pursuit of temporally extended and, in some cases, socially shared ends. Plans for weights in deliberation coordinate practical thinking at a time and over time, and potentially socially. Ordinary desires, in contrast, don’t have this coordinating role.

Rather, and to put it crudely, we seek coordinating plans in part as a way of satisfying relevant and potentially conflicting desires. It seems that it’s these coordinating roles of plans that lie behind the idea that plans are subject to a constraint of strong consistency in a way in which ordinary desires are not. One way to express this idea is to draw on an apparent parallel with belief.

Belief, many say, famously Bernard Williams, a former colleague of mine here at Berkeley, belief, many say, aims at truth. What this means, we can say, is quite roughly that an attitude does not count as belief unless it’s embedded in a psychology that shapes that attitude in a way that tends to track truth. Analogously, we can say that for an attitude to be one of planning, it is necessary that the attitude be located in a psychology that shapes that attitude in a way that tends to track compatibility with other planning attitudes and beliefs.

In this sense, attitudes of planning aim at coordination with other attitudes of planning, given one’s beliefs. Systematic violations of strong consistency manifest an underlying psychology of the wrong sort for the attitudes to be ones of planning, and so undermine interpretation of the subject as a planning agent. How would this help with Gibbard’s problem about wild contingency plans?

Well, if they really are plans, they need to be embedded in a psychology that adjusts in a way that tracks coordination with all the agent’s other plans, including other wild contingency plans. This isn’t a defense of the idea that my judgment that Caesar ought not to have crossed the Rubicon is my plan not to cross under certain conditions. But if it is a plan, that is, if my ought judgment is, as Gibbard’s theory says, my planning so to act, then it will be subject to pressures for coherence with the rest of the edifice of plans.

None of this shows, it’s important to see, that we have reason to make our plans coherent. The judgment about reasons for coherence is a normative judgment, one that goes beyond the thought that coherence for the most part is a constraint on interpretability. That’s the difference between interpretive inquiry and normative inquiry in Gibbard’s views.

It might well be that we do have reason to make our plans coherent, since, after all, we have reason to be effective agents, and coherent plans contribute to that. Indeed, that we have reason to make our plans coherent. is the view towards which the initial observation that incoherent plans can’t all be carried out seems to point.

But these are judgments within normative inquiry, albeit judgments Gibbard mi- Gi-Gibbard might well go on to make. On the theory, however, these ought judgments are not the basic answer, though they are an answer, to why coherence of plans matters.

You might say we’ve uncovered two senses in which plan

(cough)

coherence matters. One that appeals to a condition of being interpretable as a planning agent. A second that appeals to reasons to be coherent, and it’s the former that in the theory is primary.

John Broome in conversation worried that there remains a problem for the theory in its unders– This was at breakfast this morning. There re– It’s amazing how it goes from breakfast to a talk very fast. But what…

It’s the nature of this kind of work, working thing. So John, in conversation, worried there remains a problem for the theory in its understanding of my thought that I ought to be coherent in plan. So the idea is, um, let’s make sure we got it right now.

So the idea is the basic answer to the question now is you need coherence for interpretabil-interpretability as a planning agent. But to support that, we need to get clear about the coordinating role of plans and the idea– sense in which they aim at coordination. Now, like this, the way belief aims at truth.

Okay. But so far, we just have this idea that coherence is, is, is kind of a condition of in-interpretability as a planning agent. We don’t have the idea in normative inquiry that we ought to be coherent, though of course we might have that idea and, uh, um, at the end I’m going to say we need it.

Um, but, um, okay, So that’s where we are. But then the– Broome’s worry this morning at breakfast, um, was that there seemed maybe a problem for the theory in its understanding of, say, my thought that I ought to be coherent in plan.

My thought is for Gibbard, remember, something like my plan to be coherent in plan. So I think I ought to be coherent in plan, and that’s a plan to be coherent in plan, right? But do I then need yet a further thought that that plan to be coherent in plan ought to be coherent with my other plans?

Uh-oh, does regress loom?

[00:18:07] AUDIENCE MEMBER:
Right.

[00:18:09] MICHAEL BRATMAN:
Well, I think Allen can avoid a regress here by saying that my thought that my plans ought to be coherent, which thought I might have in normative inquiry, is my plan that all of my plans, including this very plan, together be coherent. Re-reflexivity staves off Lewis Carroll’s tortoise. Let me note another virtue of the appeal to coordinating roles.

It may help a Gib– The idea I have is it may help a theory like Allan’s solve a problem about the metaphysics of agency, a problem that needs to be solved by an adequate metaethics. Thoughts about oughts and reasons…

Here’s the problem. Thoughts about oughts and reasons normally are central elements of what we can call the agent’s practical standpoint. When these elements of this standpoint guide thought and action, then, at least normally, the agent guides thought and action.

The agent governs. Thoughts about oughts and reasons normally have what we can call agential authority. When these thoughts guide, the agent governs.

Appeal to ought thoughts, then, will be one part of a story of how agential guidance, guidance by the agent, can be constituted within a psychic economy of events, states, and processes. Or so it seems. For this to work, we need to know what it is about thoughts about oughts and reasons that, at least normally, gives these thoughts regential authority.

That is to say that when they guide, the agent governs. If thoughts about oughts and reasons are responsive to independent normative facts of the sort posited by the normative realists, then we could try to explain the regential authority by appeal to such responsiveness to these normative facts. This would be a kind of Platonic theory of regential authority.

I mean, think about why reason has authority in the Republic, right? In the tripartite story of the soul in the Republic. Okay.

But if thoughts about oughts and reasons are planning attitudes, as they are in Allan’s theory, why do they at least norm– h-why do they at least normally have such authority to speak for the agent? See, I think this is a problem for any metaethical theory, and it’s important to see that. Well, I think a theory like Allan’s can say, In light of the conversation, the discussion I’ve just been having, these planning attitudes speak for the agent because their roles are to organize and coordinate the agent’s thought and action over time.

When functioning properly, these planning attitudes help create the unity of agency at a time and over time that’s presupposed in talk of agential governance. And this has the further advantage that this account of agential authority can be extended to other planning attitudes that are not strictly speaking normative judgments because it’s, okay, generally applicable. Okay.

So a Gibbardian view developed along these lines has much to recommend it, I think. I do suspect that this emphasis on cross-temporal coordinating roles exerts pressure against Gibbard’s apparent identification of planning and preference. Since preference seems less tightly, to me anyway, seems less tightly tied to such cross-temporal coordinating roles.

I mean, one way to see this is the standard conditions on preference are not about cross-temporality at all. They’re time-slice conditions. But in closing, I’m going to focus on a different issue.

And this is in closing. Okay. Gibbard’s basic account of why plan coherence matters appeals to interpretation, not to reasons to be coherent.

That’s what we’ve seen. But I suspect that some of the uses to which he puts the idea of coherence in– especially in the second lecture, need the view that we have reason to be coherent. I take it, for example, that Alan would say you have reason to avoid incoherence in your plans concerning Zeckhauser’s Russian roulette example.

After all, if the only problem about incoherent plans were that they tended to undermine your interpretability, you might think, “So what?” But this takes us back to the question, why do we have reason to avoid incoherence in plans, given that many of the plans at issue are wild contingency plans? That was the original question that started this.

And Gibbard’s content theory, even when supplemented by the appeal to coordinating plans, doesn’t yet provide an answer to that question. Okay.

(crowd applauding)

[00:22:27] MODERATOR:
You have to come to a philosophy event to experience pronouncement like reflexivity staves off Lewis Carroll’s tortoise.

[00:22:33] MICHAEL BRATMAN:
Right. That was an in-joke.

[00:22:35] MODERATOR:
Right. It conjures an image of Michael walking along and about to be attacked by a tortoise, pointing to himself. Anyway, um, our next commentator today will be, uh, John Broome.

[00:22:46] JOHN BROOME:
Thank you. Um, I’ve got a handout which will be…

[00:22:55] MODERATOR:
Is that all right?

[00:22:56] JOHN BROOME:
Yeah. Um, a few copies, yeah. Um, shall I fix them here?

(paper rustling)

[00:23:02] MODERATOR:
Of course. Um, oh, I’ll, I’ll pass them out, please.

[00:23:04] JOHN BROOME:
Thank you. It’s not meant to be comprehensive, I’m sorry. Then it was originally planned to be an overhead projected thing, but, uh, it’ll be useful for you to have it.

It might be useful in, in discussion. Uh, I’m also going to talk about, uh, Alan’s, um, first lecture. I, I hope what I’ll have to say is complementary to what Michael says, and it, uh, doesn’t overlap much, but I think it fits quite well with what he’s, uh, just been talking about.

Um, take any normative sentence such as, uh, “Caesar ought not to cross the Rubicon,” or, uh, “Nobody ought to have, uh, conflicting, um, intentions.” Philosophers used to worry a lot about whether sentences like that, normative sentences, could be true, uh, or false, many of them denied that they could be, that they were the sort of thing that could be true or false. Uh, these days, we worry less about that.

Um, most of us think it’s not so hard for a sentence, uh, in a particular class, um, to be true or false, and for some sentences in a particular class, uh, to be true. Um- because we think that all that’s required for them to be true is, or false, is that they, um, participate in our, our thinking, our lives and our discourse in, uh, characteristic ways. Um, for instance, we need to do, uh, truth-functional logic, uh, with them.

Uh, we need sometimes to disagree, uh, with each other, uh, about them. Um, we need to think that just because, uh, somebody utters one of these sentences in an assertive fashion, that doesn’t mean it’s true, um, even if the person is, uh, justified in doing what she’s, uh, just done. Um, and so on.

So that there are ways that you can think of as characteristic of truth, um, uh, in which, uh, a class of sentences might, might participate. And so long as it does, then it earns the right to be, uh, treated as the sort of thing that can be, uh, true or false. And, um, it’s pretty clear that normative sentences meet those requirements.

They do meet the standards that are characteristic of, of, uh, truth and falsity. So we take it that they can be true or false, and if one of them is false, then its negation, uh, is, is true. So we don’t worry so much about that, uh, these days.

And of course, if these sentences can be true or false, it goes along with that, that we can have attitudes of belief or disbelief, uh, towards their, uh, towards them, um, or towards their contents, uh, towards, uh, what they say. So we can have cognitive attitudes towards sentences, uh, of this sort. And I don’t think that’s any longer particularly controversial.

Um, on the grounds of what we do with these sentences, we can accept that, um, that they, uh, are true, uh, or can be true or false. But that does leave us with the question of why is that, uh, so? They participate in our thinking and discourse, uh, in ways that are characteristic of, of truth, but we can wonder how come.

How come, how come that they participate in our lives, uh, in those ways. If we were dealing with sentences about, uh, natural matters, then we could give a, uh, an explanation o-of that, which would have something to do with the way that they correspond to, uh, facts in the, uh, natural world. We would say that they participate in our discourse in the way that they do because of their correspondence with, uh, uh, the facts.

Um, we could give a parallel answer for normative sentences, but many people, including Alan, would find that fantastic. Uh, he just finds it incredible that there are normative facts at this– that, that roughly correspond to natural facts that would explain why our normative sentences work in the ways that our, our natural sentences, uh, do. So he offers to provide an alternative explanation of why our discourse with these sentences works in this truth-characteristic, uh, way.

