Parfit - From Reasons and Persons

7/30/2019 Parfit - From Reasons and Persons

1/14

8From Reasons and Persons

Derek Parfit

10. How Consequentialism is Indirectly Self-defeatingMost of my claims could, with little change, cover one .gr.oup of m o r ~ ltheories. These are the different versions of Consequentzalzsm, or C. C scentral claim is

V (C1) There is one ulthnate moral aim: that outcomes be as good aspossible.

C applies to everything. Applied to acts, C claims bothWhat each of us ought to do is whatever would make theoutcome best, an dIf someone does what he believes will make the outcome worse,he is acting wrongly.

I distinguished between what we have most r e ~ s o n to do, and w ~ a t itwould be rational for us to do, given what we beheve, or ought to l?,eheve.We 1nust now distinguish between what is objectively and subjectively i ~ h tor wrong. This distinction has nothing to do with whether moral theones . .can be objectively true. The distinction is between what smne theory ,implies, given (i) what are or would have been the. effects of w ~ a t someperson does or could have done, an d (ii) what th1s person beheves,ought to believe, about these effects.

Derek Parfit, Reasons and Persons (Oxford: Oxford University Press, 1984),24-43.

From Reasons and Persons 133It ma y help to mention a similar distinction. The medical r e a t m ~ t that

is objectively right is the one that would in fact be best for the patient.The treatment that is subjectively right is the one that, given the medicalevidence, it would be most rational for the doctor to prescribe. As thisexample shows, what it would be best to know is what is objectively right.The central part of a moral theory answers this question. We need anaccount of subjective rightness for two reasons. We often do not knowwhat the effects of our acts would be. And we ought to be blamed fordoing what is subjectively wrong. We ought to be blamed for such actseven i f they are objectively right. A doctor should be blamed for doingwhat was very likely to kill his patient, even i f his act in fact saves thispatient's life.

In most of what follows, I shall use right, ought, good, and bad in theobjective .sense. But wrong will usually mean subjectively wrong, or blame-worthy. Which sense I mean will often be obvious given the context. Thusit is clear that, of the claims given above, (C2) is about what we oughtobjectively to do, and (C3) is about what is subjectively wrong.To cover risky cases, C claims

(C4) Wh1;1t we ought subjectively to do is the act whose outcome hasthe greatest expected goodness.

In calculating the expected goodness of an act's outcmne, the ~ l u e of eachpossible good effect is multiplied by the chance that the act will produce'it:' The same is done with the disvalue of each possible bad effect. Theexpected goodness of the outcome is the sum of these values minus thesedisvalues. Suppose, for example, that i f I go West I have a chance of 1 in. 4of saving 100 lives, and a chance of 3 in 4 of saving 20 lives. The expect edgoodnE1ss of my going West, valued in terms of the number of lives saved,is 100 x 1/4 + 20 x 3/4, or 25 + 15, or 40. Suppose next that, if I go East,I shall certainly save 30 lives. The expected goodness of my going East is30 x 1, or 30. According to (C4), I ought to go West, since the expectednumber of lives saved would be greater.

Consequentialism covers, not just acts and outcomes, bu t also desires,dispositions, beliefs, emotions, the colour of our eyes, the clilnate, an d

. eyerythi ng else. More exactly, C covers anything that could make out

. better or worse. According to C, the best possible climate is the onefp.at would make outcomes best. I shall again us e 'motives' to cover both

. ~ s i r e s and dispositions. C clain1sThe best possible motives are those of which it is true that, if wehave them, the outcmne will be best.


2/14

134 Derek ParfitAs before, 'possible' means 'causally possible'. An d there would be manydifferent sets of motives that would be in this sense best: there would beno other possible set of motives of which it would be true that, if we had .this set, the outcome would be better. I have described some of the waysin which we can change our n1otives. (C2) implies that we ought to try tocause ourselves to have, or to keep, any of the best possible sets ofmotives. More generally, we ought to change both ourselves, and anything else, in any wa y that would make the outcome better. If we believethat we could make such a change, (C3) implies that failing to do so wouldbe wrong.1

S( To apply C, In.ust ask wl:a_t ~ a k ~ s outc?mes better or _worse. T_he.._!! simplest answer IS given by Utzlztarzanzsm. Th1s theory combines C Withthe following claim: the best outcome is the one that gives to people the

,' greatest net sum of benefits minus burdens, or, on the Hedonistic versionof this claim, the greatest net sum of happiness minus misery.

"' There are many other versions of C. These can be pluralist theories,'l'~ ; : / appealing to several different principles about what makes outcomesbetter or worse. Thus, one version of C appeals both to the Utilitarianclaim an d to the Principle of Equality. This princip le claims that it is ba di( through no fault of theirs, some people are worse off than others. Onthis version of C, the goodness of an outcome depends both on how greatthe net su m of benefits would be, and on how equally the benefits andburdens would be distributed between different people. One of twooutcomes might be better, though it involved a smaller sum of benefits,because these benefits would be shared more equally.

A Consequentialist could appeal to many other principles. Accordingto three such principles, it is bad if people are deceived, coerced, and

J betrayed. And some of these principles essentially e f ~ r to past events.'i ; Two such principles appeal to past entitleinents, an d to JUSt deserts. The

~ " " " ' P r i n c i p l e of Equality may claim that people should receive equal shares,-l-! t .not at particular times, bu t in the whole of their lives. If t makes this claim,:!: . j; his principle essentially refers to past events. If ou r moral theory contains; - ; such principles, we are not concerned only with consequences in the narrow( ~ s e n s e : with what happens after we act. But we can still be, in a wider sense,

f 1 Consequentialists. In this wider sense our ultimate moral ahn is, not thatoutcomes be as good as possible, bu t that history go as well as possible.What I say below could be restated in these tenns.

With the word 'Consequentialism',and the letter 'C', I shall refer to allthese different theories. As with the different theories about self-interest,i t would take at least a book to decide between these different versions ofC. This book [Reasons and Persons] does no t discuss this decision. I discuss

From Reasons and Persons 135only what these different versions have in common. My r g u m ~ ~ t s andconclusions would apply to all, or nearly all, the plausible theories of thiskind. It is worth emphasizing that, i f a Consequentialist appeals to all ofthe principles I have mentioned, his moral theory is very different fromUtilitarianism. Since such theories have seldom been discussed, this iseasy to forget.Some have thought that, i f Consequentialism appeals to many differen t principles, it ceases to be a distinctive theory, since it can be made tocover all moral theories. This is a mistake. C appeals only to principlesabout what makes outcomes better or worse. Thus C might claim that it

~ m i l d be worse i f there was m?re deception or coercion. C would then o.Sgive to all of us two common arms. We should try to cause it to be truethat there is less deception or coercion. Since C gives to all agents common r.Jmoral aims, I shall call C ~ g e n t - n e u t r a l .

~ a n y ~ ? r a l t h e ~ r i e s do not take_ this form. These theories are a,8E1t- r . t ~rr{JElJ:!.!e, giving to different agents different aims. It can be claimed, for 11example, that each of us should have the aim that he does not coerce other 1?eople.?n this view, it would be wrong for me to coerce other people, eveni f by doing so I could cause it to be true that there would be less coercion.S i m i l ~ r claims might be made about deceiving or betraying others. Onthese claims, each person's aim should be, not that there be less deceptionor betrayal, bu t that he himself does not deceive or betray others. Theseclailns are not Consequentialist. And these are the kinds of claim that most?f us accept. C can appeal to principles about deception and betrayal, bu tIt does not appeal to these principles in their familiar form.I shall now describe a different way in which some theory T might be selfdefeating. Call T

i n d t r _ . e . E J l y _ 9 2 l l e c . ! _ ~ v e l Y . self-defeating hen it is true that, if several peopletry to achieve their -given aims, these aims would be worse achieved.

