Readings in Ethical Theory [2 ed.]

Citation preview

second edition

Readings iii Ethical Theory Edited by

Wilfrid Sellars University of Pittsburgh

John Hospers University of Southern Californw

PRENTICE-HALL, INC., Englewood Cliffs, New Jersey


by PRENTICE-HALL, Inc., Englewood Cliffs, NeW Jersey All rights reserved. No part of this book may be reproduced in any form or by any means, without permission in writing from the publisher.

Printed in the United States of America 1SBN: 0-13-756007-9

Library of Congress Catalog Card Number: 75-107 425 10 9 8 7 6 5 4 3


Preface The enthusiastic re,,J>onse to Readings in Philosophical Analysis has reinforced our conviction that there is a genuine need for collections of important papers and other readings in the various areas of philosophy. The present volume is an attempt to satisfy this need in the field of ethical theory. ;\fore specifically, our aim has been to pro,ide a balanced and first-hand account of the theoretical controversies that have developed in ethics since the publication in 1903 of Moore's Principia Ethica. With the exception of a very few eases in which it seemed dear from the beginning that an item belonged in our collection, we have pondered our choices seriously and long. In many cases it was extremely difficult for us to make up our minds. More than one "final" list was scrapped when excluded items made their absence felt, and had it not been for the necessity of meeting a deadline, this process might have gone on indefinitely. "Iu De!iberation, the last Appetite . . . immediately adhearing to the action that wee call the WILL. ..."• In this sense only did we "will" the following group of selections rather than any one of a number of lists between which, like Buridan's ass, we hesitated. Though the volume as a whole is organized by topics rat.1:ier than by publication date, we have followed, ceteris paribus, the chronological order of items within each section, thus reproducing the sequence in which controversy developed. However, we have occasionally seen flt to violate chronological order even here, namely when the logical or pedagogical order ·of ideas was such that an issue could be more clearly grasped by doing so. For example, in Part IV, the essay "Pree-will and Psychoanalysis," though chronologically later, has been placed before the three remaining essays of the section, inasmuch as the empirical material it contains gives body to the more abstract treatment of moral freedom in the remaining essays. Teachers who use this book should by no means feel it incumbent upon them to assign the material in the order in which it is presented in this anthology. As a matter of fact, the order in which the selections are printed is the result of highlighting only a few of the complex interrelalionships which exist between them. It is useful primarlly as a point aappui for the exploration of the contro­ versies to be found in the literature of ethical theory, but it is not the order in which the selections are necessarily to be read or taught. 'vVe can conceive of several equally valid sequences in which the readings might be studied, and have no doubt that colleagues who use the book will think of still others. Since many of the essays deal with a variety of topics, it was by no means obvious in many cases where a given essay belonged, and our grouping has been accompanied by awareness of the fact that the compartments are not water-tight and many


.... Thomas Hobbes, Le-vJathan, Part I, Chapter 6, quoted from the Clarendon Press ( Oxford) edition, 1909, p, 46. V



selections bulge their compartments. \Ve hasten to add that we did not begin with the Part headings and look for material appropriate to them; rather the headings varied with the successive lists of essays in the process of mutual accommodation. Our thanks are due to the many friends and colleagues who have aided and encouraged us in this enterprise, and particularly to Professor William Frankena of the University of Michigan, whose detailed suggestions and comments at all stages have been invaluable, though he must not be held responsible for the contents of the volume. We wish finally to tender our apologies and regrets to those whose essays, sometimes at the last moment, we were obliged to omit; particularly in those cases where permission to reprint had already been secured from both author and publisher. The bibliography at the end of the volume exhibits the richness from which we had to choose, and our publishers have indeed been generous in giving us as much - space as they have. This book should be of particular use in introductory courses in ethics at the senior college level, second courses in ethics, courses in ethical theory, and seminars in moral philosophy and theory of value. It also contains invaluable material for courses and seminars in contemporary philosophy and in philosopbical analysis. vVe wish to express our deep appreciation to the authors of the articles included in this anthology for their kind permission to reprint, either in full or by way of excerpt, the material here presented. Our gratitude is also extended to the original editors and publishers of these essays for their friendly cooperation.

In the seventeen years since the first edition of tbis book appeared, the literature of ethical theory has become so extensive as to defy encapsulation in one volume, however lengthy. Nevertheless, the passage of time has dated the first edition sufficiently to make us try. In the revised edition we have not only attempted to update the material and delete those selections wbich in the light of recent developments seemed less important; we have also attempted to alter somewhat the range of subjects treated. We have, for example, iucluded some representative twentieth-century readings in normative ethics�an area which, except for G. E. 1'Ioore�s two brief chapters on utilitarianism, was not included iu the first edition. And we have, on the other hand, eliminated all readings specifically devoted to the free-will problem, except insofar as it bears directly on the problem of moral responsibility-a policy which seemed advisable iu view of the great variety of books and anthologies now available wbich are devoted exclusively to the problems of free-will and determiuism.

w.s. J.H.

Contents Preface

I. Introduatory The Elements of Ethics-Bertrand Russell


1 3

II. The Analysis of Ethical Concepts



Moore and the Naturalistic Fallacy


The Naturalistic Fallacy-,V. K. Frankena


The Indefinability of Good-G. E. Moore

How to Derive "Ought" from "Is"-John R. Searle

How Not to Derive "Ought" from "Is"-James Thomson

and Judith Thomson





Ethical Non-Naturalism


Does Moral Philosophy Rest on a Mistake?-H. A. Prichard


Ethical Judgments-Henry Sidgwick

On the Idea of a Philosophy of Ethics�]ames Balfour

The. Meaning of "Right"-Sir David Ross

A Suggested Non-Naturalistic Analysis of Good-A. C. Ewing C. Ethical Naturalism




115 130

Hypostatic Ethics-George Santayana


The Meaning of 'Good'-Sir David Ross


Value as Any Object of Any Interest-R. B. Perry



Moral and Non-Moral Values-C. A. Campbell


Method in Ethics-Paul Henle

Ethical Absolutism and the Ideal Observer-Roderick Firth


'Good,' 'Right,' 'Ought,' 'Bad'-Brand Blanshard


Critique of Ethics-A. J. Ayer


Ethical Judg.ements-A. J. Ayer


A Suggestion About Value -W. H. F. Barnes

The Emotive Conception of Ethics and Its Cognitive ImplicationsC. L. Stevenson



254 267

Emotivism and Ethical Objectivity-Carl Wellman


Mixed Views


The Impasse in Ethics-and a Way Out-Brand Blanshard

The Meaning of "Goocf'-Patrick Nowell-Smith



vVhat Is a Value Judgement?-R. M. Hare


Relativism and Justification


A Quasi-Naturalist Definition-RichardBrandt

Ethical Relativism-Richard Brandt

The Justification of Value Judgments: Rational Choice-Paul Taylor

Aren't Moral Judgments "Factual"?-Martin E. Lean

III. Theories of Normative Ethics A.



The Emotive Meaning of Ethical Terms-C. L. Stevenson



Ethical N o n -Cognitivism

Critique of Ayer-Sir DavidRoss



What Things Have Value?

331 335

346 369




Multiple Intrinsic Goods-G. E. Moore

Why Not Hedonism?-Ralph Mason Blake

Intrinsic Valm,-Monroe C. Bearibley

The Good of Man-G. H. von Wright


vVhat Acts Are RightF'

Utilitarianism-G. E. Moore

Utilitarian Generalization-David Lyons

The Moral Point of View-Kurt Baier

What Makes Right Acts Right?-Sir David Ross

Universal Prescriptivism-R. M. Hare

The Trivializability of Universalizability-Don Locke

Generalization in Ethics-Marcus Singer


387 392 401 413 431 431 451 469

482 501 517 529

Generalization Arguments-!. Howard Sobel


JV. Rights, Justice, Punishment, and Responsibility





On. Natural Rights-Ralph Mason. B..lalf_�•.• .... Rights-Sir David Ross

B. Justice

Justice as Fairness-John Raw/,s

Problems of Distributive Justice-Nicholas Rescher C. Punishment and Responsibility Punishment-Sir David Ross

"The Justification of Punishment"-Antony Flew

567 573

578 578 596 615 615 620


Free-W ill and-Psychoanalysis-John Hospers

The Humanitarian Theory of Punishment-C. S. Lewis

The Humanitarian Theory of Punishment-]. J. C. Smart

Determinism and Moral Perspectives-Elizabeth L. Beardsley





V. Ethics and Psychology



Reasons and Causes


Reasons and Causes-A. J. Ayer



Reasons and Causes-S. I. Benn and Richard Peters

Obligation and Motivation

Remarks on Psychological Hedonism-C. D. Broad

Duty and Interest-H. A. Prichard A Criticism of Kant-G. C. Field

Obligation and . Motivation in Recent Moral Philosophy-

W. K. Frankena






704 708

Why Be Moral?


Why Should I Be Moral?-Kai Nielsen


Why Be Moral?-John Hospers

Suggested Further Readings Index




I Introducto1y

The Elements of Ethics BERTRAND R USSELL I. THE SUBJECT-MATTER OF ETHICS' 1. The study of Etlilcs is perhaps most commonly conceived as being concerned with the questions "What sort of actions ought men to perform?" and ''What sort o f actions ought men to avoid?" It is con­ ceived, that is to say, as dealing with human conduct, and as deciding what is ,irtuous and what vicious among the ldnds of conduct bet\veen which, in practice, people are called upon to choose. Owing to this view of the pro'vince of ethics, it is sometimes regarded as the practical study, Reprinted fron1 Philosophical Essays by Ber­ trand Russell by -pennissfon of George Allen & Umvin Ltd. and Simon & Schuster1 Inc, 'I'he author had :requested that the following note be printed in conjunction with this selection: " "The Elements of Ethics' was written u11der the influence of Moorets Principia Ethica. There are some important points in ,vhkh, not long after publishing it, I came to disagree ,vith the theory that it advocates. I do not now think that 'good' is u.,ry2.efinable, and I thi.'lk that whatever objectivity the coacept may possess' is politica: rather than logical. I v,ras £....-st led to tltis view by Santayana's critid.sms of my "\t•ork in his lVinds of Doctrlr.e1 hut .have since found confirmation in ma:cy other directions. I am not, however, quite satisfied with any view of ethics that I have been able lo anive at, and that is why I have abstained from \Vriting again on the subject.'' t ·what follows is largely based on Mr. G. E. Moore's Principia Ethica,_ to which the reader is referred for fuller discussions, Section:; I, and II. of the following essay• are repl'fated from the New Quarterly, February, 1910; section !JI. from the New Quarterly, May, 19l0; section IV. from the Hibbert Journal, October, 1908; and sections V. and VI. from the New Quarterly, September, l910.

to which all others may be opposed as theoretical; the good and the tme are somet'.mes spoken of as independent king­ doms, the former belonging to ethics, while the !alter belongs to the sciences. This view, however, is doubly defec­ tive. In the first place, it overlooks the fact that the object of ethics, by its own account, is to discover true propositions about virtu­ ous and vicious conduct, and that these are just as much a part of truth as true propo­ sitions about oxygen or the multiplication table. The aim is, not practice, but propo­ sitions about practice; and propositions about practice are not themselves practical, any more than propositions about gases are gaseous. One might a.s well maintain faat botany is vegetable or zoology animal. Thus the study of ethics is not something outside science and co�ordinate with it; it is merely one among sciences. 2. In the sec-ond place, the view in question 1mduly limits the province of eth­ ics. "\'Vhen we are told that actions of cer­ tain kinds ought to be performed or avoided, as, for example, that we ought to speak the truth, or that we ought not to steal, we may always legitimately ask for a reason, and this reason ,vill always be con­ cerned, not only with the actions them­ selves, but also with fue goodness or bad­ ness of fue consequences likely to follow from such actions. '\Ve shall be told that truth- speaking generates mnh1al confi­ dence, cements friendships, facilitates the dispatch of business, and hence increases the wealth of the society which practises 3

4 it, and so on. If we ask why we should aim at increasing mutual con£dence, or cement­ ing friendships, we may be told that ob­ viously tbese tbings are good, or tbat tbey lead to happiness, and happiness is good. If we still ask why, the plain man will probably feel irritation, and will reply that he does not know. His irritation is due to tbe conflict of two feelings-the one, that whatever is true must have a reason; the other, tbat the reason he has already given is so obvious that it is merely contentious to demand a reason for the reason. In the second of tbese feelings he may be right; in tbe first, he is certainly wrong. In ordinary life, people only ask wh1f when they are unconvinced. If a reason is given which tbey do not doubt, they are satisfied. Hence, when they do ask wh1J, they usually have a logical right to expect an answer, and tbey come to think that a belief for which no reason can be given is an unrea­ sonable belief. But in this they are mis­ taken, as tbey would soon discover if tbeir habit of asking wh1f were more persistent. It is tbe business of the philosopher to ask for reasons as long as reasons can legit­ imately be demanded, and to register the propositions which give the most ultimate reasons that are attainable. Since a proposi­ tion can only be proved by means of other propositions, it is obvious that not all prop­ ositions can be proved, for proofs can only begin by assuming something. And since the consequences have no more certainty than their premisses, the things that are proved are no more certain than the things tbat are accepted merely because they are obvious, and are then made the basis of our proofs. Thus in the case of ethics, we must ask why such and such actions ought to be performed, and· continue our back­ ward inquiry for reasons until we reach the kind of proposition of which proof is im­ possible, because it is so simple or so ob­ vious that nothing more fundamental can be found from which to deduce it. 3. Now when we ask for the reasons in favour of the actions which mora1ists rec-

INTRODUCTORY ommend, these reasons are1 usually, that tbe consequences of the actions are likely to be good, or if not wholly good, at least the best possible under the circumstances. Hence all · questions of conduct presuppose tbe decision as to what things other than conduct are good and what bad. What is called good conduct is conduct which is a means to other things which are good on tbeir own account; and hence the study of what is good on its ovm. account is neces­ sary before we can decide upon rules of conduct. And tbe study of what is good or bad on its o,vn account must be included in ethics, which . thus ceases to be con­ cerned only with human conduct. The first step in ethics, therefore, is to be quite clear as to what we mean by good and bad. Only tben can we return to con­ duct, and ask how right conduct is related to tbe production of goods and tbe avoid­ ance of evils. In tbis, as in all philosophical inquiries, after a preliminary analysis of complex data we proceed again to build up complex things from their simpler con­ stituents, starting from ideas which we un­ derstand though we cannot define them, and from premisses ,vhich we know though we cannot _prove them. The appearance of dogmatism in this procedure is deceptive, for the premisses are such as ordinary rea­ soning unconsciously assumes, and there is less real dogmatism in believing them after a critical scrutiny tban in employing tbem implicitly without examination.

