williams, b. morality_an introduction to ethics

8/2/2019 Williams, B. Morality_An Introduction to Ethics

1/59

BERNARD WILL IAMS

MORALITYAN I NTRODUCT ION

TO ETHICS


2/59

Published by the Press Syndicate of the University of CambridgeThe Pitt Building, Trumpington Street, Cambridge CB2 1RP

40West 20th Street, New York, NY 1001 14211, USA10 Stamford Road, Oakleigh, Melbourne 3166, Australia

OBernard Williams 1972Preface to the Canto edition @Bernard Williams 1993

ISBN 0 521 45729 7 paperbackFirst published in the USA by Harper& Row 1972Published in Pelican Books 1973Reissued by the Cambridge University Press 1976

Reprinted 1978 1980 1982 1987 1990Canto edition 1993

ToMy Mother and Father

Printed in Great Britain at theUniversity Press, Cambridge


3/59

CONTENTS

Preface to the Canto edition page xiPreface xviiThe amoralist 3Subjectivism : First thoughts 14Interlude: Rela tivism 20Subjectivism: Further thoughts 26'Good' 38Goodness and roles 48Moral standards and th e distinguishing mark o fa m an 55God, morality, andprudence 63W ha t ismorality about? 73Utilitarianism 82


4/59

PREFACE T O TH E C AN TO EDITION

THIS ext was originally intended to be part of a largerbook, which was to consist of several novella-lengthpieces by different writers, forming collectively asubstantial introduction to philosophy. When theeditor, Arthur Danto, invited me to write the sectionon moral philosophy, he made it clear that while wewere encouraged to write in an introductory way, wewere not being asked to write merely a survey, butrather to pursue the interests and questions that eachfound most interesting or fruitful. The publishers in theend decided not to put out the big book (which,granted who they were, some of us had inevitablycalled 'Harper's Bazaar'), and published each sectionseparately.

At least one of those books (Richard Wollheim's Artand Its Ob jec ts) has grown in later life, acquiring newsections in further editions. This one, on the otherhand, remains as it was. The main reason for this isthat I have subsequently written other books andpapers on some of the same subjects, and could see nopoint in loading this text with intrusive (and probablymisleading) references to that later work. In the case ofone topic, utilitarianism, this would have been par-ticularly inappropriate, since in what I wrote later Itried to take account of what I had written here, and todevelop rather different points; the relevant chapterhere perhaps summarizes the central problem, as I see


5/59

it, of the utilitarian project more compactly than Ihave done elsewhere.

For rather similar reasons, I have not tried to providean up-to-date bibliography. Recent literature on sub-jectivism, for instance, has obviously changed andextended the questions beyond my treatment of thesubject here, but to explain this so as to introduce thatliterature would have involved substantial furtherphilosophical discussion. In fact, there is no bibli-ography, even an out-of-dateone, but only a handful ofreferences, perhaps rather idiosyncratic, to some writ-ing that I had found helpful.

In one respect at least the book may seem dated, tothe extent th at i t starts by complaining of a situationwhich no longer exists, one in which moral philosophyaddressed itself to meta-ethical questions about thenature of moral judgement, the possibility of moralknowledge, and so forth, at th e expense of discussingfirst-order ethical questions. Moral philosophy still,appropriately, discusses meta-ethical questions, but itis certainly not true any longer that first-order ques-tions are not mentioned. On the contrary, issues suchas abortion, feminism, and famine are now standardlydiscussed in moral philosophy courses and textbooks. Imust confess that some of these discussions, assumingas they do that ethical thought is made more rationalby deploying ethical theory, seem to me as distancedfrom real experience as the forms of uncommittedmoral philosophy about which I complain here. Thespirit in question is, unnervingly, the same. But thecomplaint is certainly different. *

There are points at which the two complaints unite.I have criticized in some more recent work the assump-

tion often made, not just by moral philosophy but byethical reflection more generally, that we are clearenough about what count as 'moral' considerationsand sentiments, and tha t what moral philosophy mustseek is the basis and status of these considerations,taken more or less as a whole. I have wanted to ask aprior question, about what the distinction between the'moral' and the 'non-moral' is supposed to do for us;and I have suggested tha t considerations of the moralkind make sense only if they are related to otherreasons for action that human beings use, and gener-ally to their desires, needs and projects.

Those concerns are perhaps not altogether explicit inthis book, and in particular it does not observe acertain verbal distinction which I have more recentlyfound useful, between a broader conception of 'theethical', and the narrower concerns (focused particu-larly on ideas of obligation) of what may be called thesystem of 'morality'. Others may not find this termin-ology helpful, but since I have suggested it, it is perhapsworth mentioning, in particular, that the title andsub-title of this book do not use those words in thatway. It would surely be possible to discuss morality- in,my current, restrictive sense - as an introduction toethics (though I doubt that i t would be the best way toget introduced to it ); but this is not in fact wha t thisbook does. Rather, i t discusses, a lot of the time, ethics

* I have pressed the later com plain t, against the supposedpower of ethical theory, in Ethics and the Limits of Philos-ophy (London: Collins , and H arvard University Press, 1985),where I also discuss some peculiarities of the 'moralitysystem', wh ich I mention below.


6/59

as an introduction to the problems and limitations ofmorality.

The placing of morality in relation to other ethicalconsiderations and to the rest of life - in relation tohappiness, for instance - is in fact a theme here,although it is not expressed in those terms. Because itdoes contain that theme, the book has a characteristicwhich I did not consciously have in mind when I wroteit, but which was pointed out to me by a classicist whohad used it in his teaching, that the concerns fromwhich i t sets out are those more typical of the ancientworld and its philosophies than of modernity. In arecent book, * I have tried to develop more reflectivelyand on a larger scale this interest in ethical ideas of theancient world (and not only of its philosophies), and asense of their relevance to our present situation.

Near the beginning of this book, I talk about theproblems of finding a style for moral philosophy. I stillthink tha t these problems are real, and also that moralphilosophy involves such problems to a greater extentthan most other areas of philosophy. What I should notwant to accept now is an implication to be heard in thisdiscussion, that there might be one general solution tothis problem, and tha t once one had found i t one wouldknow how to write moral philosophy. That cannot beso: the problems of finding a convincing, adult, andunmechanical way of approaching the subject must befaced on each occasion. Sometimes literature or his-tory can be called upon, t o give some idea of the weightor substance of ethical concepts that we use or havebeen used by others; analytic argument, the philos-

opher's speciality, can certainly play a part in sharpen-ing perception. But the aim is to sharpen perception, tomake one more acutely and honestly aware of whatone is saying, thinking and feeling. Philosophy invitesus (perhaps more insistently now than when this bookwas written) to ask whether what we say in morality istrue. One thing I felt in writing this book, and feel evenmore now, is that it is vital not to forget anotherquestion that is to be asked both about morality andabout moral philosophy, how far what we say ringstrue.

BERNARD WILLIAMSBerkeley, March 1993.

* Shame and Necessity (California University Press, 1993).


7/59

-PREFACE

WRITINGbout moral philosophy should be a hazard-ous business, not just for the reasons attendant on writ-ing about any difficult subject, or writing aboutanything, but for two special reasons. The first is thatone is likely to reveal the limitations and inadequaciesof one's own perceptions more directly than in, at least,other parts of philosophy. The second is that one couldrun the risk, if one were taken seriously, of misleadingpeople about matters of importance. While few writerson the subject have avoided the first hazard, very manyhave avoided the second, either by making it impos-sible to take them seriously, or by refusing to writeabout anythlng of importance, or both.This sad truth is often brought forward as a partic-ular charge against contemporary moral philosophy ofthe 'analytical' or 'linguistic' style: that it is peculi-arly empty and boring. In one way, as a particularcharge, this is unfair: most moral philosophy at mosttimes has been empty and boring, and the number ofgreat books in the subject (as opposed to books involvedin one way or another in morality) can be literallycounted on the fingers of one hand. The emptiness ofpast works, however, has often been the emptiness ofconventional moralizing, the banal treatment of moralissues. Contemporary moral philosophy has found anoriginal way of being boring, which is by not discussingmoral issues at all. Or, rather, it is not so much that a