And his explanation is that these sentences, um, help us to plan, uh, our lives, plan our lives in general, and plan what to do on particular occasions. It’s a natural fact about us, he points out, that we are the sort of creatures, uh, who do make plans. Um, so Alan’s explanation of why these sentences work in the way that they do doesn’t appeal to anything apart from our na- our, our nature, uh, how we are in the, the, the, the natural, uh, world.

Um, but what we’re doing when we utter a normative sentence is, uh, expressing a partial plan, or to at least a first approximation, uh, that’s what we’re doing. Um, and on that basis, Alan explains, uh, why, um, uh, as our thinking and, uh, discourse, um, uses those sentences, um, it’s going to endow them with the characteristics that go along, uh, with truth. Uh, for one thing, he provides a semantic theory, uh, that, uh, that explains how they participate in truth-functional logic in the way that, that, uh, they need to.

So he’s got an explanation then of why we can have the attitude of belief or disbelief towards sentences, uh, of these sorts. And the explanation is that those cognitive attitudes of belief and disbelief arise from non-cognitive attitudes, which are, uh, uh, planning, uh, attitudes. In fact, our cognitive attitudes are really simply those planning attitudes in another guise.

They’re nothing other than the, uh, planning attitudes. So for instance, to believe you ought not to feel resentment in particular circumstances is to do nothing else than plan not to feel resentment in those, uh, circumstances. And Allan, uh, argues that these planning attitudes are woven together in a, a structure that, uh, mimics the structure of believing attitudes, and that’s how we can treat them as believing attitudes indeed.

Uh, that explains why their, uh, um, their contents have the logical structure of, of truth. Um, so our attitudes towards normative sentences are basically non-cognitive, but they, as, uh, as Michael just said, they earn the right to be treated as cognitive, uh, attitudes because of the way that they fit together, and that arises from the way that they participate in, in our, our thinking and discourse. Um, so in Allan’s account, each attitude of normative belief is going to be accounted for by a non-cognitive planning attitude that’s the same as it, equivalent to it, uh, in fact.

So to believe you ought not to feel resentment, as I say, is just to re– to, um, plan not to feel reject the re-resentment, or as a matter of fact, as he puts it, you reject feeling, uh, resentment. And to believe that, um, Caesar ought not to c-cross the Rubicon is just to plan when in the position of Caesar, or if in the position of Caesar before the Rubicon, to plan, uh, not to, uh, uh, cross it. Now, that though is only an approximation, uh, as Alan says, and there are at least two reasons why it’s only an approximation.

And I’m going to, uh, mention them, the, the two of them. The first is actually one that Alan himself describes. Um, even if it were true that all ought beliefs are plans, it would not be true that every plan is an ought belief.

Buridan’s ass is the illustration of that. Buridan’s ass stood between two equally attractive, uh, bales, uh, of hay. As it happened, it couldn’t decide to go to one or the other, uh, and, uh, it died.

But if it had been more sensible, it would have formed the plan of eating one or the other bale. But it didn’t think that it ought to eat one or the other bale. They were both equally desirable.

It didn’t think it ought to go left, and it didn’t think it ought to go right. But nevertheless, had it been more intelligent, it would have planned to go one way or the other. So then it would have had a plan that was not, uh, uh, an auth- an ought belief.

So that must happen. And because of that, Allan doesn’t, um, find normative belief, uh, strictly, um, on attitudes of planning, but on what he calls valenced attitudes that… or, or the– ultimately, on one valenced attitude, which might be called the attitude of okaying, of, uh, attaching an okay, of assigning an okay attitude towards the thing. So if Buridan’s ass had been more sensible, it would have assigned an okay attitude towards eating the left bale or to– and towards eating, uh, the, the right bale, but then it would have planned to go to, uh, one or the other.

Um, from– for Allan’s purposes, this attitude of okaying is not to be initially identified as an attitude of thinking it’s okay. So the attitude of think– of okaying the left bale that the ass would have had is not the attitude of believing that the left bale is okay. It’s something other than that.

It’s a primitive, non-cognitive attitude, but it will, in due course, in the working out of the theory earn the right to be treated as the cognitive attitude of believing that the, the left bale, uh, is okay. So that’s one way in which to say that normative attitudes are plans is an approximation. Actually, they’re not really plans, they’re these okay-ing sort of, uh, attitudes.

But there’s a second way in which the, uh, attitudes he’s using, um, uh, are not really planning attitudes, and this is one that he doesn’t mention. And that is that the attitude, uh, of believing you ought to do something just simply is not an attitude of planning to do it. Not only can you plan to do something without believing you ought to do it, you can also believe you ought to do something without planning to do it.

If Buridan’s, uh, uh, ass had been, um, uh, more sensible, uh, then, um, it certainly would have had the attitude of planning to do something without believing it ought to do it. But it’s also quite possible for an ass or anybody else to believe it ought to do something, but, uh, not plan to do it. If you think that, you’re in a state that’s often– that’s, um, universally really in philosophy called the state of akrasia.

Most philosophers think the state of akrasia, that’s to say believing you ought to do something without planning to do it, is irrational, but most of us think it’s possible, though, uh, irrational. However, many non-cognitivists, including Allan, find themselves in a position effectively of denying that akrasia, uh, exists. Richard Hare is another example.

Um, what Hare said is that you couldn’t sincerely believe you ought to do something without intending, uh, uh, to do it. Uh, I must say, I find that unconvincing, and a lot of other people I know find it unconvincing. It seems to me quite possible sincerely to intend to do something, uh, to believe you ought to do something, uh, without, uh, intending to, to do it.

Um, Alan doesn’t actually explicitly deny that but on the other hand, he does use language that elides the distinction between, on the one hand, thinking you ought to do something and on the other hand, planning to do it. So at one point, he says that thinking what I ought to do is thinking what to do. Now, I suppose it’s true that in a way you could say that thinking what to do is the same as thinking, uh, what you ought to do.

Um, but that’s not actually a very idiomatic usage of that term, and anyway, there are certainly two different questions that you can ask yourself. There’s the question of what to do, and there’s the question of what you ought to do. I’ve put, um, the, an example of those two quite different questions, um, on one side of that, uh, handout, uh, that I’ve put there.

There’s the question of what to do. It’s an unusual sort of question because, um, the answer to it calls for a decision. If you ask yourself what to do, and you answer yourself, then in answering yourself, you’ve decided, uh, something or other.

So you might decide to clean your teeth, and so when you ask yourself, “What shall I do?” Uh, uh, or, “What to do now?” The answer will be, “I’ll clean my teeth,” but that answer expresses an intention that you’ve got to clean your teeth, and it results from a decision that you’ve made.

So it’s an unusual sort of question, a what to do, uh, type of question. And it certainly doesn’t express a normative view that you ought to clean your teeth. You might also ask yourself the question, “What ought I to do?”

And the answer you might give yourself if you do that is quite different. It might be, “Oh, I ought to stay in bed.” Those two things are perfectly consistent with each other, and then you will end up believing you ought to stay in bed, but on the other hand, with the plan of cleaning your teeth.

That’s an example then of akrasia, and I take it for granted that akrasia, uh, exists. And this is another way in which planning attitudes are separated from normative attitudes. It doesn’t mean that Alan is wrong to treat normative beliefs as fundamentally non-cognitive attitudes.

They may still be fundamentally non-cognitive attitudes, but they’re not ordinary planning attitudes, and they’re even further from ordinary planning attitudes than he tells us. It’s not merely that we’ve gotta deal with, deal with Buridan’s ass, but there is this other thing we’ve gotta deal with, namely akrasia, that sometimes, um, uh, what we believe we ought to do is not at all what we, uh, plan to do. Alan doesn’t say they’re planning attitudes, he says to a person’s approximation, and we can still accept, or this is a possibility, we could still accept that they are, what he says, they’re like planning, uh, attitudes.

We’ve just got to recognize that they’re removed from ordinary planning attitudes. They’re not actually planning to do something, although they may be like planning to do something in a way. For instance, they might be some sort of ideal plans.

It’s, it may be some, some higher superior sort of plan that we have, which is distinct from the actual plans, uh, that we make. Now, to be honest, as it happens, I doubt that normative beliefs really could be much like planning attitudes at all, um, because despite what Alan says, uh, I think the logic of normative beliefs is not really much like the logic of planning attitudes. And on the back of that handout, the other side of that handout that I’ve given you, there is, uh, an example of a difference between the logic of planning attitudes and the logic of normative beliefs, which I’ll, uh, you’ve got now, and if it comes up in discussion, then, uh, it, it’s available, but I’ll leave that aside for now.

So that’s, uh, an as it happens remark. I do not think that normative beliefs really could be much like, uh, planning attitudes. But, uh, let me leave that aside, because I want to get on to, uh, what I think is, um, um, another difficulty for Allan’s, um, argument that arises from recognizing, uh, this.

Um, these attitudes which we’re talking about, the non-cognitive attitudes, are the ones that are at the foundation of his, his, uh, system and the foundation of what turns into the cognitive attitudes of, uh, normative, uh, beliefs. They earn the right to do it, but fundamentally, they’re the non-cognitive attitudes that are like planning attitudes. But how do we identify what these non-cognitive attitudes are?

If they were ordinary planning attitudes, we could do that pretty easily because we know what it’s like to intend to clean your teeth. Um, we, we form plans, we work with plans. They’re very– we’re very familiar with, uh, ordinary, uh, plans.

But the ones that Alan is talking about are not ordinary plans. They’re s- if anything, they’re ideal sorts of plans, something of that sort. And we can only really describe what these attitudes are through cognitive, uh, descriptions, uh, of them.

So take this attitude of okaying something, the primitive attitude of okaying something. Now, I told you you weren’t to think of this initially as the attitude of believing the thing was okay, but nevertheless, in order to describe what that okaying attitude is, I’ve got to say, to point to it by saying it’s like the attitude of believing the thing is okay, or indeed that it is the attitude, uh, of, uh, believing it’s okay. I’m pretty sure that you’re only going to be able to recognize this okaying attitude because you understand what it is to think that a thing is okay.

Because you know what the cognitive attitude is, you can go through that to pick out the non-cognitive one, but we can’t point to the non-cognitive one, uh, directly. And what about the attitude of planning ideally to stay in bed when actually what you plan to do is to get up and, uh, clean your, uh, teeth. Your mundane planning attitude is to brush your teeth, but you’ve got an ideal planning attitude of staying in bed.

Well, how do we point to that? Can you recognize this ideal planning attitude? Well, it seems to me that the only way you’re going to be able to pick out what that ideal planning attitude is, is to realize that it’s an attitude of believing that you ought to stay in bed.

Um, so these non-cognitive attitudes, if you want to know what they are, we’re going to have to pick them out for you through the cognitive attitudes that, that they underlie. Now, uh, that doesn’t mean that what Alan says is, is wrong. Um, so far as Alan is concerned, these underlying cognitive and non-cognitive attitudes are indeed the same as the the cognitive attitudes by means of which we, uh, I’m saying that we need to, uh, identify them.

They are indeed beliefs because they earn the right to be beliefs because of their structure. But they’re supposed to be fundamentally non-cognitive, but still, uh, they are the things that we can, um, identify as the beliefs. So that’s– I’m not contradicting Alan in saying what I just, uh, said, but I do think that this raises a serious difficulty over Alan’s project of providing of how these attitudes earn that.

[00:43:43] MICHAEL BRATMAN:
Okay, okay. Uh-huh.