II. THE MEANING OF GOOD AND BAD 4. Good and Bad, in tbe sense in which tbe words are here intended ( which is, I be­ lieve, their usual sense), are ideas which everybody, or almost everybody, possesses. These ideas are apparently among tbose which form tbe simplest constituents of our more complex ideas, and are therefore in­ capable of being analysed or built up out


of other · simpler ideas. \Vhen people ask "What do you mean by Good!"' the answer must consist, not in a verbal definition such as could be given if one were asked "VVhat do you mean by Pentagon?" but :in such a characterisation as shall call up fue ap­ propriate idea to the mind of the ques­ tioner. This characterisation rnay, and probably will, itself contrun the idea of .good, which would b e a fault :in a defini­ tion, but is harmless when our purpose is merely to stimulate the imagination to the production of the idea which is intended. It is :in this way that children are taught the names of colours : they are shovvn ( say) a. red book, and told that that is red; and for fear thev should think red means book, they ar; shown also a red flower, a red ball, and so on, and told that these are all red. Thus the idea of redness is conveyed to their minds, although it is quite impossi­ ble to analyse redness or to find constitu­ ents which compose it. In the case of good, the process is more difficult, both because goodness is not per­ ceived by the senses, like redness, and be­ cause there is less agreement as to the things that are good than as to the things that are red. This is perhaps one reason that has led peoPle to think that th e notion of good could be analysed into some other notion, such as pleasure or object of desire. A second reason, probably more potent, is the common confusion that makes people think they cannot understand an idea un­ less they can define it- forgetting that ideas are defined by other ideas, which must be already understood if the definition is to convey any meaning. V>'hen people begin to philosophise, they seem to make a point of forgetting everyt.1:ilng familiar and ordi­ nary; othern,iBe their acquaintance with, redness or any other colour might show them how an idea can be intelligible where definition, in the sense of analysis� .is impos­ sible. 5. To explrun what we mean by Good and Bad, we :nay say that a thing is good when on its own account it ought to existJ


and bad ,v-hen on its own account it ought not to exist. If it seems to be in our pO\ver to cause � thing to exist or not to exist, we ought to try to make it exist if it is good, and not exist if it is bad. When a thing is good, it is fitting that we should feel pleas­ ure in its existence; when it is bad, it is fit� ting that we should feel pain ln its exist­ ence. But all such characterisations really presuppose the notions of good and bad, and are therefore useful only as means of calling up the right ideas, ;,ot as logical definitions. It might be thought that goad could he defined as the quality of whatever we ought to try to produce. This would merely put ought :in the place of good as our ulti­ mate undefined notion; but as a matter of fact the good is much wider than what we ought to try to produce. There is no reason to doubt that some of the lost tragedies of Aeschylus were good, but we ought not to try to re-,vrite them, because we should certainly fail. What we ought to do, in fact, is limited by our powers and opportunities, whereas the good is subject to no such lim­ itation. And our knowledge of goods ls con­ fined to the things we have experienced or can imagine; but presumably there are many goods of which we human beings have absolutely no knowledge, because they do not come within the very restricted range of our thoughts and feelings. Such goods are still goods, although human eon­ duct can have no reference to th em. Thus the notion of good is wider and more fun­ damental than any notion concerned with conduct; we use the notion of good in ex­ plaining what right conduct i-;, but we do not use the notion of right conduct in ex­ plaining what good is. 6. A fairly plausible view is that good means the same as desired) so that when we say a thing is good we mean that it is desired. Thus anything is good which we either hope to acquire or fear to lose. Yet it is commonly admitted that there are bad desires; and when people speak of bad de­ sires, they seem to mean desires for what is

6 bad. For example, when one man desires another man's pain, it is obvious that what is desired is not good but bad. But the sup­ porter of the view that good. means desired ,vill say that nothing is good or bad in i t ­ self, but i s good for one person and per­ haps bad for another. This must happen, he will say, in every case of a conflict of desires; if I desire your suffering, then your suffering is good for me, though it is bad for you. But the sense of good and bad which is needed in ethics is not in this way personal; and it is quite essential, in the study of ethics, to realise that there is an impersonal sense. In this sense, when a thing is good, it ought to exist on its own account, not on account of its conse­ quences, nor yet of who is going to enjoy it. We cannot maintain that for me a thing ought to exist on its own account, while for you it ought not; that would merely mean that one of us is mistaken, since in fact ev­ erything either ought to exist or ought not. Thus the fact that one man's desire may be another man's aversion proves that good, in the sense relevant to ethics, does not mean the same as desired, since everything is in itself either good or not good, and cannot be at once good for me and bad for you. This could only mean that its effects on me were good, and on you bad; but here good and bad are again impersonal. 7. There is another line of argument, more subtle but more instructive, by which we can refute those who say that good means desired, or who propose any other idea, such as pleasure, as the actual mean­ ing of good. This line of argument \viii not prove that the things that are good are not the same as the things that are desired; but it will prove that, if this were the case, it could not be proved by appealing to the meaning of the word "good." So far, it might be thought that such an argument could only have a purely logical impor­ tance. But in fact this is not so. Many ethi­ cal theories have been based upon the con­ tention that ugood" means so-and-so, and people have accepted consequences of this


contention which, if they had relied upon inspection untrammelled by false theory, they would almost certainly have rejected. Whoever believes that "good" means "de­ sired" \viii try to explain away the cases where it seems as if what is desired is bad; but if he no longer holds this theory, he will be able to allow free play to his un­ biassed ethical perceptions, and will thus escape errors into which he would other­ wise have fallen. The argument in question is this: If any one. affirms that the good is the desired, we consider what he says, and either assent or dissent; but in any case our assent or dis­ sent is decided by considering what the good and the desired really are. vVhen, on the contrary, some one gives a definition of the meaning of a word, our state of mind i s quite different. I f w e are . told " a pentagon is a figure which has five,,-sides," we do not consider what we Jsn�v about pentagons, and then agree �disagree; we accept this as the meaning/of the word, and we know that we are getting information, not about pentagons, ; but merely about the word "pentagon,," \Vhat we are told is the sort of thing that we expect dictionaries to tell us. But when/ we are told that the good is the desired, we feel at once that we are being told smtiething of philosophical impor­ tance,(something which has ethical conse­ quen,tes1 something which it is quite be­ yond the scope of a dictionary to tell us. )h'e reason of this is, that we already know what we mean by the good, and what we 1 mean by the desired; and if these two meanings always applied to the same ob­ jects, that would not be a verbal definition, but an important truth. TI1e analogue of such a proposition is not the above defini­ tion of a pentagon, but rather: "A pentagon ( defined as above) is a figure which has five angles." Whenever a proposed defini­ tion sets us thinking whether it is true in fact, and not whether that is how the word is used, there is reason to suspect that we are not dealing with a definition, but ,vith a significant proposition, in which the word /,

THE ELEMENTS OF ETHICS professedly defined has a meaning already known to us, either as simple or as defined in some other way. By applying this test, we shall easily convince ourselves that all hitherto suggested definitions of t.'1e good are significant, not merely verbal, proposi­ tions; and that therefore, though they may be true in fact, they do not give the mean­ ing of the word "good." The importance of this result is that so many ethical theories. depend upon the de­ nial of it. Some have contended that "good" means «desired/� others that '"good'' rneans "pleasure,"' others again that it means "'con­ £om1itv to Nature" or "obedience to the will of God," The mere fact that so many different and incompatible definitions have been proposed is evidence against any of them being really definitions; !there have never been two incompatible definitions of the word "pentagon." None of the above are really definitions; t,':iey are all to be un­ derstood as substantial affirmations con­ cerning the things that are good. All' of them are, in my opinion, mistaken in fact as well as in form, but I shall nt here un­ dertake to refute them severally. 8. It iB important to realise that when we say a thing is good :in itse::£, and not merely as a means, we atbibute to the thing a property which it either has or does not have, quite independently of our opin­ ion on the subject, or of our wiBhes or other people's. Most men are inclined to agree with Hamlet: "There is nothing good or bad but thinking makes it so." It is sup­ posed that ethical preferences are a mere matter of taste, and that if X thinks A is a good thing, and Y thinks it is a bad thing, all we can say is that A is good for X and bad for Y. This view is rendered plausible by the divergence of opinion as to what is good and bad, and by the difficulty of find­ ing arguments to persuade people who dif­ fer from us in such a question. But diffi­ culty in discovering the truth does not prove that there is no truth to be discov­ ered. If X says A is good, and Y says A is bad, one of them must be mistaken, though

7 it may be impossible to disccver which. If this were not the case, there would b e no difference of opinion between them. If, in asserting that A is good, X meant merely to assert that A had a certain reiation to him­ self, say of pleasing his taste in some way; and if Y, in saying that A is not good, meant merely to deny that A had a like re­ lation to himself: then the,re would be no subject of debate between them. It would be absurd, if X said "I am eating a pigeon pie," for Y to answer "that is false: I am eating nothing." But this is no more absmd than a dispute as to what is good, if, when we say A is good, we mean merely to affirm a relation of A to ourselves. When Chris­ tians assert that God is good, they do not mean merely that the contemplation of God arouses certain emotions in them: they may admit that this contemplation rouses no such emotion in the devils who believe and tremble, but the absence of such emo­ tions is one of the things that make devUs bad. As a matter of fact, we consider some tastes better than others: we do not hold merely that some tastes are ours and other tastes are other people's. We do not even always consider our own tastes the best: we may prefer bridge to poetry, but think it better to prefer poetry to bridge. And when Christians affirm that a world cre­ ated by a good God must be a good world, they do not mean that it must be to their taste, for often it is by no means to their taste, but they use its goodness to argue that it ought to be to their taste. And they do not mean mereIv that it is to God's taste: for that woula'have been equally the case if God had not been good. Thus good and bad are qualities which belong to ob­ jects independently of our opinions 1 Just as much as round and square do; and when two people differ as to whether a thing is good, only one of them can be right, though it may b e very hard to know which is right 9. One very important consequence of the indefinability of good must be empha­ sised, namely, the fact that knowledge as to


what things exist, have existed, or will exist, can throw absolutely no light upon the question as to what things are good. There might, as far as mere logic goes, be some general proposition to the effect "whatever exists, is good,"' or "whatever ex­ ists, is bad," or "what "�II exist is better ( or worse) than what does exist." But no such general proposition can be proved by con­ sidering the meaning of «good," and no such general proposition can be arrived at empirically from experience, since we do not know the whole of what does exist, nor yet of what has existed or will exist. We cannot therefore arrive at such a general proposition, unless it is itself self-evident, or follo,vs from some self-evident proposi­ tion, which must ( to warrant the conse­ quence) be of the same general kind. But as a matter of fact, there is, so far as I can discover, no self-evident proposition as to the goodness or badness of all that exists or has existed or will exist. It follows that, from the fact that the existent world is of such and such a nature, nothing can be in­ ferred as to what things are good or bad. 10. The belief that the world is wholly good has, nevertheless, been widely held. It has been held either because, as a part of revealed religion, the world has been sup­ posed created by a good and omnipotent God, or because, on metaphysical grounds, it was thought possible to prove that the sum-total of existent things must be good. vVith the former line of argument we are not here concerned; the latter must be briefly dealt with. The belief that, without assuming any ethical premiss, we can prove that the world is good, or indeed any other result containing the notion of good, logically in­ volves the belief that the notion of good is complex and capable of definition. If when we say that a thing is good we mean ( for example ) that it has three other simpler properties, then, by proving that a thing has those three properties we prove that it is good, and thus \Ve get a conclusion in­ volving the notion of good, although our


premisses did not involve it. But if good is a simple notion, no such inference will be possible; unless our premisses contain the notion of good, our conclusion cannot con­ tain it. The case is analogous to the case of elements and compounds in chemistry. By combining elements or compounds we can get a new compound, but no chemical op­ eration will give an element which was not present in the beginning. So, if good is sim­ ple, no propositions not containing this no­ tion can have consequences which do con­ tain it. As a matter of fact, those who have en­ deavoured to prove that the world as a whole is good have usually adopted the view that all evil consists wholly in the ab­ sence of something, and that nothing posi­ tive is evil. This they have usually sup­ ported by defining good as meaning the same as real. Spinoza says: 2 "By reality and perfection I mean the same thing"; and hence it follows, ,vith much le�s trou­ ble than metaphysicians have usually taken in the proof, that the real is perfect. This is the view in "Abt Vogler": "The evil is null, is nought, is silence implying sound." YVhenever it is said that all evil is limi­ tation, the same doctrine is involved; what is meant is that evil never consists in the ex­ istence of something ,vhich can be called bad, but only in the non-existence of some­ thing. Hence everything that does exist must be good, and the sum-total of exist­ ence, since it exists most, must be the best of all. And this view is set forth as resulting from the meaning of evil. The notion that non-existence is what is meant by e,11 is refuted exactly as the previous definitions of good were refuted. And the belief that, as a matter of fact, nothing that exists is evil, is one which no one would advocate except a metaphysi­ cian defending a theory. Pain and hatred and envy and cruelty are surely things that exist, and are not merely the absence of their opposites; but the theory should hold that they are indistinguishable from the ,.., Ethics, pt, ii. df. vi.