8/59

MORAL ITY PREFACEstyle of moral philosophy has been evolved which cutsthe connection with moral issues altogether - hat, if itwere possible, would have the interest of being remark-able; but the desire to reduce revealed moral commit-ment to a minimum and to use moral arguments in therole of being uncontentiously illustrative leaves an im-pression that all the im portan t issues are off the page,somewhere, and that g reat caution and little imaginationhave been used in letting tiny corners of them appear.There are many reasons for this situation. A centralone is that contemporary views about morality itselfleave an unclarity about what qualities of mind orcharacter are particularly called upon in constructivemoral thought (indeed, in som e accounts of m orality itis not even clear that there can be such a thing as con-structive moral thought); they hence leave one all themore uneasy abou t whether those qualities are likely tobe the qualities of philosophers, when philosophyis largely a professional and academic activity callingprincipally, though not exclusively, on discursive andanalytica l abilities. If there were such an activity asdeducing substantial moral conclusions from a prioripremises, trained philosophers might reasonably beexpected to be rather specially good at it; but there isnot, and the fact that if there were, then professionalphilosophers would stand a specially good chance ofbeing informed about morality, is itself one of th e goodreasons for thinking that there could not be such anactivity.Certainly the trouble is not, as some pretend, that ifthe philosopher is not patently detached and evenmethodological, then he must be preaching; that can-not possibly be the only alternative. It is rather a styl-

istic problem, in the deepest sense of 'style' in whichto discover the right style is to discover what you arereally trying to do. How does one combine argument(which is after all likely to constitute the philosopher'sSpecial claim on anyone's attention) with either thelonger leaps or th e more concrete detail which providethe more interesting stuff of moral thought? Can thereality of complex mora l situations be represented bymeans other than those of imaginative literature? Ifnot, can more schematic approaches represent enoughof the reality? How much of wha t genuinely worriesanyone is responsive to general theory ?If I knew answ ers to these questions, I should nothave to ask them now.This essay takes a rather tortuous course, and whileI have tried to signpost the major bends, it may beworth sketching a plan in advance. I sta rt with a figurewho has often been of interest, indeed a cause of con-cern, to moralists, as providing a challenge to moralityand a demand for its justification: the arnoralist, wh ois supposedly immune to moral considerations. Someof the most interesting questions about him, which Ihave barely touched on, lie no t so much in wh at mightbe said to him, as in what might be said about him -what the amoralist can consistently be like. From him,we move to those who do not reject morality, but dotake certain special, and it may be almost as disquiet-ing, views about its natu re: subjectivists of vari-ous kinds, and an unashamedly crass (but common)kind of relativist. Here I try to examine carefully a p r eject very close to the heart of much modem moralphilosophy, which I have called that of defusing s u bjectivism.

xviii xix


9/59

M O R A L I T YFrom there, to some considerations about goodness;and, trying to get clear about some ways a man can begood at certain things, and still more about things hecan be good as, I seek to disentangle some purely logi-cal considerations from what seem to me more su bstantial issues about what men are and the connectionof that with goodness. Two questions in particularemerge from the many that crowd round that area : herelations of intellectual achievement to the standardsof morality, and the question whether, if God existed,that would make any difference to the situation ofmorality. This raises some important genera; questionsabout moral and other motives. These in turn lead tosome issues about the point or substance of morality,and whether it is ultimately all about human welfareor (more narrowly, perhaps) happiness. Lastly, the mostsimple-minded way of aiming morality at happiness,that of utilitarianism, is touched on, but only longenough to suggest how special and peculiar a system,properly understood, it is; and to point in the directionwhere its peculiarities are to be found. To follow themout is a task for another occasion.*One of the many ways in which this essay is not atextbook, even an introductory and outline textbook,of moral philosophy is that it offers no systematictheory. I am unashamed about that, since it seems tome that this subject has received more over-general andover-simplified systematization, while inviting it less,than virtually any other part of philosophy. I do notmean by that that one should approach moral philo-sophy without preconceptions (which would be impos-*See A Critique of Utilitarianism, which appears in Utili-

tarianism: For and Against (Cambridge University Press, 1973).XX

P R E F A C E&le), or even without theoretical preconceptions(which might well prove static and sterile). It is merelythat one's initial responsibilities should be to moralphenomena, as grasped in one's own experience andimagination, and, at the more theoretical level, to thedemands of other parts of philosophy - n particular, ofthe philosophy of mind. There is no reason why moralphilosophy, or again something in some respectsbroader, in some respects narrower, called 'valuetheory', should yield any interesting self-containedtheory at all.Another way in which this is not a textbook is thatit leaves out large tracts of the subject. That is, at least,fully obvious. But it may help to put what is here insome better perspective if I mention one or two subjectswhich a larger treatment of moral philosophy should inmy view have near its centre: what practical thought,and acting for a reason, are; what consistency in actionis, and in moral thought; relatedly, how moral conflictis a basic fact of morality; how the notion of a rule isimportant for some, but not all, parts of morality (thepresent essay has nothing to say about its importance);how shaky and problematical is the distinction betweenthe 'moral' and the 'non-moral' - above all in its mostimportant employment, to distinguish between differ-ent sorts of human excellence.That this essay should leave out most things of im-portance was inevitable; that it should follow a tor-tuous path, was not. Whether it was inevitable that itshould fail to find an answer to the problem of how towrite about moral philosophy, I do not know.

xxi


10/59

M O R A L I T YA N I N T R O D U C T I O N TO E T H I C S


11/59

T H E A M O R A L I ST

< W H Y should I do anything?' Two of the many waysof taking that question are these: as an expression ofdespair or hopelessness, when it means something like'Give me a reason for doing anything; eve-ng ismeaningless'; and as sounding a more defiant note,against morality, when it means something like 'Whyis there anything that I should, ought to, do 2'Even though we can paraphrase the question in thefirst spirit as 'Give me a reason . ', t is very unclearthat we can in fact give the man who asks it a reason -that, starting from so far down, we could argue himinto caring about something. We might indeed 'give hima reason' in the sense of finding something that he isprepared to care about, but that is not inducing him tocare by reasoning, and it is very doubtful whether therecould be any such thing. What he needs is help, or hope,not reasonings. Of course it is true that if he stays alivehe will be doing something, rather than something else,and thus in some absolutely minimal sense he has somesort of reason, some minimal preference, for doingthose things rather than other things. But to point thisout gets us hardly anywhere; he does those things justmechanically, perhaps, to keep going, and they meannothing to him. Again, if he sees his state as a reasonfor suicide, then that would be to make a real decision;as a way out of making any decisions, suicide comesinevitably one decision too late (as Camus points out in


12/59

MORALITY THE AMORALISTLe Mythe de Sisyphe). But it would be no victory for usor for him if it turned out there was after all just onedecision that he was prepared to acknowledge, thatone.

I do not see how it could be regarded as a defeat forreason or rationality that it had no power against thisman's state; his state is rather a defeat for humanity.But the man who asks the question in the second spirithas been regarded by many moralists as providing a realchallenge to moral reasoning. He, after all, acknowledgessome reasons for doing things; he is, moreover, likemost of us some of the time. If morality can be got offthe ground rationally, then we ought to be able to get itoff the ground in an argument against him; while, in hispure form - in which we can call him the arnoralist -he may not be actually persuaded, it might seem a com-fort to morality if there were reasons which, if he wererational, would persuade him.We might ask first what motivations he does have.He is indifferent to moral considerations, but there arethings that he cares about, and he has some real pre-ferences and aims. They might be, presumably, pleasureor power; or they might be something much odder, suchas some passion for collecting things. Now these endsin themselves do not exclude some acknowledgement ofmorality; what do we have to leave out to represent himas paying no such acknowledgement? Presumably suchthings as his caring about other people's interests, hav-ing any inclination to tell the truth or keep promises ifit does not suit him to do so, being disposed to rejectcourses of action on the ground that they are unfair ordishonourable or selfish. These are some of the substan-tial materials of morality. We should perhaps also leave

out a more formal aspect of morality, namely any dis-position on his part to stand back and have the thoughtthat if it is 'a l l right' for him to act in these ways, itmust be 'all right' for others to act similarly againsthim. For if he is prepared to take this stance, we mightbe able to take a step towards saying that he was not aman without a morality, but a man with a peculiarone.

However, we need a distinction here. In one way, itis possible for a man to think it 'all right' for everyoneto behave self-interestedly, without his having got intoany distinctively moral territory of thought at all: if,roughly, 'it's all right' means 'I am not going tomoralize about it'. He will be in some moral temtory if'all right' means something like 'permitted', for thatwould c a w implications such as 'people ought not tointerfere with other people's pursuit of their own inter-ests', and that is not a thought which, as an amoralist,he can have. Similarly, if he objects (as he no doubtwill) to other people treating him as he treats them, thiswill be perfectly consistent so long as his objecting con-sists just in such things as his not liking it and fightingback What he cannot consistently do i's resent it ordisapprove of it, for these are attitudes within the moralsystem. t may be difficult to discover whether he hasgiven this hostage to moral argument or not, since hewill no doubt have discovered that insincere expressionsof m t r n e n t and moral hurt serve to discourage someof the more squeamish in his environment from hostileaction.