[00:43:48] JOHN BROOME:
Um, those underlying non-cognitive attitudes are supposed to explain how we have the cognitive attitudes, but we can only recognize the underlying non-cognitive attitudes through recognizing the cognitive ones. Uh, or I should say perhaps more accurately, by means of their cognitive attitude aspect. These attitudes are Janus-faced.

On the one hand, they’re non-cognitive, but they earn the right to count as cognitive, and it’s that face through which we have to, uh, recognize them. So, uh, Alan, um, is, uh, in effect saying to us, um, that the belief that you ought to stay, Sorry, um, but yes, the belief that you ought to stay in bed is explained by that non-cognitive attitude, whatever it is, that corresponds to the belief that you ought to stay in bed, and we are only gonna be able to identify that attitude through the thing that it’s trying… that it’s, uh, uh, explaining. And that makes me worry that he hasn’t really got very much of an explanation.

When the explanans is identified through the explanandum, and when this happens so extensively throughout the explanatory story, I doubt that we’ve got much of an explanation at all. I think for a proper explanation of the thing that needs to be explained, namely how we have these cognitive beliefs in the first place, for a proper explanation of that, we need to be able to to– to identify in some independent way at least some of the underlying non-cognitive attitudes that do the, uh, explaining. And here’s a more specific point that goes along with that.

It’s supposed to be the, uh, logical relations among the non-cognitive attitudes that explain how they can be treated as cognitive, how they t– earn the right to have to to be treated as, as attitudes that have contents which are true or false. So it’s supposed to be the relations among the non-cognitive attitudes which explain the way in which what we do with these attitudes allows us to treat them as beliefs having, um, uh, contents that are, uh, true or false. But so those, those relations between the non-cognitive attitudes, they’re supposed to generate a logic that mirrors the logic of truth, and they’re explaining why the, the, the, the contents can be treated as actually being true or false.

But now we identify those, uh, underlying non-cognitive attitudes through cognitive attitudes in the first place, and that means that they just can’t help having the structure that’s characteristic of truth because we’re picking them out as the cognitive attitudes to begin with. So they’ve inevitably got the structure that, um, Alan is supposed to be explaining, not because of anything to do with their internal, uh, coherence or anything like that, but because we identify them through attitudes that inevitably have that structure in the first place. So we’re not succeeding in explaining why they, uh, have that, uh, structure.

So that’s a difficulty that, that I, um, uh, think is underlying Alan’s way of going about things. To summarize, he hopes to explain the truth and falsity of normative sentences and the fact that we can have normative beliefs by a structure that appears within non-cognitive attitudes that underlie those things. But I don’t think, but those non-cognitive attitudes are not truly planning attitudes.

They’re like planning attitudes perhaps, but they’re not truly ones, and I think k-planning attitudes, and we can only identify them through their corresponding cognitive attitudes, and that makes it rather doubtful that they can play the explanatory role that Alan wants them to play.

(applause)

You want me to go up there? All right. Or can I sit here?

[00:48:07] MODERATOR:
Uh, yes. I– you can sit there. Um, I think if you move the, um, microphone over maybe.

[00:48:11] FRANCES KAMM:
Yeah. I’ll just do that. Yeah. One second.

[00:48:16] MODERATOR:
Right. We’ll have, uh, one more set of, uh-

[00:48:19] FRANCES KAMM:
Get closer to my notes

[00:48:20] MODERATOR:
comments before, before taking a break, um, and they will be from, uh, Frances Kamm.

[00:48:25] FRANCES KAMM:
Yeah, you can hand them out, I guess. So, um, the other speakers have, uh, done a great job, uh,

(paper rustle)

commentators, sorry, have done a great job in, uh, sort of helping Alan, and I’m only gonna

(paper rustle)

ask Alan to help me. Uh, I just have a bunch of questions, really, uh, sort of, uh, I’m hoping can be put on the right track in some ways. Is this… something like that? Can I have the, the watch?

[00:48:49] JOHN BROOME:
Okay. Is there a handout?

[00:48:50] FRANCES KAMM:
I have some questions. Yeah, there’s just a little brief handout. Often it’s not gonna do anybody any good. But, um, I, um, I have some questions related both to the first lecture and to the second lecture.

(paper rustle)

Um, I’ll, I’ll do the first lecture

(paper rustle)

first, uh, just so for continuity. Um, one question I had was, u-unless I’m misunderstanding, um, the, the planning theory here, um, or something like planning, um, to account for something like, well, I– my belief that I ought to do something. I plan to have a, a preference for a certain sort of, um, act, uh, above all others, and for having certain feelings and for doing certain things.

Um, that this involves planning seems to be a, an empirical claim or subject to empirical confirmation or disconfirmation. I mean, John has been discussing, uh, it from the point of view of the, the logical concept of it, okay? Whether, um, all ought statements involve something like planning statements or ideal planning statements.

But I was just wondering whether, um, you think of your theory as something that is subject to, uh, at least empirical disconfirmation. So it seems like you’re saying that creatures who can’t plan can’t have moral judgments or make moral judgments. I mean, they can be subject to morality, but they, they wouldn’t be capable of making moral judgments.

And so I was just wondering if, uh, you know, uh, I don’t know very much about brain science, but, um, um, uh, I hear that there are in individuals who, through some sort of accident or defect of the brain or whatever, are incapable of planning. I mean, in the ordinary sense, you know, planning their lives, planning their actions, projecting their thoughts into the future about what they’ll do and whatever. And I was just wondering whether you intend your theory to be of the sort that, um, would be disconfirmed if it were the case that people who can’t plan in this other ordinary sense, right, um, can nevertheless engage in judgments about what Caesar ought to do or what they ought to do if they confront, you know, a drowning child or something of that sort.

Um, and, uh, if it were the case that they could, uh, would that show that there was, you know, some separate idea of planning other than the ordinary, uh, notion of planning your life, which they obviously lack the capacity to do, that is undergirding their moral judgments? Or is it, you know, the case that you would say, “Well, okay, planning’s not involved here”? Um, the other thing is, of course, that there are some moral judgments that these people, perhaps they, they couldn’t in fact engage in the thought, well, you know, if I saw the drowning child, I ought to do such and such, um, because they lacked the ability to plan and, you know, and this was connected, uh, to moral judgment.

Uh, when they’re in a situation, they could, could still have a moral judgment when they’re responding to, you know, the drowning child, and then in that case decide, you know, well, this is what I ought to do. And so I’m wondering whether, um, if an individual who in general lacked these, you know, planning capacities, um, could do that, Um, whether that would indicate that some sort of moral judgments, right, um, are, are making these judgments, uh, does– don’t involve, uh, don’t involve, uh, planning capacity. Uh, even if the long-term ones, uh, you know, or hypothetical case imagining judgments, uh, do.

And, um, I was just wondering whether this is really a theory that, uh, is subject, on your view, to this sort of a confirmation or disconfirmation. It’s just, it’s really a question. Now, the other thing that, the question that I had about the um, uh, the first paper, your first lecture, came as a res-as a, uh, further, uh, reflection on your response to a question that I asked at the first session, and I’m not sure that I understood your answer.

I, I raised the issue of supererogation as, um, you know, I wanted to know how you dealt with it on your view, Because when you say that, um, uh, y-you think that you ought to, you know, rescue the child, it means you plan to prefer something like planning to prefer it to alternatives. Let’s forget about the case where there are ties, and to feel guilt, you know, if you don’t, uh, and, uh, you plan to feel s-guilt if you don’t actually carry out the action. And I just, you know, raised the question of supererogation, uh, cases where, um, what you ought to do, you know, if you see the drowning child, you decide is, uh, you know, call the police.

Uh, but, um, you recognize that it might be morally more commendable and, uh, a great thing if you were to jump in the water and rescue the child yourself when there’s real– even though there’s real danger to you. And, um, the way, the way I keep on thinking of it, though it’s not the way you recommended I think of it. I think, uh, so, so far as I understood your answer, you were talking about, uh, an attitude of approbation of some sort of, the supererogatory action.

But I was just wondering, I mean, I have this little handout. I don’t have the neater version, which you have. Unbelievably, somebody had to actually write this neat version out for me, because the threat was that if I started writing it, um…

So this is, uh, thanks to Ellen Gobler that I have this. Uh, so the way I keep on thinking of this is, and you’ll correct me how, how I go wrong, is, you know, I think of sort of a scale of degrees of preference for, uh, uh, and, uh, I’ve got at the top, I’m imagining someone who really would prefer to be the person who would jump into the water. They would very much, you know, prefer to be that sort of person, and they would regret it if they found that they didn’t in fact jump into the water.

So I’ve got, you know, on, on my view of this, I’ve got the strongest preferences for the supererogatory act. Um, and, uh, regret if one doesn’t engage in it, you know, through weakness of will or, or whatever. Um, and then, of course, o-o-on this little chart, uh, I’ve got under that, you know, not as the thing for which I have the strongest preference or have a plan to prefer to the strongest degree or whatever, uh, the action of which I think, you know, someone might say, “Well, you ought to do, you ought to call the police.”

Okay, that’s, uh, something that I really ought to do. And, um, of course, if I don’t do that, I, I can recognize if I did something Z, like sit around on the, you know, the beach and enjoy the sight of the child drown, um, that I would feel, you know, I would plan to, uh, ex-ex… have the feeling of guilt or have something like a plan to have guilt. Um, and it seems to me, I mean, the way I keep on thinking of this is that, um, when I’m imagining this person who really does wanna engage in the supererogatory act, but the person would have regret if they didn’t, but they wouldn’t have guilt, or they wouldn’t plan, right, on having guilt.

They’d plan on having regret. Um, and then, uh, of course, they can do something that’s still okay, of course, namely call the police, and it’s really what people would say, “That’s what you, you know, uh, you’re…” In the ordinary sense, if you thought, What is, what is it that I ought to do in this circumstance, okay, is call the police.

Um, and one of the things, of course, that if they did the wrong thing, okay, they would, they would plan to feel the guilt. Now, as I see this then, to say that, you know, I ought to, you know, call the police, um, is not a plan to prefer it to, um, alternatives, alternative actions, because I do prefer the supererogatory act. And it’s, it’s not a plan to feel guilt if I don’t do what I ought to do because of course, if I do the supererogatory act, I’m not gonna feel–

I don’t plan to feel guilty at all, right? I mean, I avoid guilt completely. I mean, to me, it’s sort of interesting that if I, if I avoid, um, uh, uh, guilt by doing the supererogatory act, I mean, I can’t avoid guilt that way.

But if I fail to do the supererogatory act and the person I’m imagining, um, there’s no way I can avoid regret, even if I wind up doing what I ought to do, and I don’t feel guilty. So that the regret is somehow hanging there, right? As some sort of more inescapable fact of this person’s life than the guilt.

Um, now, of course, in– when you have these tie situations like, uh, you know, there are multiple things that are equally good, the okay things as you see them. My view is that, um, on this chart, I guess it shows that okay is, is, is not just, you know, I prefer it to, I prefer any of these possible things and alt- amongst the things that are okay to any other alternatives, um, because, um, I do prefer, uh, the supererogatory action to the just doing what I, you know, standardly think I ought to do and and believe that I, you know, it’s what I ought to do. Um, uh, So, um, I– I-I’m, I’m not saying that something is okay because there’s no alternative, right, that I prefer to it.