blank unconsciousness of an oyster. Indeed, it would seem that this whole theory has been advanced solely because of the un­ conscious bias in favour of opllinism, and that its opposite is logically just as tenable. We might urge that evil consists in exist­ ence, and good in non-existence; that therefore the sum-total of existence is the worst thing there is, and that only non­ existence is good. Indeed, Buddhism does seem to maintain some such view. It is plain that this view is false; but logically it is no more absurd than its opposite. 11. e cannot, then, i:µfer any results as to what is good or bad from a study of the things that exist. This conclusion needs chiefly, at the present time, to be applied against evolutionary ethics. The phrase "survival of the fittest" seems to have given rise t o the belief that those who survive are the fittest in some ethical sense, and that the course of evolution gives evidence that the later type is better than the earlier. On this basis, a worship of force is easily set up, and the mitigation of struggle by civili­ sation comes to be deprecated. It is thought that what fights most successfully is most admirable, and that what does not help in fighting is worthless. Such a view is wholly destitute of logical foundation. The course of nature, as we have seen, is irrele­ vant in deciding as to what is good or bad. A priori, it would be as probable that evo­ lution should go from bad to worse, as that it should go from good to better. What makes the view plausible is the fact that the lower animals existed earlier than the higher, and that among men the civilised races are able to defeat and often extermi­ nate the uncivilised. But here the ethical preference of the higher to the lower ani­ malsi and of the exterminators to the exter­ minated, is not based upon evolution1 but exists independently, and unconsciously in­ trudes into our judgment of the evolution­ ary process. If evolutionary ethics were sound, we ought to be entirely indifferent as to what the course of evolution may be, since whatever it is is thereby proved to be



the best. Yet if it should tum out that the negro or the Chinaman was able to oust the European, we should cease to have any admiration of evolution; for as a matter of fact our preference of the European to the negro is wholly independent of the Euro­ pean's greater prowess ·with the Niaxim gun. Broadly, the fact that a thing is un­ avoidable affords no evidence that it is not an evil; and the fact that a thing is impos­ sible affords no evidence that it is not a good. It is doubtless foolish, in practice, to fret over the inevitable; but it is false, in theory, to let the actual world dictate our standard of good and evil. It is evident that among the things that exist some are good, some bad, and that we know too little of the universe to have any right to an opinion as to whether the good or the bad prepon­ derates, or as to whether either is likely in the future to gain on the other. Optimism and pessimism alike are general theories as to the universe \Vhich there is no reason ,vhatever for accepting; what we know of the world tends to suggest that the good and the evil are fairly balanced, but it is of course possible that what we do not lmow is very much better or very much worse than what we do know. Complete suspense of judgment in this matter is therefore the only rational attitude.

III. RIGHT AND WRONG 12. The ideas of right and wrong conduct are, as we have seen, those with which eth­ ics is generally supposed to b e most con­ cerned. This view, which is unduly narrow, is fostered by the use of the one word good, both for the sort of conduct which is right, and for the sort of things which ought to exist on account of their intrinsic value. This double use of the word good is very confusing, and tends greatly to ob­ scure the distinction of ends and means. I shall therefore speak of right actions, not of


good actions, confining the word good to the sense explained in Section IL The word "right'' is very ambiguous, and it is by no means easy to distinguish the various meanings which it has in common parlance. Owing to the variety of these meanings, adherence to any one necessarily involves us in apparent paradoxes when we use it in a context which suggests one of the other meanings. Tbis is the usual result of precision of language; but so long as the paradoxes are merely verbal, they do not give rise t o more than verbal objections. In judging of conduct we find at the outset two widely divergent methods, of which one is advocated by some moralists, the other by others, while both are prac­ tised by those who have no ethical theory. One of these methods, wbich is that advo­ cated by utilitarians, judges the rightness of an act by relation to the goodness or badness of its consequences. The other method, advocated by intuitionists, judges by the approval or disapproval of the moral sense or conscience. I believe that it is necessary to combine both theories in order to get a complete account of right and wrong. There is, I think, one sense in which a man does right when he does what will probably have the best consequences, and another in which he does right when he follows the dictates of his conscience, whatever the probable consequences may be. ( There are many other senses which we may give to the word right, but these two seem to be the most important. ) Let us begin by considering the second of these senses. 13. The question we have to ask o u r ­ selves i s : What d o w e mean b y the dictates of the moral sense? If these are to afford a definition of right conduct, we cannot say that they consist in judging that such and such acts are right, for that would make our definition circular. We shall have to s�y that the moral sense consists in a certain specific emotion of approval towards an act, and that an act is to be called right when the agent, at the moment of action,


feels this emotion of approval towards the action which he decides to perform. There is certainly a sense in which a man ought to perform any act wbich he approves, and to abstain from any act which he disap­ proves; and it seems also undeniable that there are emotions which may be called approval and disapproval. Thus this theory, whether adequate or not, must be allowed to contain a part of the truth. It is, however, fairly evident that there are other meanings of right conduct, and that, though there is an emotion of ap­ proval, there is also a judgment of ap­ proval, which may or may not be true. For we certainly hold that a man who has done an action which his conscience approved may have been mistaken, and that in some sense his conscience ought not to have ap­ proved bis action. But this would be impos­ sible if nothing were involved except an emotion. To be mistaken implies a judg­ ment; and thus we must admit that there is such a tbing as a judgment of approval. If this were not the case we could not reason with a man as to what is right; what he ap­ proves would be necessarily right for bim to do, and there could be no argument against his approval. We do in fact hold that when one man approves of a certain act, while another disapproves, one of them is mistaken, which would not be the case with a mere emotion. If one man likes oys­ ters and another dislikes them, we do not say that either of them is mistaken. Thus there is a judgment of approval,3 and this must consist of a judgment that an act is, in a new sense, right. The judgment s The judgment of approval does not always coincide with the emotion of approval. For exam­ ple, when a man has been led by his reason to reject a moral code which he formerly held, it will commonly happen, at least for a time, that his emotion of approval follows the old code, though his judgment has abandoned it. Thus he may have been brought up, like 1vfohammed's first disciples, to believe it a duty to avenge the murder of rela­ tions by murdering the murderer or his relations; and he may continue to feel approval of such vengeance after he has ceased to fudge it approv­ ingly. The emotion of approval will not be again in question in what follows.

THE ELEMEJ\'TS OF ETHICS of approval is not merely the judgment that we feel the emotion of approval, for then another who disapproved would not neces­ sarily hold our judgment of approval to be mistaken. Thus in order to give a meaning to the judgment of approval, it is necessary to admit a sense of right other than ap­ proved. In this sense, when we approve an act we judge that it is right, and we may be mistaken in so judging. This new sense is objective, in the s�nse that it does not depend upon the opinions and feelings of the agent. Thus a man who obeys the dic­ tates of his conscience is not always acting rightly in the objective sense. When a man does what his conscience approves. he does what he believes to be objectively right, but not necessarily what is objectively right. \Ve need, therefore, some other crite­ rion than the moral sense for judging what is objectively right. 14. It is in denning objective rightness that the consequences of an action become relevant. Some moralists, it is true 1 deny the dependence upon consequences; but that is to be attributed, I think, to confu­ sion with the subjective sense. VVhen peo­ ple argue as to whether such and such an action is right, they always adduce the consequences which it has or may be ex­ pected to have. A statesman who has to de­ cide what is the right polky, or a teacher who has to decide what is the right educa­ tion, will be expected to consider what pol­ icy or what education is likely ta have the best results. vVheuever a question is at all complicated, and cannot be settled by fol­ lowing some simple rule, such as "thou shalt not steal," or "thou shalt not bear false witness;"' it is at once evident that the decision cannot be made except by consi­ deration o.f consequences. But even when tl1e decision can be rnade by a simple precept, such as not to lie or not to steal, the justiflcation of the pre­ cept is· found only by consideration of con­ sequences. A code such as the Decalogue, it must be admitted, can hardly be true with­ out exception if the goodnes; or badness of

11 consequences is what determines the right­ ness or wrongness of actions; for in so com­ plex a world it is unlikely that obedience to the Decalogue will always produce better consequences than disobedience. Yet ,t is a suspicious circumstance that breaches of those of the Ten Commandments which people still hold it a duty to obey do, as a matter of fact, have bad consequences in the vast majority of instances, and wodd not be considered \Vrong in a case i n which it was fairly certain that their consequences would be good. This latter fact is con­ cealed by a question-begging addition of moral words. Thus, e.g., "Thou shalt do no murder" would be an important precept if it were interpreted, a.s Tolstoy interprets it, to mean "thou shalt not take human life." But it is not so interpreted; on the contrary, some taking of human life is called "justiflable homicide." Thus murder comes to mean "unjustifiable homicide"; and it is a mere tautology to say, "Thou shalt do no unjustifiable homicide." That this should be announced from Sinai would be a.s fruitless as Hamlet's report of the ghost's message: "':l'here's ne'er a villain, dwelling in all Denmark, but he's an arrant knave." As a matter of fact, people do make a certain classification of ho.:ni­ cides, and decide that certain kinds are justiliable and certain others unjustifiable. But there are many doubtful cru;es: tyran­ nicide, capital punishment, killing in war, killing in self-defence, killing in defence of others, are some of these. And if a decision is sought, it is sought usually by consider­ ing whether the consequences of actions belonging to these classes are on the whole good or bad. Thus the importance of pre­ cepts such as the Ten Commandments lies in the fact that they give sin1ple mies, obedience to which will in almost all cases have better consequences than disobedi­ ence; and the justiflcation of the rules is not wholly independent of consequences, 15. In common language the received code of moral rules is usually presupposed, and an action is only called immoral ;,.vhen


it infringes one of these rules. 'Whatever does not infringe them is regarded as per­ missible, so that on most of the occasions of life no one course of action is marked out as alone right. If a man adopts a course of action which, though not contrary to the received code, will probably have bad consequences, he is called unwise rather than immoral. Now, according to the dis­ tinction we have made between objective and subjective rightness, a man may well act in a way which is objectively wrong without doing what is subjectively wrong, i.e. what his conscience disapproves. An act ( roughly speaking, I shall return to this point presently ) is immoral when a man's conscience disapproves it, but is judged only unwise or injudicious \Vhen his con­ science approves it, although we judge that it will probably have bad conse­ quences. Now the usual moral code is sup­ posed, in common language, to be admitted by every man's conscience, so that when he infringes it, his action is not merely in­ judicious, but immoral; on the other hand, where the code is silent, we regard an unfortunate action as objectively but not subjectively ,vrong1 i.e. as injudicious, but not immoral. The acceptance of a moral code has the great advantage that, in so far as its rules are objectively right, it tends to harmonise objective and subjective right­ ness. Thus it tends to cover all frequent cases, leaving only the rarer ones t o the individual judgment of the agent. Hence when new sorts of cases become common, the moral code soon comes to w a second most important result: namely, what is the nature of the evidence, by which alone any ethical prop­ osition can be proved or disproved, con­ firmed or rendered doubtful. Once we re­ cognise the exact meaning of the two ques­ tions, I think it, also becomes plain exactly what kind of reasons are relevant as argu­ ments for or against any particular answer to them. It becomes plain that, for answers to the first question, no relevant evidence whatever can be adduced: from no other truth, except themselves alone, can it be in­ ferred that they are either true or false. We can guard against error only by taking care, that, when we try to answer a ques­ tion of this kind, ,ve have before our minds that question only, and not some other or others; but that there is great danger of such errors of confusion I have tried to 31