This illwtrates, as do many of his activities, the0 ~ 0 ~ ~act that this man is a parasite on the moralSystem, and he and his satisfactions could not exist as


13/59

MORALITYthey do unless others operated differently. For, in gen-eral, there can be no society without some moral rules,and he needs society; also he takes more particular ad-vantage of moral institutions like promising and ofmoral dispositions of people around him. He cannotdeny, as a fact, his parasitic position; but he is veryresistant to suggestions of its relevance. For if we trysaying 'How would it be for you if everyone behavedlike that? ' he will reply, 'Well, if they did, not good, Isuppose- hough in fact I might do better in the result-ing chaos than some of the others. But the fact is, mostof them are not going to do so; and if they do ever getround to it, I shall be dead by then.' The appeal to theconsequences of an imagined universalization is anessentially moral argument, and he is, consistently, notimpressed by it.In maintaining this stance, there are several thingshe must, in consistency, avoid. One - as we noticedbefore, in effect - s any tendency to say that the moreor less moral majority have no right to dislike him,reject him, or treat him as an enemy, if indeed they areinclined to do so (his power, or charm, or dishonestymay be such that they do not). No thoughts about justi-fication, at least of that sort, are appropriate to him.Again, he must resist, if consistent, a more insidioustendency to think of himself as being in character reallyrather splendid - n particular, as being by comparisonwith the craven multitude notably courageous. For inentertaining such thoughts, he will run a constantdanger of getting outside the world of his own desiresand tastes into the region in which certain dispositionsare regarded as excellent for human beings to have, orgood to have in society, or such things; and while such

6

THE AMORAL I STthougaa need not lead directly to moral considerations,they give a substantial footing to them, since they im-mediately invite questions about what is so good aboutthose dispositions, and it will be difficult for him topursue those questions very far without thinking int- of the general interests and needs of his fellowhuman beings, which would land him once more backin the world of moral thought from which he is exclud-ing himself.The temptation to think of himself as courageous isa particularly dangerous one, since it is itself verynearly a moral notion and draws with it a whole chainof distinctively moral reflections. This man's applicationof the notion will also have a presupposition which isfalse: namely, that the more moral citizens would beamoral if they could get away with it, or if they werenot too frightened, or if they were not passively con-ditioned by society - if, in general, they did not sufferfrom inhibitions. It is the idea that they are afraid thatgives him the idea of his own courage. But these pre-suppositions are absurd. If he means that if as an indi-vidual one could be sure of getting away with it, onewould break any moral rule (the idea behind the modelof Gyges' ring of invisibility in Plato's Republic), it isiust false of many agents, and there is reason why : hemore basic moral rules and conceptions are stronglyinternalized in upbringing, at a level from which theydo not merely evaporate with the departure of policemen or censorious neighbours. This is part of what it isfor them to be moral rules, as opposed to merely legalMQuirements or matters of social convention. Theof moral education can actually be to makepeople want o act, quite often, in a non-self-interested


14/59

MORALITY

way, and it often succeeds in making it at least quitedifficult, for internal reasons, to behave appallingly.

But this, he will say, is just social conditioning;remove that, and you will find no moral motivations.- We can reject the rhetoric of the word 'conditioning';even if there were a true theory, which there is not,which could explain all moral and similar education interms of behaviourist learning theory, it would itselfhave to explain the very evident differences betweensuccessful and intelligent upbringing, which producesinsight, and the production of conditioned reflexes.Then let us say instead that all moral motivation is theproduct of social influences, teaching, culture, etc. It isno doubt true. But virtually everything else about aman is such a product, including his language, hismethods of thought, his tastes, and even his emotions,including most of the dispositions that the amoralistsets store by. - But, he may say, suppose we grant thatanything complex, even my desires, are influenced byculture and environment, and in many cases producedby these; nevertheless there are basic impulses, of a self-interested kind, which are at the bottom of it all: theseconstitute what men are really like.

If 'basic' means 'genetically primitive', he may pos-sibly be right: it is a matter of psychological theory.But even if true in this sense, it is once more irrelevant(to his argument, not to questions about how to bringup children); if there is such a thing as what men arereally like, it is not identical with what very small chil-dren are like, since very small children have no lan-guage, again, nor many other things which men reallyhave. If the test of what men are really like is made,rather, of how men may behave in conditions of great

8

THE AMORA L I S Tstress, deprivation, or scarcity (the test that Hobbes, inhis picture of the state of nature, imposed), one can onlyask again, why should that be the test? Apart from theunclarity of its outcome, why is the test even appro-priate? Conditions of great stress and deprivation arenot the conditions for observing the typical behaviourof any animal nor for observing other characteristics ofhuman beings. If someone says that if you want to seewhat men are really like, see them after they have beenthree weeks in a lifeboat, it is unclear why that is anybetter a maxim with regard to their motivations than itiswith regard to their physical condition.

If there is such a thing as what men are really like, itmay be that (in these sorts of respects, at least) it is notso different from what they are actually like; that is,creatures in whose lives moral considerations play animportant, formative, but often insecure role.

The amoralist, then, would probably be advised toavoid most forms of self-congratulatory comparison ofhimself with the rest of society. The rest may, ofcourse, have some tendency to admire him, or thosemay who are at such a distance that he does not treaddirectly on their interests and affections. He should notbe too ercouraged by this, however, since it is probablywish-fulfilment (which does not mean that they wouldbe like him if they could, since a wish is different froma frustrated desire). Nor will they admire him, still lesslike him, if he is not recognizably human. And thisraises the question, whether we have left him enough tobe that.

Does he care for anybody? Is there anybody whosesufferings or distress would affect him? If we say 'no*to this, it looks as though we have produced a psycho-


15/59

M O R A L I T Y

path. If he is a psychopath, the idea of arguing him intomorality is surely idiotic, but the fact that it is idiotichas equally no tendency to undermine the basis ofmorality or of rationality. The activity of justifyingmorality must surely get any point it has from the exist-ence of an alternative- here being something to justifyit against. The amoralist seemed important because heseemed to provide an alternative; his life, after all,seemed to have its attractions. The psychopath is, in acertain way, important to moral thought; but his impor-tance lies in the fact that he appals us, and we mustseek some deeper understanding of how and why heappals us. His importance does not lie in his having anappeal as an alternative form of life.

The amoralist we loosely sketched before did seem tohave possibly more appeal than this; one might picturehim as having some affections, occasionally caring forwhat happens to somebody else. Some stereotype from agangster movie might come to mind, of the ruthless andrather glamorous figure who cares about his mother,his child, even his mistress. He is still recognizablyamoral, in the sense that no general considerationsweigh with him, and he is extremely short on fairnessand similar considerations. Although he acts for otherpeople from time to time, it all depends on how he hap-pens to feel. With this man, of course, in actual factarguments of moral philosophy are not going to work;for one thing, he always has something he would ratherdo than listen to them. This is not the point (though itis more of a point than some discussions of moral argu-ment would lead one to suppose). The point is ratherthat he provides a model in terms of which we mayglimpse what morality needs in order to get off the

T H E A M O R A L I S Tground, even though it is unlikely in practice to get offthe ground in a conversation with him.

He gives us, I think, almost enough. For he has thenotion of doing something for somebody, because thatperson needs something. He operates with this notionin fact only when he is so inclined; but it is not itselfthe notion of his being so inclined. Even if he helpsthese people because he wants to, or because he likesthem, and for no other reason (not that, so far as theseparticular actions are concerned, he needs to improveon those excellent reasons), what he wants to do is tohelp them in their need, and the thought he has whenhe likes someone and acts in this way is 'they needhelp', not the thought 'I like them and they need help'.This is a vital point: this man is capable of thinking interms of others' interests, and his failure to be a moralagent lies (partly) in the fact that he is only intermit-tently and capriciously disposed to do so. But there isno bottomless gulf between this state and the basicdispositions of morality. There are people who needhelp who are not people who a t the moment he happensto want to help, or likes; and there are other people wholike and want to help other particular people in need.To get him to consider their situation seems rather anextension of his imagination and his understanding,than a discontinuous step onto something quite dif-ferent, the 'moral plane'. And if we could get him toconsider their situation, in the sense of thinking aboutit and imagining it, he might conceivably start to showsome consideration for i t: we extend his sympathies.And if we can get him to extend his sympathies toless immediate persons who need help, we might be ableto do it for less immediate persons whose interests have


16/59

M O R A L I T Ybeen violated, and so get him to have some primitivegrasp on notions of fairness. If we can get him all thisway, then, while he has no doubt an extremely shakyhold on moral considerations, he has some hold onthem; at any rate, he is not the amoralist we startedwith.

