Return to PHI 530 Home Page

Notes on Readings 
Moral Discourse and Practice: 
Some Philosophical Approaches

edited by Stephen Darwall, Allan Gibbard, Peter Railton
(Oxford University Press, 1997)

Notes on:

Thomas Nagel: from The Possibility of Altruism

I. 1. p. 323. Nagel starts out with the striking statement that ethics (as he sees it) is branch of psychology. This is something for which he will argue in this paper: there can be no strict separation between ethics and psychology. He does not mean simply that psychologists can study the ethical behavior of people. Rather, he means that in order to understand human behavior, one has to understand ethics. This is (partly?) because rationality includes ethics: in order to be rational, one has to be ethical. Being bad is irrational. This is a surprising claim, because as he points out in 4, we normally think that psychology is empirical and a posteriori, while ethics is a priori. But Nagel is claiming that there are a priori elements to psychology.

Nagel is clear that he is arguing for a form of moral objectivity, and against moral relativism. Rationality does not vary from person to person, and neither should morality. Of course, people do vary in how rational they are, but the ideals of rationality at which they should be aiming should remain universal.

Note that Nagel is not aiming to provide a full ethical theory, or even the bare bones of such a theory. Rather, he is merely trying to find one common element of morality. He thinks that he can find it simply by understanding what it is to be an agent - i.e., a person who performs actions. This project is therefore very similar to Kant's, although the details differ. This common element is altruism: we should care about other people. So he is arguing that it is irrational to be selfish. Clearly his primary opponent is the egoist, who argues that one should only care about oneself, or more modestly, that it is rationally permissible to care about only oneself.

Nagel says that he is opposing any demand that the claims of ethics appeal to our interests. "Interests" is a broad term, but means our particular needs or desires. The way he is using the term here, it seems to be closer to "desires." So he is saying that what is right should not depend on what one wants. This would lead to relativism, since different people will want different things.

2. p. 324. Nagel repeats what he has just said in slightly different language. The justification of morality should not depend on whether a person has particular motivations, (= wants or desires). If a person simply lacked such a motivation, then she would have no moral obligation to be moral. For example, Hobbes argued that we should be moral because it is in our best interests. If we are not moral, then society becomes anarchy. But this depends on a person wanting to live without anarchy, so her life is not nasty, brutish or short. But if a person was happy to live such a life, Hobbes could give her no reason to be moral.

3. Again, if ethics depends ultimately on desires, then there can be no further justification of those desires. (The desires justify ethics, and so could not also be justified by ethics without circularity. If anything else justified the desires, then ethics would not depend ultimately on those desires.) What desires we have is apparently a purely contingent, empirical matter. So it looks like a justification of objective ethics cannot depend on what desires people have.

4. "Intuitionism." This is the view that there can be no further defense of our ethical views than what ethical intuitions (feelings, basic beliefs) we have.

p. 325. "Motivation theory." This is basically the science of human behavior, especially what motives ultimately cause us to act.

Note that Nagel is not assuming that ethics is objective: he is trying to prove it. He thinks that this is the most desirable conclusion, because it best explains why we think ethics should have a hold on us. Only when we have failed to show that ethics can be and is objective should we resort to weaker, relativist, views of ethics.

II. 1. Internalism = the view that if S believes that X is the morally right action, then S will necessarily desire to do X.

Externalism = the view that S can believe that X is the morally right action and yet have no desire to do X.

Emotivism = the view that moral statements don't strictly say anything about the world: rather, they just express the emotions of the speaker.

"Philosophers who believe that there is no room for rational assessment of the basic springs of motivation will tend to be internalists, but at the cost of abandoning claims to moral objectivity."

Why?: because such philosophers will tend to say that moral beliefs are based on, or are constituted by, moral feelings or sentiments, and it is plausible to say that there such feelings are intrinsically motivating. But basing morality on feelings means that it will depend on what feelings a person has, and this can hardly be a matter of logical necessity: it will surely depend a great deal on empirical factors.

2. p. 326. Mill = John Stuart Mill, 19th century utilitarian, author of Utilitarianism and On Liberty.

Moore = G. E. Moore: early 20th century British philosopher, author of Principia Ethica. We will read his work later in the semester.

Both are externalists, it seems, although Nagel things there is an unrecognized assumption of internalism in Moore.

3. Hobbes, who thought that without society, life would be nasty, brutish and short, bases his theory of ethical obligation on the universal desire for self-preservation. But even if it is true that such a desire is universal, (and this is doubtful) it is at best a contingent truth.

4. Hume is an antirationalist internalist. He prioritized psychology over ethics.

5. p. 327. Plato and Aristotle saw ethics as more independent or even prior to psychology. Kant is even clearer about this.

III. 1. p. 328. Nagel's view. His account of motivation does not depend on the agent having any antecedent motives.

P. 329. Certain ethical principles are themselves propositions of motivation so fundamental that they cannot be derived from or defined in terms of previously understood motivation. The conception of self is an important part of Nagel's view.


3. Start off by looking at prudence, then turn to altruism.

IV. 1. p. 330. Begin with uncontroversial premises, arrive at general conclusions. The standard view starts off with the view that all reasons must be capable of motivating, and that all motivation has desire at its source. Nagel will attack this latter assumption. Any explanation of action must start with reference to a desire of the agent. Prudence must be explained by a desire to further one's future interests.

2. p. 331. The assumption that a motivating desire underlies every intentional act depends on a confusion between motivated and unmotivated desires. Motivated desires are apparently those which we come to through rational deliberation. Do some desires do not rely on others? If I act in my own interest, it is certainly logically true that I must have a desire to act in my own interest. But it is not necessarily true that this desire caused my action. There is no proof so far that an unmotivated desire was at the causal origin of my action. The desire that was part of my action could have been motivated by my deliberation.

3. p. 332. Similarly, there is not a belief behind every inference from one belief to another.

We don't have the extra belief in Similarly, Nagel is saying, when we go from it is true that the desire to do X must be present.

But it does not have to be

Rather, it could be Desire does not need to be at the root of the action.

Furthermore, Nagel's account is more explanatory. His opponent has to posit unmotivated desires (i.e., desires with no further rational explanation), while he can explain the presence of the desire: it is caused by the understanding of what is right. This is more intelligible.

4. p. 333. Summary.

V. 1. Altruism = a willingness to act in the consideration of other persons without the need for ulterior motives. How can we have a reason to be altruistic?

p. 334. The desire to do X which will promote someone's good may stem from the belief that X is good, not from a general antecedent desire to do good. Such an antecedent desire is not necessary.

2. p. 335. A reason for taking the welfare of others into consideration comes from the fact that you would resent it if someone acted without taking your welfare into consideration. You would think that the other person had a reason to stop. But a reason must admit of generalization, and if the other person has a reason to take you into consideration, then you have reason to take that other person into consideration.

3. p. 336. Nagel's view is opposed by egoism: each person's reasons for acting and possible motivations for acting must arise from his own interests and desires. There can be very few egoists around, because they would have to say that their own interests had no rational claim on the behavior of others.

p. 337. Nagel will argue that egoism is incoherent, in a different (and better) way than other philosophers. His argument rests on the universality of practical principles.

4. p. 338. Nagel will argue that for a reason to be reasons at all, they must express objective rather than subjective values. In order to be a rational agent, one must be acting on rational, objective principles, with a conception of oneself as merely one person among others.

Bernard Williams. "Internal and External Reasons"

p. 363. Internal reasons depend, apparently, on the agent having a motive to do X. External reasons apparently do not depend on such a motive for doing X: S can have a reason to X whether or not has the appropriate desires. The project of the paper is to understand this distinction and draw a few consequences.

p. 364. Internal reasons must depend on the subject's motivational set, by definition. It should also depend on the truth of the relevant beliefs. Williams says that a person does not have an internal reason to drink petrol if she thinks it is gin and wants a Gin & Tonic.

So a person may believe falsely that she has an internal reason to do X. She may also be ignorant of the fact that she has an internal reason to do X. If she knew some relevant fact, she know she had reason to X, but she has the reason to do it even if she doesn't know it. For instance, if one wants a G&T, and there is gin in front of one, then one has a reason to use the gin. But one might not know that the liquid in front of one is gin, and thus not know one has a reason to use it.

p. 366. It is possible for a person to not desire what she needs, and the mere fact that a person needs to get X does not mean that she has a reason (internal) to get X.

p. 367. External reasons. What can it mean to have a reason to X if one has no desire to X or for something to which would X would lead? If one believes that one has an external reason to X, then one could well have a motivation (desire) to X, but this does not tell us what is an external reason (independent of and logically prior to the agent's desires).

p. 368. Williams finds it hard to see how one could have a reason to X if one does not have an antecedent relevant desire.

p. 369. The external reasons theorist wants to say that a person who has an external reason to X but who is not disposed to X is being irrational. But Williams thinks this is a tall order. He thinks that external reasons statements are false or incoherent.

Christine Korsgaard "Skepticism About Practical Reason"

I. p. 373. Korsgaard identifies several sources of possible skepticism about the power of practical reason to guide our actions.

Content skepticism: the danger that the guidance given is so general, vague or otherwise non-guiding that it has no practical consequences.

Motivational skepticism: the danger that reason may fail to motivate an agent.

CK will argue that motivational skepticism is ultimately based on content skepticism, and has no independent force of its own.

p. 374. On Hume's view, (which CK opposes) reason cannot evaluate our ultimate ends, or rank them. Hume's argument is that rationality is about either logic and mathematics, or about cause and effect, and none of these can provide an evaluation of ultimate ends.

p. 375. But it is tempting to think that Hume has an additional argument, that all reasoning that has a motivational influence must start with a passion, that being the only possible source of motivation. This is the sort of view advocated by Bernard Williams in "Internal and External Reasons."

II. CK recaps the distinction between internalist and externalist theories, and claims that there are very few clear examples of externalist theories. She cites J.S. Mill, W.D. Ross, and H.A. Prichard as less than clear examples.

p. 376. CK characterizes an internalist as someone who believes that the reasons why an action is right and the reasons why you do it are the same. She says it is debatable whether Kant is an internalist - we would expect him to be.

p. 377. It is a requirement of practical reasons that they be capable of motivating rational persons, according to the internalism requirement.

III. p. 378. There are various forms of practical irrationality. One is to fail to desire a means to an end one desires, even when one knows it is the means to one's end. To be practically rational, one must be capable of transmitting motive force along the paths laid out by our mental operations.

p. 379. CK has no disagreement with Hume in the possibility of irrationality. It is possible to fail to be motivated by the proper considerations for a variety of reasons. (Of course, Hume's claim seemed to be stronger: that to fail to be motivated by certain facts does not even count as irrational.)

IV. Internalism does not require that we are always rational, but only that we should be moved by rational consideration when we are rational. The possibility of irrationality does not disprove internalism: it is not a counterexample to the thesis that a rational person will, in virtue of being rational, be motivated by the recognition of what is good.

V. p. 380. It may not always be possible to get an irrational person to be rational.

VI. Bernard Williams is meant to be an externalist, but his position actually leaves open the possibility of internalism. It may be that motivation must come from a person's subjective motivational set. But it might be that every rational person should have certain kinds of desires in that set in virtue of being rational. Williams must be implicitly assuming content skepticism: that ultimate ends are beyond the possible jurisdiction of practical reasoning.

VII. p. 384. Kant's view is that argued for by CK, that we will be motivated to act as the categorical imperative directs, if we are rational.

VIII. p. 385. Summary.

Christine Korsgaard, "The Sources of Normativity: Lecture 3: The authority of reflection"

In the first lecture, CK criticized moral realism as defended by Nagel. She argued that it was unsuccessful in explaining where moral authority derived from. In the second lecture, CK criticized Hume, Bernard Williams, and J. S. Mill. Both try to justify our moral dispositions through reflective endorsement: when we turn our moral scrutiny on itself, it can look on itself with approval.

Introduction p. 389. CK summarizes the theories she has rejected in the previous two lectures:

Voluntarism: The moral law is given to us.

This does not explain why we should obey it.

Realism: We should obey the law because it provides us objective reasons for action.

Why should we believe in objective reasons?

Reflective endorsement: Obligations and values are projections of our own moral sentiments and dispositions. They are justified because when we turn them on themselves, we find we approve of them.

She will now defend her Kantian position, which combines elements of voluntarism, realism and reflective endorsement.

Voluntarism is true in so far as the moral law is given to you, because you give it to yourself.

Realism is true in so far as the law provides us with real reason for action.

Reflective endorsement is true in so far as it provides reflection authority over ourselves.

But she makes clear that she will not attempt to show how her theory might provide non-conflicting moral principles, or what principles it might actually deliver.

The Problem. p. 390. CK argues that in ordinary language, "reason" means reflective success. But how do we know when we have succeeded?

p. 391. The will is causal, and everything causal must operate under laws. In being rational, the will must act for reasons of its own. But the will is also free, and so cannot be determined by outside causes. Therefore the will must have its own law or principle: i.e. it is autonomous.

The law of free will is the categorical imperative, i.e., to act only on a maxim that we could will to be a law. The moral law, as she defines it, tells us to act only on maxims that all rational beings could agree to act on together in a workable cooperative system. The categorical imperative does not imply the moral law so far.

The Solution. p. 392. When you act rationally, it is not a matter of your desires interacting with your beliefs. There is something further going on, viz., you choosing what to do. Your choice is a matter of your character as you find your life to be worth living. One's identity often determines one's actions. Furthermore, one has an obligation to not perform actions which would make one feel that one's life was no longer worth living. This identity is not a theoretical construct, but is a practical matter.

p. 393. So obligation derives from the reflective structure of human consciousness. We command ourselves to act.

p. 394. We can tell that our maxims are good when they have the correct form.