shew, and also what are the chief precau­ tions by the use of which we may guard against them. As for the second question, it becomes equally plain, that any answer to it is capable of proof or disproof- that, in­ deedi so many different considerations are relevant to its truth or falsehood, as to make the attainment of probability very difficult, and the attaimnent of certainty impossible. Nevertheless the kind of evi­ dence, which is both necessary and alone relevant to such proof and disproof, is ca­ pable of exact definition. Such evidence must contain propositions of nvo kinds and of t\vo kinds only: it must consist, in the first place, of truths ,vith regard to the re­ sults of the action in question-of causal truths-but it must also contain ethical truths of our first or self-evident class. Many truths of both kinds are necessary to the proof that any action ought to be done; and any other kind of evidence is wholly ir­ relevant. It follows that, if any ethical phi­ losopher offers for propositions of the first kind any evidence whatever, or if, for prop­ ositions of the second kind, he either fails to adduce both causal and ethical truths, or adduces truths that are neither, his reason­ ing has not the least tendency to establish his conclusions. But not only are his con-: clusions totally devoid of weight: we have, moreover, reason to suspect him of the error of confusion; since the offering of ir­ relevant evidence generally indicates that the philosopher who offers it has had be­ fore his mind, not the question which he professes to answer, but some other en­ tirely different one. Ethical discussioni hitherto, has perhaps consisted chiefly in reasoning of this totally irrelevant kind. One main object of this book may, then, be expressed by slightly changing one of Kant's famous titles. I have endeavoured to \vrite 'Prolegomena to any future Ethics that can possibly pretend to be scientific.' In other words, I have endeavoured to dis­ cover what are the fundamental principles of ethical reasoning; and the establishment

of these principles, rather than of any con­ clusions which may be attained by their use, may be regarded as my main object. I have, however, also attempted, in Chapter VI, to present some conclusions, with re­ gard to the proper answer to the question 'What is good in itself?' which are very dif­ ferent from any which have commonly been advocated by philosophers. I have tried to define the classes within which all great goods and evils fall; and I have maintained that very many different things are good and evil in themselves, and that neither class of things possesses any other property which is both common to all its members and peculiar to them. In order to express the fact that ethical propositions of my first class are incapable of proof or disproof, I have sometimes fol­ lowed Sidgwick's usage in calling them 'In­ tuitions.' But I beg it may be noticed that I am not an 'Intuitionist/ in the ordinary sense of the term. Sidgwick himself seems never to have been clearly av.rare of the im­ mense importance of the difference . which distinguishes his Intuitionism from the common doctrine, which has generally been called by that name. The Intuitionist proper is distinguished by maintaining that propositions of my second class­ propositions which assert that a certain ac­ tion is right or a duty-are incapable of proof or disproof by any enquiry into the results of such actions. I i on the contrary, am no less anxious to maintain that prop­ ositions of this kind are not 'Intuitions,' than to maintain that propositions of my first Class are Intuitions. Again, I would ,vish it observed that, when I call such propositions 'Intuitions,' I mean merely to assert that they are in­ capable of proof; I imply nothing whatever as to the manner or origin of our cognition of them. Still less do I imply ( as most intuitionists have done ) that any prop­ osition whatever is true1 because we cog­ nise it in a particular way or by the ex­ ercise of any particular faculty: I hold, on



the contrary, that in every way in which it reference to my article on 'Teleology' in is possible to cognise a true proposition, it Baldwin's Dictionary of Philosophy and is also possible to cognise a false one. Psychology. vVhen this book had been already If I were to rewrite my work now, I completed, I found, in Brentano's 'Origin of should make a very different, and I believe the Knowledge of Right and Wrong,'1 opin­ that I could make a much better book. But ions far more closely' resembling my own, it may be doubted whether, in attempting than those of any other ethical writer with to satisfy myself, I might not merely render whom I am acquainted. Brentano appears more obscure the ideas which I am most to agree with me completely ( 1 ) in re­ anxious to ·convey, without a corresponding garding alJ ethical propositions as defined gain in completeness and accuracy. HO\v­ by the fact that they predicate a single ever that may be, my belief that to publish unique objective concept; ( 2 ) in dividing the book as it stands was probably the best such propositions sharply into the same thing I could do, does not prevent me from two kinds; ( 3) in holding that the first being painfulJy aware that it is full of kind are incapable of proof; and ( 4 ) with defects. regard to the kind of evidence which is necessary and relevant to the proof of the second kind. But h.e regards the fun­ damental ethical concept as being, not the PRINCIPIA ETHICA: CHAPTER I simple one which I denote by 'good,' but THE S UBJECT-MATTER OF the complex one which I have taken to ETHICS define 'beautiful'; and he does not recog­ nise, but even denies by implication, the 1. It is very easy to point out some among principle which I have calJed the principle our every-day judgments, with the truth of of organlc unities. In consequence of these which Ethics is undoubtedly concerned. two differences, his conclusions as to what vVhenever we say, 'So and so is a good things are good in themselves, also differ man,' or 'That fellow is a villain'; whenever very materially from mine. He agrees, we ask, 'V\lhat ought I to do?' or 'Is it wrong however, that there are many different for me to do like this?'; whenever we haz­ goods, and that the love of good and beau­ ard such remarks as 'Temperance is a vir­ tiful objects constitutes an important class tue and drunkenness a vice'-it is undoubt­ among them. edly the business of Ethics to discuss such I wish to refer to one oversight, of questions and such statements; to argue which I becam·e aware only when it was what is the true answer \Vhen we ask what too late to correct it, and which may, I am it is right to do, and to give reasons for afraid, cause unnecessary trouble to some thinking that our statements about the readers. I have omitted to discuss directly character of persons or the morality of ac­ the mutual relations of the several different tions are true or false. In the vast majority notions, ,vhich are all expressed by the of cases, where we make statements in­ word 'end.' The consequences of this omis­ volving any of the terms 'virtue,' 'vice,' sion may perhaps be partially avoided by a 'duty/ 'right,' 'ought,' 'good,' 'bad,' we are making ethical judgments; and if we wish 1 'The Origin of the Knowledge of Right and to discuss their truth, we shall be discuss­ \�hong.' By Franz Brentano. English Translation ing a point of Ethics. by Cecil Hague. Constable, 1902.-I have written a review of this book, which will, I hope, appear So much as this is not disputed; but it in the International Journal of Ethics for October, falls very far short of defining the province 1903. I may refer to this review for a fuller ac­ of Ethics. That province may indeed be decount of my reasons for diSagreeing \Vith Brentano,



fined as the whole truth about that which is at the same time common to all such judgments and peculiar to them. But we have still to ask the question: What is it that is thus common and peculiar? And this is a questi.01.t to which very different answers have been given by ethical phi­ losophers of acknowledged reputation, and none of them, perhaps, completely sat­ isfactory. 2. If we take such examples as those given above, we shall not be far wrong in saying that they are all of them concerned with the question of 'conduct'- with the question, what, in the conduct of us, human beings, is good, and what is bad, what is right, and what is wrong. For when we say that a man is good, we commonly mean that he acts rightly; when we say that drunkenness is a vice, we commonly mean that to get drunk is a wrong or wicked action. And this discussion of human conduct is, in fact, that -with which the name 'Ethics' is most intimately associated. It is so associated by derivation; and conduct is undoubtedly by far the commonest and most generally interesting object of ethical judgments. Accordingly, we £nd that many ethical philosophers are disposed to accept as an adequate definition of 'Ethics' the state­ ment that it deals with the question what is good or bad in human conduct. They hold that its enquiries are properly conflned to 'conduct' or to 'practice'; they hold that the name 'practical philosophy' covers all the matter with which it has to do. Now, with­ out discussing the proper meaning of the word ( for verbal questions are properly left to the writers of dictionaries and other persons interested in literature; philosophy, as we shall see, has no concern ,vith them) , I may say that I intend to use 'Ethics' to cover more than t4is-a usage, for which there is, I think, quite sufficient authority. I am using it to cover an enquiry for which, at all events, there is no other word: the general enquiry into what is good. Ethics is undoubtedly concerned with

the question what good conduct is; but, being concerned with this, it obviously does not start at the beginning, unless it is prepared to tell us what is good as well as \vhat is conduct. For 'good conduct' is a complex notion: all conduct is not good; for some is certainly bad and some may be indifferent. And on the other hand, other things, beside conduct, may be good; and if they are so, then, 'good' denotes some property, that is common to them and con­ duct; and if \Ve examine good conduct alone of all good things, then we shall be in danger of mistaking for this property, some property which is not shared by those other things : and thus we shall have made a mis­ take about Ethics even in this limited sense; for we shall not know what good conduct really is. This is a mistake which many writers have actually made, from limiting their enquiry to conduct. And hence I shall try to avoid it by considering first what is good in general; hoping, that if we can arrive at any certainty about this, it will be much easier to settle the question of good conduct: for we all know pretty well what 'conduct' is. This, then, is our first question: What is good? and vVhat is bad? and to the discussion of this question ( or these questions ) I give the name of Ethics, since that science must, at all ev­ ents, include it. 3. But this is a question which may have many meanings. If, for example, each of us were to say 'I am doing good now' or 1 had a good dinner yesterday,' these state­ ments would each of them be some sort of answer to our question, although perhaps a false one. So, too, when A asks B what school he ought to send bis son to, B's answer will certainly be an ethical judg­ ment. And similarly all distribution of praise or blame to any personage or thing that has existed, no\v exists, or \vill exist, does give some answer to the question '\Vhat is good?' In all such cases some particular thing is judged to be good or bad: the question 'vVhat?' is answered by 'This.' But this is not the sense in which a



scientific Efaics asks the question. Not one, noticed, the sense in which this book is a of all the many million answers of this particular book, and A's friend's advice kind, which mu;t be true, can form a· part particular advice. Casuistiy may indeed be of an ethical system; although that science more particular and Ethics nwre general; must contain reasons and principles suffi­ but that means that they differ only in de­ cient for deciding on the truth of all of gree and not in kind. And this is them. There are far too 111any persons1 universally true of 'particular' and 'general,' things and events in the world, past, pres­ when used in this common, but inaccurate, ent, or to come, for a discussion of their sense. So far as Ethics allows itself to give individual merits to be embraced in any lists of virtues or even to name constituents science. Ethics, therefore) does not deal at of the Ideal, it is indistinguishable from all with facts of this nature, facts that are Casuistry. Both alike deal with what is gen­ unique, individualt absolutely particulari eral, in the sense in which physics and facts with which such studies as history, chemistry deal with -.yhat is general. Just as geography, astronomy, are compelled, in chemistry aims at discovering what are the part at least, to deal. lu1d, for this reason, it properties of oxygen, wherever it occurs; is not the business of the ethical phi­ and not only of this or that jl'trticular spec­ losopher to give personal advice or ex­ imen of oxygen; so Casuistry aims at discovering what actions are good, whenev­ hortation. 4. But there is another meaning which er they occur. In this respect Ethics and may be given to the qt:estion '\Vhat is Casuistry alike are to be classed with such good?' 'Books are good' would be an sciences as physics, chemistry and ans\:ver to it, though an answer obviously physiology, iu their absolute distinction false; for some books are very bad indeed. from those of which histo.ry and geography And ethical judgments of this kind do in­ are instances. And it is to be noted that, deed belong to Ethics; though I shall not owing to their detailed nature, casuistical deal with many of them. Such is the judg­ investigations are actually nearer to physics ment 'Pleasure is good.'-a judgment, of and to chemistry than are the in­ which Ethics should discuss the truth, al­ vestigations usually assigned to Ethics. For though it is not nearly as important as that just as physics cannot rest content with the other judgment, with . which we shall be discovery that light is propagated by waves much occupied presently-'Pleasure of ether, but must go on to discover the is good.' It is ·judgments of this sort, which particular natirre of the ether-waves cor­ are made in such books on Ethics as con­ responding to each several colour; so tain a list of 'virtues- in Aristotle's 'Ethics' Casuistry, not content 'Nith the general law for example. But it is judgments of pre­ that charity is a virtue, 1nust atten1pt to cisely the same kind, which form the sub­ discover the relative merits of every d.iffer­ stance of what is commonly sepposed to be ent form of charity, Casuistry forms, a study different froin Ethics1 and one therefore, part of the ideal of ethlcal sci­ much less respectable-the study of Casuis­ ence: Ethics cannot be complete "\\'ithout it. try. vVe may be told that Casuistry differs The defects of Casuistry are not defects of from Ethics, in that it is n1uch more de­ principle; no objection can be taken to its tailed and particular, Ethics much more aim and object. It has failed only because general. But it is rnost. important to notice it is far too difficult a subject to be treated that Casuistr; does not deal with anything adequately in our present state of knowl­ that is absolutely particular-particular iu edge. The casuist has been unable to the only sense in which a perfectly precise distinguish, in the eases which he treats, line can be drawn between it and what is th.ose elements upon which their value de­ general. It is not particular in the sense just pends. Hence he often thinks two cases to



be alike in respect of value, when in reality they are alike only in some other respect. It is to mistakes of this kind that the per­ nicious influence of such investigations has been due. For Casuistry is the goal of eth­ ical investigation. It cannot be safely attempted at the beginning of our studies, but only at the end. 5. But our question "vVhat is good?' may have still another meaning. vVe may, in the third place, mean to ask, not what thing or things are good, but how 'good' is to be defined. This is ·an enquiry which be­ longs only to Ethics, not to Casuistry; and this is the enquiry which will occupy us first. It is an enquiry to which most special attention should be directed; since this question, how 'good' is to be defined, is the most fundamental question in all Ethics. That \vhich is meant by 'good' is, in fact, except its converse 'bad,' the only simple object of thought which is peculiar to Ethics. Its definition is, therefore, the most essential point in the definition of Ethics; and moreover a mistake with regard to it entails a far larger number of erroneous ethical judgments than any other. Unless this first question be fully understood, and its true answer clearly recognised, the rest of Ethics is as good as useless from the point of view of systematic knowledge. True ethical judgments, of the two kinds last dealt with, may indeed be made by those who do not know the answer to this question as well as by those who do; and it goes without saying that the two classes of people may lead equally good lives. But it is extremely unlikely that the most general ethical judgments v.�ll be equally valid, in the absence of a true ans,ver to this ques­ tion: I shall presently try to shew that the gravest errors have been largely due to beliefs in a false ansvver. And, in any case, it is impossilbe that, till the answer to this question be known, any .one should know what . is the evidence for any ethical judg­ ment whatsoever. But the main object of Ethics, as a systematic science, is to give

correct reasons for thinking that this or that is good; and, unless this question be answered, such reasons cannot be given. Even, therefore, apart from the fact that a false answer leads to false conclusions, the present enquiry is a most necessary and im­ portan� part of the science of Ethics. 6. f'hat, then, is good? How is good to be defined? Now, it may be thought that this is a verbal question. A definition does indeed often mean the expressing of one word's meaning in other words. But this is not the sort of definition I am asking for. Such a definition can never be of ultimate importance in any study except lexicogra­ phy. If I wanted that kind of definition I should have to consider in the first place bow people generally used the word 'good'; but my business is not with its proper usage, as established by custom. I should, indeed, be foolish, if I tried to use it for something which it did not usually denote: if, for instance, I were to announce that, whenever I used the word 'good,' I must be understood to be thinking of that object which is usually denoted by the word 'table.' I shall, therefore, use the word in the sense in vvhich I think it is or­ dinarily used; but at the same time I am not anxious to discuss whether I am right in thinking that it is so used. My business is solely with that object or idea, which I hold, rightly or wrongly, that the word is generally used to stand for. What I want to discover is the nature of that object or idea, and about this I am extremely anxious to arrive at an agreement. But, if we understand the question in this sense, my answer to it may seem a very disappointing one. If I am asked 'What is good?' my answer is that good is good, and that is the end of the-matter. Or if I am asked 'How is good to be defined?' my answer is that it cannot be defined, and that is all I have to say about it. But disap­ pointing as these ans\vers may appear, they are of the very last importance. To readers \vho are familiar with philosophic terminol­ ogy, I can express their importance by say-