This model is not meant to sketch the outline of aconstruction of the whole of morality from the possi-bility of sympathy and the extensions of sympathy:that would be impossible. (Even Hume, who perhapscame nearest to it, did not attempt that. His system,among the many interesting and valuable things thatit contains, has a distinction between the 'natural' andthe 'artificial' virtues which is relevant to this point.)The model is meant to suggest just one thing : hat if wegrant a man with even a minimal concern for others,then we do not have to ascribe to him any fundarnen-tally new kind of thought or experience to include himin the world of morality, but only what is recognizablyan extension of what he already has. He is not veryfar into it, and it is an extensive territory : as we saw indrawing up the amoralist, you have to travel quite along way to get out of it. But the man with the extendedsympathies, the ability to think about the needs ofpeople beyond his own immediate involvement, isrecognizably in it.

It does not follow from this that having sympatheticconcern for others is a necessary condition of being inthe world of morality, that the way sketched is theonly way 'into morality'. It does not follow from whathas so far been said; but it is true.

Some of the considerations touched on here, aboutmoral and other motivations, we shall come back to

12

T H E A M O R A L I S Tlater. I shall now turn to someone who is also found

-upsetting by morality, but in a different way from theamoralist. This man is content that he should have a- .morality, but points out that other people have differ-ent ones - and insists that there is no way of choosingbetween them. He is the subjectivist.


17/59

S U B J E C T I V I S M : F I R S T THOUGHTS

SUBJECTIVISM : F I R S T TH O U G H TS

CONSIDERhree statements, each of which, in itsdifferent way, expresses a view that moral opinions,or moral judgements. or moral outlooks are 'merelysubjective' :

(a) A man's moral judgements merely state (or express)his own attitudes.(b) Moral judgements can't be proved, established, shownto be true as scientific statements can; they a n matt6rs ofindividual opinion.(c) There are no moral facts; there are only the sorts offacts that science or common observation can discover, andthe values that men place on those facts.

The three statements come very close to one another.and in discussions of subjectivism and objectivism oneoften finds versions of the three being used virtuallyinterchangeably. They are, indeed, genuinely related toone another. Yet they are significantly different. Thefirst, (a), expresses what might be called in a broadsense a logical or linguistic view : t purports to tell ussomething about what moral remarks are or do. Thesecond, @), introduces a set of notions not present inthe first, notions connected with the concept of know-ledge, and may be taken to express an epistemologicalview about moral judgements. The third statement, (c),is the vaguest and least tangible of the three, and showson its surface the danger of collapsing, partly or com-pletely, into one or other of the first two :which is whatmany philosophers would claim it must do. Yet in its

inadequate way it seems to gesture towards somethingwhich is closest of all to what has concerned manywho have been womed about moral objectivity: theidea that there is no moral order 'out there ' - out there,in the world, are only the sorts of things and the sortsof facts that science, and the more everyday processesof human enquiry of which science is a refinement,deal. Alles anderes ist Menschenwerk. Statement (c)can be said - using the word in an unambitious way -to express a metaphysical view.I The metaphysical view brings out most explicitlysomething that is latent in all the three statements, adistinction between fact and value. A central concern1 of much modem moral philosophy has been the dis-tinction between fact and vahe. One important form, that this concern has taken has been to emphasize the1 distinction, while rejecting the supposedly disquietingconsequences of it, by trying to show either that theyI are not consequences, or that they are not disquieting.Thii project of defusing subjectivism (as it might be

I called) can be expressed in terms of our three subjecti-I vist statements roughly as follows: that, in so far asthey are defensible, they come to the same thing; and

that what they come to is both not alarming and essen-tial to the nature of morality (the point that is essentialto the nature of morality seems sometimes to be thought,oddly, to imply just by itself the point that it is notalarming).

This project we shall follow, with interruptions, forsome whiIe. It starts as follows. Statement (a), first ofall, is either false or harmless. It is false if it claims thatmoral judgements state their utterer's attitudes. For ifthis were so, they would be simply autobiographical


18/59

M O R A L I T Yremarks, replaceable without loss by statements ex-plicitly of the form 'My attitude towards this . .'or 'Ifeel . ' But if this were so, there would be no inter-personal moral disagreements; two persons expressingwhat we would normally take to be conflicting viewswould not be expressing conflicting views at all, butwould be, rather, like two persons on a boat, one ofwhom says that he feels sick while the other says thathe, on the other hand, does not. But it is an evident factthat there are genuine moral disagreements, andthat moral views can conflict. Moral judgements must(at least to this extent) mean what we take them tomean; and what we take them to mean, the way we usethem, is such that they do not merely make autobio-graphical claims, but a sort of claim which is beingrejected by someone who utters a contrary moral judge-ment. Thus they do not merely describe the speaker'sown attitude.

This argument, however, does not dispose of the pos-sibility that moral utterances, while they do not directlydescribe, may nevertheless have the function of expres-sing, the utterer's attitude. This claim in itself is incon-testable and harmless; for in itself it comes to nomore than saying that a man who makes a moral judge-ment can (if that judgement is sincere) be said to beexpressing his attitude to a certain moral issue, and thisdoes not support any distinctively subjectivist view ofmoral judgements: a man who sincerely makes a fac-tual judgement may be said to be expressing his beliefabout a factual issue, but that does not support a sub-jectivist view of factual judgements The subjectivistinterpretation comes in when one says that one whomakes a moral judgement expresses his attitude, and

S U B J E C T I V I S M : F I R S T T H O U G H T Sthat is all there is to be said about it. In particular thesubjectivist force of (a) lies in a suggestion that there isno question of the attitudes expressed in moral judgemenu being right or wrong, whereas there is a ques-tion of the beliefs expressed in factual judgementsbeing true or false.

Now that the thesis has regained a distinctively subjectivist form it will be said that it is false. For - appeal-ing once more to the ways in which moral judgementsare actually made and treated - t is not true that thereis no question of moral attitudes being right or wrong.One of their distinguishing marks, as against mere ex-pressions of taste or preference, for instance in mattersof food, is that we take seriously the idea of a man'sbeing wrong in his moral views; indeed, the very con-cept of a moral view marks a difference here, leaning asit does in the direction of belief rather than of meretaste or preference. It is precisely a mark of moralitythat de gustibus non disputandum is a maxim whichdoes not apply to it.

To this it might be replied that the fact that moralattitudes can be called 'right' or 'wrong', and that thequestion of their rightness or wrongness is taken seri-ously, does not in any ultimate sense help to transcendsubjectivism. It shows not that moral attitudes aremore than (merely) attitudes, but that they are attitudeswhich we get disturbed about; that it matters to us tosecure similarity of attitude within society. The use ofthe language of 'right' and 'wrong' can be seen as partof the apparatus of securing agreement, marking offthose who disagree and so forth: it remains the case thatall we have are people's attitudes towards different sortsof conduct, personality, social institutions, etc.


19/59

MORALITYYet this account, in its turn, seems inadequate to the

facts: it must at least underestimate the logical com-plexity of the situation. Notably it fails to accountfor the undoubted facts that a man may be in a state ofmoral doubt, which he may resolve - that a man cannonarbitrarily change his mind about a moral matter,not merely in the individual case, but on a general issue,and for reasons. Thus a man previously convinced thata permissive attitude towards abortion was wrong,might change his view on this, and not merely (for in-stance) because he felt lonely in a group which held thepermissive attitude. No doubt many writers on moralphilosophy overestimate the extent to which people areled by rational considerations to modify their moralviews; those writers ignore the evident extent to whichattitudes are modified by factors such as the desire toconform with one group or nonconform with another- he groups themselves not being chosen in the light ofmoral reasons, but rather determined by the individ-ual's situation and needs. But this is, in one sense, be-side the point. For even if moral attitudes were rarelydetermined by reasons, and the reasons advanced intheir support were rationalizations, our model of moralattitudes and moral judgements must at least be com-plex enough to leave a place for the rationalizations. Itis only if the position to which a man is led by theseforces satisfies some conditions of being the sort of posi-tion to which reasons are relevant that we can under-stand it as a moral position at all.

However, even if moral attitudes are of a characterto admit of support and attack by reasons, and thedeployment of reasons in reaching a conclusion; never-theless it may be said that these activities are possible

SUBJECTIVISM: FIRST THOUGHTSonly within a framework of given assumptions. Ifpeople can argue one with another about an individualmoral issue or a question of principle, this is only be-cause there are moral attitudes in the backgroundwhich are not in dispute and in the light of which theargument goes on. Putting this point in a rather strongerform, it might be said that it is only about the applica-tion of accepted moral views that the argument goeson. So where there is no background of moral agreement, there can be no argument. At this point the sub-jectivist attitude can reappear, claiming now that allthat has been shown by the considerations about ex-changing reasons is that the morality of a man or asociety is to some degree general and systematic andthat general attitudes can be applied to less generalcases. When we get outside the framework of agreedgeneral attitudes, there is no further argument, and noway of showing any position to beright or wrong.