Moral Obligation. p. 395. CK still needs to rule out relativism, i.e. explain how restrictions on the form of the law we give ourselves rules out relativism.

Concepts are the meanings of words. When we have different theories of the good, they all use the same concept of the good. They have to, because otherwise they would not even be talking about the same subject and so would not even manage to disagree with each other. But the different theories do disagree with each other (otherwise they wouldn't be different) by having different conceptions of the good.

CK refers to Bernard William's use of the distinction between thin and thick ethical concepts. Thin ethical concepts are very general ones, such as "good" and "bad." This is what is used by Utilitarianism. Thick ethical concepts are much more specific, such as "honor," "gratitude," and "disgrace." They have much stronger implications about what actually counts as right or good. They are connected with different conceptions of human identity. For CK, the point of talking about thick ethical concepts seems to be to show how moral theory or normativity is related to one's practical identity.

p. 396. Some conceptions of the good and the associated conceptions of identity will be better for us than others. [Better by what standard? Is there an Archimedian point from which we can judge this?]

Communitarians are political theorists, generally taking Aristotle as their inspiration, who think that the community or society should be prior to the individual. They are more willing to place the general good over individual rights, and they tend to think that people are essentially social animals dependent on each other. They criticize liberals who place individuals over society as having an "atomic" conception of the self, and as ignoring the interrelatedness humans in society.

To be human, she recaps, is to have self-consciousness and thus to be capable of acting for reasons rather than out of instinct. Since we can act for reasons, we must. Without reasons, we do not act at all. So we need to adopt a practical identity. This so far still does not rule out relativism. But CK does think that it establishes the general result that humans are valuable in themselves.

This is how CK derives normativity from human nature. She uses slightly different language, making the same point, saying a human being "needs a practical conception of her own identity." This practical identity is normative for the person doing the acting. "Since you cannot act without reasons and your humanity is the source of your reasons, you must endorse your own humanity if you are to act at all." (397) This is CK's proof that human beings are valuable, at least to themselves.

Obligating One Another p. 397. Consistency then forces us to say that other people have their own practical identities and that their humanity is valuable to them. But it still does not force me to take your humanity as an end in itself for me, and does not show that I have any obligation to you. Reasons are still private to particular people, and are not shared. At this point, CK turns to Wittgenstein.

p. 398. It is not easy to summarize Wittgenstein's ideas. His most important work was the Philosophical Investigations, a book posthumously published in 1952. In that work he examines the nature of language and meaning, and the relation of those to experience. He rejects the old view that the meaning of a word is derived from the object it represents, and that meanings are private mental entities. Part of his argument has been given the name the "private language argument." In this he tries to show that a person on her own could not create a private language unknowable to others. But the old view of language implies that a private language would be possible, and so that old view must be mistaken. It seems that Wittgenstein thought that language is an essentially public, social phenomenon, and so incapable of being private.

CK sets out some of his argument. The basic idea is that it is a necessary part of language to be able to use a word correctly. But that requires that we be able to distinguish correct usage from incorrect usage. Yet with a private language, we would not be able to make any sense of a person using a word wrongly. It would be, in principle, impossible to have a record of incorrect usage. So a private language wouldn't meet the necessary criteria of being a language.

CK extends Wittgenstein's ideas to apply to private reasons. She points out that language and action are both normative, meaning that there are standards by which to judge them. She concludes from this that they must be relational. "Reasons are relational because reason is a normative notion: to say that R is a reason for A is to say that one should do A because of R; and this requires two, a legislator to lay it down and a citizen to obey." (398)

p. 399. One can use language rightly and wrongly, and similarly one can act rightly and wrongly. For this to be possible, reasons can no more be private mental entities in a person's head than meanings can. Reasons must be public, like relations we have with ourselves and others. We are enmeshed with each other, and we cannot ignore each other. By making you notice me, CK thinks, I can force you to acknowledge the value of my humanity. In the same way that meanings can be shared, so can reasons.

The Origin of Value and the Value of Life p. 401. CK points out that Wittgenstein's view about the relation between language and experience is tested by the case of naming pains. Pain does seem like a private sensation that is hard to communicate to others. At the same time, we generally feel very sure when we are in pain, and we are confident in our ability to use pain words correctly, independently of the corroboration of other people. That is to say, pain seems to be a private sensation, and no one but myself can know for certain whether I am in pain or not. Some would go so far as to say that we could not be wrong about whether we are in pain, and that, contrary to Wittgenstein, this is not a problem in our use of pain language.

CK argues that pain is not in fact completely private, and one can be wrong about whether one is in pain. Consider the statement "My left middle toe hurts." Can someone sincerely uttering this statement possibly be wrong? Yes. She may have mistaken which toe that hurts. Or more dramatically, she may be lying in a hospital bed with an amputated food, and although she thinks her left middle toe hurts, in fact she no longer has that toe at all.

Someone might reply as follows, "Yes, of course one can be wrong about the state of one's body, but one can never be wrong about the pain itself. I can be certain I am in pain, and to be more accurate in reporting this, I should say that I know I have a pain, which feels like my toe hurts." This reply separates the sensation from the body. It supposes that when we talk about pains, we are really talking about the contents of our minds, and "reduces all mental activity to the contemplation of sensations and ideas." (p. 402). Wittgenstein was opposed to this model of the mind and the relation between our words and the contents of the mind.

CK follows Wittgenstein's line of argument, and suggests that when someone says she is in pain, she is announcing that she has a very strong impulse to change her condition. I.e., in saying she is in pain, she is not just talking about a sensation, she is also talking about her whole self. She explains her position by saying that the impulse to change her condition does not depend just on the pain itself: it also depends on the condition of the person. A person can have a pain which does not change, but through taking a tranquilizer, the need to avoid the pain is reduced. She says that pain "is our perception that we have a reason to change our condition."

p. 403. This implies that only creatures capable of having reasons and rationality are capable of experiencing pain. CK pauses to explain how non-human animals, which we are sure can experience pain, can have reasons. Her answer to this is that the purpose of animals is to maintain and reproduce themselves. Since this is their purpose, they have a reason to do it, whether they know it or not. Animals can perceive that they have reasons to change their condition because they can perceive and revolt against threats to the preservation of their identity.

Having dealt with the privacy of pain as a sensation, she considers the pain as a reason for action. Merely being in pain can be a reason for ending the pain. The reason here seems like it could be entirely private, not shared by anyone else, in the same way that the experience seems private. CK's argument against the privacy of experience should, she hopes, also apply to the privacy of reasons for stopping pain. On her view, obligation is the reflective rejection of a threat to your identity, and pain is the unreflective rejection of a threat to your identity. So the preservation of identity is imperative. (This suggests one has an obligation not to commit suicide.)

So the perception of threats to the lives of other people does give one a reason to act, to preserve their identity. Those reasons are public and shared. Furthermore, CK argues that we can perceive the threats to the lives of animals, and thus we have an obligation to them as well. We need to value the lives of animals.


CK addresses Mackie's argument from queerness against moral objectivity. He argued that intrinsically normative entities would have to tell us what to do and force us to do it, and no objects in the natural world are like that. CK replies that there are such objects: they are humans and other animals.

G. E. Moore, from Principia Ethica

This book was published at the beginning of the century. G. E. Moore was an English philosopher, writing in the shadow of Utilitarian theory, and this was one of his first works. He was a very influential philosopher for the twentieth century, partly responsible for the preoccupation of philosophy with language and meaning, and the study of these to solve other philosophical problems.

5. p. 51. Moore sets out to find out the meaning of "good." He says this is the most fundamental question of Ethics. Without a clear answer to that question, we have no basis for any philosophical understanding of ethics.

6. p. 52 Moore says he is not simply trying to find out how people ordinarily use the term "good." He wants to find out the idea or object that we refer to when we use the word "good." [Note that this approach would be rejected fifty years later by 'ordinary language philosophers,' who would say that once we have understood the subtleties of ordinary usage of the word, there will be no more philosophical problems left. However, it was, almost paradoxically, Moore's influence that led philosophers to be preoccupied by questions of the meaning of words rather than the nature of the world.]

To understand what Moore is saying, it might be useful to think of the distinction, made by Russell, between what a word denotes and what it connotes. The connotation of a word is its meaning, which can sometimes be given by a description. The denotation of a word is the thing or things it refers to. The connotation of "horse" is, maybe, hoofed quadruped of the genus Equus. The denotation of "the horse" is an actual animal. Some words can have a connotation without having a denotation. I.e., they fail to actually refer to anything in the world. For example, "unicorn" means "horse with a horn in its head" but "unicorn" does not denote any actual objects because there are no unicorns. Moore is saying he is interested in the object referred to by "good," not the ordinary meaning of the term. (Of course, since "good" is an abstract term, it is much less clear what kind of object it might refer to. "Good actions" would denote actual actions. "Good people" would denote actual people.)

Moore will argue that we cannot define the term "good." He is rejecting any theory which pretends to find out the nature of good by examining the meaning of the word.

7. "Good" is unanalyzable into simpler terms. It is like a color: if you haven't experienced it or if you don't have any knowledge of it, then no one can explain it to you. [How then do children learn the meaning of good?] Only complex things can be reduced to simpler ones, and good is already simple.

8. p. 53. Moore distinguishes 3 kinds of definition of a word:

1) arbitrary verbal definition

2) verbal definition proper

These are both ways of explaining what one means by a word. "Good" can be defined in these ways.

3) a semantic equivalence between the word and a combination of other terms referring to simpler objects, where one phrase could replace the term being defined.

Moore says that "good" cannot be defined in this third sense.

9. We can still pick out good things, using other terms. He is not saying the good is unknowable. It might be that what gives pleasure is good. However, "good" does not mean "pleasure-giving."

10. p. 54. It would be a similar mistake to try to define "yellow" as some kind of physical property of matter that reflects light of a certain frequency. It might be that all and only yellow things do have that physical property, but still that is not what we mean by yellow. Those who think we can define "good" in such a way are committing the "naturalistic fallacy."

11. Moore attacks purported proofs of theories about the nature of good that rely on claims to know the meaning of the word.

12. p. 55. Moore goes on explaining that we do not have to suppose that we are giving the meaning of "good" when we say, truly, that pleasure is good.

13. p. 56, 57. The only alternative to Moore's view that "good" is simple is that it is complex or that it has no meaning at all. One attempt to define good as a complex is

Theory: A is good = we desire to desire A

This cannot be right, Moore argues, because it still makes sense to ask "Is it good to desire to desire A?"

If the theory was right, this would mean the same is "do we desire to desire to desire to desire?"

It would also mean the same as "Is good good?" But the latter question is obviously true in a way that the prior questions are not, and the former is more complicated than the original question. This is Moore's "open question" argument against purported analysis of the meaning of good.

[Consider the plausibility of Moore's argument here. Take an analysis of meaning that is obviously right, such as "a bachelor" = "a never married man." Does it make any sense to ask "are all bachelors never married men?" Does "bachelors are never married men" mean the same as "bachelors are bachelors." If our analysis of "bachelor" is correct, then shouldn't the two sentences have exactly the same meaning?]

The fact that these questions make sense shows that good does have meaning, and that meaning is not complex.

14. p. 58. Moore goes on to explain why Bentham's argument for Utilitarianism does not work. Bentham claimed, or implicitly assumed, that "right" means "conducive to the general happiness." Bentham says that the greatest happiness of all concerned is the proper end of human action, so it is not just a means to an end. Moore concludes that Bentham has committed the naturalistic fallacy. If it were true that "right" means "conducive to the general happiness," then this would be the same as telling us that "right" means "right." I.e., it would be entirely trivial.

p. 59. Moore points out that even if Bentham's argument for Utilitarianism does not work, Utilitarianism might still be true. But he warns that the naturalistic fallacy has been the cause of philosophers believing false principles. If one starts off with a definition of the good that already builds a moral theory into itself (as Bentham does) it will not be surprising to conclude that that moral theory is correct, although one's argument will be fallacious.

Moore suggests it is better to start out on one's Ethical enterprise without a definition of the good (which is of course necessary anyway since it is indefinable). One will then not implicitly and illegitimately presuppose a particular moral theory. Starting without a naturalistic definition leaves one much more open minded.

15. p. 60. Moore says it is essential to distinguish between what is intrinsically good and what is good because it is the means to an intrinsic good.

16. Ethics should discover universal laws about the good. But it will be very difficult to find universal laws about causal relations leading to the good, (good as means), because it is so difficult to find universal laws in general, and especially when it comes to human action. The best we will ever get is a generalization. Generalizations are not the business of Ethics.

17. p. 61. Moore explains that "to assert that a certain line of conduct is, at a given time, absolutely right or obligatory, is obviously to assert that more or less evil will exist in the world, if it be adopted than if anything else be done instead." (p. 62) This will require looking at the effects of the action and its alternatives. All practical questions in Ethics require looking at the effects of actions. Moore thinks that many of the disagreements in ethics have arisen due to a failure to distinguish intrinsic from extrinsic value. Some assume that nothing has intrinsic value, which is impossible. Others assume that what is necessary must have intrinsic value [Moore may be thinking of Kant here].

[A comment on the relevance of Moore's argument now. Few philosophers now see themselves as trying to explain the meaning of moral terms. A philosophical theory of ethics, whatever it is, cannot proceed merely by conceptual analysis. So in a sense, Moore's lesson has been taken to heart. But really Moore's opponents, those trying to give a naturalist account of morality, have bypassed his objection because they no longer see themselves as required to give a conceptual analysis of "good" in naturalistic terms. There is a parallel to philosophy of mind: there are many philosophers who want to give a reductive account of the mind in terms of the brain or its function, but most don't normally claim to be giving an account of the meaning of mental terms. Metaphysical reduction, it is assumed, can proceed without semantic reduction.