I mean, I do think it’s okay if I just do my duty. But it’s not because there isn’t some alternative that I would prefer. Um, so one of the things

I, I was thinking about on, on my model here is that to say that I think that something is okay, given that I think that either supererogation or doing what I ought to do was okay, Um, is that when I think, when I, you know, think of something as being okay, I’ve had that attitude towards it, um, so I plan to prefer, right? Um, what I’m thinking of is I’m planning to prefer to do what I never planned to have guilt if I don’t do. That’s what it– um, this chart, I’m just reading off from the chart, right?

Um, and that, um, furthermore, I mean the way I think of this is that it’s not merely that I have these plans to prefer, you know, to do some things And I plan to feel guilt if I don’t do this other thing, as if it were some sort of a, you know, if I failed in my best plan to do what’s okay or to do what’s supererogatory, not merely one of the other things that’s okay. But, um, it’s not like that planning to feel the guilt is sort of a, a backup, you know, a second-best plan because, um, there’s a demand on me that I not go below that line, right?

I’ve got this line here where, uh, I’ve got a negative, you know, don’t go into the guilt area. Uh, you know, plan not to, uh, to do what I would then plan to be guilty about doing. Uh, there’s some sort of a demand, uh, not just sort of a preference, that I stay in that okay area.

And it’s not just because I would fail to satisfy the thing that I prefer most or that I would plan to prefer the most, um, because when I plan to prefer the supererogatory action, I don’t necessarily think that I, it’s a demand on me that I do it. So it’s not just failing to do what I think is my, the plan to prefer, what I should plan to prefer the most that leads to this sense of, of demand, okay. So anyway, this is, uh, I couldn’t stop thinking about this.

Um, it sort of became obsessive, and I just was wondering what you thought about this model, and can you help me? Uh, so those are my comments about the, the thoughts that occurred to me with respect to the first, um, section. I guess I still have some time and can go on to the second.

Um, I was struck by the fact that you said at one point, uh, that you thought that the morality of respect, um, and people’s attempt to build up an i– a morality on the basis of respect, you say this in the second lecture, has been, in your view, a failure, right? Um, uh, and yet, uh, you in your view about, um, encapsulating or somehow deriving, uh, out of Scanlon or out of the idea that, uh, one ought to, uh, have a preference, right? Uh, one ought to plan to have a preference to engage in relationships of, uh, res– fair reciprocity with people on the basis of agreements that, uh, no one could reject on their own behalf.

Um, you think that that is what it amounts to, to respect persons. You, you, you adopt the language and the, the format and the framework of, of respect. And so I was, I was wondering why you felt it necessary to speak in those terms.

I mean, uh, and not just, uh, directly, you know, speak in favor of the, uh, the utilitarian solution. Not saying in, sort of, sort of putting it in the context of, well, it’s what… If you live according to those terms, the utilitarian terms, you will be engaging in respect for persons.

Um, you know, Pat Scanlon, uh, doesn’t just think he has a preference, right, for enga– for, uh, doing what he thinks is required by respect for persons. He thinks is a good reason to have it. And you may, of course, uh, may think that too, though I think that in your lectures, you simply just say that, “Well, you know, this is something that appeals to me.”

Um, but I, I was struck by the fact that, uh, you yourself have now sort of, um, find it attractive or maybe necessary, there’s some sort of demand, that we, uh, fit, right, uh, other theories into the respect for persons umbrella. Okay. And so th-that’s one question related to respect.

My other questions related to respect, if I still have time, I don’t know how much time I have. I sort of, uh, bring up the, uh, tail end of what I said last time, um, because I had been discussing cases, uh, the parent cases. I don’t know if those people who are here today You know what I’m referring to.

If a parent, uh, um, has a child that’s, uh, in an, uh, a, a canoe that’s in danger, and, uh, he has to to decide whether to save his own child or the children of two other, two other children of a parent, uh, who can’t reach his own children. And, um, the, uh, uh, the solution, uh, seemed to be, uh, Alan’s thought that the, uh, that the solution was to do what, quite independent of knowing who you would have been and which of the parents, what you would have agreed to, uh, an arrangement you would agree to behind a veil of ignorance, uh, assuming that you had an equal probability of being either in the position of the parent A who could help or a parent who couldn’t help. Um, now, John raised questions about this equal probability argument yesterday, but what I’m interested in today is another case of yours where the same sort of issues seem to come up, namely the case of Ida and Jay.

And, um, uh, Ida, uh, the, the assignments of, uh, utilities, I think, were that Ida would, uh, uh, and, in one state of the world, uh, have a nine, right? And Jay, who would be in that world too, would have a one. Um, and this would be a state in which Ida would get this very expensive funeral.

Um, and, um, another state of the world might be in one in which, uh, Ida only has, uh, you know, what’s good for her to the degree five, right? And, uh, Jay has it to the degree three. So Jay is better off and, uh, you know, Ida is not so great.

Um, and, um, y-you think that, uh, you know, what each individual… You, you raise problems for differential views about the good of a person, uh, that Jay may think that Ida is mistaken in her, uh… Even if he wa-he turned out to be Ida, given this equal chance of being himself in that world or someone with Ida’s preferences, he would think that it didn’t really matter whether the preference for having a funeral got satisfied.

Um, and you focused on that, the difference in different people’s views about what is, uh, really the good of, uh, of, of Ida or Jay. But I wanna just go back to the cases where there isn’t any disagreement about the good and, uh, you know, you do think you agree with Ida that if you want to have a big funeral, it’s really good for you to have that good funeral. Um, And, uh, just say that again.

You know, those of us who think that the maximin, uh, solution, namely bringing up the position of the person who’ll be worse off in an outcome in the world to the highest degree, even if it means lowering, uh, the outcome for the other person, you know, the solution of five and three, um, is preferable, right? Um, How do they think? Why are they getting this result?

And I wanted to know, um, my understanding has always been that there’s a different view about how to, uh, conceive of the point of the veil of ignorance, right? That, um, those who get the maximin solution think that, that, uh, the way to think of this device is not as being there in order to, um, uh, allow a person in ignorance, right? Assigning equal probability of being in every possible position, uh, outcome position, to think how they would maximize their expected utility.

But, uh, so where they think of each of these, you know, positions of life as something that they have a po-pi-possibly fall into, right? But rather, the contrasting view that’s supposed to get you the maximin is that the veil of ignorance forces you to identify, um, or take seriously the lives of the actual people who will fall into those different slots, the different separate persons, right? Uh, they’re not just possible slots that you might occupy, but there are positions that people, real people will occupy.

And to get you to take that seriously, I think that’s the way Scanlon, uh, you know, describes it. Um, at least when he’s– he doesn’t make use of the veil of ignorance when he’s describing Rawls’s use of the veil of ignorance. And it seems to me that, uh, in terms of respect, when people think of, uh, showing respect for persons as ends or whatever, that it’s that focus on, you know, what it will actually be like for each particular person in those positions that’s driving, um, the desire to use the veil of ignorance, right?

Uh, and of course, this is a completely different way in which you’re using it. And, um, uh, it connects up, of course, with this idea of what it is, uh, how we flesh out the idea of respect for persons. So just to sort of make this clearer, I mean, um, suppose Ida, you know, would have a, an i-

A nine, right? Suppose, uh, because she wants a slave. Uh, that’s what would give real meaning to her life.

And it’s in, and it’s in her interest to have all, uh, you know, these things done for her. And Jay would be the slave. I mean, that’s why she’s got a nine.

And he’s got a one in one world. In the other world, she doesn’t get Jay to be s- her slave. She gets a five, and Jay’s better off, he gets a three.

Um, and, you know, I, I suppose that, uh, you know, the person who takes the, the Rawlsian view will say, “Well, look, that desire, right, is intrinsically disrespectful of persons,

[01:07:41] JOHN BROOME:
right?

[01:07:41] FRANCES KAMM:
To have that desire to get your, uh– to want to have a slave, for example.” Um, and the view, of course, that you take about how to use this, um, the veil of ignorance, um, says nothing about that,

[01:07:54] JOHN BROOME:
right?

[01:07:54] FRANCES KAMM:
Uh, there is this desire,

[01:07:56] JOHN BROOME:
okay,

[01:07:56] FRANCES KAMM:
and of course, Jay might wind up being Ida and agree with her that having– when you want a slave, having a slave is a great thing, and you wind up with a nine,

[01:08:03] JOHN BROOME:
right?

[01:08:04] FRANCES KAMM:
And increase his expected utility by, uh, you know, um, going for an arrangement in which he might, of course, turn out to be Ida, turn out to be a Jay, and ha- and, in your view, have no complaint if he turns out to be Jay the slave. Um, and I, I also think that, you know, the inability to evaluate these sort of desires, uh, you yourself said desires might change in different contexts, but not to be able to say that there is something intrinsically disrespectful about this because after all, respect on your view only amounts to, you know, playing your part in a fair reciprocity in that, uh, utilitarian, uh, outcome. A-a-again, uh, the idea of… Am I over time?

Uh, a little bit. Yes. Okay.

Uh, okay. Treating people as, uh, as means, I mean, Sometimes I, I, um, uh, on your view, Jay would not be treated as a means because he’s been going on, uh, as a slave in the way that, uh, is demanded by the system that he would’ve, uh, uh, had, uh, you know, the highest expected utility under. And re-recall, I mean, there’s these debates I have with Derek Parfit.

Derek Parfit has the view that you’re not treating someone as a means when you would in any way limit your, um, uh, use of them out of concern for their own good. And I’ve always thought that that was too weak, okay? But this view here on, uh, what it is to fail on, you know, to satisfy the crime of treating people not as mere means,

(gasp)

of course, is even weaker. I mean, Ida need have absolutely no concern, right, for the welfare of Jay to any degree at all, modi– you know, potentially modify her conduct in light of him. He’s just a slave, you know, he’s a piece of property.

(gasp)

And, um, but he’s– she’s not treating him as a mere means on this view because it’s just in accord with the system, you know, uh, that he would’ve had the highest, uh, expected utility under. And I just want to point out that, you know, it seems that all these differences are arising from a different view about what the veil of ignorance is supposed to be there for, you know, to make you do, right? Um, and I guess I’ll end at that point.

I had wanted to discuss cases where that problem arises even when it’s not a question of deciding between those who are going to be worse off, but differences in expected outcome of people who would be equally badly off if they were not helped. But I’m over time, so I’ll stop. Thank you.

I hope… Yeah, I hope I get a lot of help.

[01:10:19] MODERATOR:
Thank you.

(applause)

[01:10:25] ALAN GIBBARD:
I think we could break briefly.

[01:10:27] MODERATOR:
Yes. Uh, so we’ll take a brief break, about five minutes, and, uh, Alan Gibbard will then, um, respond, and we’ll have a general discussion that, again, will be followed by some refreshments behind the, uh, behind the wall, for today’s comments, and, uh, and then we’ll open up into a general discussion. Um, Alan?

[01:10:50] ALAN GIBBARD:
Well, it’s such a privilege to have, uh, uh, things that I’m interested in, uh, raised by, uh, such a terrific, uh, array of commentators that, uh, I wish I had the ability to do justice to very much of this. Uh, I wanted to go backwards just because it’s, uh, easier to get started, uh, remembering, uh, what’s freshest. Uh, so I’d like to talk about, uh, uh, start out with Francis’s, uh, uh, comments, questions about lecture two.