On all or most of its different versions, this may be true of C. C impliesthat, whenever we can, we should try to do what would make theoutcome as good as possible. If we are disposed to act in this way, we are .Jr't

~ ! ! . : ! : ! ! do-gpoders. If we were all pure do-gooders, this might make the 0outcome worse. This might be true even i f we always did what, of the actsthat are possible for us, would Inake the outcome best. The bad effectswould come, not from our acts, but from our disposition.There are many ways in which, if we were all pure do-gooders, thismight have bad effects. One is the effect on the sum of happiness. On any


3/14

136 Derek Parfitplausible version of C, happiness is a lar ge p.art of what _makes o u t c o ~ e sbetter. Most of our happiness comes from acting on certam strong e s ~ r e s .These include the desires that are involved in loving certain other people,the desire to work well, and Inost of the strong desires on which we actwhen we are not working. If we become pure do-gooders, most of ouracts would be attempts to make outcomes better, not just in our owncommunity, but in the world as a whole. We would therefore seldom acton these strong desires. It is likely that this would enormously reduce thesum of happiness. This might make the outcome worse, even If we alwaysdid what, of the acts that are possible for us, Inade the outcome best. Itmight not make the outcome worse than it actually is, given what peopleare actually like. But it would make the outcome worse than it would.be,if we were not pure do-gooders, bu t ha d certain other causally possibledesires an d dispositions.

There are several other ways in which, if we were all pure do-gooders,this Inight make the outcome worse. One rests on the fact that, when wewant to act in certain ways, we shall be likely to deceive ourselves aboutthe effects of our acts. We shall be likely to believe, falsely, that these actswill produce the best outcome. C o n s i d e ~ , for examp:e, killing otherpeople. If we want someone to be dead, it IS easy to beheve, falsely, thatthis would make the outcome better. It would therefore make the outcomebetter if we were strongly disposed not to kill, even when we ~ l i e v e thatthis would make the outcome better. disposition not to kill shouldgive a only when we believe that, by killing, we would_ Il1ake theoutcome very mUC e er. Imi ar 10Ug Wea er C aimS app y to deception, coercion, ana-sever:a:rother kinds of act.2

11. Why C Does Not Fail in its Own TermsI shall assume that, in these and other ways, C is indi rectly collectivelyself-defeating. If we were all pure do-gooders, the out:ome would beworse than it would be if we had certain other sets of mohves. I f we knowthis C tells us that it would be wrong to cause ourselves to be, ortorem'ain, pure do-gooders. Because C makes this claim, it is not f a i l i n g i ~ :its own terms. C does not condemn itself.This defence of Cis like Iny defence of S ['For each person, there issupremely rational aim: that his life go, ~ i ~ , as we.ll ?ossible'] is worth pointing out one difference. S IS Indirectly Individuallydefeating when it is true of some person that, if he was never self-rl. , , . . . . , ~ r ; n , o - : : c : ,this would be worse for him than if he ha d some other set of desires

From Reasons and Persons 137dispositions. This would be a ba d effect in S' s terms. An a this b a ~ effectoften occurs. There are many people whose lives are going worse becausethey are n ~ v e r , or v e ~ . seldom, self-denying. C is indirectly collectivelys e l f - d e f e a t i ~ g when It Is true that, if some or all of us were pure dogood.ers, thi s would make the outcome worse than it would be if we hadcertain other motives. This would be a bad effect in C's terms. But thisbad effect ma y not occur. There are few people who are pure do-gooders.Because there are few such people, the fact that they have this dispositionma y not, on the whole, make the outcome worse.

T ~ 1 e b a d effect.in terms often occurs. The bad effect inC's terms ma ynot o ~ c u r . But this difference does no t affect my defence of s and c. Boththeones t ~ l l . us not to have the dispositions that would have these bad

e f ~ e c : s . This Is whyS is not, an d C would not be, failing in their own terms.It Is Irrelevant whether these bad effects actually occur.

My d e f ~ n c e of C assumes that we can change our dispositions. Itrna! be obJected: 'Suppose that we were all pure do-gooders, because we

b : h e v ~ And suppose that we could no t change our dispositions. OurdiSpositions would have bad effects, in C's terms, and these ba d effectsw o u l ~ be the result ~ e l i e f C. C would here be failing in its ownter.ms.. T h ~ r e was a sumlar obJection to my defence of S. I discuss theseobJections In [a later Section].

12. The Ethics of FantasyI have assumed that C is indirectly collectively self-defeating. I haveassumed that, i f we were all pure do-gooders, the outcome would be:vorse than it would be if we had certain other sets of motives. I f his claimIs t r ~ e , C tells us that we should try to have one of these other sets ofInohves.. Whether this claim is true is in part a factual question. An d I would

: _need to say much more if, rather than assuming this claim, I wished to:; show. that this claim is true. I shall not try to show this, for three reasons.; I beheve t l ~ a t this c l a ~ m . is probably true. Rather than arguing about the

I b e h e v ~ that It Is more worthwhile to discuss what this claimMy third reason is that assume that most of us wovld not in fact

.: ,..,,_ V.LJ.u:::; pure do-gooders, even If we became convinced that Consequen.... , . . , . . ~ ~ ~ ...... is the best moral theory.Because he makes a similar assmnption, Mackie calls Act Utilitarian'the ethics of fantasy'.3 Like several other writers, he assumes that we

reject a moral theory if it is in this sense .unrealistically demanding:


4/14

138 Derek Parfitif it is true that even if we all accepted this theory, most of us would infact seldom d ; what this theory claims that we ought to do. Mackiebelieves that a moral theory is something that we invent. If this is so, it isplausible to claim that an acceptable theory cannot be u n r e a l i ~ t i c a l ~ ydemanding. But, on several other views about the nature_ of mor ahty,claim is not plausible. We ma y hope that the best theory 1s not unreahshcally demanding. But, on these views, this can only be a hope. We cannotassume that this must be true.

Suppose that I am wrong to assume that Cis indirectly o l l e c t ~ v e l y selfdefeating. E ven if this is false, we can plausibly assume that C 1s u n r e a ~ -istically demanding. Even if it would not make outc?me worse 1fwe were all pure do-gooders, it is probably causa lly 1mposs1ble that all ormost of us become pure do-gooders. . .

Though these are quite different assumptions, they have the same !mphcation. If it is causally impossible that we become pure do-gooders, Cagain implies that we ought to try to have one of the best possible sets ofi,otives, in Consequentialist terms. This implication is t h e r e f o r ~ v:orthdiscussing if (1) C is either indirectly self-defeating or unreahshcallydemanding, or both, an d (2) neither of these facts would show that Ccannot be the best theory. Though I am not yet convinced that C is thebest theory, I believe both (1) an d (2).

13. Collective ConsequentialismIt is worth distinguishing C from another form of C o n ~ , e q u e n t i a l i s m . Asstated so far, C is individualistic and concerned with actual effects. According to C, each of us should try to do what would make the outcome best,given what others will actually do. An d each of us should try _to have one ofthe possible sets of motives whose effects would be best, g1ven the actualsets of motives that will be had by others. Each of us should ask: Is theresome other set of motives that is both possible for me an d is such that, ifI had this set the outcome would be better?' Our answers would dependon what we know, or can predict, about the sets of motives that will behad by others.

What can I predict as I type these words, in January 1983? I know thatmost of us will continue to have 1notives much like those that we havenow. Most of us will love certain other people, an d will have the otherstrong desires on which most happiness depends. Since I know this, C:may tell me to try to be a pure do-gooder. This ma y ~ n a k e the outcome}better even though, if we were all pure do-gooders , th1s would make ~

From Reasons and Persons 139outc01ne worse. If most people are not pure do-goodefs, it r n a ~ makethe outcome e t t e r if a few p e ~ p l e are. I f most people remain as they arenow, there w1ll be much suffenng,_much inequality, an d 1nuch of most ofth : other _things that make outcomes bad. Much of this suffering I couldfa1rly eas1ly prevent, and I could in other ways do much to make theoutcome better. It ma y therefore make the outcome better if I avoid closepersonal ties, and cause my other strong desires to become comparativelyweaker, so that I can be a pure do-gooder.