This is not to say that conceptual analysis is irrelevant or useless to ethics. Indeed, my report of its demise is exaggerated, and there are some philosophers who still take it as the central method of philosophy. Even those who use a greater variety of tools still need to give some account of the meaning of moral terms, whatever moral theory they hold.]

J. L. Mackie, from Ethics: Inventing Right and Wrong, Chapter One

Mackie was influenced by the British empiricists, and wrote books on Hume and Locke. Much of his work can be seen as developing themes present in those philosopher's works. His view of ethics is reminiscent of that of Hume. Mackie's final book, published posthumously, was The Miracle of Theism, in which he argued that all religious belief is irrational.

1. p. 89. Mackie says there are no objective values. Note that he is not denying (or affirming) that there are any values at all. I.e., he is leaving it a possibility that there are subjective values. He is not saying that we should reject morality altogether; he is just denying that moral values are part of the world.

2. p. 90. Mackie distinguishes his view from the one that says that "This action is right" means "I approve of this action." This is not his view, since he thinks that "This action is right," insofar as it tries to express an objective truth, is always false, while "I approve of this action" is clearly going to be true in some cases. Mackie is not trying to give a reductive account of the meaning of "This action is right."

7. p.91. Mackie explains that he is not attacking a straw position: not only have many philosophers, such as Plato, Aristotle, and Kant, held that there are objective values, but he also claims that this is a part of ordinary thought.

p. 92. Mackie refers to the debate about the meaning of moral terms engaged in by Moore.

"Noncognitivist" theories say that moral terms have no descriptive meaning, i.e., they say nothing about the world as distinct from the agent. Such theories may allow that moral terms express attitudes of the agent, such as approval or disapproval. In this sense, Hume seems to count as a noncognitivist.

"Naturalist" accounts of moral terms are those such as Bentham's, which say, for example that "X is good" simply means that X has some natural property, such as maximizing the general happiness.

Confusingly, Hume might also count as a naturalist in this sense, even though noncognitivism and naturalism are incompatible. (The confusion probably arises from the fact that Hume wasn't really trying to give an account of the meaning of moral terms, and so fitting his view into either category is anachronistic and distorting.) Tellingly, Mackie, who is in sympathy with Hume's view, thinks that there is some truth to both views, but also some falsity.

p. 93. Mackie says that the problem with these views is that moral judgements do seem to express claims about the nature of the world (and so are cognitive) that are meant to have normative implications ("a call for action") whatever desires we might have (and so are non-natural). Mackie points out that for naturalists a moral judgment only have normative claim on a person if that person has prior desires.

Mackie points out, in support of his analysis of moral terms, that people are inclined to suppose (falsely) that objective values are the only kind there could be, so that if there are no objective values, then there can be no reason to care about anything. (God is dead, everything is permitted.)

8. p. 94. Having claimed that our moral judgments are based on a mistaken belief in objective value, Mackie has to justify his claim that this belief is false. He first gives the argument from relativity.

If there were objective moral truths, then people would live according to one moral code.

But people across the world live according to many different incompatible moral codes.

So there are no objective moral truths.

Addendum: people's moral behavior should be explained by sociology and anthropology, not in terms of their ethical knowledge.

p. 95. Rebuttal. There are shared elements to all people's moral behavior, and it is this common core that is part of objective moral knowledge. These are the basic moral principles.

Mackie says that this rebuttal could be only partially successful at best. It would only justify a small part of conventional morality (the common core). Many other moral rules would be left unjustified. So Mackie seems to endorse the argument from relativity.

9. However, he places much more weight on the argument from queerness. There are two parts to this. First, metaphysical, based on the idea that we cannot make sense of objective value as being part of the world. Second, epistemological, based on the idea that even if there is objective value in the world, we could never have any knowledge of it.

p. 96. "Intuitionism" is the view that we have a special faculty of moral intuition that tells us what is morally right and wrong. Mackie says that any defender of moral objectivism must assume the truth of some version of moral intuitionism.

Mackie points out that the best defense for the moral objectivist to the argument from queerness is to say that although it is hard to say how we know the moral truth or what kind of thing it is, it is no stranger than many other things that we have no trouble believing in, such as "essence, number, identity, diversity, solidity, inertia, substance, the necessary existence and infinite extension of time and space, necessity and possibility in general, power and causation." Mackie gives a promissory note saying that it should be possible to explain all these things in empirical terms. That is to say, he thinks that we should be able to give a scientific explanation of all these things, and an explanation of how we know about them. But he is also ready to bite the bullet and say that if we can't give an empirical account of one of them, then we should conclude that it is illusory.

[It is especially hard to imagine how we might give an empirical account of logic and mathematics. We have a great deal of confidence in the truth of these and of our knowledge of those truths, but how are we to explain that knowledge? Is it truth independent of humans, or would there be no logical or mathematical truths if there were no human minds?]

p. 97. Part of the queerness of objective values is that they are not meant to be just there in the world, but they are also meant to be action-directing. They tell us what to do. But our ordinary senses don't provide us with any information about such entities.

How are objective values related to natural properties? A piece of needless cruelty is thought to be objectively wrong. But the wrongness is neither logically nor causally related to the cruelty. What other kind of relation is left?

Mackie doesn't go so far as to claim that the idea of objective values is incoherent. Rather, he says it is much more plausible to suppose that there are no such things because we can explain everything we see in the world without resorting to such mysterious entities. The alternative explanation would presumably be in terms of our subjective responses to the world. Using Ockham's razor, we should conclude that the simpler explanation is correct.

10. p. 98. Mackie still needs to explain why people make the mistake of believing in objective value. He does this by pointing out that we have a tendency to project our feelings about the world onto the world, as if they were part of the world. If a fungus disgusts us, we call it disgusting.

Furthermore, there are social pressures to justify moral claims. We want to control the behavior of other people, and when they ask "why shouldn't I do this," we reply "because it is wrong." On Mackie's view, this is just a rationalization, in the sense that we construct a false justification to make our lives easier.

p. 100. The "Euthyphro dilemma": Is X good because the Gods find it pleasing, or do they find it pleasing because it is good?

The first horn of the dilemma: if X is only good because the Gods find it pleasing, then why should we agree with the Gods? Is it only the fear that they will be displeased with us if we don't do what they want? On this horn of the dilemma, there is no intrinsic goodness to X, and so the fact that the Gods find it pleasing is just a contingent fact about them. It does not show why X should be pleasing to us.

The second horn of the dilemma: The Gods find X because it is good. This makes the Gods rational and shows why we should agree with them. But it leaves us with absolutely no explanation of why X is good.

This dilemma highlights the difficulty of providing an account of the objectivity of value.

12. Summary.

Richard Boyd, "How to be a Moral Realist"


There is a debate within philosophy of science whether we have good reason to think that the unobservable entities postulated by modern physics really exist. Do electrons really exist, or are they a useful fiction that help us to predict the results of experiments (this is known as instrumentalism)? The standard view, scientific realism, says that such entities do exist, and normally argues for this by saying that it would be too much of a strange coincidence the our scientific theories which posit them to have success in their predictions if they didn't really exist. The best explanation of the success of these theories is that the theories are literally true, and so, because we should believe in the best explanations we have of the world, we have good reason to believe in electrons.

There are parallels to the debates about the existence of God here. Some theories try to prove the existence of God by saying that the existence of God explains things that would otherwise be mysterious, such as the existence of the universe or the miracle of life. But of course there is controversy whether these "God theories" are really explanatory (postulating a mysterious unobservable entity but making no bold predictions with the theory) and whether they are any more explanatory than alternative theories (such as "no God theories" and Evolutionary theory).

Here Richard Boyd, a stalwart scientific realist, draws parallels between the debates about scientific realism and ethical realism, and argues that a more sophisticated understanding of what it takes to defend scientific realism could be helpful to those who want to defend ethical realism.

1.1 p. 105. Boyd characterizes moral realism with three properties:

1. Moral statements are meaningful and have a truth value (either true of false).

2. Morality is not relative to human thought: something is right or wrong independent of whether we happen to think it right or wrong.

3. We can have moral knowledge using normal methods of moral reasoning.

1.2 p. 106. Moral antirealists often draw a distinction between science and ethics. Science is meant to be hard-nosed, in that it deals with observable entities and can be tested. Ethics is thought of as soft, making no testable predictions about what will happen. Then antirealists say there is no good reason to believe in ethics, because it is soft, on the assumption that it is only the hard of methods of science that can really lead to knowledge.

p. 107. Boyd wants to argue that the contrast between science and ethics is not as stark as these antirealists suppose. Some philosophers, such as Thomas Kuhn, in The Structure of Scientific Revolutions, have argued that science is not neutral or objective, but in fact must always be approached from some point of view, or paradigm, which is in some sense subjective. In this way, Kuhn argued that science is theory-laden and value-laden. But rather than bringing science "down" to ethics, Boyd wants to lift ethics "up" to the level of science. He wants to show that it is more objective and empirical than is often supposed.

Boyd says that he has in the past argued for scientific realism, and has argued that this requires a naturalistic and realist conception of knowledge, natural kinds and reference. "Natural kinds" are the categories we use in science to classify the objects it talks about. For example, atoms, humans, salt, gold and cells are all natural kinds. All members of a natural kind have something in common which is shared by nothing outside that natural kind. This is sometimes known as the essence that goes along with the natural kind. For example, the essence of water is being composed of H2O. The essence of humans is having a certain genetic code. In saying that he is a realist about natural kinds, he is saying that he thinks that there are real differences in nature which are captured by our natural kinds. They "carve nature at its joints." In saying that he is a naturalist about naturalist about natural kinds, he means that we can explain how we come to have these categories in natural, scientific terms. In being a naturalist and a realist, Boyd is taking a position against the common argument that since our ideas are the products of our minds, they must be subjective.

"Reference" is the ability of words to denote objects in the world. In using the word "Moses," I manage to refer to Moses, the person. Philosophers have found this rather mysterious. For example, Moses has been dead a long time. He no longer exists. How can I manage to refer to something which does not now exist? In being a realist about reference, Boyd is saying that we do really manage to refer to objects with our words. Our words do really pick out the objects we want them to, for the most part. And in being a naturalist about reference, he is saying we can explain the relation of reference in natural, quasi-scientific terms.

2. Boyd goes through some of the antirealist arguments in greater detail.

2.1 Science is based on observation, which is an objective way of finding information about the world. Ethics seems to be based on intuition, which seems to be a subjective reflection of our feelings. Aren't intuitions a weak foundation for a system of thought, compared to scientific observation? Is there anything else that ethics could be based on?

2.2 p. 108. Rawls has argued that we come to our ethical opinions by going through a process of reflective equilibrium. When our moral intuitions conflict with each other, we trade off between them until we have reached a stable compromise position. But isn't this just rationalization, not a process to discover truth? At best, it is a construction.

2.3 Any form of knowledge of the world will progress, getting closer to the truth or learning more truths. But ethics doesn't progress. Therefore it can't be a form of knowledge of objective facts.

2.4 Ethics contains hard cases about which no agreement can be achieved. Science solves its hard cases, and does not contain unanswerable questions. Either this shows an important difference in kind between ethics and science, or else it shows that the ethics is not about facts.

2.5 Either ethical properties are meant to be natural, in which case science gives no evidence for their existence, or else they are supernatural, and belief in them is a form of superstition.

2.6 p. 109. Factual judgments don't provide intrinsic motivation, but moral judgments should always provide a reason for action. So moral judgments cannot be factual.

2.7 Different people mean very different things by "good," so they can't all be referring to the same objective property when using the word.

2.8 Boyd notes that the seven previous arguments do not rely on traditional verificationist worries -- we can't prove ethical statements, so they can't really mean anything or tell us about the world. He says that are more like Kuhn's arguments against the objectivity of science. Kuhn looked at the history of science and argued that so-called scientific "progress" was not as rational as philosophers of science thought. He argued that scientists clung to a "paradigm" that ruled their thought, until the number of problems for that paradigm grew to be serious, and when a more promising alternative paradigm became available, there would be a "scientific revolution." One of his most important claims was that different paradigms are "incommensurable," which means that there is no absolute standard by which we can compare two paradigms, and so there is no rational way to show that one paradigm is "closer to the truth" than another. Boyd has previously argued against both verificationism and Kuhnian anti-realist arguments, and here he plans to extend his anti-anti-realist arguments to show how to defend moral realism.

3. Boyd explains how recent philosophy of science has dealt with anti-realist arguments.

3.1 p. 110. Boyd says that philosophers of science have become more realist in the last thirty years, partly because they have seen how tightly theory and observation (or more generally, methodology) are intertwined. It is not feasible to say that we can believe our observations but disbelieve our theories, as philosophers used to think. One example is the "observation" or sub-atomic particles. This in fact involves many theoretical assumptions. It is very hard, if not impossible, to draw any clear distinction between pure observation and theory. So if we are going to believe our observations, we have to believe our theories. This means that our observations are less certain than we thought, and our theories are more certain than we thought.

This change in philosophical attitude has not been restricted to philosophy of science. It is matched by development in philosophy of language, epistemology and metaphysics. Boyd will explain how these are relevant to ethics.

3.2 p. 111. Boyd says that realists believe that science steadily gets closer to the truth, and that this is an important part of their explanation of scientific measurement. Second, there is a dialectical interaction between theory and methodology, working for their mutual improvement. Our confidence in one enhances our confidence in the other. So the theory dependence of methodology enhances its reliability. Only realism explains how our measurement and observation is so reliable.