Um, well, is respect-based morality a failure? Well, I think what I was claiming was that the attempt to get a, a non-utilitarian morality, uh, from a general rationale of respect, uh, is a failure. Now, of course, that’s a very broad claim, and, uh, and we, uh, I’m, uh, as she says, trying to get a, um, trying to get an account of morality, uh, starting with, uh, a way of understanding respect, but I’m not trying to get a, uh, find respect as the reason that, uh, utilitarianism is wrong and something incompatible with utilitarianism is right.

Um, so, uh, one thing I had in mind was, uh, attempts to get, uh, uh, Kantian sorts of, uh, uh, tests, uh, the test of the first version of the categorical imperative to, uh, to work properly or attempt to get the second version of the categorical imperative to, uh, have definite non-utilitarian content. Uh, specifically on Rawls, uh, well, uh, Rawls rules out equal probability reasoning, but we have to ask, well, what’s the rationale within the general theory of doing that? Uh, well, he says, well, uh, he’s ruling out, uh, knowledge of what kind of society, uh, one will be in, and we can’t make any sense of the idea of what kind of, uh, society, what the chances are of being in a particular kind of society.

But that would still allow the, uh, conditional judgment that, uh, if I’m in a particular kind of society with a population of a million, say, then I have a one in a million chance of being anybody. Uh, is that not taking seriously the positions that people will actually be in? Well, Rawls was trying to do it with, well, Well, he always distinguished, uh, uh, the parties be-behind the veil of ignorance, uh, people in a well-ordered society and, uh, us.

Uh, the parties behind the veil of ignorance were supposed to be mutually disinterested. They didn’t respect, uh, uh, but the thought experiment about them was supposed to, uh, tell us something about respect. Uh, and my, my view is we– well, we have to scrutinize the, the whole way in which things are set up to try to, uh, try to give an account of what respect requires.

Now, uh, I, I should say if the, uh, in a situation where, uh, maximin choice is rational, I doubt that there are such situations, but in, in such a situation, we have to generalize the, uh, the idea of utility, uh, because, uh, that’s, uh, going to satisfy some of the ax– most of the axioms of decision theory, but it will violate some sort of, uh, Archimedean condition. So if we, uh, I think if we generalize, uh, the idea of, uh, of utility, we still get something that’s, uh, utilitarian in a way, but has this strange kind of, uh, utility where, uh, you would never cross the street for a chocolate bar because, uh, uh, because getting hit by a car is so bad. On the, uh, is my talk about, uh, my attempt to explain normative judgments in terms of planning and empirical claim?

Well, I hadn’t been thinking of it that way, and let– instead of trying to, uh, think whether that would be a good test, let me say a bit about how I am thinking of it. Uh, we can think of, uh, the, the logic of decision in a way that, uh, decision theorists often do. Uh, there’s a situation where various different And alternatives are, are open.

Uh, I classify some of those alternatives as, uh, as okay, some of them as, uh, as not okay. That may lead to a further, uh, situation in which I’ll— that is, any choice can lead to a further situation where, uh, again, there are a number of alternatives, uh, and, like, I allow some, reject others. I said, well, that has to be the valence part of— uh, now, my talk of, uh, planning maybe was, uh, uh, a bad term to use because we could think of, uh, picking a strategy, uh, uh, just picking, uh, uh, in the sense of what we might call a plan that, uh, just gets a contingency plan.

It picks a particular alternative in each situation one might be in, and that’s not what I had in mind. So I’ll use the term strategy for something that just picks out one alternative in each decision situation, and then I’m using plan as a sort of valence classification of the, uh, of the, uh, alternatives in each decision situation. Now my claim was that, uh, that, that, uh, kind– that planning in that sense, well, first, uh, we, uh, can ask what planners in that sense are committed to, and then my claim has been that, uh, planners in that sense are committed to something that turns out to, uh, look just like, uh, like normative thought.

Now, uh, John has some, uh, some arguments that it doesn’t work that way, uh, which I’ll address, uh, shortly. But, uh, let me, uh, let me say a bit about supererogation again. Uh, so there are two kinds of oughts in play.

There’s, uh, the moral ought, which is attached to, Uh, to guilt and, uh, resentment, And then there’s the, what we might call the, uh, the, the rational ought, uh. Uh, so thinking about the, the thing to do is supposed to, uh, supposed to explain, uh, thinking about the, the rational ought, uh, in my first book, I talked about what it makes, uh, what it makes sense to do. Okay.

So, uh, this is somebody who, uh, uh, doesn’t… I, I think the situation that Francis describes isn’t one where, uh, somebody thinks that, uh, morally, uh, he ought to, uh, ought to call the police, because to think that would be to think that it’s morally wrong to jump in and rescue, uh, the person. This, this is somebody who thinks, uh, morally it’s, uh, permissible to, uh, call the police, uh, permissi– but permissible also, and in fact, uh, morally admirable to, uh, jump in and save the person.

And, uh, this is somebody who, uh, who because of the, uh, moral admirableness of it or because, uh, more– thinking more or less narcissis-cistically, uh, uh, the effectiveness in saving the, the child, uh, somebody who, uh, thinks that, uh, what it really makes sense to do in the situation is jump in and save the child but is too, uh, uh, well, is overcome by, uh, akrasia in this case. So, uh, and then we’ll have to talk about that, uh, later. Uh…

Okay, uh, now, uh, uh, So is, uh, in fact, let’s talk about akrasia next. Uh, so I, I wish I could, uh, think myself into the, uh, the example. All, all my, uh, experiences that are candidates for akrasia are, uh, matters of excessive sloth, not matters of excessive energy.

And this is, uh, somebody who can’t, uh, can’t resist getting up and brushing teeth while, uh, uh, well, thinking that he ought to stay in bed, uh, but, uh

[01:20:41] JOHN BROOME:
Well, when I, when I think about situations like that, I’m, uh, I guess I’m not very, uh, clear about how to describe them. After all, it has to be the thought that, uh, at this very instant, uh, I ought to such and such. Now, uh, T. M. Scanlon, for example, uh, has a vivid example of, uh, uh, thinking that I’ve got to call the doctor.

I keep walking to the telephone and then walking away from it, uh, because I dread the news I’m going to get. Well, it seems to me that what, uh, uh, what typically happens is I get to the telephone, I think, uh, \”But, uh, shouldn’t I get a drink first?\” Uh, then I walk back to the telephone, and then I have some other such thought.

I’m the sort of process that, uh, uh, Shakespeare, uh, illustrated in Hamlet, Um, but, uh, I, I do want to allow that, uh, uh, I mean, I’m ambivalent enough about this that I’d certainly like to allow what lots of other people insist on, namely that, uh, it’s not only that I don’t, uh, hold steady my view of what I ought to do, uh, when the moment comes from, for, uh, actually executing it, uh, but that, uh, uh, but that even when I hold steady my view of what I ought to do, uh, uh, I may not do it. Uh, okay, well, if we say, well, then there’s planning is one thing and uh, and ought beliefs are another thing, uh, there’s something peculiar about that, namely, well, we have the kinds of concepts that I’ve described. Uh, now, uh, whether the, uh, whether my description is coherent or not, of course, is, uh, is debated.

But, uh, and if the– whether the logic works as– I’ll talk shortly about whether the logic works as I claim that it does. But okay, so we’ve got these, these, planning concepts in this, in this, in this, uh, s-somewhat, uh, special sense of planning. Okay.

Uh, these are concepts that, uh, that any, uh, anybody who thinks what to do, anyone who, uh, faces what amounts to a decision, the decision tree of life, uh, is committed to, although they may not, uh, uh, they may not explicitly, uh, use these concepts. And then we’ve got this other set of, uh, normative concepts, and if you’re perfectly rational, then the two just coincide. So why do we need these, uh, these normative concepts, these, uh, uh, ought and, uh, and, uh, and, uh, rationally permissible or something, uh, concepts of, Well, just for the sake of being irrational, So they can come apart in the case of being irrational.

Well, I’m somewhat drawn to think that, uh, irrationality is a matter of, uh, some sort of, uh, incoherence of one’s, uh, of one’s thoughts that, uh, I want to… I’m drawn to explaining it, uh, either, as I said before, as, uh, as lacking steadiness, uh, in, in one’s, uh, convictions or plans. Uh, but, uh, also as, uh, but if I, uh, try to, to explain it not that way, then I want to, uh, then it seemed to me that the, these are going to be situations where in a way I plan and to do something, and in a way I, uh, don’t plan to do it.

So in my, uh, uh, in my first book, I thought of, uh, Cornelius and Babar, uh, standing on the, uh, standing on his balcony. The firemen have the, uh, net held out, uh, and he thinks, uh, “Well, what, what to do?” He can’t figure out what to do, whether, whether to jump, and then he, uh, decides to jump and stands there, uh.

Uh, well, it seems to me that’s a case of, I mean, he is planning right now to jump, except he finds his legs won’t move or something like that. Well, we all– So it seems to be that those are, uh, those are cases where, uh, in a way, in a way he plans to jump right now, but, uh, but then part of him certainly plan, uh, uh, the, the, uh, part of him that responds to terror, uh, uh, doesn’t go through with the plan.

Uh… Okay.

[01:26:08] ALAN GIBBARD:
So is the, uh, is the logic actually different in the way that, uh, John’s handout indicates? Well, I don’t quite, uh, uh… See, I don’t seem to have the… I, I guess I left the handout over here. But, uh, okay. Um,

[01:26:33] JOHN BROOME:
No, uh, um,

[01:26:37] ALAN GIBBARD:
I guess I need to know more about the, uh, uh, the way he’s conceiving of believing that, that you ought if you F to G. But, uh, if we put it in terms of planning, there, uh, there seem to be three quite distinct conditional states here. Uh, one is that, uh, if in fact, uh, P then do Q.

So I suppose, in fact, I’m going to drink the second glass of wine, uh, uh, what’s my, uh, what’s my plan for that circumstance where, uh, the second glass of wine will make me, uh, will give me a terrible hangover the next day or something like that? Well, my plan for the situation where, in fact, I’m going to take another glass of wine, uh, is not to take the other glass of wine. That’s another way, of course, in which, uh, planning as I’m speaking of it, uh, differs from planning in the ordinary sense, because planning in the ordinary sense involves a belief that one’s going to carry it out and, uh,

(coughs)

and, of course, planning, I talk about planning for the case of being Caesar at the Rubicon, and, uh, certainly I don’t think that my planning will have any effect on, uh, what Caesar does at the Rubicon. Uh, now, for that sense, uh, of the,

[01:28:08] JOHN BROOME:
uh, inference at the top is valid, I think. I’ve, uh… But, uh, there are a couple of other senses, uh, couple of other conditional constructions.

One is, uh, if unavoidably P, then Q. So– And that’s a matter of, uh, uh, if, uh, my planning takes the form of maximizing, then that would be maximizing under the constraint that P. Because the truth of P doesn’t mean, uh, that P is true, doesn’t mean that I’m constrained by P, right? And it’s true that I’m not going to lie down on the floor, uh, in the next few minutes, but I’m not under the constraint of being unable to lie down on the floor.

Okay. So if, uh, unavoidably P, then Q, and, and for that, uh, then do Q. Uh, so, uh, I mean, suppose, uh, I’m not even capable of jumping into the fireman’s net.