THE INDEFINABILITY OF GOOD ing that they amount to this: That proposi­ tions about the good are· all of them syn­ thetic and never analytic; and that is plainly no trivial matter. And the same thing may be expressed more popularly, by right, then nobody can saying that, if I foist upon us such an axiom as that 'Pleas­ ure is the only good' or that 'The good is the desired' on the pretence that this is 'the ,cj verv .; meaning of the word.' 7, Let us, then, consider th�s· position. My point is that 'good' is a simple notion, Just as 'yellow' is a simple notion; that> just as you cannot, by any manner of means, ex­ plain to any one -..vho does not already knov,r it1 ,vhat yellow fa, so you can_not ex­ plain what good is. Definitions of the kind that I was asking for, definitions which de­ scribe the real nature of the object or no­ tion denoted by a word, and which do not merelv tell us what the word is used to rnean,... are only possible when t.'½ e object or notion in question is something complex. You can give a definition of a horse, be­ cause a horse has many different properties and qualities, all of which you can enumer­ ate. But when .you have enumerated them all, when vou have reduced a horse to his simplest t�rms, then you can no longer de­ flne those terms. They're simply something which you think of or perceive, and to any one who cannot think of or perceive them, you can never, by any definition) make their nature known. It may perhaps be objec'i:ed to this that we are able to describe to oth­ ers� objects which they have never seen or thought of. VVe can, for instance, make a man understand what a chimaera is; al­ though he has never heard of one or seen one. You can tell him that it is an animal with a lioness's head and body, with a goat's head growing from the middle of its back, and with a snake in place of a tail. But here the object which you are describ­ ing is a complex object; it is entirely com­ posed of parts, with which we are all per­ fectly familiar- a sna�e, a goat, a lioness; and ,ve know, too, the manner in which those parts are to be put together, because


37 we know what is meant by the middle of a lioness's back, and where her tail is wont to grow. And so it is with all objects, not pre­ viously knov,/n, which 'We are able to de­ fine: they are all complex; all composed of parts, which may themselves, iJ1 the first i n ­ stance, be capahie of similar definition, bnt whieh must in the end be redc:cihle to sim plest parts, which can no longer be defined. But yellow and good} we say, are not com­ plex: they are notions of that simple kind, out of which definitions are composed and witli which the power of further defining ceases. 8. W1ien we say, as v\lebster says, 'The definition of horse is "a hoofed quadruped of the genus Equus/' ' \Ve may, in fact, mean three different tilings. ( 1) We may mean merely: "When I say "horse," you are to understand that I am talking about a hoofed quadruped of the genus Equus.' This might be called the arbitrary verba! definition, and I do not mean that good is indefinable in that sense. (2) We may mean, as Webster ought to mean: When most English people say "horse," they mean a hoofed quadruped of the genus Equus.' This may be called the verbal definition proper, and I do not say that good is inde­ finable in this sense either; for it is cer­ tainly possible to discover how people use a word: otherwise., we could never have known that 'good' may be translated by 'gut' in German and by 'hon' in French. But ( 3) we may, when ;,ve define horse, mean something much more ixnportaut. We may mean that a certain object, which we all of us la:1ow, is composed in a cer+...ain manner: that it has four legs) a head, a heart, a liver, etc., etc., all of them arranged in definite relations to one another. It is in this sense that I deny good to be definable. I say that it is not composed of any parts, which ,ve can substitute for it in our minds when ,ve are thinking - of it \,Ve might think I;ust as clearly a.."'1.d correctly about a horse, if we thought of all its parts and tl1eir arrange­ ment instead of thinking of the whole: we could, I say, how a horse differed 0



from a donkey just as well, just as truly, in this way, as now we do, only not so easily; but there is nothing whatsoever which we could so substitute for good; and that is what 1 mean, when I say that good is inde­ finable. 9. But I am afraid I have still not re­ moved the chief difficulty which may pre­ vent acceptance of the proposition that good is indefinable. I do not mean to say that the good, that which is good, is thus indefinable; if I did think so, I should not be writing on Ethics, for my main object is to help towards discovering that definition. It is just because I think there will be less risk of error in our search for a definition of 'the good,' that I am now insisting that good is indefinable. I must try to explain the difference between these two. I sup­ pose it may be granted that 'good' is an adjective. Well 'the good,' 'that which is good,' must therefore be the substantive to which the adjective 'good' will apply: it must be the whole of tllat to which the a d ­ jective will apply, and the adjective must always truly apply to it. But if it is that to which the adjective will apply, it must be something different from that adjective itself; and the whole of that something dif­ ferent, whatever it is, will be our definition of the good. Now it may be that this some­ thing will have other adjectives, beside 'good,' that will apply to it. I t may be full of pleasure, for example; it may be intelli­ gent: and if these two adjectives are really part of its definition, then it will certainly be true, that pleasure and intelligence are good: And many people appear to think that, if we say 'Pleasure and intelligence are good/ or if ,ve say 'Only pleasure and intelligence are good,' we are defining 'good.' Well, I cannot deny that proposi­ tions of this nature may sometimes be called definitions; I do not know well enough how the 1.vord is general!y used to decice upon this point I only wish it to be understood that that is not ,vhat I mean when I say there is no possible definition of good, and that I shall not mean this if I use

the word again. I do most fully believe that some true proposition of the fonn '.Intelli­ gence is good and intelligence alone is good' can be found; if none could be found, our definition of the good would be impossible. As it is, I believe the good to be definable; and yet I stil l say tllat good itself is indefinable. 10. 'Good,' then, if we mean by it that quality which we assert to belong to a thing, when we say that the thrng is good, is incapable of any definition, in the most important sense of that word. The most i m ­ portant sense o f is essential to its whether tn1e or false, it implies both a being a proper end of action. Accordingly proposition as to the degree of goodness of neither our judgments as to what actions the action in question, as compared with we ought to perform, nor even our judg­ other things, and a number of causal propo­ ments as to the ends which they ought to sitions. For it cannot be denied that the ac­ produce, are pure judgments of intrinsic tion ,vill have consequences: and to deny value. \Vith regard to the former, an action that the consequences matter is to make a which is absolutely ohligatory may have no judgment of their intrinsic value, as com­ intrinsic value whatsoever; that it is per­ pared ,,,.;th the action itself. In asserting fectly virtuous may mean merely that jt that the action is the best thing to do, we causes the best possible effects. And with assert that it together with its c;nseq uences regard to the latter, these best possible re­ presents a greater sum of intrinsic value sults which justify our action ca.n; in any fixan any possible alternative. And this con� case, have only so rriUch of intrinsic value dition may be realised by any of the three as the laws of nature aHow us to secure; cases:-( a) If the action itself has greater and they in their tum may have no intrin­ intrinsic value than any alternative, where­ sic value whatsoever; but may merely be a as both its consequences and those of mear1s to the attainment ( in a still further the alternatives are absolutely devoid ei­ future) of something that has such value. ther of intrinsic merit or intrinsic demerit; \Vhenever, therefore, we ask VVhat ought or ( b ) if, though its -consequences are inF we to do?' or \Vhat ought we to try to get?' trinsically bad, the balance of intrinsic we are asking questions ,-:vhich involve a value is greater than would be produced correct answer to bvo others� completely



different in kind from one another. We mnst know both what degree of intrinsic value different things have, and how these different things may be obtained. But the vast majority of questions which have ac­ tually been discnssed in Ethics-all practi­ cal questions, indeed-involve this double knowledge; and they have been discussed without any clear separation of the two dis­ tinct questions involved, A great part of the vast disagreements prevalent in Ethics is to be attributed to this failure in analysis. By the use of conceptions which involve both that of intrinsic value and that of causal re­ lation, as jf they involved intrinsic value only, two different errors have been ren­ dered almost universal. Either it is assumed that nothing has intrinsic value which is not possible, or else it is assumed that what is necessary must have intrinsic value. Hence the primary and peculiar business of Ethics, the determination what things have intrinsic value and in what degrees, has received no adequate treatment at an. And on the other hand a thorough discus­ sion of means has been also largely ne­ glected, ovvjng to an obscure perception of the truth that it is perfectly irrelevant to the o:uestion of intrinsic values. But hmv­ ever this may be, and however strongly any particular reader may be convinced that some one of the mutually contradictory sys­ tems which hold the field has given a cor­ rect answer either to the question what has intrinsic value, or to the question what we ought to do, or to both, it must at least be admitted that the questions what is best in itself and what will bring about the best possible, are utterly distinct; that both be­ long to the actual subject-matter of Ethics; and that the more clearly distinct questions are distinguished; the better is our chance of answering both correctly. 18. There remains one point which must not be omitted in a complete descrip­ tion of the kind of questions which Efaics has to ansvver. The main division of those questions is, as I have said1 into two; the question what things are good in them-

selves, and the question to what other things these are related as effects. The first of these, which is the primary ethical ques­ tion and is presupposed by the other, in­ cludes a correct comparison of the various things which have intrinsic value ( if there are many such ) in respect of the degree of value which they have;. and such compari­ son involves a difficulty of principle which has greatly aided the confusion of intrinsic value with mere 'goodness as a means/ It has been pointed out that one difference between a judgment, which asserts that a thing is good in itself, and a judgment ,vhich asserts that :it is a means to good, consists in the fact that the first, if true of one instance of the thing in question, is necessarily true of all; whereas a thing which has good effects under some circum­ stances may have bad ones under others. Now it is certainly true that all juc.gments of intrinsic value are in this sense univer­ sal; but the principle which I have now to enunciate may easily make it appear as if they were not so but rese,nbled the judg­ ment of means in being n1erely general. There is 1 as will presently be maintained, a vast number of different things, each of which has intrinsic value; there are also very many which are positively bad; and there is a still larger class of Hth1gs, which appear to be indifferent. But a foing be­ longing to any of these three classes may occur as part of a whole, which includes among its other parts other things belong­ ing both to the same and to the other two classes; and these \Vholes:o as such, may also have intrinsic value. The paradox} to �vhich it Is necessary to call attention> is that the value of such a whole bears no regular pro­ portion to the sum of the values of its parts. It is certain that a good thing may

exist in such a relation to another good thing t::lat the value of the whole thus formed is immensely greater than the sum of the values of ilie two good things. It is certain that a whole formed of a good thing and an indifferent thing may have im­ reensely greater value than that good thing


itself possesses. It is certain that two bad things or a bad thing and an indifferent thing may form a whole much worse than the sum of badness of its parts. And it seems as if indifferent things may also be the sole constituents of a whole which has great value either positive or negative. vVhether the addition ,of a bad thing to a good whole may increase the: positive value of the whole, or the addition of a bad thing to a bad may produce a whole having posi­ tive value, may seem more doubtful; but it is, at least, possible, , and this possibility must be taken into account in our ethical investigations. However we may decide particular questions, the principle is clear.