If _mn lucky, it may not be bad for me to become like this. My life willbe t ~ ~ p p e d of n:ost of the sources of happiness. But one source of happiness _lf? the bel_1ef that one is doing good. This belle ma y give mehappmess, making my austere life, not only morally good, bu t also a goodlife for me.

_I ma y be less lucky. It ma y be true that, though I could come close tobe1ng a pure do-gooder, this would not be a good life for me. And therema y be many other possible lives that would be much better for me. Thiscould be true on most of the plausible theories about self-interest. Thedemands made on me by C may then seem unfair. Why should I be theone who strips his life of most of the sources of happines s? More exactly, ...

should I be among the few who, according to C, ought to try to do \Jth1s? Would it not be fairer if we all di d more to make outcomes better?

This suggests a form of Consequentialism that is both collective an d con- J-'''-'""-'- . _,_v_._L.u. The two theories disagree about what we ought to do. Considerthe question of how much the rich should give to the poor. For most Consequentialists, this question ignores national boundaries. Since I know

.}hat most other rich people will give vety little, it would be hard for me:tb: deny that it would be better if I gave away almost all my income. Even

.J gave nine-tenths, some of my remaining tenth would do more goodspent by the very poor. Consequentialism thus tells me that I ought toaway almost all my income.


5/14

140 Derek ParfitCollective Cons e uentialism is much less demandigg. It does not tello 1ve t e amount that would in fact make the outcon1e best. Ittells me to give the amount which is such that if we all gave this amount,

the outcome would be best. More exactly, it tells me to give what wouldbe demanded by the particular International Income Tax that would 1nakethe outcome best. This tax would be progressive, requiring larger pro-portions from those who are richer. But the de1nands made on each personwould be much smaller than the de1nands made by C, on any plausibleprediction about the amounts that others will in fact give. I t might be bestif those as rich as me all give only half their income, or only a quarter. Itmight be true that, if we all gave more, this would so disrupt our owneconomies that in the future we would have much less to give. An d itmight be true that, if we all gave more, our gift would be too large to beabsorbed by the econmnies of the poorer countries

.._, f The difference that I have been discussing arises only within what isIS G ~ ' - . }t called partial compliance theory. This is the part of a moral theory t 1atZ :- "covers cases where we know that some other e o ~ l e will not do what theyought to do. C might require that a few people give away almost all therrmoney, an d try to make themselves pure do-gooders. But this would only

be because most other people are not doing what C claims that they oughtto do. They are not giving to the poor the amounts that they ought to give.In its partial compliance theory, C has been claimed to be excessively

demanding. This is not the claim that C is unrealistically demanding. AsI have said, I believe that this would be no objection. What is claiinedis that, in its partial compliance theory, C makes unfair or unreasonabledemands. This objection may not apply to C's full compliance theory.C would be 1nuch less demanding if we all ha d one of possible setsof motives that, according to C, we ought to try to cause ourselves tohave.4 .

14. Blameless WrongdoingThough C is indirectly self-defeating, it is not failing in its ow n tenns. Butit may seem open to other objections. These are like those I raised whendiscussing S. Suppose that we all believe C, and all have sets of motivesthat are among the best possible sets in Consequentialist tenns. I haveclaimed that, at least for most of us, these sets would not include being apure do-gooder. If we are not pure do-gooders, we shall sometimes.do what we believe will make the outcome worse. According to C, weshall then be acting wrongly.

From Reasons an d Persons 141Here is one example. Most of the best possible sets''of m o t i ; ~ s would

include strong love for our children. Suppose that Clare has one of thesesets of motives. ConsiderClare's Decision. Clare could either give her child some benefit, or givemuch greater benefits to some unfortunate stranger. Because she lovesher child, she benefits him rather than the stranger.

As a Consequentialist, Clare ma y give moral weight, not just to how muchc h i l ~ r e n are benefited, bu t also to whether they are benefited by their ownparents. She ma y believe that parental care and love a re intrinsically, orin themselves, part of what makes outcomes better. Even so, Clarema y believe that she is doing what makes the outcome worse. Shema y therefore believe that she is acting wrongly. And this act is quitevoluntary. She could avoid doing what she believes to be wrong, ifshe wanted to. She fails to do so simplybecause her desire to benefit herchild is stronger than her desire to avoid doing what she believes to bewrong .

I f son1eone freely does what she believes to be wrong, she is usuallyopen to serious moral criticism. Ought Clare to regard herself as open tosuch criticism? As a Consequentialist, she could deny this. Her replywould be like Kate's [in an earlier section] when Kate claimed that shewas not irrational. Clare could say: 'I act wrongly because I love my child.But it would be wrong for me to cause myself to lose this love. This badeffect is part of a set of effects that are, on the whole, one of the best possible sets of effects. It would be wrong for me to change my motives sothat I would not in future act wrongly in this kind of way. Since this i s so,when1 do act wrongly in this way, I need not regard myself as morallybad. We have seen tha! there can b


6/14

142 Derek Parfitin Consequentialist terms, the ba d effects are, in the relevant sense, partof one of the best possible sets of effects.'

It may be objected: 'I f it is not causally possible that you act differently,given your actual motives, you cannot make clailns about what you oughtto do. Ought implies can.' Kate answered this objection in [an earl ier Section]. It cannot be claimedthat Clare ought to have acted differently if she could not have done so.This last clause does not mean 'i f this would have been causally impos-sible, given her actual motives'. It means 'i f this would have been causallyimpossible, whatever her motives might have been'.Like Kate, Clare may be wrong to assume Psychological Determinism.If this is so, her claims can be revised. She should cease to claim that, ifshe has one of the best possible sets of Inotives, this will inevitably causeher to do what she believes to be wrong. She could claim instead: 'I f I wa sa pure do-gooder, it would be easy not to do what I believe be :vrong.Since I have another set of motives, it is very hard not to act In this way.And it would be wrong for me to change my motives so that it would beeasier not to act in this way. Since this is so, when I act in this way, I ammorally ba d only in a very weak sense.'

Consider nextThe Imagined Case. It might have been true that Clare could either saveher child's life, or save the lives of several strangers. Because she lovesher child, she would have saved him, and the strangers would havedied.

If this had happened, could Clare have made the same claims? T]Je deathsof several strangers would have been a very ba d effect. Could Clare. haveclaimed that it was part of one of the best possible sets of effects? Theanswer ma y be No. It might have made the outcome better if Clare hadnot loved her child. This would have been worse for her, and much worsefor her child. But she would then have saved the lives of these severalstrangers. This good effect might have outweighed the bad effects,making the outcmne, on the whole, better.

If this is so, Clare could have said: 'I ha d no reason to believe that mylove for my child would have this very ba d effect. It was subjectively i g l ~ tfor me to allow myself to love my child. An d causing myself to lose thislove would have been blameworthy, or subjectively wrong. When I savemy child rather than the strangers, I ain acting on a se t of motives that itwould have been wrong for Ine to cause myself to lose. This is enough to

From Reasons and Persons 143justify my claim that, when I act in this way, this is a case of b l ~ e l e s swrongdoing.'