Uh, if, uh, unavoidably I don’t jump, then what to do? So that’s another… a second kind of conditional. And for that kind of conditional, this, uh, pattern at the stop, top would not be valid as far as I can see.

And then a third very interesting one that, uh, that, uh, Hare, uh, discusses in a, in a wonderful paper called, uh, Wanting: Some Pitfalls. I think he makes some mistakes there, but has a, a marvelous set of devices for thinking about these things. Uh, something ungrammatical, but, uh, if imperative do P, then imperative do Q.

And we c– I can think of, uh, means-end reasoning, uh, in that way. So if leave the room right now and go back to where I’m staying, then follow the path up the stream, understanding both of those as imperatives. So to evaluate that, I assume a goal or assume, uh, some sort of policy and then, uh, and then, uh, in a, in a kind of Ramsey test, uh, uh, Taking on the policy, I, uh, sort of offline as it were, I, uh, decide how to carry it out.

And that kind of, uh, conditional also, uh, doesn’t satisfy the, uh, doesn’t make the pattern at the top ballot. So I think there are lots of resources. I’m, I’m not quite sure which, uh, reading of, uh, of the conditional ought and the second is intended, but, uh, there are some… there are two alternatives for, uh, how to realize it in planning language.

Um… Okay. Uh, so, uh, I’m, I’m, I’m never quite sure what cognitive means

but it can mean, uh, explained in terms of truth, and of course, I’m trying a different kind of explanation. My– The explanation is that an agent plans, and, uh, Michael has some, uh, uh, intriguing thoughts about agency, which I won’t be able to get very far with.

But, uh, let, uh, let’s start with, uh, coherence and, uh, why c-coherence matters. Uh, so we both agree there’s, uh, one question, uh, whether, uh, a way of thinking is undesirable, and another question whether it has logical defects. So, uh, so there’s nothing particular– particularly undesirable about uh, uh, logical defects in my, uh, in my thinking about some-something that it doesn’t matter having correct beliefs about anyway, uh, but we’re interested in logical defects.

Now, uh, I think I made a blunder in the, uh, in my discussion of why coherence matters. Uh, I s- went pretty carelessly from the, uh, term coherent to the term, uh, consistent, and I defined coherent as satisfying the standard, uh, axioms of decision theory. Uh, my example of, uh, incoherence was something that’s really more like inconsistency.

It would violate the, uh, formal constraints of decision theory, but, uh, in a special way, uh, in that, uh, it’s a contingency plan that, uh, that, uh, can’t be carried out in the contingencies to which it applies. Okay, so is, uh, is what’s wrong with incoherence a matter of, uh, not being interpretable? Well, uh, not for consistency, I think, like, uh, for consistency, the problem really is that, that, uh, there’s, uh, uh, no, uh, no specific contingency plan that fits, uh, an incoherence, uh, or an inconsistent set of, uh, beliefs and plans.

Uh, is that undesirable? Well, no, inconsistency isn’t always, uh, undesirable. Often it doesn’t matter.

Uh, but then there’s the rest. Well, what’s the rest? Uh, in the Zeckhauser, uh, uh, Russian roulette example, uh, people like Ned McClennen have given, uh, consistent policies, uh, for, uh, violating the, uh, uh, standard constraints of decision theory.

Uh, and so the s- uh, the standard constraints of decision theory that are violated aren’t matters of consistency. And there, I think, uh, interpretability has something to do with it, but I have a different picture of what interpretability has to do with it. Um, if I’m, uh, If, say, I respond differently if the, uh, problem is what to do in a series of actions, uh, with certain end probabilities and what to do, uh, with a single action that has those same end probabilities, well, it’s, uh, consistent to respond to things obviously, uh, not even the end probabilities.

Uh, and then, uh, it’s, uh, we might say, in… incoherent unless the process matters. Uh, so we have to, uh, now we have to interpret my plan to see what I’m treating as mattering and what’s, uh, what’s incoherent in a way. Well, you can represent the policy, uh, as long as you represent the, the process as mattering.

Uh, the constraints of decision theory, uh, don’t make any difference. Uh, you can always, uh, fudge things to fit the constraints by saying, “Well, uh, what he, uh, really cares about is how the game goes, not whether he ends up alive or dead, or rich or poor.” But then it’s just fantastic that those are the things to care about in, in that situation.

So we have to interpret in order to, uh, uh, assess the, uh, we have to interpret the, the plan, uh, Uh, as having a rationale in order to see whether the, uh, in order to assess the rationale, in order to see whether the rationale is, uh, is credible. Uh, so that’s the role I picture interpretation as having. Not that I reject what Michael says.

I find it intriguing, but I, I don’t, uh, I don’t know how to think it through in those terms. And, uh, and so I’m, uh, eager to learn more. Um, Okay, so uh, preferences aren’t plans because they have different constraints.

Well, I’m trying to explain, uh, everything in terms of what I’m calling plans. And in the case of, uh, in the case of, uh, desires or preferences, Well, first I’d take it that talk about desires is really talk about preferences. Uh, I desire something in that I, uh, prefer it, uh, to something else that’s implicit, uh, maybe to– maybe just prefer it to the world never having existed.

But, uh, uh, but, uh, And economists have, uh, have systematic ways of talking about preferences. What do preferences do? Well, as what’s feasible varies, uh, your, uh, choices, uh, vary in a systematic way given that you have certain preferences.

So we, uh, so we read off the preferences by the way the choices, Well, actually the way the, uh, uh, uh, the valence choices, uh, vary as the feasible set varies. Well, that’s, uh, uh, that’s a feature of a contingency plan that covers quite a different, uh, quite a lot of different, uh, uh, feasible sets of outcomes. So in the case of, uh, wanting both to be scholarly and, uh, fabulously rich, uh, I, uh, if that’s my desire, uh, that is, I, presumably, I prefer being scholarly and fabulously rich to being, uh, fabulously rich and non-scholarly, to being scholarly and, uh, and not fabulously rich, to being neither.

Uh, okay, well, if, uh, uh, if those are my preferences, then, uh, my contingency plan for a situation in which it’s feasible to be, uh, both scholarly and rich is to choose it. Now, of course, actually, we’d have to say everything’s equal and, uh, and I think what’s actually going on here is that we think of preferences as somehow resulting from what we might call preference tendencies. So I have a preference tendency to be rich.

I, uh, I have a preference tendency to be scholarly. Uh, of course, the two might interact so that, uh, but, uh, what– But one way in which, uh, those preference tendencies might act is just by, uh, adding up straightforwardly.

So, uh, so I think that we can read preferences off of a contingency plan, and then preferences, and uh, uh,

[01:40:11] ALAN GIBBARD:
do have, uh, constraints. I, I would take it that they have the constraints that, uh, economists work out for, uh, utilities, even though, as John was saying yesterday, economists, uh, standardly define utilities in terms of preferences, and I was trying to use the same formal apparatus to, uh, talk about, uh, uh, prospective goodness for a person or prospective goodness, uh, all told. Uh, uh, do plans aim at coordination?

Well, uh, again, I’m, I’m intrigued, but, uh, I’m looking at plans as aiming at, uh, uh, being, uh, well, plans aiming at being carried out. But, uh, of course, I’m talking about contingency plans, so what contingency plans aim at is being carried out, uh, should the contingency arise. Uh, and I’m trying to explain the, uh, the logic of planning, uh, starting with, uh, starting with that.

Um, well, I’d love to say more things about, uh, Francis’s, uh, fascinating comments yesterday, but I think I’d better let the proceedings go on.

(applause)

[01:41:41] MODERATOR:
All right. We have, uh, opportunity now for, for general discussion. Sam? Oh, could you, could you also use the microphone, please?

[01:42:00] JOHN BROOME:
Um, I just wanna raise two extremely simple-minded questions, and, uh, uh, y- I’m sure you’ve answered them, but, uh, I would find it helpful to have the answers restated. Um, so the first really is related to a question that Jay raised on the first day.

Um, and that has to do with this idea of, uh, um, justified resentment or warranted resentment. So you say that if I judge that resentment is warranted, we should think of that as being like a kind of plan to feel resentment in the relevant circumstances. So a kind of quasi-planning, we might call it.

[01:42:43] ALAN GIBBARD:
Mm-hmm.

[01:42:44] JOHN BROOME:
So if I actually, I mean, y– leaving aside the awkwardness of thinking about planning what to feel and waiving worries about the, um, possibility of emotions being responsive to the will, um, if I actually imagine myself sitting down and trying to plan when shall I feel resentment, Um, I, I c- I can imagine myself thinking along two lines. I mean, the first thought I think I’m likely to have is, well, I should– I plan to feel resentment when it’s warranted. Um, and I might have some views about when it’s warranted, when I’ve been wronged, when I’ve been treated unfairly, when someone has taken unfair advantage of me, something like that.

Um, another thought I might have is, um, well, I should feel less resentment than I do. I plan to feel less resentment than I do. I’m going to feel resentment, um, you know, I’m, I’m gonna ep– I’m g-

I’m gonna avoid feeling resentment even on many occasions when it, when it would be warranted, when I have been treated unfairly or been wronged. Um, maybe at the limit case, I decide I’m never gonna feel resentment, um, although I, of course, continue to feel that people wrong each other, and I can be wrong. But for various reasons, I think it would sort of… it would just, you know, life will be meaner if I am always, you know, go– or feeling resentment even when it’s warranted, and I’m just gonna, I’m just gonna give up resentment.

Um, I don’t– I, I take it on your view, neither of those is a, uh, is a way of, uh, is a kind of plan that, uh, that would be appropriate somehow. Um, that is, I mean, it seems circularly to involve notions of warrant that they’re supposed– that you’re supposed to be explaining.

And the second case also involves a kind of divergence between the view of the plan of when to feel resentment and the idea of when its moral judgments are, um, are in place. So I guess… But if I, if I’m not allowed to think in those ways, and I try to imagine, “No, no, you can’t– don’t think those thoughts.

Just plan when to feel resentment, but don’t rely on judgments of warrant.” And I, I, I com– feel completely at sea. I mean, I don’t know how I would begin actually to go about making that plan, even if I thought I could put it into effect, that my will would somehow be magically responsive to my planning.

So, you know, maybe this is just a problem about me, but after all, these are, among other things, my judgments of warranted resentment that are g- being, um, explained here. So I, I, I just would like to know w-how to think about, um, uh, what I should be taking into account when I’m making these plans or when I imagine myself making these plans.

The other question is, is from the second lecture, and it, i-in a way, it’s related to something Francis said. Uh, you– In this path from the hypothetical choice under uncertainty to sort of s- more roughly utilitarianism, um, you appeal a lot to the force of the criticism, uh, the rebuke, you would have chosen it.

Um, how much force that has, of course, depends on what the choice, you know, the circumstances under which you’ve cho- would have chosen it, um, what the thing is we’re talking about, uh, the object of the choice. Um, it doesn’t always have force. I mean, sometimes it would be out of place.

And, um, so the question is, is really about w-w– Could you say a little bit more about what you think the authority of the relevant kind of hypothetical choice is? I mean, Rawls faces the same problem. Why should we care about choice in the original position?

And he gives a variety of answers, you know, one of which appeals to reflective equilibrium, and the fact that what comes out of the choice will match our intuitive judgments. Um, I take it you don’t want to say that, but I’m wondering if you could just say a little bit more about why this particular rather under-described, I think, in the way you, um, in, in the way you, you use it here, choice sh-should be able to bear the weight that you place on it. It’s also a much more general kind of choice than you’re imagining than, than, than Rawls does.