This comes very close to our second formulation, @),of subjectivism; we have been led to it, it seems, bymodifications of the first formulation. One might, how-ever, want to change one element that came naturallyinto the formulation of @), and, with that, one implica-tion of the term 'subjectivism'. For when one turns tothe issue of ultimate disagreements, it is natural to takeas the unit which holds a set of moral attitudes, thesociety rather than the individual - not in order tohypostasize societies, but to draw attention to thepoint that there are limits to the degree of ultimate dis-agreement that can exist within a society (for withoutsome degree of moral homogeneity it would not be asociety); but there are no limits, a t least of that kind, ondisagreement beween societies.


20/59

INTERLUDE: RELATIVISM

LET US at this stage of the argument about subjectivismtake a brief rest and look round a special view or assem-blage of views which has been built on the site of moraldisagreements between societies. This is relativism, theanthropologist's heresy, possibly the most absurd viewto have been advanced even in moral philosophy. In itsvulgar and unregenerate form (which I shall consider,since it is both the most distinctive and the most influ-ential form) it consists of three propositions: that'right' means (can only be coherently understood asmeaning) 'right for a given society'; that 'right for agiven sodety' is to be understood in a functionalistsense; and that (therefore) it is wrong for people in onesociety to condemn, interfere with, etc., the values ofanother society. A view with a long history, it was popular with some liberal colonialists, notably British ad-ministrators in places (such as West Africa) in whichwhite men held no land. In that historical role, it mayhave had, like some other muddled doctrines, a bene-ficent influence, though modem African nationalism maywell deplore its tribalist and conservative implications.

Whatever its results, the view is clearly inconsistent,since it makes a claim in its third proposition, aboutwhat is right and wrong in one's dealings with othersocieties, which uses a nonrelative sense of 'right' notallowed for in the first proposition. The claim thathuman sacrifice, for instance, was 'right for' the

INTERLUDE: RELATIVISM

Ashanti comes to be taken as saying that human sacrificewas right among the Ashanti, and this in turn as sayingthat human sacrifice among the Ashanti was right; i.e.,we had no business to interfere with it. But this last iscertainly not the sort of claim allowed by the theory.The most the theory can allow is the claim that it wasright for (i.e., functionally valuable for) our society notto interfere with Ashanti society, and, first, this is cer-tainly not all that was meant, and, second, is very dubi-ously true.Apart from its logically unhappy attachment of anonrelative morality of toleration or non-in erferenceto a view of morality as relative, the theory suffers in itsfunctionalist aspects from some notorious weaknessesof functionalism in general, notably difficulties thatsurround the identification of 'a society'. If 'society'is regarded as a cultural unit, identified in part throughits values, then many of the functionalist propositionswill cease to be empirical propositions and become baretautologies: it is tediously a necessary condition of thesurvival of a group-withcertain-values that the groupshould retain those values. At the other extreme, thesurvival of a society could be understood as the sur-vival of certain persons and their having descendants,in which case many functionalist propositions aboutthe necessity of cultural survival will be false. When inGreat Britain some Welsh nationalists speak of the sur-vival of the Welsh language as a condition of the sur-vival of Welsh society, they manage sometimes toconvey an impression that it is a condition of the sur-vival of Welsh people, as though the forgetting ofWelsh were literally lethal.

In between these two extremes is the genuinely inter-


21/59

MORALITYd n g territory, a province of informative socialscience, where there is room for such claims as that agiven practice or belief is integrally connected withmuch more of a society's fabric than may appear on thesurface, that it is not an excrescence, so that discourage-ment or modification of this may lead to much largersocial change than might have been expected; or, again.that a certain set of values or institutions may be suchthat if they are lost, or seriously changed, the people inthe society, while they may physically survive, will doso only in a deracinated and hopeless condition. Suchpropositions, if established, would of course be of firstimportance in deciding what to do; but they cannottake over the work of deciding what to do.

Here, and throughout the questions of conflict ofvalues between societies, we need (and rarely get) somemildly realistic picture of what decisions might be beingmade by whom, of situations to which the consid-erations might be practically relevant. Of variousparadigms that come to mind, one is that of conflict, suchas the confrontation of other societies with Nazi Ger-many. Another is that of control, where (to eliminatefurther complications of the most obvious case, colon-ialism) one might take such a case as that of the rela-tions of the central government of Ghana to residualelements of traditional Ashanti society. In neither casewould functionalist propositions in themselves provideany answers at all. Still less will they where a majorissue is whether a given group should be realistically ordesirably regarded as 'a society' in a relevant sense, orwhether its values and its future are to be integrallyrelated to those of a larger group - as with the case ofblacks in theUnitedStates.

22

INTERLUDE: RELATIVISMThe central confusion of relativism is to try to con-

jure out of the fact that societies have differing atti-tudes and values an a priori nonrelative principle todetermine the attitude of one society to another; thisis impossible. If we are going to say that there are ulti-mate moral disagreements between societies, we mustinclude, in the matters they can disagree about, theirattitudes to other moral outlooks. It is also true, how-ever, that there are inherent features of morality thattend to make it difficult to regard a morality as applyingonly to a group. The element of universalization whichis present in any morality, but which applies undertribal morality perhaps only to members of the tribe,progressively comes to range over persons as such. Lessformally, it is essential (as was remarked earlier) to mor-ality and its role in any society that certain sorts ofreactions and motivations should be strongly interna-lized, and these cannot merely evaporate because one isconfronted with human beings in another society. Justas de gustibus non disputandum is not a maxim whichapplies to morality, neither is 'when in Rome do as theRomans doD,which is at best a principle of etiquette.Nor is it just a case of doing as the Romans do, butof putting up with it. Here it would be a platitude topoint out that of course someone who gains wider ex-perience of the world may rightly come to regard somemoral reaction of his to unfamiliar conduct as parochialand will seek to modify or discount it. There are manyimportant distinctions to be made here between thekinds of thoughts appropriate to such a process in dif-ferent cases: sometimes he may cease to regard a cer-tain issue as a moral matter at all, sometimes he maycome to see that what abroad looked the same as some-

23


22/59

M O R A L I T Y I N T E R L U D E : R E L A T I V I S M

thing he would have deplored at home was actually, inmorally relevant respects, a very different thing. (Per-haps - though one can scarcely believe it - there weresome missionaries or others who saw the men in apolygamous society in the light of seedy bigamists athome.) But it would be a particular moral view, and oneboth psychologically and morally implausible, to in-sist that these adaptive reactions were the only correctones, that confronted with practices which are foundand felt as inhuman, for instance, there is an a prioridemand of acceptance. In the fascinating book by Ber-nal de Diaz, who went with Cortez to Mexico, there isan account of what they all felt when they came uponthe sacrificial temples. This morally unpretentious col-lection of bravos was genuinely horrified by the Aztecpractices. I t would surely be absurd to regard this reac-tion as merely parochial or self-righteous. I t rather in-dicated something which their conduct did not alwaysindicate, that they regarded the Indians as men ratherthan as wild animals.It is fair to press this sort of case, and in general thecases of actual confrontation. 'Every society has itsown standards' may be, even if confused, a sometimesuseful maxim of social study; as a maxim of socialstudy it is also painless. But what, after all, is one sup-posed to do if confronted with a human sacrifice?- nota real question for many of us, perhaps, but a real ques-tion for Cortez. 'I t wasn't their business,' it may besaid; 'they had no right to be there anyway.' Perhaps -though this, once more, is necessarily a nonrelativemoral judgement itself. But even if they had no rightto be there, i t is a matter for real moral argument whatwould follow from that. For if a burglar comes across

the owner of the house trying to murder somebody, ishe morally obliged not to interfere because he is tres-passing 1None of this is to deny the obvious facts that manyhave interfered with other societies when they shouldnot have done; have interfered without understanding:and have interfered often with a brutality greater thanthat of anything they were trying to stop. I am sayingonly that it cannot be a consequence of the nature ofmorality itself that no society ought ever to interferewith another, or that individuals from one society con-fronted with the practices of another ought, if rational,to react with acceptance. To draw these consequences isthe characteristic (and inconsistent) step of vulgar rela-tivism.