A Consequentialist might have claimed: 'When. Clare learns that shecould save the strangers, it would not be subjectively wrong for her tocause herself not to love he r child. This would be right, since she wouldthen-save s t r a n g e ~ s . ' Clare could have answered: 'I could not possiblyhave lost this love with the speed that would.have been required. Thereare ways in which we can change ou r motives. But, in the case of ourdeepest motives, this takes a long time. It would have been wrong for meto try to lose my love for my child. I f I had tried, I would have succeeded

o n l ~ after the strangers had died. After they ha d died, this change in Inymotives would have made the outcome worse.'As this answer shows, Clare's claims essentially appeal to certain

f ~ c t u a l assumptions. It might have been true that, if she ha d the disposition of a pure do-gooder, this would on the whole have made the outcomebetter. But we are assuming that this is false. We are assuming that theoutcome would be better if Clare has some set of motives that will sometimes cause her to choose to do what she believes will make the outcomeworse. And we are assuming that her actual set of motives is one of thebest possible sets.We could imagine other motives that would have made the outcomeeven better. But such motives are not causally possible, given the factsabout human nature. Since Clare loves her child, she would have savedhim rather than several strangers. We could imagine that our love for ourchildren would 'switch off' whenever other people's lives are at stake. Itmight be true that, if we all ha d this kind of love, this would make theoutcome better. If we all gave such priority to saving more lives, therewould be few cases when our love for our children would have to switcho:f. This love could therefore be much as it is now. But it is in fact impos-sible that our love cou.ld be like this. We could not bring about such 'fine-tuning'. If there is a threat to the life of Clare's child, her love for himcould not switch off merely because several strangers are also threatened.5Clare claims that, when she does what she believes will make the outcome

, worse, she is acting wrongly. But she also claims: 'Because I am acting ona set of motives that it would be wrong for me to lose, these acts are blame. l e s ~ . When I ac.t this way, I need ;not regard myself as bad. I f PsychoIogical Determinism was not true, I would be bad only in a very weak

. When I act in this way, I should not feel reinorse. No r should Ito try not to act in this wa y again.'


7/14

144 Derek ParfitIt may now be objected that, since she makes these claims, Clare cannot

really believe that she is acting wrongly. But there are s ~ f f i c i ~ n t g r o u n ~ sfor thinking that she does have this belief. Consider the Imagined case Inwhich Clare saves her child rather than several strangers. 'Though sheloves her child, Clare would not have believed that his death would be aworse outcome than the deaths of the several strangers. His death wouldhave been worse for him and her. But she would have believed that thedeaths of several strangers would, on the whole, be much worse. In savingher child rather than the strangers, she would have done what she believes will make the outcome much worse. She would therefore havebelieved that she is acting wrongly. He r moral theory directly impliesthis belief. She would also have believed that she should not feel remorse.But her reason for believing this would not have casted doubt on her beliefthat she is acting wrongly. He r reason would have been that she is actingon a motive - love for her child - that it would have been wrong for herto cause herself to lose. This supports the claim that she deserves noblame, bu t it does not support the claim that her act is not wrong.

It might be said(G4) If smneone acts on a motive that he ought to cause himselfto have, and that it would be wrong for him to cause himself to

lose, he cannot be acting wrongly.If (G4) was justified, it would support the claim that Clare's act would nothave been wrong. An d this would support the claim that she cannot reallybelieve that her act would have been wrong. But in Section 16 I describea case where (G4) is not plausible. ,

Clare could add that, in many other possible cases, if she believed thather act was wrong, she would believe herself to be bad, and she wouldfeel remorse. This would often be so if she did what she believed wouldmake the outcome worse, an d she wa s not acting on a set of motives thatit would be wrong for her to cause herself to lose. Consequentialism doesnot in general break the link between the belief that an act is wrong, andblame and remorse. This link is broken only in specia l cases. We have beendiscussing one of these kinds of case: those in which someone acts on aInotive that it would be wrong for him to cause hhnself to lose.There is another kind of case where the link is broken. C applies toeverything, including blame and remorse. According to C we ought toblame others, and feel remorse, when this would make the outcome better.This would be so when blame or remorse would cause our motives tochange in a way that would make the outcome better. This would not be ,

From Reasons and Persons 145true when, like Clare, we have one of the best possiblesets of Iitotives.And it might not be true even when we do not have such motives. If weare l ~ m e d too often, blame ma y be less effective. C may thus imply that,even If we do not have one of the best sets of motives, we should beblamed only for acts that we believe will make the outcome much worse.

15. Could it be Impossible to Avoid Acting Wrongly?C l a r ~ ' s claims imply that she cannot avoid doing what she believes to bewrong. She might say: 'I t is not causally possible both that I have one ofthebest possible sets of motives, and that I never do what I believe tobe wrong. If I was a pure do-gooder, my ordinary acts would never bewrong. But I v:ould be acting wrongly in allowing myself to remain a puredo-gooder. If Instead I cause myself to have one of the best possible setsof motives, as I ought to do, I would then sometimes do what I believe tobe wrong. If I do not have the disposition of a pure do-gooder, it is notc a u s ~ l l y possible that I always act like a pure do-gooder, never doing whatI beheve to be wrong. Since this is no t causally possible, and it would be

w ~ ~ n _ g for me _cause myself to be a pure do-gooder, I cannot be morallycnhcised for faihng always to act like a pure do-gooder.' -

I t may now be said that, as described by Clare, C lacks one of theessential features of any moral theory. It may be objected: 'No theory candemand what is impossible. Since we cannot avoid doing what C claimsto be wrong, we cannot always do what C claim.s that we ought to do. Weshould therefore reject C. As before, ought implies can.'

This objection applies even if we deny Psychological Determinisin.Suppose that Clare ha d saved her child rather than several strangers. Shewould have acted in this wa y because she does not have the dispositionof a pure dc_>-gooder. He r love for her child would have been stronger thanher d e s ~ ~ to avoid doing what she believes to be wrong. I f we denyDeternnnism, we shall deny that, in this case, it would have been causally

. ~ ~ l l i . / V D C U . U ' l . c : for Clare to avoid doing what she believes to be wrong. By aneffort of will, she could have acted against her strongest desire. Even ifwe claim this, we cannot claim that Clare could always act like a pure do,gooder without having a pure do-gooder's disposition. Even those who'deny Determinismcannot completely break the link between our acts and

. ~ . o u r dispositions.. :'If we cannot always act like pure do-gooders, without having a pure

gooder's disposition, the objection given above still applies. Even ifdeny Determinism, we must admit the following. We are assuming


8/14

146 Derek Parfitthat we believe truly that the outcome would be worse if we were all puredo-gooders. If we have this belief, it is not possible that we never do whatwe believe will make the outcome worse. If we cause ourselves to be, orallow ourselves to remain, pure do-gooders, we are thereby doing whatwe believe will make the outcotne worse. If instead we have other desiresand dispositions, it is no t possible that we always act like pure dogooders, never doing what we believe will make the outcome worse. Theobjector can therefore say: 'Even if Determinism is not true, it is not possible that we never do what we believe will make the outcome worse. Inclaiming that we ought never to act in this way, Cis demanding what isimpossible. Since ought implies can, C' s claim is indefensible.'

Clare could answer: 'In most cases, when smneone acts wrongly, hedeserves to be blamed, an d should feel remorse. This is what is mostplausible in the doctrine that ought implies can. It is hard to believe thatthere could be cases where, whatever someone does, or might have earlierdone, he deserves to be blamed, and should feel remorse. It is hard tobelieve that it could be impossible for someone to avoid acting in a wa ythat deserves to be blamed. C does not imply this belief. If I saved mychild rather than several strangers, I would believe that I am doing vvhat

&.) will make the outcome tnuch worse. I would therefore believe that I am~ acting wrongly. But this would be a case of blameless wrongdoing. Accord

~ i n g to C, we can always avoid doing what deserves to be blamed. This isenough to satisfy the doctrine that ought implies can.'We ma y believe that these claims do not sufficiently meet this objection. There was a similar objection to S. It is impossible that we never dowhat S claims to be irrational. I began to meet that objection by appealing to the case in [an earlier Section]: Schelling's Answer to ArmedRobbery. In this case, on any plausible theory about r a t i ~ n a l i t y , I couldnot avoid acting irrationally. To meet the objection to C, Clare tnightappeal to other cases where we cannot avoid acting wrongly. That thereare such cases has been claimed by some of the writers who are mostopposed to C.6