I mean, he’s concerned with just principles for the basic structure, and this is all of morality, you know, and I’m in this very sketchy way supposed to imagine myself choosing in ignorance of who I am, all of the principles of morality. And, you know, uh, not obvious to me that that should– that how much should rest on that or if so why. So I’d just like to hear a little more about that.

Thanks.

[01:47:28] ALAN GIBBARD:
Yeah. That, uh, that is, we might think of, uh, this in a sort of, sort of Brandian way of, uh, choosing the, the, uh, moral code or moral ethos of my society or so. Uh, well, uh, the conditions I put on, and maybe these aren’t enough, but, uh, one is this only applies when the, uh, when the system is the going system.

So, uh, so that’s the way we do things, uh, that’s the established, uh, uh, system seems right. Uh, and, uh, the condition for choice of a system has to be, has to be fair. Uh, so that, that’s what I was, uh, en-envisaging.

Now, what, uh, what is it for it t-to be fair? Well, uh, I’m just going to have to say something pretty, uh, circular. To regard it as fair, uh, is to be regarded as having this, uh, sort of, uh, uh, uh, upshot for, uh, for our, for what our, what moral attitudes are, are warranted and, uh, and that’s, uh, uh, that’s a planning question.

Uh, now, on, on warranted emotions. So, uh, seems to me I can, uh, think about, uh, what sorts of things to resent and what sorts of things not to resent. And I understand this not as, uh, requiring, uh, resent, resenting things at will.

In fact, I, uh, regard it as, uh, as, uh, talk about something different. Uh, so is, uh, uh, being expected to stand in line at a store where there’s a sudden onslaught of, uh, more customers than there normally are that couldn’t have been foreseen. Uh, something to resent.

Uh, know what else can they do. Uh, so I see that as being in a kind in a state that, uh, may not, uh, may not be effective in getting me not to resent it. But if it is, uh, effective in getting me not to, uh, not to resent standing in line, uh, then it’s, uh, then it’s not effective through, uh, my willing not to resent it.

Uh, it’s, uh, uh, I’m in a state that I’d, uh, that I’d express to myself as saying, “Well, resenting it, uh, doesn’t make sense.” And, uh, and I think that, uh, any, uh, feeling like resentment, uh, does respond to such thoughts. Uh, now, uh, I would, uh, on, uh, on the, uh, on thinking I, uh, should, uh, resent less than, uh, uh, Uh, feel warranted resentment less, uh, less often.

Um, seems to me that’s, uh, a kind of view that, uh, even warranted resentment, uh, isn’t desirable. So I would be, uh, planning there to, uh, to desire or want to feel less warranted resentment. So I think, well, there’s some things that are, uh, things to resent, uh, but there are also things to, uh, want not to resent because it’s bad for the digestion to resent or something like that, or, uh, or bad for relations with other people.

Uh, uh, okay. I think that’s, uh, what I’ve got to say.

[01:52:10] MODERATOR:
Maybe, maybe I’ll ask a question.

[01:52:11] JOHN BROOME:
Uh-huh. Um, I, I… Actually, this is just a follow-up to one of, uh, Frances’s questions.

Mm-hmm. And I think it, it, it’s, it’s on that issue that she had more to say about, so perhaps she’d like to follow up as well, um, or, you know, you know, um, take advantage of this opportunity to, to elaborate a little bit on some of her own remarks. But, but I, um, I, I too was a little bit puzzled, um, in the second lecture about how, how quickly we moved to the interpretation of the choice situation in Harsanyi terms as a choice where we’re, we’re assuming an equal probability of occupying each of these, uh, each of these positions.

You know, the Rawlsians I talk to, for instance, they, you know, they all of course concede that if you const- const- construe the choice situation in that way, then some kind of average utilitarianism or something like that comes out of it. Uh, and so, you know, all the, all the, if we’re, if we’re trying to imagine a route from um, broadly contractarian ideas to utilitarianism, you know, it seems like, uh, boy, it’s going to be a huge crux to justify how you, how you, um, set up and describe that initial choice situation, and whether you assume it’s a, a choice under the condition that you, uh, take yours- take there to be an equal chance of occupying each of these positions, or, uh, or rather in Rawlsian’s ter-terms, where you don’t make that assumption. Uh, now, um, in re-in response to, to Francis’s question about this, you, you, you noted that Rawls’s– that, that one argument that Rawls gives, uh, having to do with not knowing which society you’re, you’re choosing for, uh, doesn’t by itself rule out, um,

[01:53:56] MICHAEL BRATMAN:
mm-hmm,

[01:53:57] JOHN BROOME:
you know, uh,

(cough)

um, assuming an equal probability of equal, you know, um, occupying each of the positions in any of the possible societies that, that, uh, that might come into existence. And, and I think that’s true. Uh, so maybe that argument just by itself in Rawls isn’t conclusive, but, but there are other…

So, so the, the friends of Rawls a-a-appeal to other considerations. I think Francis was adumbrating or adverting to some of them, um, you know, having to do with the di-distinctive nature of this choice and the importance of, uh, y-you know, the, the, the w- the worst possible outcomes would be, you know, outcomes for your whole– uh, for the whole of a life, and, and these are the kinds of considerations that make it, uh, they, they would suppose, uh, perhaps unreasonable to build in this equal probability assumption in the initial choice situation. Um, and, and maybe those considerations are, are also inconclusive, but it, but it strikes me that this is really the crux, and it seems like unless we have more to say on one side of the issue or the other, uh, the, the– there’s really not gonna be anything like a linear route to utilitarianism.

And, and it, and, um, it’s– In, in particular, it would be nice if there– something more could be said in favor of the equal probability assumption than, than just that some of the things Rawls says aren’t completely convincing. So I’m, I’m wondering if you could, if you could say something more about why that really is the default assumption. I, it– when I read Harsanyi, it looks like there’s a kind of economist assumption that this is the only way we have of dealing with rationality under uncertainty, um, conditions of this sort.

But, um, um, but it’s, it’s not obvious to me how compelling that consideration is. So I’d, I’d be curious to hear, uh, whether there are other things that can be said about that. Uh,

[01:55:38] ALAN GIBBARD:
well, there, I mean, one, one question is just, uh, are there convincing things on the, on Rawls’s side? So yeah, so that it can affect, uh, that it will affect me for my entire life, uh, doesn’t seem convincing.

If I get, uh, maimed in an earthquake, that will, uh, affect me for my entire life. Uh, if a child gets maimed by an earthquake, it, uh, affects the child for the, uh, uh, child’s entire life. Uh, I came into California.

Uh, the, uh, we deal with low probabilities of, uh, terrible things all the time, and I don’t think a different kind of decision theory applies to that. Uh, and, uh, um, now, when, uh, Rawls was writing against a background of a, a decision theory that tried to do without probabilities and, uh, and stressed maximin, uh, Rawls, uh, doesn’t think that maximin is appropriate, uh, all the time. He thinks…

He argues that it’s appropriate in the– for the kind of choice that he’s making but gives various conditions, and they amount to conditions under which maximin and, uh, and expected utility maximization would coincide as far as I can see. So, uh, so I’m– I prefer the more, uh, the more general, uh, account of rational choice. Uh, so, I mean, another reason is just going to be the, uh, the, uh, the a matter of how plausible I find the, uh, uh, the standard constraints of decision theory as constraints on rational choice.

Uh, of course, the, uh, in this case, the, uh, there’s only one constraint that’s, uh, that’s relevant to whether you maximin or, uh, or do normal expected utility maximizing, namely something like an Archimedean, uh, constraint that, uh, there, uh, that there’s nothing such that, uh, a sufficiently small probability wouldn’t be outweighed by, uh– I mean, there’s no case of something that’s, uh, uh, so immensely important and something else that’s, uh, that matters but is so immensely unimportant that a sufficiently small probability, uh, of the immensely important thing wouldn’t be outweighed by certainty of the less important thing. But, uh, crossing the street to get a chocolate bar just seems like a, um, the sort of case where it’s, uh, implausible to say that the Archimedean condition is violated.

[01:58:35] MODERATOR:
Barry Stroud. Oh, oh, um, could—is this on this point or?

[01:58:39] FRANCES KAMM:
Yeah, this is on this point. I, I won’t go into the other case that I was concerned with. Okay.

But it’s just that when I think about the way, um… I mean, Scanlon— I think about Scanlon’s view about why Rawls eliminated probability. I mean, you know, he mentions all the reasons. I’m thinking about the earlier article, Contractualism and Utilitarianism, which I had an opportunity to read parts, reread parts of, thanks to Sam Scheffler’s lending me the book again.

But, um, the idea that, you know, you want to, um, get people… I mean, it’s true Rawls speaks about, you know, imagine yourself, you know, having Could wind up anywhere. You, you know, could wind up anywhere.

But that that is really, you know, he eliminates the probability, right? Because he wants it to be the case that no matter where you wind up, you could consent, you could agree to this outcome. No matter where you wind up, you could agre- uh, consent to everything,

[01:59:34] JOHN BROOME:
okay,

[01:59:34] FRANCES KAMM:
that’s going on. And it’s, it’s not s-that, um, and it’s not that the reason when you’re in any of these positions you can consent to what’s going on because, well, I would have chosen this from an original position. It, it’s not– that’s not the reason he, I think, you know, ways, uh, that you give for why you can consent to this.

And so the, the focus, I mean, it’s, it’s, it is result-driven. I mean, you, you want… You said that, you know, when I said you’re being forced by the veil of ignorance to take the position of each individual seriously and consider whether someone in that position could consent to this arrangement, okay?

And not just because he’s thinking, “Well, I would’ve, you know, would’ve maximized my expected utility, so how can I not consent,” right? Um, the thing is that, um, uh, you said, well, you know, behind the veil of ignorance, you’re not supposed to be thinking about other people. You’re a self-interested individual and so forth.

But it’s so constructed, right, that the self-interested individual, right, is gonna give us the answer to the question, you know, are we– do we come up with a system that everybody, when they’re actually in those positions, right, can accept? Um, so, well, that’s– I don’t know if that helps,

(laughter)

but it doesn’t probably help. But I just wanted to make clearer that, you know, accepting the idea that you are thinking of yourself behind the veil of ignorance. The point is, what is this, you know, what is the function of this veil?

Uh, what, what is it trying to get you, even as an individual imagined thinking about only yourself behind the veil of ignorance to come out with, right? Um, and it’s not supposed to be the answer that, you know, any position is acceptable given that I would have chosen it behind the veil of ignorance. Uh, it’s, it’s something quite different than that.

I, I- I’m sure that’s not helpful, but- Uh, I’ve tried.

[02:01:22] JOHN BROOME:
Well, but, uh-

[02:01:23] FRANCES KAMM:
I’ve tried to be helpful.

[02:01:25] JOHN BROOME:
Okay. So the, the people behind the original position– I mean, behind the veil of ignorance aren’t, uh, aren’t-

[02:01:30] FRANCES KAMM:
There’s someone behind the original position.

[02:01:32] JOHN BROOME:
Sorry, sorry.

(laughter)

Uh, the, uh, the, the people-

[02:01:37] MICHAEL BRATMAN:
there’s strings.