3.3 p. 112. Although most epistemology has been foundationalist, foundationalism is fundamentally mistaken and should be rejected. There is no class of epistemically privileged beliefs. To put this more simply, we cannot hope to build our knowledge, from the bottom up, on a foundation of experience, because the beliefs related to experience just as fallible as any other beliefs. Furthermore, there is no rational non-deductive method for finding knowledge which we can know a priori.

p. 113. Boyd says that we should instead focus on the regulation of belief, not its production. We should regulate our beliefs in the way that seems best in the light of our knowledge. Our methods for finding knowledge will change as our knowledge grows. They will be influenced by social and other factors. Our epistemology will be influenced by our theories of the world and our relation to it. The reliability of our methods will depend on the approximate truth of the background knowledge which our methods rely on. So epistemology is just as much a job for science as for philosophy.

3.4 "non-inferential perception": this is the knowledge you get directly from your sensory experience. It is the immediate content of your experience, not what you can infer from that experience. An inference is a step in reasoning.

p. 114. Scientists have to undergo training in order to become good scientists. This requires the acquisition of good scientific intuitions, getting a feel for the practice of the science. Making the right sorts of inferences needs to be second nature to the scientist. This is one way in which background knowledge is built into tacit judgments. These intuitions are not guesses: they are grounded in a theoretical tradition which is approximately true.

3.5 Causality is an important part of science, which cannot be given an analytic reduction to other concepts. We should not want to reduce causality away. (So Hume was wrong about causality.) More generally, although it is true that in a sense science provides a way to reduce everyday phenomena to scientific ones, this does not mean that we should expect to reduce the meanings of all our descriptions of these phenomena to scientific terms. Loose example: we can explain a sunset in scientific terms, but we cannot reduce to meaning of a sunset to those terms, and we should not expect to. This is not a problem for the scientific approach, since it is not trying to replace our whole vocabulary.

3.6 p. 115. Our scientific categories need to match real fundamental features of the world. We cannot decide our categories a priori: we must mold them to our knowledge of the world.

3.7 Our categories should refer to natural kinds in the world. Our understanding of that ability to refer to the world should be causal. Boyd gives an analysis of reference.

3.8 p. 116. Some scientific properties and kinds can be defined with necessary and sufficient conditions. For example, something is of the natural kind water (in either solid, liquid, or gaseous forms) if and only if it is made of H2O molecules. But there are other scientific properties and kinds which cannot be defined so simply. For these, being part of the non-natural kind in question will depend on have enough of a variety of properties. For example (maybe), an object is a chair if it has enough of the following properties. Some of these properties will have more importance than others. What exactly counts as a chair will be somewhat vague or indeterminate.

you can sit on it

you can sit in it

you cannot lie on it or in it

it was designed to be sat upon

it is used for sitting on

it has four legs

it has three legs

it has five legs

p. 117 Boyd gives 11 conditions for non-natural-kind terms based on the notion of a homeostatic property cluster. He gives health as an example. He also says that what is normally thought of as a natural kind, in fact the paradigm of a natural kind, viz., a biological species, is best characterized by a homeostatic property cluster. I.e., there is no simple necessary and sufficient criterion for what is to be a member of a particular species; a species is characterized by "imperfectly shared and homeostatically related morphological physiological, and behavioral features." (118). It will be inevitable and that exactly counts as a member of the species will be indeterminate. It would be scientifically counterproductive to have an exact definition. Indeterminacy is a necessary and desirable feature in referring to complex phenomena.

4.1 p. 119. Boyd will now use this model of scientific realism to show the errors of several arguments for moral anti-realism, (which he often refers to as "constructivism," which is the idea that we create morality rather than discover it, and which abandons any claim for our moral beliefs being objectively true). He mentions one off the cuff, apparently: the fact that different people and cultures have very moral concepts does not prove that they are not all managing to refer to the same thing, in the way that different cultures may have very different beliefs about the nature of water, yet all refer to the same stuff. This depends on the success of a naturalist theory of reference with respect to moral language. [Of course, it is very unclear how a causal theory of reference could be applicable to morality. It would depend on moral properties having causal powers, and this runs into precisely the sort of objections that antirealists like Mackie have given.]

Similarly, the moral realist can argue against Moore that good might be a natural property even if we can give no naturalist reduction of the meaning of good. As he mentioned previously scientific reductionism does not require the possibility of analytic reduction.

Another possible use of the new understanding of science concerns reflective equilibrium. In 2.2, we saw the accusation that reflective equilibrium is just a way of finding a compromise between our conflicting intuitions. But now we can see it as parallel to the interplay of observations, theory and methodology in science, which we are ready to accept constitutes the discovery of facts rather than the creation of fiction. The fact that moral belief is influenced by culture is then also not a decisive argument against its objectivity, since science is also influenced by culture, and yet remains objective in an important sense.

Finally, the fact that there are hard cases in ethics does not prove that ethics is subjective, because there are hard cases for classification in science, due to the indeterminacy of scientific concepts.

4.2 p. 120. For any of these rebuttals of moral antirealism to work, there must be constraints on moral realism, as there are constraints on scientific realism. The defense of scientific realism relied on the assumption that science is mostly true. Furthermore, the subject matter of science has to be able to allow nonconventional definitions (i.e., the development of a posteriori definitions and a causal theory of reference). Also the very project of scientific inquiry must be able to turn its attention to itself. Finally, indeterminacy in the extension of concepts is necessary and useful for scientific inquiry.

Note the technical language that Boyd uses needs a little explaining.

The "extension" of a concept is the class of objects that the concept refers to. The extension of "cow" is the class of all cows. It is sometimes contrasted with the "intension" of a concept, which is its meaning. The distinction between extension and intension is basically the same as that between denotation and connotation.

The indeterminacy of concepts implies the failure of bivalence. Bivalence means "two values." A concept is bivalent if all objects either do or do not fall under the concept. "Prime number" is a bivalent concept because all numbers are either prime or they are not. But is "cow" bivalent? Only if all animals are clearly either a cow or not a cow. I am told that cows cannot breed with any other animals, so, at least so far, "cow" has been bivalent. But it cannot be long before genetic engineering enables us to create a creature which has a mixture of genes from a cow and another animal. Imagine a cow is bred with a sheep. The resulting animal is called a "shcow". The female provide both milk and wool. Are shcows cows? There may be no simple yes or no to this question: it might be that bivalence would fail in this case.

Boyd takes these constraints and gives their counterparts for moral theory.

1. It must be possible to explain how the moral beliefs set subsequent moral thought in the right direction.

2. There must be some counterpart to scientific observation.

3. It must be possible to explain why moral properties require natural rather than conventional definitions.

4. Our use of moral terms must be guided to some extent by the reality to which they refer.

5. It must be possible to show how indeterminacy in the extension of moral terms as useful to moral inquiry.

p. 121. All Boyd is doing is trying to show that moral realism is plausible, not that it is true. He will do this by using one account of moral realism as an example.

4.3 p. 122. Boyd sets out his conception of morals.

1. Humans have needs. They do better when their needs are met.

2. Human needs are homeostatically clustered. Human needs are mutually supporting; i.e., getting one met generally helps to get others met. Our psychological traits in society generally help our needs get met. [Note that Boyd seems to be assuming that we don't have antisocial needs, such as naturally violent urges. He has said previously that his view of human nature is optimistic.]

3. Moral goodness is defined by this cluster of goods and the homeostatic mechanisms which unify them. I.e., something is good if it helps to meet our needs.

4. Concern for moral goodness helps to guide our action. This is done by meeting human needs and balancing the meeting of needs.

p. 123. Knowledge of the good is largely experimental. It is almost technical. Boyd seems to want to avoid a distinction between pure and applied ethics. He says our knowledge of the good has grown through political and social experiments. For example, social and technological developments have enabled us to learn more about our artistic needs and abilities. We have learned more about the nature of democracy, and this helped us see the wrongness of slavery. Knowledge in the social sciences helps us to understand the homeostasis unity of the good.

On this view, goodness is a property of actions, policies, and states of affairs. Moral facts are natural facts and good is a homeostatic cluster property. [Don't confuse the idea that moral facts are natural with the idea that moral categories are natural kinds. If good is a homeostatic cluster property, then good is not a natural kind in the sense that water is a natural kind, although it might be a natural kind in the sense that the concept of species is. Remember that in 3.8, p. 118, he said that species is a homeostasis cluster property.] Since Boyd holds a consequentialist theory in that the worth of something is measured in its ability to lead to good things.

4.4 p. 124. Boyd that ethics is based on observation in the same way that social science is. It uses a process of reflective equilibrium to determine what is right in similar ways to science. The observation of human needs and their fulfillment in ethics is reliable enough to defend moral realism, in the same way that scientific observation is reliable enough to defend scientific realism. We have come to understand humans better, although it is admittedly a difficult project.

p. 125. Intuitions are not assigned a foundational role in this theory. They do not substitute for observations. They are merely a manifestation of our moral understanding, similar to the intuitions of a scientist.

p. 126. Boyd says that, evolutionary speaking, it is not at all surprising that humans should be capable of knowing each other's needs, or the interaction of these needs. It is more surprising that we are able to do science, since that would have been of less evolutionary use.

4.5 p. 127. Boyd addresses objections 2.5 and 2.7 together. He says a moral realist can assimilate moral terms to naturalistic ally and nonreductively definable terms in the sciences. He needs to show that moral terms should be defined naturalistically, (a posteriori) rather than stipulatively (a priori), and that moral terms manage to refer to moral properties. Our use of the term 'good' should be partly regulated by reality, i.e., good things. And Boyd thinks it is, because all human use of the term 'good' does revolve around human goods and harms. This is not to say that many ideas of the good are seriously flawed, but they are close enough, just as many scientific theories of the past were seriously flawed, but they were good enough for us to have confidence that science was heading in the right direction.

4.6 p. 129. The fact that there is so much disagreement in ethics has been taken as a reason for moral antirealism. In particular, the antirealist conclusion considered by Boyd is that there is not objective subject matter about which all these different views are actually about. The different theories aren't even managing to disagree with each other. But Boyd says that we should expect much disagreement since doing ethics is like doing the controversial social sciences (e.g., sociology, anthropology, and these days, geography). We can agree that the different theories here are all about the same subject matter, human life, even though there is no way to find common ground between them. In cases where there really seems to be no way to rationally decide a moral question, the moral realist can say that there is no single answer to the question: it is genuinely indeterminate. It is interesting and unusual that Boyd's realism allows there to be some actions/state of affairs which simply cannot be assigned a determinate value of good or bad.

p. 130. Boyd says that on his view, we can explain why most moral theories agree on most cases. They are all about the same thing, and their use of the term 'good' is constrained by what is good, and the homeostatic character of the good.

4.7 p. 130. Finally, Boyd considers the antirealist argument that moral judgments should have a logical link to action, i.e., they should be intrinsically motivating, but a naturalist approach to morality which says that moral facts are natural facts will fail to motivate. Boyd agrees that there is no strict logical connection: natural facts do not necessarily motivate any being that perceives them.

p. 131. The standard naturalist realist response is to say that moral facts will motivate psychologically normal people. Boyd agrees with this but thinks it does not go far enough. He wants to explain in what way people who are not motivated by moral facts are not normal. He says it is not just a defect of their will or motivation: it must be a defect in their cognition. There is a logical connection between the judgment and motivation in that anyone who really understands the moral facts will be motivated by them. The cognitive deficit is like a perceptual deficit. It is the inability to imagine themselves in the situation of others (voluntarily or even involuntarily). This is the faculty of sympathy, and it is motivationally important. [Boyd seems to be assuming that even the motivational part of sympathy can be classed as a cognitive capacity.] Anyone who lacks the capacity to sympathize with others is cognitively deficient. They may still be able to act morally, using some other route for motivation of their action.

4.8 p. 132. Boyd concludes that acceptance of scientific and moral realism may go hand in hand. He also points out that his arguments depend on an optimistic view of human nature and the unity of human goods (e.g. the welfare of individuals and that of society are mutually reinforcing). If such a view is wrong, and people are not naturally sympathetic to each other or concerned for the good of society, this will be an argument for ethical antirealism. It should therefore be part of philosophy to study human nature.

Peter Railton, "Moral Realism"

Unfortunately Railton has an especially bad tendency to lapse into technical jargon, even when it is not helpful. This makes his writing especially hard to comprehend the first time through. His general strategy in arguing for moral realism is to say that we can best explain some people's behavior by referring to moral reality, i.e., what is actually good or bad. The main opponent he faces is someone who replies that we never need to refer to moral reality in explaining people's behavior, and that we can always explain behavior just as well by reference to people's psychological states and non-moral facts about their environment. Using Ockham's razor, Railton's opponent will argue that there is no reason to think that moral reality exists because it adds nothing to our explanations.

p. 138. Railton will argue for a moral realism that has clear relevance to social and psychological theory. He is not trying to prove beyond all doubt that moral realism is true, but just defend it from criticism.

I. He will argue for a moral realism in which moral judgments are objective although relational, and have a truth value (cognitivism) except when indeterminate. We have moral knowledge but many of our moral beliefs are wrong. Some people may have a reason not to be moral. Moral inquiry is similar to empirical inquiry. Moral properties supervene on (are intimately tied with) natural properties, and may even be reducible (identical) to them.

II. p. 139. Moral realists cannot accept a distinction between facts and values. If moral judgments could not be factual judgments, then there would be no moral facts.