16. Could it be Right to Cause Oneself to Act Wrongly?Since Cis indirectly self-defeating, it tells us to cause ourselves to do, orto be more likely to do, what it claims to be tnorally wrong. This is not adefect in C's terms. We can ask a question like the one I asked about theSelf-interest Theory. C gives us one substantive moral aim: that history goas well as possible. Does it also give us a second substantive aim: that we

From Reasons and Persons 147never act wrongly? On the best known form of C, Utilitarianis"rh, theanswer is No. For Utilitarians, avoiding wrong-doing is a mere means tothe achievement of the one substantive moral aim. It is not itself a substantive aim. An d this could also be claimed on the versions of C thatjudge the goodness of outcomes in terms no t of one bu t of several moralprinciples. It might be claimed, for instance, by the theory that appealsboth to the Utilitarian claim and to the Principle of Equality. All thesetheories give us the formal aim of acting morally, an d avoiding wrongdoing. But these theories could all claim that this formal aim is no t partof our substantive moral aim.r Though this claim might be made by any Consequentialist, it wouldnot be made on several other moral theories. On these theories, the avoidance. of wrong-doing is itself a substantive moral aim. If we accept one ofthese theories, we may object to C in at least two ways. We ma y say, 'Anacceptable theory cannot treat acting morally as a mere means.' We ma yalso say;'An acceptable theory cannot tell us to cause ourselves to do whatthis theory itself claims to be wrong.'We should ask whether, if we raise this objection, we ourselves believethat the acts in question would be wrong. We are considering cases wherea Consequentialist believes that, though he is acting wrongly, he is notmorally bad, because he is acting on motives that it would be wrong forhim to cause himself to lose. In such cases, do we ourselves believe thatthis Consequentialist is acting wrongly?

This is unlikely in the imagined case where Clare saves her child ratherthan several strangers. If we are no t Consequentialists, we shall be likelyto believe that Clare's act would not have been wrong. We ma y thinkthe same about some of the other cases of this kind. Suppose that Clarerefrains from killing me, though she has the true belief that killing mewould make the outcome better. Clare would think that, in refraining~ r o m killing me, she would be acting wrongly. But she would regard thisas a case of blameless wrongdoing. She acts wrongly because she isstrongly disposed not to kill, and, for the reason given at the end of Section10, she believes that this is a disposition that it would be wrong for herto cause herself to lose. We ma y again believe that, in refraining fromkilling me, Clare is not acting wrongly.

I,f tbj..._is what we believe about these cases, it is less clear that we shouldo b j e . c L t u . . . t h J ~ a r t of c::--we-aecept C:' s Clatm that, m tll.ese cases, Clarewould not show herself to be morally bad, or deserve to be blamed. Overthis there is no disagreement. We ma y object to C' s claim that, thoughClare is blmneless, her acts would be wrong. But perhaps we should notobject to this claim, if it does not have its usual implications.


9/14

ij

148 Derek ParfitWe ma y still object that an acceptable moral theory cannot tell us to

cause ourselves to do what this theory claims to be wrong. But considerMy Moral Corruption. Suppose that I have some public career thatwould be wrecked i f I was involved in a scandal. I have an enemy, acriminal whom I exposed. This enemy, no w released, wants revenge.Rather than simply injuring me, he decides to force n1e to corruptmyself, knowing that I shall think this worse than most injuries. Hethreatens that either he or some member of his gang will kill all mychildren, unless I act in some obscene way, that he will fihn. If he latersent this film to some journalist, my career would be wrecked. He willthus be able later, by threatening to wreck Iny career, to cause me tochoose to act wrongly. He will cause me to choose to help him commitvarious minor crimes. Though I am morally as good as most people, Iam not a saint. I would not act very wrongly merely to save my career;bu t I would help my enemy to commit minor crimes. I would here beacting wrongly even given the fact that, if I refuse to help my enemy,my career would be wrecked. We can next suppose that, since I knowIny enemy well, I have good reason to believe both that, if I refuse tolet hiln make his film, my children will be killed, and to believe that,if I do not refuse, they will not be killed.

I ought to let this man make his film. We can plausibly clailn that govern-ments should not give in to such threats, because this would merely exposethem to later threats. But such a clailn would not cover this threat madeto me by my enemy. It would be wrong for me to refuse his demand, withthe foreseen result that my children are killed. I ought to let him make hisfilm, even though I know that the effect will be that I shaH later often actwrongly. After my children are freed, I shall often, to save Iny career, helpmy enemy commit Ininor crimes. These later acts will be quite voluntary.I cannot claim that my eneiny's later threats force me to act.in these ways.I could refuse to act wrongly, even though this would wreck Iny career.I have claimed that I ought to let this man make his fihn. This wouldbe agreed even by most of those who reject Consequentialism. Thesepeople would agree that, since it is the only way to save my children'slives, I ought to cause it to be true that I shall later often act wrongly. Thesepeople thus believe that an acceptable moral theory can tell someoneto cause himself to do what this theory claims to be wrong. Since theybelieve this, they cannot object to Consequentialism that it can have thisimplication.

From Reasons and Persons 149If I le t my enemy make his film, I would become disposed to heitp him

commit minor crimes. Let us no w ad d some features to this case. I couldcause .myself to lose this disposition, by abandoning my career. But myenemy has threatened that, if I abandon my career, his gang will kill mychildren. It would therefore be wrong for me to cause myself to lose thisdisposition. In contrast, i f I refuse to help my enemy commit his crimes,he will merely wreck my career, by sending to some journalist the film inwhich I act obscenely. My enemy assures me that, if he wrecks my career,my children will not be killed. He gets perverse pleasure from causing meto do what I believe to be wrong, by threatening to wreck my career. Thi spleasure would be lost if his threat was to kill my children. I f I help himto commit his crimes because this is the only wa y to save my children'slives, I would not believe that I was acting wrongly. Since my enemywants me to believe that I am acting wrongly, he does not make this threat.. _Knowing my enemy, I have good reason to believe what he says. SinceIt Is the only way to save my children's lives, I ought to let him make hisfilm. I ought to make myself disposed to help him commit his minorc ~ ~ e s . -:And would be wrong for me to cause myself to lose this dispo-sition, since, If I do, my children will be killed. But, when I act on this dis-position, I am acting wrongly. I ought not to help this ma n to commit hiscrimes, merely in order to save my career.

This case shows that we should reject what I called (G4). This is theclaim that, if I ought to cause myself to have some disposition, and it.would be wrong for me to cause myself to lose this disposition, I cannotbe acting wrongly when I act upon this disposition. In the case justdescribed, when I act on such a disposition, I am acting wrongly.1 shall now state together four similar mistakes. Some people claimthat, if it is rational for me to cause myself to have some disposition, it

~ , a n n o t be irrational to act upon this disposition. This was shown to befalse by the case I called Schelling's Answer to Armed Robbery. A secondclaim is that, if it is rational for me to cause myself to believe that someact is rational, this act is rational. This was shown to be false by the casethat I called My Slavery. A third claim is that, if there is some disposition

. I ought to cause myself to have, and that it would be wrong for meto cause myself to lose, it cannot be wrong for me to act upon this dispo-sition. The case just given shows this to be false. A fourth clailn is that, if........ ought to cause myself to believe that some act would not be wrong, this. act cannot be wrong. In Section 18 I show that this is false. These four$: cli:tims assume that rationalityand rightness can be inherited, or transferred.


10/14

150 Derek ParfitIf it is rational or right for me either to cause myself to be disposed toact in some way, or to make myself believe that this act is rational orright, this act is rational or right. My examples show that this is not so.Rationality an d rightness cannot be inherited in this way. In this respectthe truth is simpler than these claims imply. These claims cannot showthat, if we believe some act to be irrational or wrong, we are making amistake.