[02:01:38] JOHN BROOME:
The people- the people. Okay. So the people behind the veil of ignorance aren’t thinking, well, uh, what, uh, what will I accept if I’m, uh, in this circumstance?

And they’re assuming that they, uh, they will, uh, accept, uh, whatever’s chosen as, uh, as legitimate, and that that will motivate them in the way that’s, that, uh, regarding a system as legitimate motivates people. Uh, so the– as you, as you said, uh, the, the real point is that the original position is supposed to, uh, the original position thought experiment is supposed to, uh, tell us something, uh, about the sort of social morality, the justice of basic structures, uh, if, uh, uh, by thinking, by running a thought experiment ab-about people who aren’t concerned with such things. Uh, now, uh, so what will people accept?

Well, that could– we could give that an, an empirical interpretation. What, uh, what sorts of social orders are such that if those social orders are instituted, uh, people, um, people will accept them as legitimate and, uh, uh, and not work to undermine them? Well, that’s a, that’s a very important question, and it has, uh, I think it has complex and important bearings on, uh, questions of social justice.

But, uh, take, uh, take one thing that, uh, we observe, that if people are in the, uh, in unjustly privileged positions and, uh, people try to take away those positions, they, uh, they resist furiously. Um, and that’s part, I guess, of a general, uh, tendency to, uh, uh, especially value and, uh, regard as legitimate what, uh, what you already f– uh, frame as, uh, as yours. So, uh, so people with, uh, uh, Say slave own–

Oh, well, the American Civil War is a, a fine example. Okay, so what were the– what would the slave owners accept? Uh, nothing just, certainly.

Uh, so presumably, uh, but then the other word was, uh, acceptable. So the question then is, uh, something like, uh, what, uh, what ought I to accept, uh, if it’s the going social order and if I’m in a, um, if I’m in a, not in one of the best positions? Um, well, that’s a, uh, that’s a very important question in, in social philosophy.

Uh, of course, Rawls himself didn’t really use a maximin, right? He, uh, uh, he worried about, uh, what at one point he was calling, uh, basket cases, uh, pe-people with, uh, injuries such that, uh, enormous resources would be needed to make their lives, uh, slightly better, and he, and he, uh, decided that he needed something, uh, like, say, the, uh, prospects of the, uh, people in the, uh, lower half of the starting positions or, or something like that, uh, well, I, I, uh, approve of the, of mitigating it, it like that, uh, but I, I still find it, uh, curious that the, uh, that the arguments that lead to it, uh, I mean, the argument he actually seems to depend on there, uh, depend, uh, just depends on circumstances being such that this coincides with utilitarianism. And utilitarianism, I think, is a pretty egalitarian view when you put it, uh, uh, together with, uh, facts about how, uh, how, uh, how one’s life goes depends on one’s income and what the possibilities are of, uh, alternative, uh, uh, arrangements of income.

Sarah Stroud?

[02:06:24] SARAH STROUD:
Well, I have a question very far from this sort of detail. Uh, I have a question, a huge question about just your project in general. It picks up on some of the things John Broome was asking about.

Um, It’s good for us to have plans, and we look at, uh, us in action, and we see that the way to understand our normative beliefs is to start with by attributing to us non-normative beliefs and then explain in some way how it can come to be that we have normative beliefs or has they earned the right to- Mm-hmm… be sort of thought of as, as cognitive. And it’s the idea of that explanation, explaining the cognitivity of those beliefs that I’m puzzling about, and I don’t know that what I’ve got to ask will help.

But suppose, uh, we look at ourselves in the world, and we have these attitudes, um, and we have what you are happy to start with accepting as beliefs about the natural world, cognitive attitudes.

[02:07:42] JOHN BROOME:
But suppose we try to explain them in terms of some kind of non-cognitive attitude, like something like planning to behave as if it were the case that P, as it were in the natural world, something like that. Now, I don’t know first whether that would count or something like that could be described as a non-cognitive attitude, and we could put that together with the other plans. Now, if we got a full account of human goings-on in that way, would we feel that we had explained the presence of cognitive beliefs about the natural world on the basis of starting with something non-cognitive, which gives us the right to think of ourselves eventually now with cognitive beliefs about the world?

And if not, what, what’s the– what do you think is the disanalogy, um, on the–

[02:08:40] SARAH STROUD:
granting what John is objecting to, that you could actually identify the non-cognitive, uh, beliefs first. But let’s waive that. I mean, what– Could you say something about the lack of parallel or whether there is a parallel?

[02:08:53] ALAN GIBBARD:
Uh, Well, there might, uh, there might possibly be a parallel. Uh, uh, Bob Brandom’s project in, in a way is to, uh, uh, has lots of strains, but, but a central strain is kind of, uh, parallel. Oh, I think, uh, I mean, he starts with social, whereas I’m talk-

I, I’m starting with the, uh, with the, uh, thinking about the individual and thinking of the social as interactions of individuals, and he’s not always, uh, consistently what I’d call an expressivist, but uh, but sometimes says things that I, I sound expressivist. He sounds more expressivist than anything else to me. So, uh, so we’ve got the idea there of explaining logical relations in terms of plans, so, uh, or in terms of, uh, of oughts.

So, uh, what is it to say that, uh, uh, snow is white entails something is white? Well, that, uh, if you ought to think that snow is white, then you ought to think that something is white. Uh, or, uh, to put that in planning terms, uh, if imperative, uh, plan to think that snow is white, then, uh, imperative, uh, plan to think that something is white.

I’m, uh, it has to have the, uh, imperative and the antecedent as well as the consequent, uh, ’cause we don’t have, uh, if, uh, uh, if you do think that, uh, uh, uh, uh, that Michael is a unicorn, then, uh, then you ought to think that something is a unicorn. If you do think that Michael is a unicorn, you ought to change your mind. Uh, okay.

So I th- I think possibly you could, uh, uh, explain, uh, the factual beliefs in, in that sort of way. Uh, now, if you, if you did, then as Brandom says, you’d, uh, you’d have norms all the way down and, uh-

uh, well, does, uh, does this then explain? Well, in the, Uh, in the way of, Uh… Well, of course, we– I mean, we have other modes of explanation.

We can, uh, I can talk about beliefs just by, uh, having them or simulating them in, in other people. And, uh, uh, and if I talk about normative beliefs in that way, I, uh, come out sounding like G.E. Moore, I think, uh, the straight explanations, uh, that fit the, uh, phenomena fit, uh, I mean, seem to fit the non-naturalistic, uh, pattern of. Uh, but, uh, thinking about what sort of, uh, uh, frame of mind it is to have a normative belief, and that, uh, anybody who, uh, did valenced planning for life, uh, would have such, uh, uh, states of mind, it seems to be explains it in, in another way.

[02:12:27] JOHN BROOME:
Well, um, but what you just described sounded like planning to have certain cognitive attitudes, planning to think that P and so on. I was trying to get a non-cognitive attitude, like planning to behave in a certain way.

[02:12:45] ALAN GIBBARD:
Mm-hmm.

[02:12:46] JOHN BROOME:
Um, and then all of one’s behavior could then be accounted for, um, in the parallel with your sort of explanation, without attributing even a plan to have a cognitive attitude, but just to have a non-cognitive attitude— Mm-hmm, which would then make sense of everything, and then we would have earned the right then to introduce the idea of our having cognitive attitudes. Not just that we’re planning to have them. As you described, well, Brandom, but really

that now they suddenly

(laughter)

show up. Now, I, I see, my guess, my feeling was, and it’s hard to distinguish that position from saying, “Well, you see,” the way to understand what they’ve been doing all along, what we’ve been doing all along, is holding cognitive attitudes. There’s no difference when we elaborate

(laughter)

the whole thing between believing that P and plan– and carrying out the plan- Mm-hmm to behave as if P were true.

[02:13:39] ALAN GIBBARD:
Um, and so that’s where I felt we would lose any sense of explana-explaining the second in terms of the first.

(coughing)

Uh-huh. It would just turn out to be another way of describing it, as I think John said-

(laughter)

[02:13:51] JOHN BROOME:
Mm-hmm

[02:13:52] ALAN GIBBARD:
with respect to the-

[02:13:52] JOHN BROOME:
(mm-hmm)

[02:13:53] ALAN GIBBARD:
um, your ideas that the– It’s supposed to be that the– the cognitive normative attitudes are just

[02:14:01] JOHN BROOME:
Mm-hmm.

[02:14:02] ALAN GIBBARD:
non-cognitive attitudes in a different guise. But, um-

[02:14:06] JOHN BROOME:
Mm-hmm.

[02:14:06] ALAN GIBBARD:
the guise that they have in their more pristine state-

[02:14:10] JOHN BROOME:
Uh-huh.

[02:14:10] ALAN GIBBARD:
is supposed to explain their having,

[02:14:12] JOHN BROOME:
uh, this later guise, and I wondered whether you thought that would be true if we started with a non-cognitive attitude- Mm- with respect to- No, as it were, non-normative belief.

[02:14:22] ALAN GIBBARD:
Well, when I wrote my first book, I said it was a non-cognitivist theory, and, uh, and then people kept, uh, stamping their feet and telling me that they were cognitivists, and I tried to figure out what the difference was between what they believed and what I believed. And it was hard to put my finger on it, and the things they said, uh, made them cognitivists seemed to be things that I was open to believing. So I began to wonder what the term meant, and I began to wonder what psychologists meant by the term as well.

And I asked one distinguished psychologist who said, “Well, I think when I say that something is cognitive, I mean that it’s complex.” Uh, um, well, I don’t, I don’t think that’s what we philosophers mean, But I don’t think it’s straightforward what we philosophers mean either. One, one thing we could mean is, well, is to be explained on the pattern, uh, say what the subject matter is, and then just say it’s a belief in that subject matter.

But of course, uh, we can conduct that kind of explanation of normative beliefs. Uh, uh, they, uh, start with the, uh, state of affairs of, uh, one, uh, uh, one ought to try to respond to questions. Uh, and then, uh, then there’s the belief that one ought to try to respond to questions, and we’ve given a–

Maybe we’ve given a cognitivist explanation. Uh, the question is what kind of further explanation to, uh, to give and, uh, and I’m not very clear what, uh, its being a cognitivist explanation amounts to. But note that, uh, uh, decision theory, uh, in its standard, uh, uh, sorts of ways of doing things is supposed to explain both, uh, uh, what David Lewis calls credences or degrees of credence, uh, uh, degrees of belief, subjective probabilities, and, uh, and, uh, utilities, preferences, uh, and Uh, and standardly it’s supposed to be a, uh, uh, supposed to be an empirical theory, but then the empirical theory doesn’t fit the phenomena very well, and we might s– uh, we might better take it as a, as a normative theory.

So maybe you can think of, uh, of s-standard decision theory turned into a normative theory as a, as a, an, uh, explanation of cognitive states in, uh, in normative terms.

[02:17:08] MODERATOR:
Okay. I think at this point, it’s, um, it’s fitting that we should thank our, um, Tanner lecturer and our distinguished commentators for a, for a very stimulating set of events, and continue our discussions over, over refreshments.

[02:17:20] ALAN GIBBARD:
And, and we should thank the, uh, the organizers very much, the Tanner Committee and, uh, and, uh, Tanner Foundation.

[02:17:30] JOHN BROOME:
Uh- Let’s thank everyone. And, uh- And William.

(applause)

Sure, yes. I was expecting it.

(laughter)