Railton considers an epistemological argument for the fact/value distinction. We can do all the empirical investigation of a situation we want, but we will still leave open the question of what we should do. For instance, I can be watching one person going up to another to hit her with a baseball bat. I can learn what each person has done in the past, what their relationship is, and so on. None of this will by itself determine whether I should intervene or not. Two people could be in possession of the same facts and reasonably come to different conclusions about what to do. So value judgments about what is best to do cannot be factual judgments, it is argued.

Railton counters this argument by saying that two people could come to different conclusions about the scientific facts given the same total evidence. The rationality of belief-formation does not depend just on the evidence, but also on the desires of the investigator.

Note that when Railton talks about instrumental reason, he is using the idea of reason being the slave of our desires. It is the view that there is no way to reason about our ends, only the means by which we should achieve those ends.

p. 140. Another argument for the fact value distinction.

Moral facts would, if they existed, provide a reason for people to act in a certain way whatever their desires. But scientific and logical facts are not like that. So moral facts cannot be of the same kind as scientific or logical facts. [Does this succeed in showing that there cannot be moral facts, or only that moral facts are different from other kinds of facts?]

Railton says that this antirealist argument depends on the assumption that morality is essentially practical. Railton denies that there is an essential connection between the values embraced by an agent and her reasons for action. He proves this with an example.

The sensible Knave. This is also what is known as the free rider. Hume's argument for justice basically says that it is better for society if we live according the principles of ethics or justice, and so we should do so. The sensible Knave understands this and agrees that society is better off when people live ethically. But he argues that there are cases where his particular actions, although breaking promises, cheating, stealing, and so on, will have a negligible effect on society, certainly not enough for it to rebound onto him. He doesn't care about anyone but himself, and he doesn't care about his own honor, so he argues that he has no reason to act ethically. Nevertheless, he still accepts that the things he is doing are wrong. The wrongness of the actions need not motivate him to act otherwise.

p. 141. Although Railton has said that scientific belief formation can and should depend on our desires, he denies that he has opened the door to epistemic relativism, which is, very roughly, the view that we can believe whatever we like. He says that the evidence for a belief is independent of our desires.

III. Railton says that not only are there moral facts, but they are part of the world in the same way that scientific facts are. They are not in a corner on their own, sui generis. We can learn moral facts through the same mechanisms as we learn other facts.

p. 142. A moral realist can argue for moral facts by saying that they explain our experience. For moral facts to be capable of such explanation, they must be independent of our beliefs in them, and we must be able to interact with moral reality. We may not be able to directly see moral facts, but they can still influence our experience.

Railton will give an account of the nature of moral facts. But he first addresses non-moral value.

Something is sweet to me if it normally takes sweet to me. The sweetness of sugar depends on its primary qualities (chemical composition), on my sensory system, and the surrounding environment. He calls this set of relational , dispositional, primary qualities the reduction basis of the secondary quality, sweetness in this case.

Remember the distinction between primary and secondary qualities: primary qualities are those fundamental properties that are intrinsic to an object and independent of any observer, such as mass, size, shape, and chemical composition. Secondary qualities are those properties that do depend on observers, such as color, felt temperature, texture, and beauty.

The objectified subjective interest for an individual A can be defined as follows. A+ is what A would be with full rationality and information. We ask A+ what A should want. This is A's objectified subjective interest (OSI). There will be a reduction basis for this OSI. Basically an OSI is what would be best for a person, from her own point of view, and this depends on a reduction basis, which is the primary qualities of her constitution and her circumstances. This gives us a way to define what the objective interest of a person is.

p. 143. X is non-morally good for A if and only if S would satisfy an objective interest of A. This basically captures what we mean by non-morally good.

p. 144. X is intrinsically non-morally good for A if and only if X is in A's objective interest without reference to any other objective interest of A. I.e., it is not just a means to another good.

This notion of non-moral goodness can have explanatory powers. It explains why people who act in their own interest do better than people who do not. (Merely referring to what people believe to be in their best interest will not have the same explanatory power.)

p. 145. Our desires need to evolve to fit out best interests in order for us to do well. But we can also fail to know what is in our best interest, and we can want things that are not good for us.

p. 146. This conception of objective value can explain why we are generally the best judges of our own interests; why knowledge of our own interests will increase with our general knowledge; why similar people will have similar values; and why there will be most agreement about values in the aspects of life where people are most similar, such as the level of basic motives.

This objective value is dependent on humans in the sense that it would not exist if humans didn't. But being relational does not make a fact subjective. Objective interests depend on natural and social facts.

IV. p. 147. Railton now turns to directly defending moral realism, and sets himself the hard task of defending the existence of moral norms, as opposed to the more plausible idea of moral value. (He is supposing that it is plausible to suppose that some objects, such as humans, have intrinsic moral value.)

The puzzle he is trying to solve is how moral norms can do any explaining of our experience. He notes that we often use "ought" language in everyday explanation, as in "the car crashed because the driver had been drinking too much. He ought not to have drunk a pint of bourbon." [This seems a weak example of "ought" explaining. The ought clause is tacked onto the end of the explanation, and doesn't add much. The same explanation could be given by saying "the car crashed because the driver had drunk half a pint of bourbon."] He calls this a criterial explanation, explaining why something happened by showing that some important criterion was met or exceeded.

p. 148. We explain behavior also by reference to someone's beliefs and desires or to account for the way an individual's beliefs change with experience. Some people are more rational than others. Some people are self-defeating. Some of our rationality is not in our conscious thought, but in the habits of thought or behavior that we unintentionally acquire.

p. 149. We also explain behavior by reference to her objective interests. If a person becomes extra sure about the rationality of her action, she will look at it in an even more positive light. So the rationality of her behavior can explain her subsequent behavior. [But here, isn't it the belief in the rationality of her behavior that explains her subsequent behavior. Would the explanation be just as good if her belief was false?] Railton says that reflection on the explanation of behavior shows that people have reasons for behavior independent of their beliefs about those reasons.

p. 150. "non-indexical reasons" are ones that are general, as opposed to indexical reasons, which apply only in the particular case they are operative. Railton says that moral norms and reasons should be impartial in the sense of not applying to any particular individual, but rather being from a social point of view. He is saying that morality is part of rationality, so immorality is irrational.

Railton suggests that the rightness of an action is given by taking into account, equally, the objective interests of all people affected by that action. He takes this as his definition, and now considers whether this could play a useful part in explaining people's behavior. He says that a group of people may grow dissatisfied with their social situation if their objective needs are not being met. The behavior of the group will be explained by the wrongness of the society. He claims that the explanation will not (always) work just by referring to the group's belief in the wrongness of the society. For example, a society in which all citizens believe that they are being treated justly may still have some groups being discriminated against, which will make the society unstable because those groups will feel unhappy or alienated.

p. 151. A society that does not meet the needs of its citizens equally, i.e., one which fails to be good, will likely experience some feedback , and may well learn better norms through the experience of the discontent of some of its citizens. For example, this happened through the civil rights movement of the 1960s in the US.

p. 152. Railton is not claiming that all societies will always improve in their goodness. Obviously there have been periods of time where societies have got worse. But the parallel phenomenon is true in science too, so this cannot be a good argument against moral realism. We acknowledge that we can learn from our mistakes in science, and the same is true for morality.

p. 153. Learning from feedback is more difficult in the case of society than it is for an individual person. However, Railton thinks that it will lead to certain historical trends. He lists three.

Generality. A society will, as it evolves, come to grant personhood even to people different from itself.

p. 154. Humanization. Moral principles come to be seen as deriving not so much from God or the supernatural, but from human interests.

Patterns of Verification. There will be more agreement about morality in areas where the feedback mechanisms work best, such as when almost everyone has importantly similar or mutually satisfiable interests. For example, this is seen in the agreement about the prohibition of aggression and theft, and of the violation of promises. There is more disagreement where the feedback mechanism works less well, such as slavery and gender inequalities. Railton acknowledges that these predictions are very sketchy and simple-minded.

V. p. 155. This is not a transcendental account of morality: it is firmly grounded in human needs and psychology. Furthermore, it does not provide categorical imperatives. But it is not relativist either, since rational motivation is not a precondition of moral obligation. One can have reason to be moral even if it would not help one personally, because what one should do is dependent on what is rational from a point of view that includes, but is not exhausted by my own. Similarly, one might have a logical reason to make one's beliefs self-consistent, even though doing so would not benefit one because it would take so much careful thought (i.e., it would not be instrumentally rational for one to make one's belief consistent).

p. 156. This does not rob morality of its authority over people. There can still be moral imperatives for people who have no instrumental reason to follow them. Variations in personal desires, and in particular having knavish desires, cannot license exemption from moral obligation. It does suggest that we should change the ways we live so that moral conduct would be more regularly instrumentally rational given our actual ends. [This sounds somewhat similar to Bernard Williams' comments on the importance of moral education in "Internal Reasons".]

VI. p. 157. What is a good moral theory? One that meets with most if not all of our linguistic or moral intuition concerning goodness and rightness. It should also permit plausible connections between what is good and what will characteristically motivate individuals who are rational. It should be usable in conjunction with our empirical theories, so that we can know what is good and right, and so it can have explanatory power. Railton says that his theory meets these criteria. Other theories do not meet these criteria as well.

Simon Blackburn: How to be an Ethical Antirealist


Blackburn defends a form of "non-cognitivism." I.e., he thinks that calling something good is not so much a way of describing it as expressing one's approval of it, without attributing any special properties to it.

p. 168. "projectivism": the view that although we talk as if ethical properties were really part of the world, it is we who color our vision of the world to make it look as if it contained ethical properties.

Blackburn's position is "quasi-realism." It is a view which explains why projectivism is right.

I. p. 169. Blackburn shows some skepticism about the useful of the realist/antirealist distinction. Some of his comments are reminiscent of what has been called the "redundancy theory of truth." A theory of truth is meant to explain what it is for a sentence to be true. The redundancy theory says there is no difference between saying "P" and "It is true that P." I.e., the "It is true" part is redundant. So he tends to think that someone saying that she is a realist about ethics is not saying anything more than she thinks ethics is important, (or something like that). On this view, being a realist about ethics does not even manage to be about the fundamental nature of the world: it is simply an endorsement of doing good. This is (part of) what Blackburn means when he says that the realist commitment is not a belief.

On this view, to say that "X is good" is not to say something about the world in itself, but rather to say something about one's attitude towards how we should behave. Blackburn's view is a form of emotivism. We can explain our moral attitudes with evolutionary theory, without ever supposing that values-as-part-of-the-world play a part in evolutionary biology.

p. 170. Blackburn criticizes the theory of McDowell and Wiggins (which we will discuss later in the semester) that moral properties are (like) secondary qualities such as color, and we can see them. Of course, color is in some ways dependent on humans, and so on their theory morality is dependent on humans. But Blackburn does not think that this makes sense. He does not think it makes sense to talk as if morals are properties of objects or actions in the same say that colors are properties of surfaces.

p. 172. Blackburn goes on to say that the secondary quality view of morality does not explain why cruelty is wrong. But his own projectivist theory can explain why cruelty is wrong. Cruelty is not wrong because we tend to think it is wrong. On his view cruelty is not some separate entity whose moral properties (wrongness) need explaining.

p. 173. Blackburn rejects the metaphysical question "what makes cruelty wrong?" He writes, "Talk of dependence is moral talk or nothing." Talk of ethics is just discussing or voicing ethical opinion. Again, one can't even succeed in making a metaphysical claim about ethics, it seems.

This view is strongly influenced by Wittgenstein. He argued that much of our use of language is not aimed at describing the world, but is rather a form of action, a way of living in the world. He thought that philosophy created philosophical problems which didn't really exist, by supposing that our language is describing an independent reality, when that is not what language is doing. (Wittgenstein held this not just about ethics, but also the mind and mathematics.) When we explain why something is bad or good, we don't do so though metaphysics, but rather through explanation of the concept. If someone understands the concept of friendship, she will understand that it is good. We won't be able to explain ethics to people who can't understand it.

p. 174. The question which some think crucial of whether the correctness of norm depends on the existence of human activities is one which cannot even be coherently posed, according to Blackburn. The realist/antirealist debate is based on a misconception. Of course, projectivism is closer in spirit to antirealism, since it denies the existence of a metaphysical morality.

p. 175. Can projectivism explain the depth of our moral commitment? No, says Blackburn. A lover's passion does not decrease on the realization that it is only his passion, and grief does not diminish when we remember that it is we who are grieving. In these cases, we are generally ready to acknowledge that our emotions stem from ourselves. Our moral sensibility, that by which we project morality onto the world, so to speak, does not figure in the picture it paints, any more than an eye figures in our visual experience, except when we look in mirror. So being a projectivist need not diminish one's moral life.

p. 176. But doesn't this view lead to relativism? Doesn't the wrongness of cruelty depends on one's point of view? Blackburn says no, because this is to make the same metaphysical error that he argued against before. It is true that our feelings about fashion will change with time in a way that those about ethics should not, but it is a defining feature of ethics that it is relatively invariable.

II. p. 177. Blackburn considers the viability of projectivism with respect to other things whose existence has been disputed by philosophers, such as color and modality.

 Alan Gibbard: Wise Choices, Apt Feelings
Gibbard's style of philosophizing might be described as careful. Others would call it pedantic. He certainly has a fondness for definitions and complicated distinctions. The view for which he argues is, like Blackburn's, a form of non-cognitivism. He thinks that moral judgments express a state of mind that is not a belief in a moral fact, but neither is it a simple feeling. He says that a moral judgment is a complex state of mind that consists in accepting certain norms. His task, as an antirealist, is to give an account of moral norms that does not rely on any objective moral theory.

1. p. 179. Gibbard is interested in our discussions of how to live, and what makes some ways of life better than others.