The value of a whole must not be assumed to be the same as the sum of the values of its parts. A single instance "111 suffice to illus­ trate the kind of relation in question. It seems to be true that ·to be conscious of a beautiful object is a thing of great intrinsic value; whereas the same object, if no one be conscious of it, has certainly compara­ tively little value, and is commonly held to have none at all. But the consciousness of a beautiful object is certainly a whole of some sort in ,vhich we can distinguish as parts the object on the one hand and the being conscious on the other. Now this lat­ ter factor occurs as . part of a different whole, whenever we ::ire conscious of any­ thing; and it would , seem that some of these wholes have at all events very little value, and may even be indifferent or posi­ tively bad. Yet we cannot always attribute the slightness of their Value to any positive demerit in the object · which differentiates them from the consciousness of beauty; the object itself may approach as near as possi­ ble to absolute neutrality. Since, therefore, mere consciousness does not always confer great value upon the whole of which it forms a part, even though its object may have no great demerit; we cannot attribute the great superiority Of the consciousness of a beautiful thing , over the beautiful thing itself to the mere' addition of the value


of consciousness to that of the beautiful thing. VVhatever the intrinsic value of con­ sciousness may be, it does not give to the whole of which it forms a part a value pro­ portioned to the sum of its value and that of its object. If this be so, we have here an instance of a whole possessing a different intrinsic value from the sum of that of its parts; and whether it be so or not, what is meant by such a difference is illustrated by this case. 19. There are, then, wholes which pos­ sess the property that their value is differ­ ent from the sum of the values of their parts; and the relations which subsist be­ tween such parts and the whole of which they form a part have not hitherto been distinctly recognised or received a separate name. Two points are especially worthy of notice. ( 1 ) It is plain that the existence of any such part is a necessary condition for the existence of that good which is consti­ tuted by the whole. And exactly the same language will also express the relation be­ tween a means and the good thing which is its effect. But yet there is a most impor­ tant difference between the two cases, con­ stituted by the fact that the part is, \vhereas the means is not, a part of the good thing for the existence of which its existence is a necessary condition. The ne­ cessity by which, if the good in question is to exist, the means to it must exist is merely a natural or causal necessity. If the laws of nature were different, exactly the same good might exist, although what is now a necessary condition of its existence did not exist. The existence of the means has no in­ trinsic value; and its utter annihilation would leave the value of that which it is now necessary to secure entirely un­ changed. But in the case of a part of such a whole as we are now considering, it is oth­ erwise. In this case the good in question cannot conceivably exist, unless the part exist also. The necessity which connects the two is quite independent of natural law. \Vhat is asserted to have intrinsic value is the existence of the whole; and the exist-



ence of the whole includes the existence of its part. Suppose the part removed, and what remains is not what was asserted to have intrinsic value; but if we suppose a means removed, what remains is just what was asserted to have intrinsic value. And yet ( 2 ) the existence of the part may itself have no more intrinsic value than that of the means. It is this fact which constitutes the paradox of the relation which we are discussing. It had just been said that what has intrinsic value is the existence of the whole, and that this includes the existence of the part; and from this it would seem a natural inference that the existence of the part has intrinsic value. But the inference would be as false as if we were to conclude that, because the number of two stones was two, each of the stones was also two. The part of a valuable whole retains exactly the same value when it is, as when it is not, a part of that whole. If it had value under other circumstances, its value is not any greater, when it is part of a far more valu­ able whole; and if it had no value by itself, it has none still, however great be that of the whole of which it now forms a part. We are not then justified in asserting that one and the same thing is under some cir­ cumstances intrinsically good, and under others not so; as we are justified in assert­ ing of a means that it sometimes does and sometimes does not produce good results. And yet we are justified in asserting that it is far more desirable that a certain thing should exist under some circumstances than under others; namely when other things will exist in such relations to it as to form a more valuable whole. It ,vill not have more intrinsic value under these cir­ cumstances than under others; it will not necessarily even be a means to the exist­ ence of things having more intrinsic value: but it will, like a means, be a necessary condition for the existence of that which has greater intrinsic value, although, unlike a means, it will itself form a part of this more valuable existent. 20. I have said that the peculiar rela­ tion between part and whole which I have

just been trying to define is one which has received no separate name. It would, how­ ever, be useful that it should have one; and there is a name, which might well be ap­ propriated to it, if only it could be di­ vorced from its present unfortunate usage. Philo;ophers, especially those who profess to have derived great benefit from the writ­ ings of Hegel, )lave latterly made much use of the terms 'organic whole,' 'organic unity,' ..'Perience of Value," Journal of Philosophy, 1938. 0 Principia Ethica, pp, 73, 77.


turalistic fallacy the attempt to derive the Ought from the Is. 6 \Ve may begin, then, by considering this bifurcati?_n, emphasis on which, by Sidgwick, Sorley, and others, came largely as a reaction to the proce­ dures of Mill and Spencer. Hume affirms the bifurcation in his Treatise: "l cannot forbear adding to these reasonings an ob­ servation, which may, perhaps, be found of some importance. In every system of mor­ ality which I have hitherto met with, I have always remarked, that the author pro­ ceeds for some time in the ordinary ,vay of reasoning, and establishes the being of a God, or makes observations concerning human affairs; when of a sudden I am sur­ prised to find, that instead of the usual cop­ ulations of propositions, is, and is not, I meet vvith no proposition that is not con­ nected with an ought, or an ought not. This change is imperceptible; but is, hov;rever, of the last consequence. For as this ought, or ought not, expresses some new relation or affirmation, it is necessary that it should be observed and explained; and at the same time that a reason should be given, for what seems altogether inconceivable, how this new relation can be a deduction from others, which are entirely different from it. But as authors do not commonly use this precaution, I shall presume to re­ commend it to the readers; and am per­ suaded, that this small attention would subvert all the vulgar systems of morality, and let us see that the distinction of vice and virtue is not founded merely on the re­ lations of objects, nor is perceived by reason."7 Needless to say, the intuitionists have found this observation of some im­ portance.8 They agree with Hume that it subverts all the vulgar systems of morality, though, of course, they deny that it lets us see that the distinction of virtue and vice is 0

"Ethics as Pure Postulate," Philosophical R e ­

view, 1933. See also T. \Vhittaker, The Themy of

Abstract Ethics, pp. 19 f. 1 Book III, part ii, section i. B See J. Laird, A Study in Moral Theory, pp. 16 f.; \1/hittaker, op. cit., p. 19.



not founded on the relations of objects, nor is perceived by reason. In fact, they hold that a small attention to it subverts Hume's own system also, since this gives naturalis­ tic definitions of virtue and vice and of good and evil.9 Hume's point is that ethical conclu­ sions cannot be draVi'Il validly from prem­ ises which are non-ethical. But when the intuitionists affirm the bifurcation of the 'ought' and the 'is,' they mean more than that ethical propositions cannot be de­ duced from non-ethical ones. For this diffi­ culty in the vnlgar systems of morality could be remedied, as we shall see, by the introduction of definitions of ethical no­ tions in non-ethical terms. They mean, fur­ ther, that such definitions of ethical notions in non-ethical terms are impossible. «The essential point," says 11r. Laird, "is the ir­ reducibility of values to non-values."10 But they mean still more. Yellow and pleasant­ ness are, according to :tvfr. 1vfoore, indefin­ able in non-ethical terms, but they are nat­ ural qualities and belong on the 'is' side of the fence. Ethical properties, however, are not, for him, mere indefinable natural qual­ ities, descriptive or expository. They are properties of a different kind-non-descrip­ tive or non-natural. 1 1 The intuitionist bifur­ cation consists of three statements:( 1 ) Ethical propositions are not deducible from non-ethical ones, 12 ( 2) Ethical characteristics are not de£.nable in terms of non-ethical ones. ( 3 ) Ethical characteristics are different in kind from non-ethical ones.

Really it consists of but one statement, namely, ( 3 ) since ( 3 ) entails ( 2 ) and (2) entails ( 1 ) . It does not involve saying that any ethical characteristics are absolutely indefinable. That is another question, al­ though this is not always noticed.

See C. D. Broad, Five Types of Ethical Theory, ch. iv. 10 A Study in Moral Theory, p, 94 n. 11 See his Philosophical Studies, pp,' 259, 273 f. l:l See J. Laird1 op. cit., p. 318. Also pp. 12 ff. 9

What, now, has the naturalistic fallacy to do with the bifurcation of the 'ought' and the 'is'? To begin with, the connexion is this: many naturalistic and metaphysical moralists proceed as if ethical conclusions can be deduced from premises all of which are non-ethical, the classical examples being Mill and Spencer. That is, they vio­ late ( 1 ) . This procedure has lately been re­ ferred to as the "factualist fallacy" by Mr. Wheelwright and as the "valuational fal­ lacy" by Mr. Wood. 13 Mr. Moore some­ times seems to identify it with the natural­ istic fallacy, but in the main he holds only that it involves, implies, or rests upon this fallacy.14 ,Ve may now consider the charge that the procedure in question is or in­ volves a fallacy. It may be noticed at once that, even if the deduction of ethical conclusions from non-ethical premises is in no way a fallacy, Mill certainly did commit a fallacy in drav.ring an analogy between visibility and desirability in his argument for hedonism; and perhaps his committing this fallacy, which, as Mr. Broad has said, we all learn -' about at our mothers knees, is chiefly re­ sponsible for the notion of a naturalistic fallacy. But is it a fallacy to deduce ethical conclusions from non-ethical premises? Consider the Epicurean argument for he­ donism which Mill so unwisely sought to embellish: pleasure is good, since it is sought by all men. Here an ethical conclu­ sion is being derived from a non-ethical premise. And, indeed, the argument, taken strictly as it stands, is fallacious. But it is not fallacious because an ethical term oc­ curs in the conclusion which does not occur in the premise. It is fallacious because any argument of the form "A is B, therefore A is C" is invalid, if taken strictly as it stands. For example, it is invalid to argue that

P. E. Wheelwright, A Critical. Introduction to Ethics, pp, 40-51, 91 f.; L. \\food, "Cognition and Moral Value," Journal of Philosophy, 1937, p. 231. 1 4 See Principia Ethica, pp. 114, 57, 43, 49. la

\Vhittaker identifies it with the naturalistic fallacy and regards it as a "logical" fallacy, op. cit., pp. 19 £.


THE NATGRALISTIC FALLACY Croesus is ric..._h because he is wealthv. Such arguments are; however1 not m:tend�d to be taken strictly as they stand, They are enthymemes and contain a suppressed premise, And, when this suppressed prem­ ise is made explicit, they are valid and in­ vo:ve :10 log,cal fa'.'.acy,15 Thus the Epicu­ rean inference from psychological to ethical hedonism is valid when the suppressed premise is added to the effect that what is sought by all men is good, Then the only question left is whether the premises are true. It is clear: then} that the naturalistic fallacy is not a logical fallacy, since it may be involved even when the argament is valid. H01-v does the naturalistic fallacy enter such "mixed ethiqal arguments11 16 as that of the Epicureans? Whether it does or not depends on the nature of the sup­ pressed premise. This may be either an in­ duction, an intuition, a deduction from a "pure ethical argument," a definition, or a proposition whicb is true by definition. lf it is one of the first three, then the naturalis­ tic fallacy does not enter at a!L In fact, the argument does not then involve violating ( 1 ) , since one of its premises will be ethi­ cal, But if the premise to be supplied is a definition or a proposition which is true by deflnition, as it probably was for the Epi­ cureans;, then the argument, while still valid, fovolves the llilturalistic fallacy, and will run as follows:(a) Pleasure is sought by all men, ( b) What is sought by all men is good ( defi­

nition), ( c) Therefore, pleasure is good.

Now I am not greatly interested fr, de­ ciding whether the argument as here set up violates ( 1 ) , If it does not, then no 'mixed ethical argument' actually commits any fac­ tualist or valuational fallacy, except when it is t:nfairly taken as complete in its enthySee ibid., pp. 50, 139; \Vheelw:right1 loc. cit. See C. D. Broad, The Mind and Its Place fa Nature, pp, 488 f.; Lai.rd, loc. cit. 1�


mematic form, If it does, then a valid argu­ ment may involve the deduction of an ethl­ cai conclusion from non-ethical premises and the factualist or valuational fallacy is not really a fallacy, The question depends on whether or not ( b ) and ( c) are to be regarded as ethical propositions, Mr. Nloore so to regard them, contend­ ing that, by hypothesis, ( b ) is .analytic or tautologous, and that ( c) is psychological, since it really says only that pleasure is sought by all men.1 7 But to say that ( b ) is analytic and not ethical and that (a) is not ethical but psychological is to prejudge the question whether 'good' can be de::ined; for the Epicureans \VOuld contenc! preciseiy that if thefr definition is correct then ( b ) is ethical but analytic and ( e) ethical though psychological. Thus, unless the question of the definability of goodness is to be begged, ( h) and ( a ) must be re­ garded as ethical, in which case our argu­ ment does not violate ( l). 1Iowever, sup­ pose, if it be not nonsense, that { b) is non­ ethical and ( c) ethical, then the argument will violate ( 1), but it will still obey all of the canons of logic, and it is only confusing to talk of a 'valuational logic' whose basic rule is that an evaluative conclusion cannot be deCuced from non-evaluative premis­ es.18

For the only vvay in which either the in­ tuitionists or poslulationists like Mr. ·wood can cast doubt upon the conclusion of the argument of the Epicureans ( or upon the conclusion of any parallel argument) is to attack the premises, in particular ( b ) , "1ow, according to Mr, :',foore, it is due to the presence of ( b) that the argument in­ volves the naturalistic fallacy. ( b ) invo!ves the identi:Heation of goodness with 'being sought by all men,' and to make this or any other such identification is to commit the naturalistic fallacv, The naturalistic fallacy is not the procedure of violating ( l ) , It is the procedure} implied in ma::iy mixed ethi11See op, cit,j DP• 11 f.; 19, 38, 73, 139. P• See L. Wood, loc. Git.



cal arguments and explicitly carried out apart from such arguments by many moral­ ists, of q.efining such characteristics as goodness or of substituting some other characteristic for them. To quote some pas­ sages from Principia Ethica:( a ) " . . . far too many philosophers have thought that when they named those other properties [belonging to all things which are good] they were actually deflning good; that these properties, in fact, were simply not 'other,' but absolutely and en­ tirely the same with goodness. This view I propose to call the 'naturalistic fallacy.' . . ."HI

( b ) '1 have thus appropriated the name Naturalism to a particular method of approaching Ethics . . . . This method con­ sists in substituting for 'good' some one pro­ perty of a natural object or of a collection of natural objects . . . ."20 ( c ) " . . . the naturalistic fallacy [ is J the fallacy which consists in identifying the simple notion which we mean by 'good' ,vith some other notion."21 Thus, to identify 'better' and 'more evolved,' 'good' and 'desired/ etc., is to com­ mit the naturalistic fallacy.22 But just why i s such a procedure fallacious or erro­ neous? And is it a fallacy only when ap­ plied to good? We must now study Section 12 of Principia Ethica. Here Mr. Moore makes some interesting statements:". . . if anybody tried to deflne pleasure for us as being any other natural object; if anybody were to say, for instance, that pleasure means the sensation of red. . . Well, that would be the same fallacy which I have called the naturalistic fallacy. . . . I should not indeed call that a naturalistic fallacy, although it is tl1e same fallacy as I have called naturalistic vvith reference to Ethics. . . . VVhen a man confuses two natural objects with one another, defining p, 10, 20 p, 40. lO


p. 58, cf. pp. xiii, 73.