23/59

SUBJECTIVISM : F U R T H E R T H O U G H T S

T H E enticements of vulgar relativism are not verytempting. But its central error is quite important; andthat it is an error has a significance beyond th at partic-ular doctrine, concerned as that is with the relationsbetween societies. It is possible for someone persuadedof subjec tivist views to cease to c are a bou t mora l issues.(This is d ifferent from ceasing to regard som ething as,in itself, a moral issue. Thus it is both possible and rea-sonable to suppose that there is no distinctively sexualmorality, in the sense of moral considerations thatgovern sexual relationships and nothing else; while ad-mitting the extremely obvious fact that sexual relation-ships are profoundly and specially liable to give rise tomoral issues, of trust, exploitation, unconcern for theinterests of third parties and so forth.) A man con-fronted with some monstrous political injustice, for in-stance, may feel no confidence in protesting or fightingagainst it because, as he says, 'Who's to judge?,' or 'It'sonly m y feelings against theirs,' or som ething like that.In so far as there is any traceable intellectual link be-tween the subjectivism and the indifference, it mustinvolve something like the relativist error. For thethought must be something like this : Because subjec-tivism is true, I am not justified in protesting.' If this isright, it must be so either because, if subjectivism istrue, no one is justified in doing anything; or because,if subjectivism is true, he is specially not justified in

S U B J E C T I V I S M : F U R T H E R T H O U G H T Sprotesting. If the first of these, then the inflicters of theinjustice are not justified in wha t they are doing either,nor is this man justified in not protesting, and theseconsiderations remove any basis he was supposed tohave for his indifference. In any case, the argument inthis version has surely gone to o far, even for subjectiv-ism; since subjectivism did not claim it to be impos-sible to think that anything was justified, but only thatwhen a given man thinks something is justified, he can-not in the end be proved wrong. The man we are argu-ing with has somehow got from that to a position inwhich no one (however subjectively) can think some-thing justified, and that must be a mistake; unless, ofcourse, subjectivism is inconsistent, in which case s u bjectivism is false, and the argument is over anyway .Let us try the second alternative, then, that becausesubjectivism is true, he is specially not justified in pro-testing. W hy should this be so ? 'Well', he may say,'they think they are right, and who am I to say thatthey are wrong?' But the apparent force of this is en-tirely gained from its subtly moving out of the subjec-tivist arena and importing the idea that there is such athing as objective rightness, only he is not sure whe-ther these other people's actions possess it or not.Sticking to the sub jectivist path, he mu st recognize tha tif he chooses to think t hat they are wrong and that he isright in protesting, then no one can say he is wrongeither, and he can be no less justified in protesting thanthey are in doing w hat they are doing. Another wa y ofputting this point is this : perhaps they are right' m ustbe one of his moral thoughts. If he also has the thou ght'They are wrong (only I am not justified in protesting)',he has inconsistent moral thoughts within his ow n sys-


24/59

MORA L I T Ytem, and subjectivism never required him to do that. Infact, of course, the thought 'Who am I to say that theyare wrong?' is one that is had, not within his own sub-jective compound, but in mid-air above his own and theother people's; it tries to stand outside all moral posi-tions (including the thinker's own) and yet still be amoral thought. But this mid-air place, by subjectivismitself, is not a place in which anyone can have a moralthought.

These sorts of reasons show why the defusers of subjectivism say that it leaves everything where it was andcannot possibly logically imply indifferentism or anyother practical attitude. (They move perhaps ratherrapidly from that to the claim that their views cannotencourage any such attitudes, perhaps on the question-able ground that I cannot be responsible for muddlesthat people get into as a result of what I say.) But doessubjectivism leave everything where it was? Surely noteverything. The dialectical skeins we just went throughare, I think, perfectly valid, as showing that indifferen-tism could not follow from subjectivism. But the reasonwhy this is so is that the argument for indifferentismrequires, and subjectivism forbids, the mid-air position.-But in forbidding the mid-air position, subjectivismseems to have taken something away, for we at leastseem to recognize the mid-air position (for instance, insome of subjectivism's own statements). Another wayof putting our bafflement is that we seem in these argu-ments to have been given no special reason why themid-air position is debarred to morality. It was said thatthe mid-air position was no place for a moral thought.But it does seem to be a place for some sorts of thought,

SUBJECTIVISM: FURTHER THOUGHTS

in particular, factual thought. Indeed, subjectivismitself (see (b) and (c) of our original formulations, givenabove), insists on the contrast between moral attitudesand factual beliefs, regarding the latter as 'objective' ina way that the first are not. The subjectivist leaves us,and may mean to leave us, with an uneasy feeling thatfactual beliefs have got, and moral beliefs lack, some-thing that i t is nice to have: that factual beliefs andscience are somehow solider than morality.But still, it will be said, subjectivism leaves every-thing where i t was, so far as morality is concerned,though not perhaps so far as muddled feelings aboutmorality are concerned. Even granted the contrast ofsolidity we cannot draw any practical conclusions.We cannot, in particular, conclude (as some today areobviously inclined to conclude) that since science isobjective and morality is not, we are objectively justi-fied in devoting ourselves to science, while only sub-jectively justified in protesting against injustice. Fordevoting oneself to science is as much a practical activ-ity as any other, and there is no more reason why that

one should be objectively justified rather than anyother. Justifications for doing objective subjects are notobjective justifications for doing those subjects, anymore than the fact that there are deductive justifica-tions of the theorems of Principia Mathemqtica meansthat there are deductive justifications of the projects ofreading, rehearsing, or discovering the theorems ofPrincipia Mathematica. All these are instances of thatstrangely tempting fallacy, the 'fat oxen' principle:who drives fat oxen must himself be fat.Equally, though more subtly, the fact that 'prag-matic' political policies (i.e., policies which apply29


25/59

MORALITYsophisticated technical considerations to the pursuit ofself-interest) can be evaluated one against another withgreater expectations of expert agreement than moreidealistic policies often achieve, does not, even if true,show that we are on more objective and solid ground inadopting pragmatic rather than idealistic policies. Oncewe have adopted them we may be on more objectiveground in working them, i.e., in deciding what to donext; but we never get on to more objective groundwith reference to the question whether we ought to bedoing this sort of thing at all. We are merely more com-fortable, and if among the sophisticated experts, inwork.

So subjectivism, even with its unsettling contrastbetween morality and science, is still not logically com-mitted to producing practical consequences. It cannoteven yield the conclusion that we are more solidly justi-fied in having factual beliefs than moral ones; or onmore objective ground in pursuing a factual questionthan pursuing a moral one; or objectively justified inseeking the truth about anything; or in trying to findscientific explanations of phenomena rather than rest-ing content with superstition. For all it gave us was thatfactual and scientific beliefs were objective; that weshould seek factual or scientific beliefs is not itself afactual or scientific belief.

Perhaps the subjectivist will readily grant thesepoints. The unsettling contrast which he insists onbetween factual enquiry and moral thought is not a con-trast in the grounds for starting on or pursuing thoseactivities - rather, it is a contrast in what those twosorts of activities are, in what goes on wher, one is en-gaged in them. In particular, it is a contrast concerning

SUBJECTIVISM: FURTHER THOUGHTSthe nature and extent of disagreement which people en-gaged in those activities may, in the nature of the case,encounter. So let us look more closely at what the sub-jectivist has to say, in this central respect, about theunsettling contrast.

He may start by saying that if we engage in factualor scientific enquiry, then, facts being as they are, weare bound to reach some agreed scientific or factual be-liefs; but we are not bound, because we engage in moralthought, and the facts are so, to reach some agreedmoral beliefs. There is one element in this answer whichhas to be corrected straight away. For it is of course not

1 true that if the facts are so, we are bound (granted fac-tual thought) to reach some agreed factual beliefs: theI facts may be hidden from us. The most we can say isthat if we recognize that the facts are so, we are boundto reach some agreed factual beliefs. And this is a tauto-logy, since our recognizing the facts to be so entailsour reaching some agreed factual beliefs. Perhaps thesubjectivist can improve on this unexciting proposi-tion by saying something like this: that if two obser-

b vers are in the same observational situation, and haveL the same concepts and are not defective as observers,

etc., then they will reach the same factual beliefs aboutthat situation. If the 'etc.' in this can be handled so as tomake the statement true, i t will almost certainly emergeas necessarily true as well. All right, says the subjec-tivist; but this is not necessarily true, indeed is not trueat all: that if two observers are in the same observa-tional situation, have the same concepts, etc., then theyare bound to reach the same moral beliefs - and thereis the contrast. But, we may say, are we sure that thislatter is not true if we posit that they have the same


26/59

MORALITY

moral concepts - which is surely the fair parallel? No,it is not true even then, the subjectivist says; for if theyhave, as they might, rather minimal moral concepts,such as merely the concept of what one ought to do,they can agree on all the facts and disagree morally.This is a central position of many philosophers whoinsist on the fact/value distinction. They represent dif-ferent moral outlooks as all using some common,skeletal, moral concept and giving it different fillings orcontents. Now we might point out that a lot of moralthought does not operate with such impoverished con-cepts, that we standardly think in terms of more complexideas of virtues, types of wrong action, etc., notions oftheft, for instance, or cowardice, or loyalty, or the dutiesof one's job. With those more substantial concepts, thereis much greater hope that if we use the same concepts,we will reach agreement, or at least the sorts of dis-agreement which we may reasonably expect, and whichare less unnerving. But the subjectivist will say thatthere is this asymmetry, that in scientific or factualcontexts, if two observers have different concepts theymay ultimately find either that the concepts are infact equivalent, or else find reason for preferring one tothe other in terms of predictive success, explanatorypower, and so forth, and there is no parallel to this in themoral case.Even if we abandon the ndive view (which some sub-jectivists, in working their contrast, are attached to)that science 'proves' things; even if we accept thatwhat science does is eliminate hypotheses and thatthere are infinitely many hypotheses which have neverbeen eliminated because they are too dotty for anyoneto bother to test them (a point I have heard Hilary Put-