2. p. 180. On a narrow definition of "morality," it is not the same as rationality. But on a broad definition of morality, it is the same thing as rationality.

p. 181. Gibbard will use the narrow sense of morality, where one can do something which is irrational but not immoral. He defines it as follows:

An agent is to blame for X iff it is rational for her to feel guilty for doing X, and for others to resent her for doing it.

An action is wrong iff it violates standards for ruling out actions, such that if an agent in a normal frame of mind violated those standards because he was not substantially motivated to conform to them, he would be to blame.

The point of this convoluted definition is to acknowledge that there can be extenuating circumstances. Note that Gibbard is reversing what we normally think as the order of logical priority between wrongness and guilt. We normally think a person should feel guilty when what she has done is wrong. But Gibbard is saying that an action is wrong when a person should feel guilty about it.

3. p. 182. What is it to take something to be rational? Gibbard suggests it is to accept norms that, on balance, permit it. A norm is a rule or a prescription. Moral norms are norms for the rationality of guilt and resentment.

p. 183. We still need some account of responsibility. When is an agent not responsible for what she has done? An act can be wrong even if the agent is not to blame for it. This is so when the agent, due to her psychological condition, is not responsible for her action, but the action itself, if it had been performed by the agent when normal, would be blameworthy.

4. What is it to accept a norm? This is not well understood. Gibbard will try to point to the kind of psychological state involved.

p. 184. Sometimes we accept norms but do not act on them, such as in cases of weak-willed behavior. If one can't stop doing something, such as eating nuts at a party, then this is a clash between one's normative motivation and one's animal control system. Normative motivation seems particularly linked to language, a motivation that evolved because of the advantages of coordination and planning through language.

5. p. 185. Some motivations are social, such as embarrassment, or not wanting to displease a person in authority. A person can experience a conflict between two norms.

p. 186. Gibbard draws a distinction between accepting a norm and being in the grip of a norm. Being rational involves acceptance of a norm rather than being gripped by one. [Are there no counterexamples to this claim?]

6. Acceptance of norms is a natural biological phenomenon. The function of norms is to coordinate our social behavior. Coordinated expectations are essential to a social life, both in cooperation and even in some hostile behavior.

p. 188. Evolutionary selection should have led humans to adopt stable strategies. Gibbard suggests that our emotional propensities and normative capacities are largely the result of selection pressures to coordinate human behavior.

p. 189. For example, feelings of guilt can enable acknowledgment of wrongdoing and reconciliation. This reduces the chances of damaging conflict.

7. If one accepts a norm, then one is not in its grip. To be in the grip of a norm, one has to not accept it. But apart from that, there are likely to be great similarities between accepting a norm and being in its grip. When one accepts a norm in a similar way to being in its grip, Gibbard says one has internalized it.

p. 190. We have many behavioral patterns without explicit rules (similar to animals). To internalize a norm is to have a motivational tendency of a particular kind to act on the pattern of behavior prescribed by that norm. Animals can internalize norms.

8. p. 191. Language allows us to communicate about what is not directly in front of us, and this greatly opens the scope for coordination. Shared evaluation also helps in a complex social life. Normative discussion might coordinate acts and feelings if (a) it tends towards consensus, and (b) the consensus moves people to do or feel accordingly.

p. 192. People in society engage in normative discussion and try to influence each other through considerations such as logical consistency. Gibbard suggests that to accept a norm is, in part, to be disposed to avow it in unconstrained normative discussion, as a result of the workings of demands for consistency in the positions one takes in normative discussion. [Doesn't this imply that the only people who really accept any norms are academics in the humanities who run graduate student seminars? Even not many of them would qualify.] Engaging in this sort of discussion requires a strong imaginative life.

9. p. 193. Why would having norms be useful in terms of natural selection? Using them could provide social coordination. But there are other cases where it would be useful to say one thing and do another, thus confusing others. It may not always be best for personal survival for a person to act according to her norms.

p. 194. How would norms lead to consensus in a society? People must be persuadable in discussion. Someone who was too stubborn would risk ostracism. But people must also be firm, because if they were not, they would be easily manipulated by others. So they need a balance between flexibility and rigidity. The best norms for a person to avow will be ones that can attract others yet which will also help her. Some people will have a great deal of integrity, others will be self-deceiving, and others will be cunning.

10. p. 195. Gibbard calls his view the Norm-Expressivistic Analysis:

Calling something rational is expressing a normative judgment, which is a state of mind.

To judge something rational is to accept a system of norms than on balance permits it.

11. p. 196. The meaning of a normative statement is given by what it rules out.

p. 197. A person cannot have consciously in mind everything that a norm rules out (an infinite number of things). To understand someone's meaning we can use two constraints to narrow down the possibilities:

i) the inferences she makes with confidence are to be explained by relations of logical entailment among the pieces of normative content we attribute to her.

ii) her propensities toward normative governance (what she is motivated to do right now) are to match the content we attribute to her normative judgments concerning herself.

 John McDowell: "Values and Secondary Qualities"

I find McDowell's prose style dense and convoluted. It takes persistence to understand his ideas. McDowell argues that values are analogous to secondary qualities, enough so for us to acknowledge their reality, as colors are mind-dependent yet real. In order to understand this and evaluate the claim, we will need to spend some time on our knowledge of secondary qualities. This involves metaphysics and the philosophy of mind. He argument works, at a fundamental level, by reconceiving how we relate to the world and understand it, and thus what ways in which we can legitimately talk about reality. His point is that we can legitimately talk about reality in way which includes values.

1. p. 201. McDowell sets out his own view by discussing that of Mackie. He says make that ordinary evaluative thought has a phenomenology of being a sensitivity to aspects of the world. He is thinking especially of §7: The claim to objectivity of Chapter One of Ethics: Inventing Right and Wrong. McDowell thinks it is important to pay attention to the "lived character of evaluative thought or discourse." [He is of course making the questionable assumption that there is some common factor in all such experience for all people.] The idea is that the world, when we consider it evaluatively, presents itself to us as morally loaded. It is not, as non-cognitivists would suggest, that we look at the neutral world and then experience internal feelings or reactions to the world. Actions and events in the world present themselves to us as being good or bad in themselves. Mackie thinks that this experience of the world is an illusion. However, McDowell wants to say that our moral experience can indeed provide us with moral knowledge.

2. If the world presents itself to us as morally loaded, then we need to give an account of how we get this impression. If the experience is correct, then it would seem that we have some kind of perception of the moral. Mackie said that we know that none of the primary qualities of objects are moral, and he assumed that such moral perception would have to be of primary qualities.

p. 202. Such a moral sense would be very mysterious, and it would hard to prove that it gave us knowledge. But why couldn't moral perception be awareness of secondary qualities. Mackie would say that this is of no help to the moral cognitivist, because our knowledge of secondary qualities is illusory as well. There are not really any colors in the world; we just think there are because our minds make it looks as if the world is colored. We project color onto the world, but science tells us that colors are not really intrinsic properties of physical objects. McDowell thinks Mackie is mistaken about our knowledge of secondary qualities.

3. Red is a secondary quality. An object is red if it looks red under certain circumstances. Red objects may not look red under certain sorts of light, or in the dark. "Red" does not mean having certain microscopic properties, although it may be that an object is red in virtue of having such properties. Redness refers to how an object looks to us. Red objects look as if they genuinely possess the property of being red. When we look at a tomato, we do not normally think of the experience as if we are looking at a colorless object that is causing us to have red experiences. We think of the tomato as red, independently of our experience of it. McDowell wants to say our normal way of thinking is correct here: a tomato really is red, it is not just a matter of looking red.

p. 203. Mackie thinks that tomatoes are not really red, although they really are round. He takes from Locke the idea that our experience of the roundness of the tomato resembles the actual roundness of the tomato (roundness is a primary quality that can be described by science), but our experience of the redness does not resemble any actual redness of the tomato (because science tells us that the tomato really has no color). McDowell suggests that the idea that our experience of redness could resemble the actual redness of the tomato is incoherent. (The argument will be that Mackie is wrong in saying that our experience of redness fails to resemble the actual redness of the tomato, because we can't even make sense of what this would be. It is not like saying that a portrait fails to resemble the actual person depicted.) To be coherent, it would require

i) that we could conceive of color as we can conceive of shape, as making sense independently of how things look. Objects have a certain shape which is independent of how they look to people. Colors are not like that.

ii) that we could make sense of the idea of a primary quality of an object resembling our experience of it, e.g., the redness of the tomato resembling our experience of its redness.

McDowell does not think that (ii) is possible. It would require us to conceive of redness independently of how red things look. We have no idea how else we might conceive of redness.

McDowell agrees that secondary qualities are subjective in the sense that we cannot understand the concept of them without reference to our own subjective experience. But he says that they are not subjective in these sense that they are figments of our experience.

p. 204. We cannot think of the relation between our experience and the world as parallel between that of a picture and what it depicts. In understanding our experience, we need to focus on its content, "intentional object". Then, if an experience is veridical, the intentional object of the experience does not resemble the thing in the world; it simply is the thing in the world. The relation is identity. So if I have an experience of a red tomato, and the experience is veridical, then the content of my experience is the thing in the world. This is part of a more general view that we actually manage to think about things in the world, and not just our ideas of them. If the content of an experience could only be an idea, then we could never think about anything except ideas (as Berkeley and later phenomenologists thought we are). This is another Wittgensteinian theme relating to the private language argument: our thoughts do not get their content from their intrinsic mental features.

We could give up the primary/secondary quality distinction, and say that there is no epistemological difference between colors and shapes. But then we would have no idea how the scientific nature of objects related to our experience of the world. The perceived shape of an object would be no related to its actual shape than the perceived color of an object would be related to its surface texture. McDowell does not want to give up distinguishing between primary and secondary qualities, but he cannot distinguish between then on the basis of the resemblance (or lack of it) between a quality and our idea of that quality.

McDowell wants to keep hold of the view that primary qualities are distinctive in being both objective and perceptible. In order to do this, he says we need to give up a view about what it is for a quality to figure in experience. It cannot be for the experience to have a certain intrinsic feature. Colors and shapes figure in experience simply as properties that objects are represented as having, distinctively phenomenal in the one case, and not so in the other. Then colors and shapes are equally real, although colors are intrinsically phenomenal qualities.

p. 205. McDowell says that his position does not have to reject the idea that an experience intrinsic features, but that there are good reasons to reject the idea anyway. He seems to be saying that the only features of an experience are its representational features. [This view is hardest to defend for experiences that do not seem to contain representations of anything, such as pains, itches, and the colors you see when you press your eyeballs.]

McDowell summarizes his argument so far. Mackie divided our understanding of the world into what might be called the "Scientific Image" and the "Manifest Image." Mackie thought that in fact only the Scientific Image is true, and the Manifest Image contains a great deal of illusion. Objects and events in the world appear to us to be colored, beautiful, disgusting, bad or praiseworthy. These qualities are part of our Manifest Image. McDowell argues that this picture of our relation to reality makes it hard to explain how we have any subjective or phenomenal knowledge of scientific properties. On McDowell's view, there is an essential subjectivity to our experience, but at the same time, our experience is not (intrinsically) misleading or inaccurate.

4. p. 206. Mackie defended his view that there is no good reason to think that objective reality contains colors by arguing that we can explain all our experience without supposing that colors objectively exist. This was generalized to skepticism about any objective properties which resemble our ideas of secondary qualities. McDowell has already argued that it is not a contingent fact about the world that there are no objective properties which resemble our ideas of secondary qualities, but that the very idea is incoherent.

This leaves the question for McDowell whether irreducibly subjective features can be a part of the world. His argument is not that we need these features to explain our experience, but that someone explaining our experience cannot consistently deny the existence of irreducible subjectivity. He concedes that it would be implausible to suggest that values have a causal influence on the world. Furthermore, there is a crucial disanalogy between values and secondary qualities. We see something as red if it is disposed to look red to us. We see something as good if it tends to look virtuous to us. But there is more to virtue than that: A virtue does not just cause our approval of it: it deserves our approval.

p. 207. McDowell elaborates this point with the example of danger. We fear something if it is dangerous. Our fear is not just caused by the dangerous thing, it is justified by it. The explanation of our color experience is different from our explanation of our fear. We want to make sense of the fear. The best way of doing this is to say that the object feared is objectively (in the sense of really, not as an opposite to subjectively) dangerous (or fearful, as McDowell says).

Similarly, we can best explain some of our other beliefs and behavior by saying that some things really have value. If we try to do without values in our explanations of our responses to the world, our explanations will be less intelligible. [Why? Presumably a lot of weight rests on the analogy with secondary qualities such as color. It is less easy to make sense of our experience if we suppose the objects in the world have no color. We can talk about what color an object really is, and of people being more or less able to tell what color it is. Similarly, it is harder to explain our behavior if we don't allow that some things are really dangerous, and of some people's ability to discern danger.] We try to understand ourselves and improve our behavior, and these two projects are connected. We improve ourselves through better understanding of ourselves and the world. It makes sense to suppose that some of our responses to the world are reasonable. Just as we can talk of the real color of an object, we can talk of the real value of an action, and of the variation in different people's ability to understand that value. There will of course be disagreement about values, but that does not invalidate the argument. In a footnote, McDowell says that the contentiousness of values may be ineliminable.

5. p. 208. McDowell is ready to concede that values depend on us as much as colors do, but insists that this does not impugn their reality. But colors, danger, and values all help to explain our experience and responses to the world. They are mind-dependent but still real.