" Cf. pp. 49, 53, 108, 139.

the one by the other . . . then there is no reason to call the fallacy naturalistic. But if he confuses 'good,' which is not . . . a natural object, with any natural object whatever, then there is a reason for calling that a naturalistic fallacy. . . ."23 Here Mr. Moore should have added that, when one confuses 'good,' which is not a metaphysical object or quality, with any metaphysical object or quality, as me­ taphysical moralists do, according to him, then the fallacy should be called the meta­ physical fallacy. Instead he calls it a natur­ alistic fallacy in this case too, though he re­ cognises that the case is different since me­ taphysical properties are non-natural24-a procedure \Vhich has misled many readers of Principia Ethica. For example, it has led Mr. Broad to speak of "theological naturalism."25 To resume: "Even if [goodness] were a natural object, that would not alter the nature of the fallacy nor diminish its im­ portance one \Vhit."26 From these passages it is clear that the fallaciousness of the procedure which Mr. Moore calls the naturalistic fallacy is not due to the fact that it is applied to good or to an ethical or non-natural characteristic. When Mr. R. B. Perry deflnes 'good' as 'being an object of interest' the trouble is not merely that he is defining good. Nor is the trouble that he is deflning an ethical characteristic in terms of non-ethical ones. Nor is the trouble that he is regarding a non-natural characteristic as a natural one. The trouble is more generic than that. For clarity's sake I shall speak of the definist fallacy as the generic fallacy which under­ lies the naturalistic fallacy. The naturalistic fallacy will then, by the above passages, be a species or form of tl1e deflnist fallacy, as would the metaphysical fallacy if Mr. Nioore had given that a separate nam.e .27 :m p, 13. u See pp. 38-40, 110....112. u Five Types of Ethical Theory, p, 259. 26 p. 14. 27 As '(Vhittaker has, Zoe. cit,

THE NATURALISTIC FALLACY That is, the naturalistic fallacy, as illus­ trated by Mr. Perry's procedure, is a fal­ lacy, not because it is naturalistic or con­ fuses a non-natural quality with a natural one, but solely becase it involves the defin­ ist fallacy. VVe may, then, confine our at­ tention entire1y to an understanding and evalution of the de!inist fallacy. To judge by the passages l have just quoted, the definist fallacy is the process of confusing or identifying two properties, of defining one property · by another, or of substituting one property for another. Fur­ ther...nore, the fallacy is always simply that two properties are being treated as one, and it is irrelevant, if it be the case, that one of them is natural or nonethical and the other non-natural or ethical. One may commit the definist fallacy without in­ fringing o n the bifurcation· of the ethical and non-ethicai as when one identifies plea.santness and redness or rightness and goodness. But even when one infringes on that bifurcation in committing the definist fallacy, as when one identifies goodness and pleasantness or goodness and satisfac­ tion, then the mistake is still not that the bifurcation is being infringed on, but only that two properties are being treated as one. Hence, on the present interpretation� the de6nist fallactJ does not, in any of its fonns, consist in violating ( 3), and has no essential connexion with the bifurcation of the 'ought' and the 'is.' This formulation of the definist fallacy e,c-p!ains or reflects the motto of Principia Eth/ca, borrowed from Bishop Butler : "Everything is what i t is, and not another thing." It follows from this motto that goodness is vvha t it is and not another thing. It follows that views which try to identify it with something else are making a mistake of an elementary sort. For it is a mistake to confuse or identify two proper­ ties. If the properties really are two, then they simply are not identical. But do those who define ethical notions in non-ethical terms make this mistake? They will reply


to Mr. Moore that they are not identifying two properties; what they are saying is that two words or sets of words stand for or mean one and the same property. Mr. Moore was being, in part, misled hy the material mode of speech, as Mr. Carnap calls it, in such sentences as ''Goodness is pleasantness," "Knowledge is true belief," etc. 'vVhen one says instead, '"The word 'good' and the word 'pleasant' mean the same thing," etc., it is clear that one is not identifying two things. But Mr. Moore kept himself from seeing this by his disclaimer that he -..vas in'!:erested in any statement about t:Ie l.!Se of words. 2s The definist fallacy, then, as we have stated it, does not rnle out any naturalistic or metaphysical definitions of ethical tenns. Goodness is not identifiable with any 'other' characteristic ( if lt is a characteristic at all ) . the question is : which character­ istics are other than goodness, which names stand for characteristics other than good­ ness? And it is begging the question of the definability of goodness to say out of hand that Mr. Peny, for instance, is identifying goodness with something else, 'The point is that goodness is what it is, even if it is de­ finable. That is why Mr. Perry can take as the motto of his naturalistic Moral Econ­ omy arrother sentence from Bishop Butler: "Things and actions are what they are, and the consequences of them ,vill be what they will be; why then should we desire to be deceived?" The motto of Principia Ethica is a tautology, and should be expanded as fol­ lows: Everything is "�vhat it is, and not an­ other thing, unless it is another thing, and even then it is what it is. On the other hand, if Mr. i\foore's motto ( or the deflnist fallacy ) rules out any definitions, for example of 'good,' then it rules out all definitions of any term what­ ever. To be effective at all, it must be un­ derstood to mean, ''Every term means what it means, and not what is meant by any zs See 011. cit,, pp, 6; 8 12, ,.



other term." Mr. Moore seems implicitly to understand his motto in this way in Section 13, for he proceeds as if 'good' has no meaning, if it has no unique meaning. If the motto be taken this way, it will follow that 'good' is an indefinable term, since no synonyms can be found. But it will also fol­ low that no term is definable. And then the method of analysis is as useless as an Eng­ lish butcher in a world without sheep. Perhaps we have misinterpreted the definist fallacy. And, indeed, some of the passages which I quoted earlier in this paper seem to imply that the definist fal­ lacy is just the error of defining an indefin­ able characteristic. On this interpretation, again, the definist fallacy has, in all of its forms, no essential connexion with the bi­ furcation of the ethical and the non-ethical. Again, one may commit the definist fallacy without violating that bifurcation, as when one defines pleasantness in terms of redness or goodness in terms of rightness ( granted Mr. Moore's belief that pleasantness and goodness are indefinable ) . But even when one infringes on that bifurcation and de­ fines goodness in terms of desire, the mis­ take is not that one is infringing on the bi­ furcation by violating ( 3 ) , but only that one is defining an indefinable characteris­ tic. This is possible because the proposition that goodness is indefinable is logically in­ dependent of the proposition that goodness is non-natural: as is shown by the fact that a characteristic may be indefinable and yet natural, as yellovmess is; or non-natural and yet definable, as rightness is ( granted i'vfr. Moore�s vie,vs about yellowness and rightness ) . Consider the definist fallacy as w e have just stated it. It is, of course, an error to de­ fine an indefinable quality. But the ques­ tion, again, is: which qualities are indefin­ able? It is begging the question in favour of intuitionism to say in advance that the quality goodness is indefinable and that, therefore, all naturalists commit the de:S.nist fallacy. One must know that goodness is in­ definable before one can argue that the de-

finist fallacy is a fallacy. Then, however, the definist fallacy can enter only at the end of the controversy between intuition­ ism and de:Bnism, and cannot be used as a weapon in the controversy. The definist fallacy may be stated in such a way as to involve the bifurcation be­ tween the 'ought' and the 'is.'" It would then be committed by anyone who offered a definition of any ethical characteristic in terms of non-ethical ones. The trouble "1th such a definition, on this interpretation, would be that an ethical characteristic is being reduced to a non-ethical one, a non-natural one to a natural one. That is, the definition would be ruled out by the fact that the characteristic being defined is ethical or non-natural and therefore can­ not be defined in non-ethical or natural terms. But on this interpretation, too, there is danger of a petitio in the intuitionist ar­ gumentation. To assume that the ethical characteristic is exclusively ethical is to beg precisely the question which is at issue when the definition is offered. Thus, again, one must know that the characteristic is non-natural and indeRnable jn natural terms before one can say that the definists are making a mistake. Mr. Moore, McTaggart, and others for­ mulate the naturalistic fallacy sometimes in a ,vay somewhat different from any of those yet discussed. They say that the de­ £nists are confusing a universal synthetic proposition about the good with a defini­ tion of goodness.30 Mr. Abraham calls this the "fallacy of misconstrued proposition."31 Here again the difficulty is that, while it is true that it is an error to construe a univer­ sal synthetic proposition as a definition, it is a petitio for the intuitionists to say that what the de:6.nist is taking for a definition is really a universal synthetic proposition. 32

See J. 'Wisdom, Mind, 1931, p. 213, note 1. See Principia Ethica, pp. 10, 16, 38; The Nature of Existence, vol. ii, p. 398. 31 Leo Abraham, "The Logic of Intuitionism," 20


International Journal of Ethics, 1933. $� As Mr. Abraham points out, lac. cit,


A t last, ho\vever, the issue between the logical confusion. It is not even, properly intuitionists and the defhtlsts ( naturalistic speaking, an error. It is rather a kind of or metaphysical) is becoming clearer, The blindness, analogous to colour-blindness. definists are all holding that certain propo­ Even this moral blindness can be ascribed sitions involving ethical tenns are analytic, to the definlsts only if they are correct in tautologo;:s, or true by de!:nition, e.g., :Mr. their claim to have no awareness of any Perry so rege.rds the statement, "All ob;ects unique ethical characteristics and if the of desire are good." The intuitionists hold intuitionists are correct in affirming the that such statements are synthetic. \�1hat existence of such characteristics, but cer­ underlies this difference of. opinion i.s that tainly to call it a 'fallacy,' even in a loose the intuitionists claim to have at ·least a sense, is both unamiable and profitless. On ilie other hand, of course, if there dim awareness of a simple unique quality or relation of goodness , or rightness which are not such characteristics in the objects appears i,i the region ' which our ethical to which we attach ethical predicates, then tenns .:ougb..:y ind!cate, whereas the de-B.n­ the intuitionists, if \Ve may take them at ists claim to have no awareness of any such their word, are suffering from a corre­ quality or relation in that region, whic.½ is sponding moral hallucination, Definists different from all other qualities and rela­ might ilien call iliis the intuitionlstic or tions which belong to the same context but moralistic fallacy, except that it is no more are designated by words other than 'good' a 'fallacy' than is the blindness just and 'right' and their obvious synonyms,33 described. Anyway, they do not believe the The definists are in all honesty claiming to claim of the intuitionists to he aware of fh1C but one characteristic where the fotui­ unique ethica! characteristics, a:1d conse­ tionists claim to find two, as Mr. Perry quently do not attribute to them this hallu­ claims to find only the property of being cination, Instead, they simply deny that the desired where Mr. Moore claims to find 'intuitionists really do find such unique both it and the property of being good, The qualities or relations, and then they try to issue1 the::::i, is one of inspection or intcition1 find some plausible way of accounting for and concerns t..½ e awareness or discernn1ent the fact that very respectable and trustwor­ of qualities and relations. 34 That is why it thy people think they find them.35 Thus cannot be decided by t.lle use of the notion they charge the intuition.ists with verbaI­ ism, hypostatisation, and the like. But this of .a fallacy. If the deflnists may be taken at their half of the story does not concern us now. v\lhat concems us more is tl1e fact that word, ilien they are not actually confusing two characteristics with each other, nor the intuitionists do not credit the claim of defining an indefinable characteristic, nor the definists either. They would be much confusing definitions and universal syn­ disturbed, if they really thought that their thetic propositions-in ,short they are not opponents were morally blind, for they do co:nmittfrig the naturalistic or definist fa!­ not hold that we must be regenerated by lacy in any of the interpretations given grace before we can have moral insight, above. Then the only fallacy which they and they share the common feeling that commit- th e real naturalistic or definist fal­ morality is something democratic even lacy- is the failure to descry the qualities though not all men are good. Thus they and relations which are central to morality. hold that "we are all aware" of certain But th:s ls neither a logical fallacy nor a unique characteristics when we use the ter:::ns �good/ �right/ etc.i only due to a lack n. See R. B. Perry, General Theory of Value: p, 30i cf. Journal. of Philosophy, 1931, p, 520. See H. Osbome, Foundation of the Philoso� Cf. R, B. Perry, Journal of Philosophy, 1931, phy of Value, pp. J.5, 19, 70. pp. 52.0 ff. ' :;a.