17. How C Might be Self-effacingIt 1night be claimed that, if Consequentialism sometimes breaks the linkbetween the belief that our act is wrong and the belief that we are bad,we would not in fact continue to regard 1norality with sufficient seriousness. Our desire to avoid wrongdoing ma y be undermined if we haveother desires which are often stronger. This d esire 1nay survive only if webelieve that it should always be overriding, an d feel remorse when it isnot. It might be claimed, on these or other grounds, that it would makethe outcome better if we always keep the link between our moral beliefsand our intentions and emotions. I f this is so, it would make the outcomebetter if we did not believe C.I doubt these claims. But it is worth considering what they wouldimply. According to C, each of us should try to have one of the best possible sets of desires and dispositions, in Consequentialist terms. It mightmake the outcome better if we did no t merely have these desires an d dispositions, but had corresponding moral emotions and beliefs. Consider, for example, theft. On some versions of C, it is intrinsicallybad if property is stolen. On other versions of C, this is not On theseversions, theft is bad only when it makes the outcome worse. Avoidingtheft is not part of the ultimate moral aim given to us by C. It might betrue that it would make the outcome better if we are strongly disposednot to steal. And it 1night make the outcome better if we believed stealing to be intrinsically wrong, and would feel remorse when we do steal.

" Similar claims 1night be made about many other kinds of act.~ If these claims are true, C would be self-effacing. It would tell us thatf we should try to believe, not itself, bu t some other theory. We should tryto believe the theory which is such that, if we believe it, the outcomewould be best. On the claims made above, this theory 1night not be C.Jtmight be s01ne version of what Sidgwick called Common-Sense Morality.

If C told us to believe some version of this 1norality, this would not beCommon-Sense Morality as it is now, bu t an improved version. Common-

From Reasons and Persons 151Sense Morality is not the moral theory belief in which would make theoutcome best. Such a theory would, for example, demand much morefrom the rich. I t might make the outcome best if those in the richer nationsgave the at least a quarter or even half of their incomes every year.The nch now give, an d seem to believe that they are justified in giving,less than one pe r cent.

Suppose that C told us to believe some other theory. As I have said, itwould be hard to change ou r beliefs, if our reason for doing so is not areason which casts doubt on our old beliefs, bu t is merely that it would

~ 1 a ; e ~ o o d effects if we ha d different beliefs. But there are various waysrn wh1ch we might bring about this change. Perhaps we could all behypnotized, and the next generation brought up differently. We wouldhaye to be 1nade to forget how and why we acquired our new beliefs, andthe process would have to be hidden from future historians.It would make a difference here if we accept, not C, but Collective Cons e ~ u e n t i a l i s m . I f accept C, we might conclude that C ought to bereJected by most eo le, but should still be believed b!;/ few. Our theoryw o ~ l d t ~ 1 e n b a r t l ~ self-effacing, and partly e s o t e r ~ c ; t e l l i n g those whobelieve It not to en Ighten the ignorant majority On Collective Consequentialism, we ought to believe the moral theory which is such thatif we all believe it, the outcome would be best. This theory cannotesoteric.

~ o r n e find it s p e c i a l l y objectionable that a moral theory might be eso,Jenc. we beheve that deception is morally wrong, deception aboutmorality ma y seem especially wrong. Sidgwick wrote: 'it seems expedient that the doctrine that esoteric morality is expedient should itself be

k e p ~ esoteric. Or, if this concealment be difficult to maintain, it may be desirable that Common Sense should repudiate the doctrines which it isexpedient to confine to an enlightened few.'7 This is what Williams calls

' G o v e r ~ m e ~ t House' Consequentialis1n, since it treats the majority like.the natives In a colony. 8 As Williams claims, we cannot welcome such aconclusion. Sidgwick regretted his conclusions, bu t he did not think regreta ground for doubt.9. I have claimed that it is unlikely that Cis wholly self-effacing. It would

. most be partly self-effacing and partly esoteric. It might make the outcome better if some people do not believe C; but it is unlikely that itmake the outcome better if C was believed by no one.Here is another ground for doubting this. Suppose that we all come to C. (This will seem less implausible when we remember that C can

a pluralist theory, appealing to many different moral principles.) Wedecide that C is wholly ~ e l f - e f f a c i n g . We decide that it would make


11/14

152 Derek Parfitthe outcome best if we caused ourselves to believe so1ne improved e r s i o ~of Common-Sense Morality. We might succeed in bringing about th1schange in our beliefs. Given changes in the world, and in our e c h n ~ l o g y ,it might later come to be true that the outcome_ would be better if_ werevised our moral beliefs. But if we no longer believe C, and now believesome version of Common-Sense Morality, we would not be m a l ~ ethese needed revisions in ou r morality. Our reason for b e ~ I e v i n g _ th_Ismorality would not be that we now believe it to be the morality behef 1nwhich would make the outcome best. This would be why we caused ourselves to believe th is morality. But, in order to believe this 1 n o r a l i ~ y , wemust have forgotten that this is what we did. We w o u l ~ now s1mplybelieve this morality. We would therefore not be led to revise morality even i f it came to be true that our belief in this morality would Increasethe chances of some disaster, such as a nuclear war.These claims should affect our answer to the question whether it wouldmake the outcome better i f we all ceased to beli eve C. We might e l i e ~ ecorrectly that there is some other moral theory belief in which _w?uld, Inthe short run, make the outcome better. But once C o n s e q u e n t m l i ~ m haseffaced itself, and the cord is cut, the long-term consequences rrught bemuch worse.This suggests that the most that could be true is that C is partly ~ e l f -effacing. It might be better if Inost people cause themselves to behevesome other theory, by some process of self-deception that, to s u c c ~ e d ,must also be forgotten. But, as a precaution, a few people should ~ n b n u eto believe c, and should keep convincing evidence about this selfdeception. These people need not live in Government House, or_haveother special status. If things went well, the few would do nothing. But Ifthe moral theory believed by most did become disastrous, the_few couldthen produce their evidence. When most people learn ~ h a t their m?ralbeliefs are the result of self-deception, this would underrrune these beliefs,and prevent the disaster. Though I have claimed that this is unlikely, suppose that C whollyself-effacing. Suppose that it told all of us to make o u r ~ e l _ v e s beheve, notitself but some other theory. Williains claims that, 1f this 1s so, the theory

c e a s ~ s to deserve its name, since it 'detennines nothing of how thoughtin the world is conducted' 10 This claim is puzzling since, as Williams alsoclaims, c would be demanding that the way in which we think aboutmorality, and our set of desires and dispositions, ~ ~ . u s t be for thebest' _n This is demanding something fairly specific, and whollyConsequentialist.

From Reasons and Persons 153Williams makes the third claim that, if C was wholly s E ! l f - e f f a c i n ~ , it

would cease to ?e effective.12 This need no t be so. Suppose that thingshappen as descnbed above. We all cmne to believe in some form of C. Wethen believe truly that, if we all believed some other theory, this wouldproduce the best possible outcome. C tells us all to believe this othertheory. In some indirect way, we cause ourselves to believe thisother theory. No one now believes C. This does no t justify the claim thatC has ceased to be effective. It has ha d the effect that we all now believeso_me other particular theory. Because we believe this other theory, thisWill affect w ~ a t we do. And our belief in this other theory will producethe best possible outcome. Though no one believes C, C is still effective.There are two continuing facts that are the effects of our earlier belief inC; our ne w o r a l beliefs, an d the fact that, because we have these beliefs,the outcome IS as good as it can possibly be.

