Wilfrid Sellars, "'Ought' and Moral Principles," February 14, 1966 (unpublished). Transcribed into hypertext by Andrew Chrucky in June 2005 from a copy saved by Willem deVries.

'OUGHT' AND MORAL PRINCIPLES

1. My aim this evening is to explore a central issue -- indeed the central issue -- in moral philosophy; the relation of scientific knowledge to moral principles. By 'scientific knowledge' I mean science in so far as it tells us what is the case. For even if there is a sense in which science tells us what ought to be the case -- and there is such a science -- we can distinguish, as we shall see, between science qua telling us which is the case and science in its role of telling us what ought to be the case. The original question could, indeed be reformulates in terms of the relationship between these two roles.

2. The Sciences I have in mind include everything from logic and mathematics on one end of the spectrum to psychology and the social sciences on the other. And, since the issues I have in mind is a conceptual one which does not hinge on the current status of any of these sciences, I shall not hesitate to make thought experiments which amount to philosophical fiction about science.

3. Before I get the substantive part of my argument under way, I must introduce some terms and draw some distinctions. The terms will be based in a technical sense, though I hope that they will group together in a perspicuous way things which can be seen to belong together from the high altitude of abstraction in which our problem exists.

4. First a few words about moral principles. At this stage it will be useful to give what might be called the standard account. Whether or not or to what extent the features of this account apply to anything in our experience will be left open for the time being. In philosophy one rarely begins by being absolutely clear about one piece of a conceptual puzzle and unclear about the rest. Philosophical perplexities tends to spread and often the things which seem obvious when one begins become even more puzzling than the initial problem.

5. According to the standard account, a moral principle is a statement which says what people ought to do in a certain kind of circumstance. It is added that to be a moral principle the people in question must be all of us, and to be a principle in the sense of first principle it must not be an application of a more general principle in the way in which, for exainple,

everybody ought to keep his written promises
follows from
everybody ought to keep his promises
or in which, with the additive premise
measles is a distress
everybody ought to help people with measles
follows from
everybody ought to help people in distress

6. That other features belong to the standard account, at least in its more sophisticated form

(a) Science principles of the kind I have formulated can conflict, they must be regarded as abstractions from more complicated principles which specify what ought to be done in combinations of the kinds of circumstances mentioned in these simpler principles. This qualification has the consequence that principles, properly speaking, are far more complicated than the sort of thing we usually give as examples -- for these, on reflection, always carry the rider "other things being equal" -- which they rarely are. Indeed, it commits us to the idea that a moral principle, properly speaking, specifies the circumstances in which a certain kind of action ought to be performed in a very complex way. Yet they remain principles in that they apply to anybody in that kind of circumstance.

(b) It has the further consequence that instead of moral principles (unlike copybook maxims) being already at hand, our task being merely to apply them, serious moral thinking starts from the complexities of the actual case and arrives at a conclusion concerning the principle which applies (i.e. concerning what anybody ought to do in that kind of case in the very process of arriving at a conclusion concerning what I ought to do in that case.

7. Yet we can abstract, for the moment, from this latter feature and suppose that however difficult it may be to arrive at them, they are in some sense "there" to be arrived at. This brings me to another feature of the standard account. Moral principles, however complex, are objective in the sense that there is truth and falsity with respect to them -- that thinking doesn't make them so -- that there are false beliefs about moral principles just as there are false beliefs about scientific laws.

8. Of course, if moral principles were scientific laws, or if they were uniquely determined by scientific, there would be nothing paradoxical about this, nor about the difficulty of discovering what we ought to do. Moral truth would be either a special case of scientific truth -- or -- consequence of scientific truth.

9. But against this it is argued that science tells us what is the case and moral principles what ought to be the case and one can't get from 'is' to 'ought'.

10. Finally, according to the standard account, not only can we be mistaken about what we ought to do; even when we are quite confident that we know what we ought to do, we may well fail to do so. Duty may fall before inclination.