32

SUBJECTIVISM: FURTHER THOUGHTS

nam make); nevertheless, there are eliminatory proce-dures which must be respected by persons within thescientific framework, and even the conceptual disagree-ments may with luck yield to impersonally acceptedcriteria. Yet for some disagreements between personsor groups both engaged in recognizably moral thought,no such procedures exist. The contrast - some contrast- exists.But why should it not? This, finally, might be thepoint at which a philosopher who had been arguingwith the subjectivist all this time might at last turnround and say: of course the contrast exists; moralityis not just like science or factual knowledge, and i t isessential that it should not be. The point of morality isnot to mirror the world, but to change it; it is concernedwith such things as principles of action, choice, respon-sibility. The fact that men of equal intelligence, factualknowledge, and so forth, confronted with the samesituation, may morally disagree shows something aboutmorality - that (roughly) you cannot pass the moralbuck on to how the world is. But that does not show (assubjectivism originally seemed to insinuate) that thereis somethingwrong with it.Some such statement - and I have only blocked it invery roughly - will express the culmination of the pro-ject of what I called 'defusing' subjectivism. It has, Ithink, to be granted some success. The most obviousways in which somebody might be unnerved by sub-jectivism seem to have been blocked. Thus, to revertonce again to the indifferentism we discussed before,the 'defuser' can arrange some of the subjectivist'smaterials into the following argument. We observe thatwhen men of equal scientific or historical competence,


27/59

MORAL ITYequal perceptual and intellectual abilities, etc., disagreestrongly about some scientific or historical matter, thereis good reason for them to stop disagreeing so strongly,and recognize something which their very disagreement,granted their knowledge and skills, reveals, namely thatthe matter is uncertain : t is rational for them, and thirdparties, to suspend judgement. One may be tempted tothink that the same should obtain in cases of moraldisagreement; but this will be a mistake. I t depends onfirst contrasting morality and factual knowledge, andthen assimilating them. For the vital difference is that thedisagreement in morality involves what should be done,and involves, on each side, caring about what happens;and once you see this difference, you see equally that itcould not possibly be a requirement of rationality thatyou should stop caring about these things because some-one else disagrees with you.

This, and the similar arguments, do seem to me toshow that the defusing operation, in certain vital re-spects, has succeeded. Has it totally succeeded? If it has,then we were wrong, some time back, in feeling uneasybecause subjectivism banned, for morality but not forfactual beliefs, something we called the 'mid-air posi-tion'. But, I think, we were not entirely wrong, in feel-ing that unease - perhaps we can now see why.

If I have a factual disagreement with a man, I mightthink: 'I believe that p, he believes that not-p. Perhapshe is right.' The natural way of taking this is as an ex-pression of doubt, of somewhat shaken confidence; theway I mean 'perhaps he is right' implies that while Istill believe that p, I do not believe it as strongly as allthat. In this way of taking it, it can be painlessly paral-leled for the moral case; for if I think 'perhaps he is

SUBJECTIVISM: FU RT H E R THOUGHTSI right,' this will naturally be taken as an expression of

somewhat shaken confidence also. But now, in the fac-tual case, there is a possible thought which seems rathersimilar, but is not the same: the thought 'I am con-vinced that p, but it is possible for all that that not-p,'

I where this is not an expression of doubt at all, butrather registers the impersonal consideration that how) things are is independent of my belief; however theyI are, they are, whatever I believe. We do not know exact-

I

ly what content to assign to this thought, but unlesswe have the most drastic philosophical views, we areconvinced that it has a content: and the defuser willagree. But even defused subjectivism leaves no parallelthought possible on the moral side: for subjectivism,however defused, there just is no content to 'I am con-vinced that racial discrimination is intrinsically wrong,but it is possible for all that, that it is not,' except thingslike 'How convinced am I ? ' or 'I suppose somebodymight make me change my mind.'

Such a contrast (and it, and related contrasts, needproper investigation, which we cannot attempt here)might make us agree with the third subjectivist formulawe originally introduced: there are no ethical facts.Yet once more the defuser will say : his is just one moreformulation of what I said already, and of somethingessential, not detrimental, to morality. For I have al-ready said that moral thought is essentially practical; itis not its business to mirror the world. - But now wemight reply: you said that it was not its business tomirror the world of empirical facts, and we agreed. Butdid we agree that it mirrored no facts at all? And herethe locus of our dissatisfaction may become clearer inthe thought that the reason why even defused subjectiv-


28/59

MORA L I T Y

ism seems to have left something out is that moralthinking feels as though it mirrored something, asthough it were constrained to follow, rather than befreely creative. When we see further that many defus-ing philosophers express the essential difference be-tween factual and moral thought in terms of a contrastbetween the intellect and the will, and represent theresponsibilities of morality in terms of our deciding oncertain moral principles - then we have reason to bedissatisfied, either with them, or, if they are right, withmoral thought. For certainly the consciousness of aprinciple of action as freely decided upon is very unlikethe consciousness of a moral principle, which is ratherof something that has to be acknowledged. If it is thensaid that there is just a psychological explanation ofthat - then moral thought seems a cheat, presentingitself t o us as too like something which it is not.These remarks only gesture towards a centre of dis-satisfaction. They leave almost everything to be done:and not perhaps first in moral philosophy. For instance,we need to distinguish two things run together in whatwas just said, the idea of realism - hat thought has asubject matter which is independent of the thought -and the idea of thought being constrained to certainconclusions. Thus mathematical thought has the latterproperty, but it is a deep and unsettled question in thephilosophy of mathematics how far we should or canthink in realist terms about the subject matter of math-ematics.Here I shall leave the direct discussion of subjectivismand issues that it raises, with the conclusion that de-fused subjectivism does not leave everything where itwas, but that it does leave more where it was than we

36

SUBJECTIVISM: FURTHE R THOUGHTS

might have thought when we started. If subjectivism,however, defused, is true, things are with morality notquite as they seemed; but the fraud, we might say, just-ifies at most resentment rather than panic. We shall nothowever lose sight of the idea of constraints on moralthought, the limitations on the creation of values. Weshall come round to them by another route, throughthe idea of goodness. To talk about goodness, we shallstart with 'good'.


29/59

'GOOD'

THEuse of the word 'good' has provided a focus formany discussions of basic issues in moral philosophy;while it would be a mistake to think that this word, orits approximate equivalents in other languages, couldpossibly bear by itself the weight of the issues, never-theless the consideration of it provides a useful leverfor lifting up some of them. We shall start with somelogical considerations: these will lead us to things ofgreater moral substance.

As Aristotle observed, 'good' is used of many differ-ent sorts of things, of things indeed in different cate-gories. While in one way we do not mean the samewhen we apply it to these Mer en t sorts of things - nthis sense, that what makes a general a good general isdifferent from what makes a doctor a good doctor -nevertheless the word is not just ambiguous: we couldnot tidy language up and say just what we want to saynow by replacing 'good' with a different expression ineach of these occurrences.

More than one theory in recent times has tried to pro-vide a model to show that 'good' is genuinely unam-biguous. One such attempt was that of G. E. Moore,who claimed that goodness was a simple indefinableproperty like yellowness, but that, unlike yellowness,it was nonnatural - hat is to say (roughly), it was notthe sort of property whose presence or absence could beestablished by empirical investigation, although (in a

38

way left very obscure by his theory) observation of athing's empirical characteristics was doubtless relevantto the apprehension of goodness.

Apart from its evidently mysterious and unexplana-tory character, there is a logical objection to Moore'saccount. A very important feature of 'good' is that, inmany of its occurrences, it functions as an attributiveand not as a predicative adjective (as P. T. Geach hasexpressed the distinction).* 'Yellow', for instance, is apredicative adjective, because a sentence suchasThat isa yellow birdadmits of the analysisThat is a bird and it isyellow.By the same token, from the two sentencesThat is a yellow birdA bird is an animalwe can infer the conclusionThat is a yellow animal.But the sentenceHe is a good cricketercannot be analyzed asHe is a cricketer and he isgoodnor can we validly infer fromHe isa good cricketerA cricketer isamantoHe is a goodman.*P.T.Geach, 'Good and Evil'. Analysis, Vol. 17 1956).