. Williams rightly claims that, if C was wholly self-effacing, it would notr,e clear what this shows. We would have to decide whether it showed

~ h ~ ! t h i ~ theory 'is unacceptable, or merely that no one ought to acceptIt.; . It IS _clear that, on our last assumptions, no one ought morally to

~ c c e p t t h i ~ theory. If anyone does accept this theory, the theory wouldI t s ~ l f tell hrm that he ought morally to try to reject the theory, and instead~ e l i e v e s?me other ~ e ~ r y . But, as Williams suggests, there are two questions. It IS one q u e ~ t w n w ~ e t h e r some theory is the one that we oughtmorally to try t? believe. It Is a ~ o t h e r question whether this is the theorythat we ought zntellectually or m truth-seeking terms to believe - whetherthis t ~ e o r y is the true, or best, or best justified theory. I claimed earlierthat, If a th.eory about rationality was self-effacing, this would notshow that th:s ~ h e o r y _cannot be the true or the best justified theory. Canwe make a sinular clarm about moral theories?

Our answer to this question will depend in part on our beliefs aboutnature of ~ o ~ a l r e a s o n i n ~ . If a moral theory can be quite straight

~ o r w a r d l y true, It IS clear that, if it is self-effacing, this does not show thatIt. cannot be true. But we may instead regard morality as a social product,either actually or in some 'ideal constructivist' way. We ma y then claimthat, to be acceptable, a moral theory must fulfil what Rawls calls 'the

. publicity condition': it must be a theory that everyone ought to accept. an d publicly ackno;o/ledge to each other.14 On these meta-ethical views,

moral theory cannot be self-effacing. On other views, it be. I t wouldtake at least a b o o ~ to decide between these different views. I must there

f o r ~ , in this book, leave this question open. This does not matter if, as Ibelieve, C would not be self-effacing.


12/14

154 Derek ParfitNotes

1 Cf. Adams.2 This is argued in Sidgwick, pp. 431-9, and Adams, throughout. Adamsdiscusses what he calls Act and Motive Utilitarianism, claiming that MotiveUtilitarianism is not just a special case of Act Utilitarianism. A Motive Utilitarian claims that everyone should have the dispositions having which willmake the outcome best, even if there is nothing that this person could havedone to cause himself to have these dispositions. Adams is right to claimthat Motive Utilitarianism is different from Act Utilitarianism, just asRule Utilitarianism, at least in some versions, is not a special case of ActUtilitarianism. I shall not discuss these distinctions here.

3 Mackie, Ch. 4, Section 2.4 Collective Consequentialism is not the Cooperative C o n s e q u e n t i a l i s 1 ~advanced in Regan.

5 Cf. Hare, p. 36.6 We should first note the sense of 'can' in the doctrine that ought implies can.Suppose it is claimed that, in some case, I ought to have acted in some otherway. On the doctrine, I ought to have acted in this other way only if I couldhave done so. If I could not have acted in this other way, it cannot be claimedthat this is what I ought to have done. As I argued in [an earlier section], theclaim (1) that I could not have acted in this other wa y is not the claim (2) thatacting in this way would have been impossible, given my actual desires anddispositions. The claim is rather (3) that acting in this way would have beenimpossible, even if my desires and dispositions ha d been different. Acting inthis way would have been impossible, whatever my desires and dispositionsmight have been.We are sometimes right to claim that ought implies can. Suppose thatsomeone believes:

(A) It is always wrong to fail to save someone's life.Assume that I can either save one life, or save another life, bu t that I cannotsave both. Whichever life I save, I am failing to save so meone' s life. According to (A), I cannot avoid acting wrongly. I would avoid acting wrongly onlyif I save both lives, and this is impossible. We can plaus ibly reject (A) by claiming that ought implies can. I could not save both these lives. This is not theclaim that this is not possible, given my actual desires and dispositions. Icould not save both these lives, whatever my desires and dispositions were.Since it is in this sense impossible for me to save both lives, it cannot defensibly be claimed that this is what I ought to do, and that in failing to saveboth lives I am acting wrongly.

Return now to Consequentialism. C claims that it is wrong for anyone todo what he believes will make the outcome worse. We are assuming that, if

From Reasons and Persons 155we were all pure do-gooders, the outcome would be worse _than it w5tlid beif we had certain other desires and dispositions. If we believe this, it is impossible that we never do what we believe will make the outcome worse. Weshall believe that, if we are pure do-gooders, we have made the outcomeworse by causing ourselves to be, or allowing ourselves to remain, pure dogooders. If we are not pure do-gooders, bu t have the desires and dispositionsthat we believe would make the outcome better, it is not possible that wealways act like pure do-gooders. It is no t possible that we never do what webelieve will make the outcome worse. On either of these alternatives, it isimpossible that we never act in this way.

Is this impossible in the sense that justifies an appeal to the doctrine thatought implies can? Is it impossible that we never act in this way, whateverour desires and dispositions are, or might have been? This is true, bu t misleading. It suggests that this impossibility has nothing to do with what ourdesires and dispositions are. This is not so. This impossibility essentiallyinvolves claims about our desires and dispositions. Why is it impossible thatwe never do what we believe will make the outco me worse? This is impossible because there is only one disposition given which it would be causallypossible never to do what we believe will make the outcome worse, andcausing ourselves to have or to keep this disposition would itself be a case ofdoing what we believe will make the outcome worse. Because this impossibility essentially involves these claims about our desires and dispositions, itis not clear that this is the kind of impossibility that justifies an appeal to thedoctrine that ought implies can. It can at least be said that this case is verydifferent from the case where it is impossible for me to save both of a pair oflives. T hat impossibility ha d nothing to do with my desires and dispositions.In that case we could plausibly appeal to the doctrine that ought implies can.This does not show that we can plausibly appeal to this doctrine in this very

, different case. Perhaps we can. But I believe that we cannot support an appeall,to this doctrine here by claiming that this appeal is plausible in the case whereI cannot save both of a pair of lives. Because they are so different, I believe that case to be irrelevant.

The irrelevance ma y be denied. If we believe that the outcome would beworse if we were all pure do-gooders, it is impossible that we never do whatwe believe will make the outcome worse. I have called it misleading to claimthat this is impossible whatever our desires and dispositions are, or mighthave been. This claim suggests, falsely, that this impossibility has nothing todo with our desires and dispositions. But, even if this claim is misleading, itis still true. An d this might be claimed to make this case sufficiently similarto that in which I cannot save both of a pair of lives.

Are there any cases where we can deny that ought implies can, even when'the impossibility has nothing to do with our desires and dispositions? Somewriters claim that there are. Nagel claims that there can be moral tragedies,cases where, whatever someone does, he will be acting wrongly. Nag el admits


13/14

156 Derek Parfitthat, in making his claims, he is denying that ought implies can. In the caseshe describes, someon e ought to avoid acting wrongly, though he could nothave avoided acting wrongly, whatever his desires and dispositions mighthave been. It is natural to hope that there cannot be such cases. But Nagelwrites, 'it is naive to suppose that there is a solution to every moral p r o b l ~ mwith which the world can face us'. The world may give us problems to wh1chthere is no solution which avoids wrong-doing. (Nagel, P 79.)Nagel suggests that there can be such cases because_the best moral h e _ o r ycontains conflicting principles. He might claim: 'A smgle moral prmciplecannot have such implications. If such a principle implies that we cannotavoid wrong-doing, this principle is indefensible.' This claim is_ p l a u ~ i b l ewhen applied to cases like that where I cannot save both of a ~ a i r of hves.Principle (A) implies that, in failing to save both lives, I am actmg wrongly.The fault here lies in this principle, not in the world.