II

11. But before we beard the venerable problem of the relation of 'is' to 'ought' -- which is simply a restatement of our original topic, more distinctions are in order. Let me first introduce the expression 'to value' as in the context

x (a certain person) values y
when y is a state of affairs e.g, x's having a substantial income -- or, perhaps, the Washington Senators winning the pennant. The latter example is intended to bring out the fact that (a) the state of affairs valued need not exist, nor need it include the person who does the valuing. Indeed, I shall use the expression that the state of affairs valued is always in the future, for we are concerned with action and it seems unlikely that action can influence the past.

12. The states of affairs a person 'values' in this technical sense can be very complex. Indeed, what people value is clearly a function of its context and implications, and that valuing involves preferring some states of affairs to others. But there is, I believe, no need to spend time in botanizing the complications which any attempt to classify valuings would involve. Such distinctions as I will draw arise naturally, I hope, out of the broader requirements of our problem.

13. As a final preposition, let me call attention to the difference between the expression of a valuing by someone and the ascription of a valuing to someone. Suppose that Jones values the Senators winning the pennant this year. The statement in our terminology which ascribes this valuing to Jones is siinply:

Jones values the Senators winning this year.
Other statements in more ordinary speech which ascribe valuings of various kinds to Jones would be

Jones wishes that . . .
Jones aims at . . .
Jones intends that . . .
Jones has such and such a goal.

14. Anybody with the relevant information, including Jones himself, can ascribe a valuing to Jones. On the other hand, only Jones himself can express, his valuings. This he may do, for example, by saying

Would that the Senators won this years pennant!
There are many ways of giving exression to valuings, and, indeed, different valuings call for different sentences to express them. But the pattern of my argument as a whole will stand out more clearly if we rise above detail and use the above as our standard sentence for the expression of a valuing. Thus even the valuing of an action a person is about to perform will be supposed to have as its expression
Would that I shortly do A

15. I pointed out that anybody, including Jones himself, can ascribe a valuing to Jones. Thus Jones can say

I wish that the Senators would win this year
or, in our straightjacket,
I value that the Senators win this year
Notice that this self-ascription can serve as a secondary expression of the valuing.

For

Would that . . . but I don't value that . . .
is as conceptually odd as
It is raining, but I don't think that it is

16. Two final points: (a) to value something is to be (disposed to think thoughts properly expressed by

Would that . . . . .
(b) to value, all things considered, an action which one can do right here and now is to do the action (or at least try to do it).

III

17. Let us now introduce a team of scientists on the scene. First, let them be quite ordinary and let them do something simple. They see Jones near a heavy object (O), a long rod (R) and a stone. They mobilize their science and agree that

A necessary condition of Jones raising the stone is his using the rod as a lever
They then rephrase this as
Jones raising O implies Jones using R as lever.
So far so good. Then we suddenly find them saying:
If Jones wants to raise O (values raising O) he ought to use the rod as lever.
Here the word 'ought' has crept in -- and yet seems to belong in a tough-minded, scientific way.

Suppose they now hear Jones say

Would that I raised O
They infer, naturally enough, then he values the state of affairs of his raising O. He wants to raise O. Can they conclude
Jones ought to use the rod as lever?
i.e. can they determine what Jones ought to do? The answer seems clearly 'no.' Consider its parallel case where they pool their scientific knowledge and decide
If Jones wants to poison his Aunt quickly, he ought to give her prussic acid
and they hear him say
Would that I poisoned my Aunt quickly
Suppose they come to Jones and say
If, other things being equal, you value poisoning your Aunt quickly. Then the same things being equal you ought to give her prussic acid.
There seems to be no absurdity in their adding (or anybody who shares their scientific knowledge asserting) "but, of course, you ought not to give your Aunt prussic acid!" The 'ought' does not seem to be a moral 'ought.'

18. The point is a familiar one, but its exact nature must be spelled out -- for after a few turns and twists it will enable us to clarify and, finally, solve our problem.