McDowell considers how his view differs from projectivism, which says that values are not real, but that we can explain our behavior because we think that they are real.

p. 209. He says that projectivism relies on a thin conception of reality, with no justification for doing so. Furthermore, the justification of value-responses depends on their functionality. Whether something is functional depends on a value judgment, and according to the projectivist, this is a projection of internal approval. The projectivist mechanism must contemplate itself. This tends towards a systematic theoretical approach to value. But McDowell thinks that some cases will forever remain contentious, not capturable in a theory. His approach to value allows for more of a patchwork of values.

 John McDowell: "Projection and Truth in Ethics"
As we saw at the end of "Values and Secondary Qualities," it is important for McDowell to distinguish his view from projectivism, for it may seem to some that there is a merely verbal difference between them. Projectivism says that our moral judgments are not objectively true of the world independent of us, but are rather projections of our feelings about the world. McDowell's sensibility theory says that moral judgments are objectively true in the sense of being real, but are not independent of us. Rather, they are essentially subjective, in being from our point of view. Is there a significant difference between these views? Neither is recommending that we abandon our ethical feelings or our ethical point of view. The difference between the two seems metaphysical, and understanding it requires inquiring into the nature of ethical truth.

1. p. 215. McDowell wants to defend the idea that our value judgments, or "valuations," have truth-conditions, i.e. are capable of being true or false. Projectivism about some of our value judgments is plausible. For instance, our belief that something is disgusting can be explained in purely psychological terms, and we do not need to suppose that the disgusting object has some intrinsic property of being disgusting. If our judgment that "X is disgusting" is interpreted as meaning "X has the property of being intrinsically disgusting," then it is false. The projectivist says that we should better understand the judgment "X is disgusting" as an expression of our feelings of disgust. It is a sophisticated way of saying "yucky," or vomiting, when presented with X.

p. 216. McDowell describes Blackburn's form of projectivism, which is allows that ethical statements can be true in a sense. Blackburn calls his position "quasi-realism," and he wants to say that our moral feelings are projected onto the world in a sufficiently complex way that we can justify the steps in our ordinary moral reasoning, or as McDowell expresses it, the projection "can be sufficiently robust to underwrite the presence of the trappings of realism."

2. Blackburn's quasi-realist theory says that, strictly speaking, our ethical commitments do not have truth-conditions. Rather they express attitudes or sentiments. Saying that "X is good" is a sophisticated form of clapping in the presence of X. But this approach also needs to explain why it at least seems plausible that value judgments have truth-conditions. How can it do this without contradicting itself?

p. 217. Blackburn says that we can turn our ethical attitudes towards themselves, and see whether we approve of our values. A value judgment seems true when we approve of it, and false when we don't. This is how our ethical projections earn the right to use the notion of truth. Blackburn seems to think that we should approve of our ethical valuations when they help us to achieve social order and co-operation.

3. p. 218. A substantial notion of ethical truth requires "a conception of better and worse ways to think about ethical questions," and the justifiability of ethical opinions. Ethical truth should not be arbitrary or a matter of convention. We need to be able to distinguish between rational and irrational ethical thought. This is what it is for our ethical talk to "earn the right to the notion of truth." McDowell suggests, in opposition to Blackburn, that this right is independent of metaphysics and does not lead us to projectivism.

4. Some might think that projectivism is the only alternative to intuitionistic realism, which says that we some how cognize valuations facts that are metaphysically independent of us. But McDowell says this would be a mistake. He argues for this through asking what justifies our value judgments. Recall that he allowed that projectivism could work for the case of disgust. That is to say, we explain our judgment that "X is disgusting" by saying that we project our feelings of disgust onto X. We can understand the concept of disgust without supposing that X is intrinsically disgusting.

p. 219. Consider the case of judging something to be comical. This looks like it would be plausible to give a projectivist account of the judgment. But what feeling is it that is projected onto the world? It can't simply be laughter, or an inclination to laugh, because that would not distinguish the comical judgment from the judgment that the thing is embarrassing. Maybe the projected feeling is amusement. But what is amusement apart from finding something funny? If this is so, and suppose it is for the sake of argument, then how do we explain why we find the thing funny? The projectivist answer seems to be that we just do. This is no explanation at all. McDowell's point is that we can't even understand the concept of amusement without thinking of it as a reaction to the object of our amusement. That is to say, we can't understand the concept of the internal "feeling" without referring to something outside the person. Or to put it another way, "amusement" is not a feeling in the same way that disgust is a feeling. Presumably the feeling of disgust can be described in non-circular ways without referring to something outside the person. One should be able to explain disgust fully to someone who has never experienced disgusting things. (If one can't, so much the worse for projectivism about disgust.)

McDowell says we can avoid this problem if we allow for another option besides projectivism, which says that our reactions to the world explain our judgments about it, and intuitionistic realism, which says that the intrinsic nature of the world explains our judgments about it. He calls this option the no-priority view. On this view, we explain our judgments by reference to other similar judgments. Such explanations are circular, it is true, but that does not necessarily mean that they are uninformative. (This is a point that Wiggins addresses in "A Simple Subjectivism.") He proceeds to sketch this view further.

p. 220. On McDowell's view, we explain our comical judgment by referring to the comical properties of the object, without supposing that we can describe those properties in mind-independent ways. It is hard to see how we could talk about something being really funny, and the comic judgment being true, if there were no way to rank different people's judgments of funniness in terms of their justification. If we can rank different judgments of funniness, then we can explain this by saying it is possible to understand what it is for something to be really funny, and different judgments exhibit greater or less of this understanding. (Remember that McDowell is not committed to comic judgments actually being capable of truth-value. He is just supposing that they are for the sake of argument.)

p. 221. On a no-priority view of ethics, which is what McDowell is really set on defending, says that moral sentiment and moral properties are mutually dependent. He would says that we should be able to rank ethical judgments according to their rationality. We adopt an ethical point of view and rank ethical judgments according to it. We may or may not achieve satisfactory justifications of our judgments within this point of view, but it is at least possible that we might.

5. p. 222. McDowell ends by criticizing an extended quotation of Blackburn's writing. He points out that it begs the question to assume, until proven otherwise, that values are not part of the world. He also suggests that there are some things that a scientific point of view cannot well explain, and so we should take seriously any other point of view, such as the comical or the ethical, which helps us to explain our lives.

p. 223. Ethics does not have to be subjective in the sense of being illusion. Nor does it have to be objective in the sense of scientific truth. McDowell thinks that subjective and objective can form an "interlocking complex" so that reality can contain ethical truth.


John Rawls: "Kantian Constructivism in Moral Theory"

Rawls' most important work is A Theory of Justice (1971). In that book, Rawls defends a liberal theory of justice, which focuses on individual freedom and the responsibility of society to help those who need it. Such a theory can be used to justify taxation for welfare programs, universal education and health care, and the right of individuals to be free from government control of their personal lives. It also gives us a justification for condemning racism and sexism. These lectures by Rawls do not, however, focus on the practical implications of his theory, but instead on the fundamental principles by which such a theory can be justified. The basic idea is that we decide on the principles of justice by considering a hypothetical social contract. He derives the following principles of justice which are to govern the basic structures of society:

(1) each individual is to have a right to the greatest equal liberty compatible with a like liberty for all (this is called the principle of equal liberty);

(2) (a) social and economic inequalities are to be attached to offices and positions open to all under conditions of fair equality of opportunity; and

(2) (b) such inequalities are justified only if they benefit the worst off (this is called the difference principle).

The first principle has priority over the second.

(This is adapted from Tom Nagel's entry on John Rawls in The Oxford Companion to Philosophy, edited by Ted Honderich, Oxford University Press, 1995)

Lecture I p. 247. Rawls is building on and explaining further his ideas in A Theory of Justice. He is not attempting to prove competing views are wrong, although he does hope to show that his view is simpler, and so more attractive.

I. Kantian constructivism "specifies a particular conception of the person as an element in a reasonable procedure of construction, the outcome of which determines the content of the first principles of justice." It is basic to this theory that these principles are what rational agreement would agree to.

p. 248. There has to be public agreement about what is sufficient justification of a theory of justice, that is justifiable to all citizens. This serves as an important constraint on what theory of justice we can adopt. It is assumed that people are free, equal, and capable of reasonable and rational action. This conception of persons must be acceptable to all citizens, once sufficiently explained. This process does not so much discover prior and independent moral facts as construct a theory of justice that all can live with.

II. p. 249. Justice as fairness tries to uncover the fundamental conception of freedom and equality that are already implicit in common sense. Rawls calls these "model-conceptions." The two basic ones are a well-ordered society and a moral person. Then there is the original position. This models how citizens in a well-ordered society would ideally select their first principles of justice. These citizens, in this idealization, are assumed to have rational autonomy. All citizens will know, understand and accept the principles of justice for their society. They regard each other as persons who are free, in that they are entitled to make claims on each other, equal, in that they have an equal right to determine and assess the first principles of justice, and moral, in that they are capable of having a conception of justice and good. The principles of justice are based on reasonable beliefs as established by the society's generally accepted methods of inquiry.

III. p. 250. In order to ensure fairness in the deliberations deciding the principles of justice, the people doing the deliberating must work behind a veil of ignorance. They cannot know their place in society, their class position, social status, nor their natural talents or abilities. Neither do they know their own particular views about what is good, nor their own psychological characteristics. This means that the deliberators will be not able to press for a view of justice that they know they will personally benefit from at the expense of others. This procedural view of justice does not require a viewpoint external to the deliberators' own perspectives.

V. p. 251. For a person to have full autonomy, she must be living with fair terms of cooperation, which involves reciprocity and mutuality, i.e. benefiting or sharing in common burdens. This is being reasonable. This is guaranteed by the public nature of the principles of justice, their universality and generality. She must also be rational, which means their deliberations are guided by principles of rational choice, which in turn means that she must have a conception of the good for which she is aiming. Rawls calls these constraints The Reasonable and The Rational.

p. 252. The Reasonable presupposes the Rational because without a conception of the good, there is no point to social cooperation. But the Reasonable also constrains the Rational, because it rules out certain conceptions of the good from being possible pursuable in a fair society.

VI. It follows that the principles of justice that are agreed to have priority over conceptions of the good in their application in a well-ordered society. Once these principles are agreed to, they cannot be overridden by considerations of the welfare of society. An example of this might be the priority of freedom of speech over the good of society: if such freedom has been agreed on as a fundamental part of justice, then even if allowing people to say whatever they want results in damage to the society, such damage is not a reason to reduce freedom of speech.

Lecture III. p. 253. Rawls now considers the objectivity of his theory. His theory is objective in the sense that it is reasonable for us to adopt, as opposed to being true.

I. Rawls says that the first work of modern moral philosophy was Henry Sidgwick's The Methods of Ethics (1874), which has had tremendous influence on subsequent moral theory. It had two limitations. First, it neglects the conception of the person and the social role of morality. Instead Sidgwick focused on the epistemological problem of finding moral truth.

p. 254. Second, he did not recognize the power of Kant's method of ethics to deliver substantial moral truth. Both of these faults led Sidgwick to overlook moral constructivism.

II. p. 255. Rawls contrasts Kantian constructivism with rational intuitionism, which he says has dominated all of moral philosophy. It can be summed up by two theses:

i) the basic moral concepts of the right and the good, and the moral worth of persons, are not analyzable in terms of nonmoral concepts; and

ii) first principles of morals are self-evident propositions about what kinds of considerations are good grounds for applying one of the three basic moral concepts (i.e., something is good, an action is right, or a character trait is virtuous).

People will agree about morality because they will agree about what is morally self-evident. Morality is prior to and logically independent of our views about personhood and the social role of morality. We learn about morality through rational intuition.

p. 256. Kant would have thought that both Hume's psychological naturalism and rational intuitionism do not provide a basis for autonomous action. This is because for Kant autonomy consists in laying down the moral law for oneself, and in both these views the moral law is determined by sources external to the person. This is why Kantian moral theory requires a more complex view of a person than do these alternative approaches.

III. p. 257. Another difference between Kantian constructivism and rational intuitionism is that the former is constrained by human powers of understanding, reflection and judgment and the requirements of public agreement. Moral fairness is what we agree it to be (and what we agree to may not be very precise). While for rational intuitionism, morality is prior to our knowledge of it, and our knowledge can at best approximate an independent truth.

It is because Kantian constructivism is constrained by our cognitive powers that we need to make schematic and practical distinctions. As we have seen, this leads to considerations of justice in society trumping considerations of efficiency or overall welfare. It also explains the priority of the principle of equal liberty over the difference principle (although Rawls does not here spell out what the explanation is).

p. 258. The basic structure of a well-ordered society is established by background justice. Further complexities are dealt with later on. Comparisons of different person's well-being should be done in terms of primary goods, which are "certain public features of social institutions and of people's situations with respect to them, such as their rights, liberties, and opportunities, and their income and wealth, broadly understood." This has the pragmatic advantage of making comparison simpler and less controversial. This is necessary if any workable conception of justice is to be achieved.

In this way, constructivism places pragmatic considerations before everything else. It does not insist that all moral questions must have answers, because the pragmatic considerations dictate that it is the fundamental principles of justice that need identifying. By way of contrast, classical utilitarianism asserts that every moral question has an answer, even if we cannot know what that answer is.

p. 259. Even though some forms of rational intuitionism, such as that of W.D. Ross, allow that some moral facts are too complicated to be determined by the first principles of morality, there is still a fundamental difference from constructivism because they still hold that there is an independent moral order that can be known by our powers of moral intuition. Constructivism does not hold that there is any independent of what rational humans can agree to.