If it is true that the outcome would be worse if we were all pure do-gooders, Consequentialism implies that we c a n n ~ t avoid acting ':ro.ngly.Though a Consequentialist may appeal many ~ I f ~ e r e n t ~ ~ r a l ~ n n c i p l e s ,this particular conclusion is implied by a smgle pnnc1ple. It IS nnphed by theclaim that it is wrong to do what we believe will make the outcome worse.But if there can be cases where we cannot avoid acting .wrongly, as Nagelclahns, the explanation ma y not have to be that there is a conflict betweentwo different principles. The explanation ca nnot simply appeal to the factit is causally impossible to act in both of two different ways. !h e explanatiOnmust appeal to something deeper - something :he conflict b e t w ~ e n twodifferent principles. And, in the case we are considermg, the x p l a n a t l ~ n maybe claimed to be of this kind. There is a conflict, not between two differentprinciples, bu t between what would be the best set of ~ t s , and : V ~ a t wouldbe the best set of desires and dispositions. We are assummg that It IS causally .impossible, g iven the facts about human nature, both that we p e r f o n ~ the se t .of acts that would make the outcome best, and that we have the desues and .dispositions having which would make the outcome best. This k ~ n d of c o ~ ~ .flict may be held to be sufficiently similar to that produced by c o n f l ~ c t .between two different principles. An d the fault here may be clmmed to he,,not in the principle that it is wrong to do what we believe will make . ,outcome worse, bu t in the world. A Consequentialist could repeat p a ~ t Nagel's claim. He might say: 'I t may be true that, if we have thethat will enable us never to do what we believe will make the outcomewe have made the outcome worse by causing ourselves to have or tothis very disposition. If his is true, the world has ~ v e n us a_problem tothere is no solution. More exactly there is a solutiOn, bu t It does notus to avoid wrong-doing. There is something that we ought to do, allconsidered. We ought to cause ourselves to have, or to keep, one of theof desires and dispositions having which will make the outcome better.if we do have one of these sets, it is impossible that we never do

From Reasons an d Persons 157believe will make the outcome worse. If we do what, all things consilered,we ought to do, it is impossible that we never act wrongly.'

There is another difference between this case and those discussed byNagel. He suggests that, in cases where someone cannot avoid actingwrongly, he should be blamed for acting wrongly, and should feel remorseand guilt. This is the claim that conflicts most sharply with the doctrine thatought implies can. As I have said, what is hardest to believe is that, whateversomeone does, or might have done, he deserves to be blamed, and shouldfeel remorse and guilt. Consequentialism does not imply this claim. WhenClare saves her child rather than the strangers, she deserves no blame, andshe should not feel remorse and guilt. Perhaps she should feel what Williamscalls agent-regret. Perhaps she should have this feeling when she thinks of thedead strangers whom she could have saved. But this feeling is not remorseor guilt. C does not imply that, whatever Clare does or might have done, she

d e _ s ~ r v e s to be blamed, and should feel remorse and guilt. This is whatW I ~ m m s and Nag:l claim about agents who face moral tragedies. If webeheve that ought Implies can, we may reject this claim. But this would notbe a r ~ a s o n to reject C, since it does not imply this claim.If a Consequentialist rejects these remarks, he must revise his moral theory.Suppose that he concedes that, if we could not possibly always avoid doing

. what we believe will make the outcome worse, it cannot be claimed that thisis what we ought always to do. It cannot be claimed that it is always wrongto do what we believe will make the outcome worse. Since it is no t possiblethat we never act in this way, it cannot always be wrong to act in this way.There must be some cases where, though we act in this way, we are not actingwrongly. According to (C3) it is always wrong to do what we believe willmake the outcome worse. A Consequentialist might now abandon (C3) and

S ~ b s t i t u t e(C3') We ought always to avoid doing what we believe will make the

outcome worse, if this is possible in a way that does not itself make theoutcome worse. If this is impossible, we ought to avoid doing whatwe believe will make the outcome worse, whenever we could haveacted differently, in a way that would not have made the outcomeworse.

As I have argued, 'impossible' does not here mean 'causally impossible, given,our actual desires and dispositions'. It might mean 'causally impossible,whatever our e ~ i r e s and dispositions might have been'. If it would make the

. outcome worse if we were all pure do-gooders, it is causally impossible thatwe never do what we believe will make the outcome worse. This would havebeen causally impossible, whatever our desires and dispositions might have

. been. Since it is causally impossible, in this sense, that we never act in thisy, (C3') implies that, when we act in this way, we cannot always be acting


14/14

158 Derek Parfitwrongly. If Consequentialisrn claims (C3') rather than (C3), it takes a revisedform that satisfies the doctrine that ought implies can. This revised versionof C meets the objection that appeals to this doctrine.

How much difference do es it make if Cis revised in this way? Supposethat we all accept (C3'), and that we believe truly that the outcome would beworse if we were all pure do-gooders. The outcome would be better if we hadone of many other possible sets of desires and dispositions. Assume next thatit is causally possible that we cause ourselves to have, or allow ourselves tokeep, one of these other sets of desires and dispositions. (Only on thisassumption does this ob jection to C arise.) (C3') implies that we ought tocause ourselves to have, or to keep, one of these othe r sets of desires and dispositions. Since it is possible to act in either of these wa ys, (C3') implies thatit would be wrong not to do so.

Assume next that we have one of these other sets of desires and dispositions. Because this is true, we often do what we believe will make the outcomeworse. (C3') implies that when we act in this way we are acting wronglywhenever we could have acted differently, in a wa y that would not have madethe outcome worse. Would it have been possible that, in all of these cases, weacted differently, in a way that would not have made the outcome worse? Weare not here asking whether this would have been causally possible, givenour actual desires and dispositions. We are asking whether this would havebeen possible, if our desires and dispositions ha d been different. The answerto this question is No. It would not have been possible for us to have acteddifferently, in all of these cases, in a wa y that would not have made theoutcome worse. We could have acted differently, in all of these cases, only if we were all pure do-gooders, and this would have made the outcome worse.

We should now ask, 'Could we have acted differently in some of thesecases, in a way that would not have made the outcome worse?' Theis Yes. It is impossible that we always act like pure do-gooders without havingthe disposition of a pure do-gooder. But it would have been possible for usto have sometimes acted in this way, without having this disposition. If Determinism is no t true, we could often have acted in this way. (C3') would thenimply that, though we are not always acting wrongly when we act in thisway, we are often acting wrongly. By substituting (C3') for (C3) a Consequentialist meets the objection that appeals to the doctrine that ought impliescan. And this revision does not make much difference.

7 Sidgwick, p. 490.8 Sen and Williams, p. 16.9 'Feeling that the deepest truth I have to tell is by no means ~ ' g o o d tidings . .. I would not, if I could . . . say anything which would make philosophy .

- my philosophy - popular.' Sidgwick, pp. 395-6.10 Williams, p. 135.11 Williams, p. 135, in the same sentence.12 Sen and Williams, p. 15.

From Reasons an d Persons 15913 Williams, p. 135.14 Rawls, pp. 133 and 182, especially footnote 31.

References\Adams, R. M,; ' r ~ ~ o t i v e Utilitarianism/' The Journal of Philosophy 73 (1976): 467-81.Hare, R. M., Ethical Theo ry and Utilitarianism" in A. Sen and B w111 Ut'l-.t . . I !aiDS, lt a r ~ a m s m and ~ e y o n d (Cambridge: Cambridge University Press, 1982).Mackie, J. L., Ethzcs (Harrnondsworth: Penguin Books, 1977).Naget Thomas, "War and Massacre/' Philosophy and Public Affairs 1 (1972): 123-44.

:. Rawls, John, A T h e ~ r y o ( J u ~ t i c e (Cambridge, MA: Harvard Universit y Press, 1971).Donald, Utzlztanamsm and Co-operation (Oxford: Clarendon Press, 1980).A r n ~ r t y a Bernard Williams, Utilitarianism and Beyond (Cambridge:Carnbndge Umversity Press, 1982).

Henry, The Methods of Ethics (London: Macmillan, 1967).. . .Be:nard, "A r i t i q ~ e of Utilitarianism/' in J. J. C. Smart and B. Williams,

Utzhtanamsm: For and Agamst (Cambridge: Cambridge Universi ty Press, 1973).

Date post:	14-Apr-2018
Category:	Documents
Upload:	karl-andersson
View:	220 times
Download:	0 times

Parfit - From Reasons and Persons

Documents