19. Clearly the scientific information could have been directly communicated to Jones before it was put in terms of 'ought.' In that case, if he was convinced of its truth, he too could have made the move to the 'ought,' thus

If, other things being equal, I want to poison my Aunt quickly, I ought (the same things being equal) to give her prussic acid.
and it is clear that what this amounts to is that the valuing expressed by
Would that I poisoned my Aunt quickly
is logically tied to the valuing expressed by
Would that I gave her prussic acid
by virtue of, the scientific facts of the case.

20. In general terms

If x wants A he ought to bring about B
tells us that if x were to think
Would that A but would that I not bring about B
his valuings would be inconsistent.

21. Of course people often hold inconsistent values; and even when they are persuaded that these valuings are inconsistent they do not always change their mind and substitute consistent valuings. Thus, suppose our scientists approach Jones and persuade him that

Bringing about A implies bringing about B
He may move with alacrity to the thought
If (other things being equal) I value bringing about A, then I ought (the same thinge being equal) to bring about B.
And yet continue to think both
Would that I brought about A
Would that I not bring about B

22. The situation, thus, is the same with respect to our valuings as with respect to our beliefs. We may well continue to hold inconsistent beliefs even after we have been shown that they are inconsistent,

23. Yet logic has some grip on our minds and if, for our present purposes, we exaggerate, we may say, that a person who is convinced that

Bringing about A implies bringing about B
and hence, that
(ceteris paribus) if I value bringing about A then I ought to bring about B
will abandon either
Would that I brought about A
or
Would that I not bring about B

24. We can sum up the apparent moral of our discussion to date by saying that although the 'ought' of

if X wants A, he ought to do B
looks as though it concerned the propriety of doing B on a certain hypothesis, it actually concerns the logical propriety or impropriety of certain complex valuings. To offer advice of the form
If you want A you ought to do B
is not to offer substantive advice as does
You ought to do X
It is to give logical advice.

IV

25. We seem to have shown that the 'ought' which our team of scientists can generate is not the 'ought' which tells us what to do, but rather an 'ought' which pertains to the coherence of our valuings. Since the 'ought' of moral principles does tell us what to do, it would seem that the 'ought' which scientists generate is not the ought of morals.

26. Yet is this so clear? After all, even if the 'ought' which our team of scientists have come up with do in a sense pertain to the coherence of Jones' valuings, they are formulated in such a way that they take actions as their objects

If X wants A he ought to do B
Thus the mere fact that moral ought tells us what we ought to do is no decisive argument against the idea that they too pertain to the coherence of our valuings. Let me develop this thought.

27. It will not have gone unnoticed that I have been careful to sprinkle my story to date with such phrases as 'ceteris paribus' and "other things being equal." But is not

If, all things considered, I value A, then I ought to do B
closer to what we want? Perhaps. Yet what does ''all things considered mean?

28. Without attempting a direct answer, let me sketch a classical theory -- essentially Plato's in modern dress -- which can be called the "informed or enlightened self-interest model. It conceives of a ground floor of human valuations, which arise naturally out of bodily, mental or social need. These specific valuings need not be ego-centered. Only the tough-minded like Hobbes thought that the valuing expressed by

Would that he not be suffering
is really what would be candidly expressed by
Would that I not be in his shoes.
But whatever the character of these ground floor or specific valuings, they are dominated by an over-arching ego-directed valuing expressed, in our terminology by
Would that I led a happy life
To the extent that people are logical, this valuing, combined with beliefs about the world and the propensities of our fellow men, and. also what constitutes a happy life, generates a life plan valuing which alots subordinate places to ground-floor valuings, and is carried out in action. Most people lead blurred and fluctuating lives because of their changing beliefs, the faltering grips of logic and the breaking in of impulse.

29. Let us re-introduce our team of scientists. Let us not idealize them and suppose that they really know what things and persons are like. They turn their attention to a budding young genius (Jones, again) and say among themselves.