30/59

MORAL I TYAn adjective which has this latter characteristic, thatit is logically glued to the substantive it qualifies, maybe called an attributive adjective; or, more precisely, ause in which it is so glued may be called an attributiveuse of it. Now Moore's account claimed that 'good'was like 'yellow' in standing for a simple quality,though unlike it in that the quality was nonnatural;and, mysterious as this is, it must at least imply thatthe logical behaviour of 'good' as an adjective must be

like that of 'yellow'. But it is not, and so Moore's ac-count must be rejected not just as unilluminating, butas radically misguided.Another important attributive adjective is 'real' - anassertion that something is real can be understood onlyif we can answer the question 'a real what?'. This isillustrated by the situation in the art world, in whichcollectors are interested in acquiring the work of cer-tain forgers, so that i t can be in people's interests toforge forgeries: thus the question might arise whetherthis picture was a real Van Meegeren, everyone know-ing that at any rate it was not a real Vermeer.The characteristic of attributiveness needs, however,deeper exploration if we are to get an understanding of'good'. We can see by the tests that 'large' is attribu-tive : hus there is no valid inference from

(a) This is a large mouse@)A mouse isan animalto(c)This isa large animal.The explanation of the failure of this inference, andof the attributiveness of 'large', is clear - it is that

40

'GOOD''large' is a comparative term, and 'This is a largemouse' means something like 'This is a mouse largerthan most mice'. By a similar analysis the conclu-sion (c) means something like 'This is an animal largerthan most animals', and we can see why the inferencefails. What can be validly inferred from these premises,on the comparative analysis, is(d)This is an animal larger than most miceand indeed this conclusion does follow.

Is the attributiveness of 'good' to be explained in thesame way? I t can hardly be that all attributivenesscould be so explained - a real Van Meegeren is not onemore real than most Van Meegerens. But it is more plau-sible to suggest that 'good' in 'good F' is attributivebecause 'good F' means something like 'better thanmost Fs'. But further consideration shows that this willnot do. We have just seen that on the comparative ana-lysis of 'large' we can get validly from the premises(a) and (b) to the conclusion (d). If 'good' were attribu-tive because it is comparative, then similarly we couldget fromHe is a good cricketerA cricketer isa mantoHe is a man better than most cricketersbut this conclusion does not follow, and is just as ob-jectionable as the original one, 'He is a good man'. Weshould get nearer to an acceptable conclusion only withsomething likeHe is a man better at cricket (or better as a cricketer) thanmost cricketers

4 1


31/59

MORALITYand even this is dubious, since there seems no contra-diction in the idea that, so flourishing is the state of thegame, most cricketers are pretty good; if this is possible,the comparative analysis in anything like its presentform disappears altogether. But apart from that, thecomparative analysis in any case has not unstuck'good' from its substantive; in the conclusion above,the connection persists, with 'good' now in its compar-ative form. So the attributiveness of 'good' demands amore intimate coilnection with its substantive than isdemanded in the case of a merely comparative adjec-tive like 'large'.Since 'good' in this sort of construction is intimatelyconnected with the substantive that it qualifies, themeaning of a phrase of the form 'a good x' has tobe taken as a whole; and its meaning is partly deter-mined by what fills the place of 'x'. Can we go furtherthan this and say that in phrases of this form, the mean-ing of the whole is essentially determined by the mean-ing of what takes the place of 'x' ? In many cases, itlooks as though we might take this further step. For ifwe consider functional descriptions of artefacts, suchas 'clock' or 't in opener', or again descriptions of hu-man beings which refer to their roles or jobs or skilledactivities, such as 'gardener' or 'general' or 'cricketer',it does seem that if one understands these expressions(at least in the strong sense that one understands whata tin opener is, for instance, or what a general does),then one has understanding, within limits, of what agood thing of that sort is.This understanding may be at a very general and ab-stract level and there will be a lot of room for disagreement, and for sophisticated comparisons of merits,

'GOOD'

within the broad and abstract framework. In particular,there can be differing or changing views of what aspectsof a human activity are to have what sort of weight inthe evaluation: hus one person, or one age, may take af diBerent view from another about how important itis for a good general to gain his victories at a minimumloss of life. Nevertheless, an understanding of what anx is does seem, in these cases, to contain a general un-derstanding of the criteria appropriate to saying thatsomething is a good x: we are not just free to inventcriteria of goodness. The clearest cases of all are, ofcourse, those of technical descriptions of artefacts; ifsomeone went into an aircraft factory and said 'that'sa good aerofoil', with reference to a rejected prototypewhich was in fact ill-designed or ill-executed, he wouldjust have made a mistake; and if he then explained thathe said this because its shape or its polish took his fancy,this would not make his remark any better, becausethese are not criteria of being a good aerofoil, thoughthey may well be appropriate to some other evaluationof this piece of metal, for instance as an aesthetic object.(This illustrates once more the importance, with evalua-tion, of the question, under what concept is the thingbeing evaluated.)There is a powerful tradition in contemporary philo-sophy of resistance to the idea that criteria of value.what makes a thing of a certain sort a good thing ofthat sort, could ever be logically determined by factualor conceptual truths : his is a central application of the

I distinction between fact and value which we have al-ready referred to. This resistance was influentially en-couraged by Moore, who invented the phrase 'the1 naturalistic fallacy' for a mistake allegedly committed


32/59

MORA L I T Y

by any view which held that the goodness of a thingcould be identified with some set of empirical, or in-deed metaphysical, characteristics. Moore exposed thismistake in terms of his own view that goodness was anonnatural property, a view which we have alreadyseen to be, in so far as it is comprehensible at all, a saderror. Many modern philosophers who agree thatMoore was in error in that view, nevertheless agree withhim that 'the naturalistic fallacy' is a genuine and im-portant fallacy; they give a new explanation of its na-ture. This explanation, crudely put, comes to somethinglike this: that the function of statements of the form'that is a good x' is to prescribe or commend, or to per-form some such linguistic purpose in the general rangeof the normative or evaluative, whereas merely to de-scribe the characteristics of x is not to perform such apurpose; and no set of statements which do not performsuch a purpose can logically entail any statement whichdoes. To prescribe, commend, etc., is to do something,which (to put it roughly) the facts by themselves cannotmake us do; we must have some evaluative or prescriptive attitude which favours certain characteristics, ifthose characteristics are to count with us as grounds ofapproval. Merely knowing about the world, or under-standing concepts, cannot in itself be enough to bringthis about.A full examination of this position requires some-thing that would take us too far in this essay, an en-quiry into an important and developing field in thephilosophy of language, the theory of speech-acts, thevarious things that we can do by making utterances.Three points can be briefly made. First, there cannotbe any very simple connection between speech-acts

44

such as commending or prescribing on the one hand,and the meaning of sentences such as 'this is a good x'on the other. At best, the utterance of those sentenceswill constitute an act of commendation, etc., only ifthe sentences are actually asserted; but the sentence hasthe same meaning whether it is asserted or not. Thus weunderstand the sentence 'This is a good film' in the con-text 'If this is a good film, it will get an Oscar'; but inthat context i t is not asserted, and no actual act of com-mendation has occurred.* Thus the connection betweenmeaning and commendation must at least be indirect.Second, the theory seems too readily to assume thatthe functions of commending, etc., and of describing,are exclusive of one another. Yet one and the same ut-terance may carry many speech-acts at once: if I say'Tomorrow will be wet', I may at once have describedtomorrow's weather, made a prediction, given you awarning, etc. Further, the question of whether I do, inmaking a certain utterance, perform any of thesespeech-acts, may be determined by what I say togetherwith the facts of the situation. Thus if I say 'The ice isthin'. I will have described the ice, no doubt, but also,in the light merely of your interests and purposes, mayhave done something that counts as giving you awarning. Rather similarly (though not exactly simi-larly), if I say, in a description of this clock, that itkeeps the time exactly, needs no winding, never breaksdown, etc., I will have done something like commend-ing it as a clock quite independently of some supposedchoice of criteria for clocks on my part. Of course, the

*See J. R. Searle, 'Meaning and Speech-Acts', PhilosophicalReview, Vol. 71 (1962). and Speech Acts (Cambridge UniversityPress, 1969),Chapter6.

Moore: a falcia

naturalista


33/59

MORALITYfacts about this clock, and the general nature of clocks,cannot bring it about that 1 make the remark - I mayremain silent: but they determine that if I give a truedescription of this clock in these respects, then I dosomething in the range of commending or giving a fav-ourable evaluation of the clock.

This last point brings us to the third consideration.Activities such as commending, etc., are essentiallyovert activities, connected with actual utterances; thisis why I have said that their study belongs to the theoryof speech-acts. But no account of sentences containing'good' could possibly be adequate which remainedmerely at this level; for i t is possible merely to think, orbelieve, or reach the conclusion that something is goodof its sort, without making an utterance to that effect atall. Hating Bloggs, I

williams, b. morality_an introduction to ethics

Documents