IV. So according to constructivism, the people in the original position are to decide for themselves how complex they want their moral facts to be. The first principles depend on our understanding of human nature. If that understanding changed, (either because of a change in our powers of understanding human nature, or because of a change in human nature), then the first principles of morality would change. Rawls soon qualifies this claim.

p. 260. Constructivism does not depend on any particular theory of truth. The two main theories of truth are correspondence and coherence. The correspondence view says that what it is for a proposition to be true is for it to correspond with reality. The coherence view says that the truth of one proposition can only be assessed in the context of a large set of related propositions that we believe. The proposition is true if it coheres with that large set, false if it does not. On this view, there is no attempt to define truth in relation to an external reality. Rawls refers to a realist view, which presumable is his name for a correspondence theory of truth. He also refers to an idealist or verificationist account of truth. I don't know what he means by 'idealist' here, but a verificationist view would seem to be related to a coherence view. Presumably it would say that a proposition is true if our methods of verification can confirm it, and false if they cannot. It would be circular to explicate the concept of confirmation in terms of 'proving true,' so confirmation would have to be explained simply in terms of our methods of testing propositions yielding a positive result, without referring to an external reality.

Rawls says that our principles of justice will be largely independent of our theory of human nature. The only part of the theory of human nature that is relevant is that which can justify our conception of the person. That conception is "a companion moral ideal paired with the ideal of a well-ordered society." What knowledge we have of human nature does justify this conception of the person, and it is hard to imagine that any additional discoveries about human nature could undermine that justification.

V. p. 261. Why should we think that constructivism provides an objective view of morality? Doesn't simply choosing what morality should be make it arbitrary? Rawls points out that it is not people in the real world who are doing the choosing, but rather those in the idealized original position. It is assumed that those in the original position are constrained by the requirements of rationality and reasonableness, but are ultimately motivated by their own selfish interests, rather than a search for justice itself. The argument is that if people in the original position agree on principles to govern their society, then those principles must be fair. If this procedure "does yield the first principles of a conception of justice that matches more accurately than other views our considered convictions in general and wide reflective equilibrium, then constructivism would seem to provide a suitable basis for objectivity."

p. 262. The agreement that people in the original position is not arbitrary, since it derives from our prior ideals that are part of the shared culture. These include rationality, reasonableness, and the conception of what kind of people they want to be. There are probably only a few conceptions of justice that they could arrive at. It may even be that there is only one possible conclusion that the people in the original position could come to. Or there may be no conception of justice which is both reasonable and workable. We cannot whether there are one, many or no viable conceptions of justice ahead of the deliberations of those in the original position.

VI. p. 263. Rawls wants to have shown that the rational intuitionist conception of objectivity is unnecessary, since Kantian constructivism provides a perfectly adequate conception of objectivity. His conception of objectivity does not come from, as Sidgwick said, "the point of view of the universe" (or as Thomas Nagel would say, the view from nowhere) but rather comes from a suitably constructed social point of view. "It is the publicly shared point of view of citizens in a well-ordered society, and the principles that issue from it are accepted by them s authoritative with regard to the claims of individuals and associations." These principles regulate the basic structure of society, furthers all citizen's interests, and defines the fair terms of social cooperation.

p. 264. Although the concept of a person has often suffered from vagueness, constructivism has the advantage that it provides a way to define the concept as exactly as necessary. In the original position, there is a specific problem to solve, and the deliberators can work to come to an agreement about defining the concept as precisely as if needed to solve the problem.

T. M. Scanlon, "Contractualism and Utilitarianism

Scanlon is very sympathetic to Rawls' point of view concerning justice. In this paper, he argues for a slightly different version of contractualism, and he contrasts it both with Rawls' view and utilitarianism.

p. 267. Scanlon refers to the distinction between act and rule utiliatianism. Recall that act utilitarianism is the simple idea that we should maximize utility with each action. Rule utilitarianism is superficially more like a view which says that there are several moral rules by which we should live. Rule utilitariaism does endorse moral rules. However, these rules are not absolute here. Rather, they are rules which are meant to maximize utility in the long run. Rule utilitarianism arose as a response to several criticisms of act utilitarianism. One of these is that it is not at all practical or feasible to perform calculations about which of the infinitely many possible actions we might perform will maximize utility, especially if we are including in our calculations not just the immediate effects of the action, but also the long term effects. The rule utilitarian says that it we should use simple moral rules in order be clear about what we should do. Of course, when our different rules deeply conflict, we should then resort to calculations of utility.

Scanlon says that utilitarianism has a strong appeal to many appeal in comparison to rival views, even if they wouldn't describe themselves as utilitarians. He will explain why his view is actually more attractive than utilitarianism in that it solves the problems in question better than utlitarianism.

I. p. 268. Moral philosophy exists for similar reasons as philosophy of mathematics. Scanlon gives two main reasons. First, it is not clear what justifies moral claims. Here Scanlon is particularly concerned with the subject matter of ethics and mathematics. It is unclear how their content is epistemically related to their truth. Compare empirical claims about ordinary objects. We can justify these because we can sense ordinary objects. But we don't have any similarly simple account of how the claims of ethics or mathematics are justified. The other reason for the existence of moral philosophy is that we seem to be able to discover the truths of morality through pure thought.

p. 269. A philosohical account of morality has to explain why anyone should care about it. Caring about morality cannot simply be a matter of personal taste. A philosophical theory of morality has to do more than simply show how our first-order moral beliefs can be made to cohere together. (The coherence is called 'narrow reflective equlibrium.') It takes more for a moral statement to be true than simply to cohere with other moral statements.We need to know why those statements matter.

p. 270. The metatheory we accept concerning ethics will tend to have an effect on which first-order systems of moral views seem most plausible. Scanlon is interested in utilitarian metatheory, or what he calls "philosophical utilitarianism." It is the view that "the only fundamental moral facts are facts about individual well-being."

p. 271. Many people find this view very plausible, and it tends to lead people to normative utilitarianism, that it is alway best to maximize the amount of individual well-being in the world. People find it hard to see how there could be any other fundamental moral considerations to take into account. It seems that moral facts which are not reducible to individual well-being are hard to know, and could only be known by a mysterious intuintionist sense.

Of course, utilitarianism, (like all other global moral systems), doesn't agree with all of our prior moral beliefs. If we think that our metatheory forces us to accept normative utilitarianism, the we will have to give up some of our prior moral beliefs. Moral theories is to be assessed

on the basis of their success in giving an account of moral belief, moral argument and moral motivation that is compatible with our general beliefs about the world: our beliefs about what kinds of things there are in the world, what kinds of observation and reasoning we are capable of, and what kinds of reasons we have for action (271-2).

II. p. 272. Contractualism provides an alternative conception of the subject matter of morality to philosophical utilitarianism. An example of a contractualist account of moral wrongness is:

"An act is wrong if its performance under the circumstances would be disallowed by any system of rules for the general regulation of behavior which no one could reasonably reject as a basis for informed, unforced general agreement."

Like most other ethical metatheorists, Scanlon is not going to spell out the details of what is morally right and wrong: he is only concerned to give the general outline of a moral theory. His main focus is on how this theory can be justified. He explains that the idea of informed agreement is meant to rule out supertition and false beliefs about the consequences of actions. The idea of reasonable rejection is to rule out rejections that would be unreasonable, or irrational. Of course, exactly how he spells out the notion of reasonable is crucial, and he promises that he will explain it in greater detail later in the paper.

p. 273. The idea of unforced agreement is meant to rule out people agreeing through coercion or through being in a weak bargaining position. The only factor that should come into play in the hypothetical agreement are "the desire to find and agree on principles which no one who had this desire could reasonably reject." These principles can then apply to the real world, because anyone who was reasonable could not reject them, and the opinions of unreasonable people do not count in the discussion of what it morally right.

There is an important difference between being unable to reasonably reject a principle and being able to reasonably accept it. The former allows in more principles than the latter. Even if reasonable people are not be able to accept a principle which means that they become worse off for the greater good, they might not be able to reject it.

Principles will not be accepted or rejected one by one, and the actions the permit or forbid will certainly not be accepted or rejected one by one. For example, just because one non-rejectable set of principles says that it is permissable to cheat on one's taxes when in dire need does not mean that all non-rejectable sets of principles will allow this. Sets of principles have to be judged rejectable or not as packages coming as a whole. There will have to be elements of convention that come into a set of principles. Individual differences between people will also be relevant, as will the social conditions.

p. 274. Who should be governed by the moral rules that contractualism provides? Other versions of contractualism have been criticized for giving no answer to this question, because it must begin with some set of contracting parties taken as given, or else for being too restrictive, applying only to people who are able to keep and make agreements, and who are able to offer each other mutual benefits through cooperation. Scanlon claims his view suffers neither of these problems. His theory will apply to beings for whom the notion of justification makes sense. This requires the following necessary conditions:

(i) there is a clear sense in which those beings can be better or worse off;

(ii) these beings have to be sufficiently similar to us to provide a basis of comparability between us and them;

(iii) these beings have a point of view of life.

p. 275. These help to explain why the capacity to feel pain is so often taken to provide a being with moral status. But it is not clear whether or not they jointly provide sufficient conditions of moral status. As they stand, they would allow most animals to have moral status. But some would insist that the ability to use language and understand morality is another necessary condition. This requires further discussion, but not here.

III. p. 276. Contractualism must account for moral motivation. According to this approach, "the source of motivation that is directly triggered by the belief that an action is wrong is the desire to be able to justify one's actions to others on grounds they could not reasonably reject." Scanlon finds this psychologically plausible. This is related to the power to imagine other people's point of view, but the thought experiment of changing places with others is only a rough guide. What is most important is to be able to construct a genuinely interpersonal form of justification which each individual could agree to.

p. 277. One of the reasons contractualism is attractive is that rival theories have to explain moral motivation by saying that "there are moral properties which have justificatory force quite independent of their recognition in any ideal agreement." It is puzzling how such properties could exist "in the world."

p. 278. Contractualism can account for the apparent importance of individual well being which utilitarianism takes to be fundamental. Contractualism does not take well-being to be fundamental, but it is important because an individual could reasonably reject an argument that completely ignored her well-being. "The same contractualist framework can also account for the force of other moral notions such as rights, individual responsibility, and procedual fairness."

IV. It is unlikely that contractualism will entail act utilitarianism where one should always maximize the total well being of individuals in society.

p. 279. Scanlon considers a contractarian argument for utilitarianism and finds fault with it, which comes from Harsanyi. He is especially concerned with how to understand impartiality. Here is the argument:

(i) Moral principles must be impartially acceptable.

(ii) To choose principle impartially is to choose them in ignorance of one's position (the veil of ignorance).

(iii) Behind the veil of ignorance, one has an equal chance of being anyone in the society.

(iv) So one would choose the circumstances which provided the greatest expected overall utility, because then the average utilities of the people in society would be highest.

p. 280. Scanlon questions the move from (i) to (ii). The main way we can understand choosing behind the veil of ignorance is by choosing principles which everyone in society would accept. This leads to Scanlon's idea of reasonable rejectability, not to the ideas of the veil of ignorance and of having an equal chance of being anyone in the society. Scanlon says he is not trying to explain impartiality, but rather suggest a way we can make sense of the idea. He thinks that the veil of ignorance does not make sense of the idea of impartiality.

Scanlon suggests that the appeal of ideas like the veil of ignorance, the golden rule, and imagining yourself in other people's positions as guides to what is moral comes from the need to rule out personal biases in moral reasoning. But he says that what they are aiming at is to find what everyone could reasonably agree to, and they are at best only rules of thumb. In important cases they will produce different, and wrong, results about what is morally right.

He shows this with an example comparing two principles, A and E. Harsanhi's contractarian method would choose A, but Scanlon says that some people in society could reasonably reject A. The average utility of people under A is greater than under E. But in A there is a small group of people, the Losers, who are considerably worse off under A than under E. All the non-Losers are slightly better off under under A than under E.

p. 281. "Under contractualism, when we consider a principle our attention is naturally directed first to those who would do worst under it. This is because if anyone has reasonable grounds for objecting to the principle it is likely to be them." This does not necessarily lead to the maximin principle. Here the crucual question is, is it unreasonable for someone to refuse to put up with the Losers' situation under A in order that someone else should be able to enjoy the benefits which he wouldhave to give up under E?

p. 282. Rawls, like Harsanyi, uses the veil of ignorance to guarantee impartial judgments. Behind the veil of ignorance, what offers the best prospects for one offers the best prospects for all. "Thus the choice of principles can be made, Rawls says, from the point of view of a single rational [self-interested] individual behind the veil of ignorance." This it seems that Rawls makes the same mistake as Harsanyi. But Scanlon still agrees with the two principles of justice that Rawls derived. In the remainder of the paper, he argues that his method can be used to derive Rawls' results.
Scanlon discusses how to interpret some of Rawls' comments explaining his argument. Rawls cites three features of the decision faced by those in the Original Position which make it rational for them to use the maximin rule and thus select the Difference Principle.

(1) the absence of any objective basis for estimating probabilities;

(2) the fact that some principles could have consequences for them which "they could hardly accept while

(3) it is possible for them (following maximin) to ensure for themselves of a minimum prospect, advances which, in comparison, matter very little.

p. 283. Furthermore, it would be easier for people to live with the principles that they have previously accepted in a society which is ruled by the Difference Principle, as opposed to a society which requires greater self-sacrifice from some individuals under utilitarian principles. This is an important consideration in deciding what kind of society for create. But this is no more than a rule of thumb.

Rawls says that his method is more Kantian than utilitarianism in that it does not treat people as means to an end. But Scanlon says that his method is more Kantian than Rawls' in this respect, since choosing from behind the veil of ignorance is to treat these lives a "just so many possibilities."

p. 284. Scanlon also thinks that his method can deliver just as precise results as Rawls', and Rawls' principles are only meant to apply to the large scale structure of society, not individual moral dealings between people.

Return to PHI 530 Home Page