If Jones wants to lead a happy life he ought to lead a life of kind L
(The neo-Freudian-Pavolians among them, we suppose, have conclusively shown that for budding geniuses L is the happy life.) They go to Jones and say
If you want to lead a happy life you ought to lead a life of kind L.
And when he questions it, they then persuade him to come to the Academy of Sciences. After several years study Jones says "I see!" and proceeds to think along the following lines.
If x want to lead a happy life I ought to lead a life of kind L.
Ex hypothesi he has built unto him the valuing
Would that I led a happy life
being logical he now conjoins with this the valuing
Would that I led a life of kind L
Drawing on additional knowledge he has gained in the Academy he continues
If I want to lead a life of kind L, I ought to do such and such things in such and such circumstances
always logical, he adds to his battery of valuations
Would that I do such and such things in such and such circumstances
where the policies he has in mind take account of all the subtle ways in which circumstances can vary, yet seem on casual inspection to be the same.

Confronted with a particular circumstance C, he reasons

If I want to do A whenever I am in C I ought to do A now, for C now obtains
as before his valuings remain coherent and he adds the valuing
Would that I now do A
which, since it constitutes his being favorably disposed to doing A generates the action in question.

30. Plato's basic analogy here, is that of an art or craft. He develops in a number of dialogues the concept of an art of living which he conceives of as the art of achieving happiness. Assigning that a shoemaker values, for whatever reason, the making of shoes, and that no other values become relevant during shoemaking hours, that he knows what the various stages of making a shoe require him to do to the leather and that he knows how to do it, then he will turn to and make a good shoe. Within the limitations of his material. So, too, a man in possession of the art of living will turn to and create a happy life within the limits of his opportunities.

31. But what, it will be asked, guarantees that our genius, Jones, however meticulously he plans his life and carries out his plans will behave in the ways we think of as moral. In other words what guarantees that the rules of the art of living coincide in their content with the content of the principles of morality? Mightn't the rules include such items as

Would that whenever I wear a ring which makes me invisible, I take all available loose cash to spend on research
Not to mention more exotic pursuits?

32. Plato has a lengthy argument designed to prove that the rules of the art of living coincide in content with what we think of as sound moral principles. The argument has more bite to it than is usually allowed, but this is because Plato tacitly introduces some themes from the next and concluding section of my lecture. Even if he were successful in showing this coincidence, however, he would not have succeeded in showing that the conclusion of the reasoning

Sure I want to lead a happy life
                    .
                    .
                    .
                    .
I ought to do such and such in such and such circumstances
is identical in sense with the statement
I morally ought to do those things in those circumstances
We find it quite meaningful to say
I morally ought to do such and such actions in such and such circumstances even if it would result in my unhappiness and other courses of action would make me happy.

V

33. Yet we still have not shown that the 'ought' of moral principles is not what we have called the 'ought' of coherence; when the principles of coherence are supplied by logic and the other sciences. For the trouble may not be with the pattern of Plato's theory, but with the valuing which he postulates as the unbuilt motor of the well lived life. In particular, the trouble may be with the egoistic or self-centered nature of this valuing. An altruistic goal might, it would seem, do the job.

34. Our team of scientists returns and this time, we suppose, say with good reason

If Jones wants to promote the general welfare he ought to do such and such things in such and such circumstances.
And they now allow that a person can seek to promote the general welfare for its own sake; as before they come to Jones and say
If and to this extent that you want to promote the general welfare you ought to do such and such things in such and such circumstances
And take him to the Academy. He returns to the world convinced and reasons
If (and to this extent that) I want to promote the general welfare I ought to do such and such actions is such and such circumstances
He is convinced, in other words of the coherence of
Would that I promoted the general welfare

would that I do such and such, etc.

and the incoherence of
Would that I promoted the general welfare
Would that I not do such and such, etc.
and, other things being equal, he values promoting the general welfare. This certainly has a more moral ring to it. Yet it is not completely satisfying. And when we see why, we will have the answer to our problems.

35. Our team of scientists might equally have pointed out that

If, and to the extent that, Jones wants to promote the welfare of hummingbirds, he ought to do such and such in such and such circumstances.
A coherent system of valuings along these lines is not egoistic. Yet it does not have the ring of morality. Even if we replace it by
If Jones wants (ceteris paribus) to promote the general welfare of language using featherless bipeds
which comes a bit closer, we can picture Jones, who, other things being equal, also values his own happiness, confronted by two coherent sub-systems of valuings, thus

If I want(ceteris paribus) to be happy I ought to pursue such and such policies If I want (ceteris paribus) to promote the general welfare, I ought to pursue such and such other policies

And it is not clear how these can be resolved into one embracing coherent system of valuings so that the 'ceteris paribus' could be dropped and a single 'ought' remain.

36. We feel that moral 'ought' is in principle embracing and unequivocal, i.e. that if we had ideal knowledge, what we ought to do would be uniquely determined Egoism plus ideal knowledge did uniquely determine an 'ought' -- but not the 'ought' of moral reasoning. While if we recognise a duality of basic valuings, the moral pattern seems closer, but the uniqueness disappears.

37. Is there any way in which the uniqueness desired can be restored while retaining the idea that the moral 'ought' is the 'ought' of coherence? The first of the above objections gives the clue. People are not hummingbirds -- not even language using featherless bipeds. They are us; or, to come to the point, the moral point of view is not that in which we do good to the language-using featherless bipeds which surround us. It is that in which each of us seeks, as one of us, the common good of all of us.

38. There is a sense in which we all understand this, but how is it to be understood in the abstract categories of philosophy? The terminology developed at the beginning will prove, I believe, invaluable. It will be remembered that I distinguished between the sentence which expresses a valuing

would that . . .
and the sentence which ascribes a valuing to someone.
Jones values--------
Now in the examples I have been considering, it has been tautly assumed that Jones values X, so to speak, "as himself" or "from a personal point of view". Let me, therefore, use a device, lurking in but not highlighted by ordinary language, to bring this out in the very expression of the valuing.
I would that X were the case
The corresponding ascription would be
Jones, from a personal point of view, values X
(where "from a personal point of view" doesn't mean selfishly (nor does it preclude it).

39. What are we to contrast with this? (The very terminology should give a clue.) I spoke above of valuing as being a member of ones community. And since I am concerned with the moral point of view, I have in mind the most embracing community to which we belong; those who are in the biblical sense of the term our brothers. To value in this way is to think in terms of we, in the embracing sense, thus

We would that X were the case
Notice that it is still the individual who is valuing. The corresponding ascriptive sentence is
Jones values X as one of us
indeed
. . . from a moral point of view
valuing has, if I may so put it, a subjective form as well as a content.

40. Let me emphasise that from the fact that Jones values X as one of us it doesn't follow that

we value X
we may well -- indeed often do -- differ in what we value as members of the community. Yet in principle there would be agreement.

41. Let me give a parallel from our discussion of self-interest. The specific things Smith values may differ from time to time as his beliefs about human nature and the world change. But in relation to valuing of a happy life on the whole, then in principle i.e. if he had had at all times the knowledge provided by our team of scientists, he would have agreed with himself at all times about the valuings consistent with this overarching aim. There is in principle agreement between, so to speak, Smith at t1, Smith at t2, etc. His disagreements with himself at different times are in principle resolvable.

42. The same is true of the members of a community in so far as they aim at the common good. Thus, if our team of scientists considers the community, they can, we suppose, by virtue of their knowledge of individuals and groups, come up with something of the following form:

If the members of the community value promoting the common good, they ought to support such and such (intricate) practices
and, addressing themselves to the group.
If we value promoting the common good as members of the community, you ought to support such and such practices
and, convinced, let us suppose, by scientific arguments, each member of the community, e.g. Jones, will reason.
If I value as one of us the common good, I ought to support doing such and such things in such and such (intricately specified) circumstances
which tells him that the valuings
we would that the common good was promoted
and
we would that the practice of doing such and such things in such and such circumstances was supported
cohere. Thus, from the moral point of view they are confronted by the logical 'ought'
Since I value (as one of the promotion of our common good I ought to support the practice of doing such and such in such and such circumstances
and since the same considerations would confront each member, and they share the same primary valuing with which their other valuing are to cohere, they would share valuings of the form
We would that if any of us in C he does A
which is, I submit, the valuational content of a moral principle, the 'ought' in the more familiar formulation expressing the coherence of this valuing with the valuing of our common good in the light of the relevant facts about men and the world.

43. These valuings are, though valuings by individuals, universal in three dimensions

(a) in their content
. . . that if any of us is in C he do A
(b) in their subjective form
we would that . . .
(c) in their acceptance, i.e, in the fact that all agree in so valuing

44. Alas! The ideal knowledge of our philosophical fiction is not shared by all, nor is it even at hand. Thus the ideal (if I may use the word) consensus sketched above is only in principle "there", and reasoning in the moral point of view proceeds in a context of disagreement and diversity of opinion.

45. Yet the question

What ought I to do in these circumstances (from the moral point of view)?
is at heart the question
What would we that I do in these circumstances?
and to answer the question is to answer the question
What would we that any of us do in these circumstances?
and this again
On what valuings of the form "we would that any of us did in such and such in these circumstances"

Would we agree if we had ideal knowledge?
and this again
What practices would we agree to be conducive to our common good -- if we had Ideal Knowledge?
Yet, since the one who asks the question lacks the ideal knowledge, he must make the best of such knowledge as he has (though we know more than we think about human nature, its propensities and needs).

46. If my argument is sound, the moral 'ought' is the 'ought' of coherence once again -- but this time in the moral point of view. Thus, where inclination -- whether selfish or sympathetic -- points in one direction and moral valuation in another, the tension is between too different ways of valuing, not simply between two values which we have from our own point of view.

Yet we are, more frequently than we are willing to allow, confronted by a choice between actions one of which is called for by one way of thought, the other by the other. It is notorious that the outcome does not always have its source in the moral point of view.

47. Does our argument not commit us to the idea that the moral point of view and the personal point of view are unrelated alternatives? Two alternatives which simply present an either-or? The either-or aspect is not to be ignored, yet at the level our argument has reached we can see how there might be a harmony between the two points of view, a harmony more basic than the slipshod attempt to fit self-interest and sympathy into a coherent framework of valuings from the personal point of view, and call the result morality.

48. Most people live on second hand moral thinking, and the momentum of childhood training. Suppose such a person were to meditate as follows:

Granted then I and my fellows have been brought up to have such and such commitments as to what one is to do in various kinds of circumstances. Why shouldn't I let these commitments wither away and encourage, self-regarding attitudes? Is there any reason why I should nourish my commitment to the moral point of view?
What could we say? In effect we have seen that the only frame of mind which is the living source of moral commitment is what Josiah Royce called Loyalty, and what Christians call Love (Charity). This is a commitment deeper than any commitment to abstract principle. It is this commitment to the well-being of our fellows which stands to the justification of moral principles as the purpose of acquiring the ability to explain and predict stands to the justification of scientific theories. This concern for others is a precious thing, the foundation for which is laid in early childhood, though it can come about, in adult years, through the little understood phenomena of conversion and psycho-therapy.

I will close with the remarks that recent psychological studies make clear what has in a sense been known all along, that the ability to love others for their own sake is as essential to a full life as the need to feel ourselves loved and appreciated for our own sakes -- unconditionally, and not as something turned on or off depending on what we do. This fact provides, for those who acknowledge it, a connection which can justify a course of action designed to strengthen our ability to respond to the needs of others. Thus really intelligent and informed self-love supports the love of ones neighbor which alone directly supports the moral point of view when, as the Eighteenth century British divine, Bishop Butler, put it, we sit down in a cool hour and ask why we should do our duty.

Wilfrid Sellars

Pittsburgh
February 14, 1966