of analytic clearness of mind, abetted per­ haps by a philosophical prejudice, we may not be aware at all that they are different from other characteristics of which \Ve are also aware.36 Now, I have been arguing that the intuitionists cannot charge the definists with committing any fallacy unless and until they have sho'-'Til that \Ve are all, the definists included, aware of the dis­ puted unique characteristics. If, however, they ,vere to show this, then, at least at the end of the controversy, they could accuse the definists of the error of confusing two characteristics, or of the error of defining an indefinable one, and these errors might, since the term is somewhat loose in its habits, be called 'fallacies,' though they are not logical fallacies in the sense in which an invalid argument is. The fallacy of mis­ construed proposition depends on the error of confusing two characteristics, and hence could also on our present supposition, be ascribed to the definists, but it is not really a logical confusion,3 7 since it does not actually involve being confused about the difference betv.reen a proposition and a definition. Only it is difficult to see how the intui­ tionists can prove that the definists are at least vaguely aware of the requisite unique characteristics. 38 The question must surely be left to the inspection or intuition of the definists themselves, aided by whatever suggestions the intuitionists may have to make. If so, we must credit the verdict of their inspection, especially of those among them who have read the writings of the Principia Ethica, pp. 17, 38, 59, 61. But see H. Osborne, op cit., pp, 18 f. For a brief discussion of their arguments, see ibid., p. 67; L. Abraham, op. cit. I think they are all inconclusive, but cannot show this here. 36

31 38

intuitionists reflectively, and, then, as we have seen, the most they can be charged with is moral blindness. Besides trying to discover just what is meant by the naturalistic fallacy, I have tried to show that the notion that a logical or quasi-logical fallacy is committed by the de£.nists only confuses the issue betvveen the intuitionists and the definists ( and the issue bet,veen the latter and the emotists or postulationists ) , and misrep­ resents the way in ,vhich the issue is to be settled. No logical fallacy need appear anywhere in the procedure of the de£.nists. Even fallacies in any less accurate sense cannot be implemented to decide the case against the definists; at best they can be ascribed to the definists only after the issue has been decided against them on independent grounds. But the only defect which can be attributed to the definists, if the intuitionists are right in affirming the existence of unique indefinable ethical characteristics, is a peculiar moral blind­ ness, which is not a fallacy even in the looser sense. The issue in question must be decided by whatever method we may find satisfactory for determining whether or not a word stands for a characteristic at all, and, if it does, whether or not it stands for a unique characteristic. What method is to be employed is, perhaps, in one form or another, the basic problem of contempo­ rary philosophy, but no generally satisfac­ tory solution of the problem has yet been reached. I shall venture to say only this: it does seem to me that the issue is not to be decided against the intuitionists by the application ab extra to ethical judgments of any empirical or ontological meaning dictum.39 39

See Principia Ethica, pp, 124 f., 140.

How to Derive "Ought'' from ' 'ls ''


JOHN R. SEARLE I It is often said that one cannot derive an "'ought'� from an "is." This thesis, which comes from a famous passage in Hume's Treatise, ,vhile not as clear as it might be, is at least clear in broad outline: there is a class of statements of fact which is logically distinct from a class of statements of value. No set of statements of ' fact by themselves entails any statement of value. Put in more contemporary terminolo'gy, no set of de­ scriptive statements can entail an evaluative statement without the addition of at least one evaluative premise. To believe other­ wise is to commit what has been called the naturalistic fallacy. I shall attempt to demonstrate a coun­ terexample to this thesis.2 It is not of course to be supposed that a single counter­ example can refute a philosophical thesis, but in the present instance if we can pre­ sent a plausible counterexample and can in addition give some accO;unt or explanation of hO\v and why it is . a counterexample, Reprinted from The Philosophical Review, 73, January 1964, by permission of the author and The Philosophical Review. 1

Earlier versions of this paper were read before the Stanford Philosophy Colloquium and the Pacific Division of the American Philosophical Association. I am indebted to many people for helpful com­ ments and criticisms, especially Hans Herzberger, Arnold Kaufmann, Benson Mates, A. I. Melden, and Dagmar Searle. 2 In its modem version. I shall not be concerned writh Hume's treatment of the problem.

and if we can further offer a theory to back up our counterexample-a theory which will generate an indefinite number of coun­ terexamples- we may at the very least cast considerable light on the original thesis; and possibly, if we can do all these things, we may even incline ourselves to the view that the scope of that thesis was more re­ stricted than we had originally supposed. A counterexample must proceed by taking a statement or statements which any propo­ nent of the thesis would grant were purely factual or "descriptive" ( they need not actually contain the word '"is" ) and show how they are logically related to a state­ ment which a proponent of the thesis would regard as clearly "evaluative." ( In the present instance it will contain an "ought". ) 3 Consider the following series of state­ ments: ( 1 ) Jones uttered the words "I hereby promise to pay you1 Smith.1 £ve dollars." ( 2 ) Jones promised to pay Smith five dollars. ( 3 ) Jones placed himself under 3 If this enterprise succeeds, we shall have bridged the gap between "evaluative" and "de­ scriptive" and consequently have demonstrated a weakness in this very terminology, At present, however, my strategy is to play along with the ter­ minology1 pretending the notions of evaluative and descriptive are fairly clear. At the end of the paper I shall state in '\Vhat respects I think they embody a muddle.




( undertook) an obligation to pay Smith five dollars. ( 4 ) Jones is under an obligation to pay Smith five dollars. ( 5 ) Jones ought to pay Smith five dollars. I shall argue concerning this list that the relation behveen any statement and its suc­ cessor, while not in every case one of "entailment," is nonetheless not just a con­ tingent relation; and the additional state­ ments necessary to make the relationship one of entailment do not need to involve any evaluative statements, moral principles, or anything of the sort. Let us begin. How is ( 1 ) related to ( 2 )? In certain circumstances, uttering the words in quotation marks in ( 1 ) is the act of making a promise. And it is a part of or a consequence of the meaning of the words in ( 1 ) that in those circumstances uttering them is promising. "I hereby promise" is a paradigm device in English for performing the act described in ( 2 ) , promising. Let us state this fact about English usage in the form of an extra premise:

( la ) Under certain conditions C anyone ,vho utters the words ( sentence ) «I hereby promise to pay you, Smith, five dol­ lars" promises to pay Smith five dollars.

vVhat sorts of things are involved under the rubric "conditions C"? What is involved will be all those conditions, those states of affairs, which are necessary and sufficient conditions for the utterance of the words ( sentence ) to constitute the success­ ful performance of the act of promising. The conditions will include such things as that the speaker is in the presence of the hearer Smith, they are both conscious, both speakers of English, speaking seriously. The speaker knows what he is doing, is not under the influence of drugs, not hypno­ tized or acting in a play, not telling a joke or reporting an event, and so forth. This list will no doubt be somewhat indefinite because the boundaries of the concept of a promise, like the boundaries of most con-

cepts in a natural language, are a bit loose.4 But one thing is clear; however loose the boundaries may be, and however difficult it may be to decide marginal cases, the conditions under which a man who utters "I hereby promise" can correctly be said to have made a promise are straight­ forwardly empirical conditions. So let us add as an extra premise the empirical assumption that these conditions obtain. ( lb ) Conditions C obtain. From ( 1 ) , ( l a ) , and ( l b ) we derive ( 2 ) . The argument is of the form: If C then ( if U then P ) : C for conditions, U for utterance, P for promise. Adding the prem­ ises U and C to this hypothetical we derive ( 2 ) . And as · far as I can see, no moral premises are lurking in the logical wood­ pile. More needs to be said about the rela­ tion of ( 1 ) to ( 2 ) , but I reserve that for later. What is the relation between ( 2) and ( 3 ) ? I take it that promising is, by definition, an act of placing oneself under an obligation. No analysis of the concept of promising will be complete which does not include the feature of the promiser placing himself under or undertaking or accepting or recognizing an obligation to the promi­ see, to perform some future course of action, normally for the benefit of the promi­ see. One may be tempted to think that promising can be analyzed in terms of creating expectations in one's hearers, or some such, but a little reflection "\\rill show that the crucial distinction between state­ ments of intention on the one hand and promises on the other lies in the nature and degree of commitment or obligation under­ taken in promising. I am therefore inclined to say that ( 2 ) entails ( 3 ) straight off, but I can have no -t- In addition the concept of a promise is a member of a class of concepts which suffer from looseness of a peculiar kind, viz. defeasibility. Cf H, L. A, Hart, "The Ascription of Responsibility and Rights," Logic and Language, First Series, ed. by A. Flew ( Oxford, 1951 ) .


objection if anyone wishes to add-for the purpose of formal neatness- the tautologi­ cal premise:

( 2a) All promises are acts of placing oneself under ( undertaking) an obligation to do the thing promised.

How is ( 3 ) related to ( 4 ) ? If one has placed oneself under an obligation, then, other things being equal, one is uuder an obligation. That I take it also is a tauto­ logy. Of course it is possible for all sorts of things to happen which will release one from obligations one has undertaken and hence the need for the ceteris paribus rider. To get an entailment between ( 3 ) and ( 4 ) w e therefore need a qualifying statement to the effect that: ( 3a) Other things are equal.

Formalists, as in the move from ( 2 ) to ( 3 ) , may wish to add the tautological

premise :

( 3b) All those who place themselves under an obligation are, other things being equal, under an obligation. The move from ( 3 ) to ( 4 ) is thus of the same form as the move from ( 1 ) to ( 2 ) : If E then (if PVO then VO ) : E for other things are equal, PVO for place under obligation and VO for under obliga­ tion. Adding the two premises E and PVO we derive UO. Is ( 3a ) , the ceteris paribus clause, a concealed evaluative premise? It certainly looks as if it might be, especially in the formulation I have given it, but I think we can show that, though questions about whether other things are equal frequently involve evaluative considerations, it is not logically necessary that they should in every case. I shall postpone discussion of this until after the next step. ,Vhat is the relation between ( 4 ) and ( 5 ) ? Analogous to the tautology which explicates the relation of ( 3 ) and ( 4 ) there is here the tautology that, other things being equal, one ought to do what one is under an obligation to do. And here, just as


in the previous case, we need some premise

of the form:

( 4 a ) Other things are equal.

We need the ceteris paribus clause to eli­ minate the possibility that something extra­ neous to the relation of "obligation'' to "ought" might interfere.' Here, as in the previous






appearance of enthymeme by pointing out that the apparently suppressed premise is tautological and hence, though formally neat, it is redundant. If, however, we \Vish to state it formally, this argument is of the same form as the move from ( 3) to ( 4 ) : If E then ( if VO then O j ; E for other things are equal, VO for under obligation, 0 for ought. Adding the premises E and VO we derive 0. Now a word about the phrase "other things being equal" and how it functions in my attempted derivation. This topic and the closely related topic of defeasibility are extremely difficult and I shall not try to do more than justify my claim that the satis­ faction of the condition does not necessar­ ily involve anything evaluative. The force of the expression "other things being equal" in the present instance is roughly this. Unless we have some reason ( that is, unless we are actually prepared to give some reason ) for supposing the obligation is void ( step 4 ) or the agent ought not to keep the promise ( step 5 ) , then the obliga­ tion holds and he ought to keep the prom­ ise. It is not part of the force of the phrase "other things being equal" that in order to satisfy it we need to establish a universal negative proposition to the effect that no reason could ever be given by anyone for � The ceteris paribus clause in this step excludes some'\vhat different sorts of cases from those excluded in the previous step, In geineral we say, "He undertook .an obligation, but nevertheless he is not (now) under an obligation" when the obliga­ tion has been removed, e.g., if the promisee says, "I release you from your obligation," But we say, "He is under an obligation, but nonetheless ought not to fulfill it" in cases where the obligation is overriden by some other considerations, e.g., a prior obligation.



supposing the agent is not under an obliga­ tion or ought not to keep the promise. That would be impossible and would render the phrase useless. It is sufficient to satisfy the condition that no reason to the contrary can in fact be given. If a reason is given for supposing the obligation is void or that the promiser ought not to keep a promise, then charac­ teristically a situation calling for an evalua­ tion arises. Suppose, for example, ,ve con­ sider a promised act wrong, but grant that the promiser did undertake an obliga­ tion. Ought he to keep the promise? There is no established procedure for objectively deciding such cases in advance, and an evaluation ( if that is really the right word ) is in order. But unless we have some reason to the contrary, the ceteris paribus condi­ tion is satisfied, no evaluation is necessary, and the question whether he ought to do it is settled by saying ''he promised." It is always an open possibility that we may have to make an evaluation in order to derive ''he ought" from "he promised," for we may have to evaluate a counterargu­ ment. But an evaluation is not logically necessary in every case, for there may as a matter of fact be no counterarguments. I am therefore inclined to think that there is nothing necessarily evaluative about the ceteris paribus conditioni even though deciding whether it is satisfied will fre­ quently involve evaluations. But suppose I am wrong about this: would that salvage the belief in an unbridgeable logical gulf between "is" and "ought"? I think not, for we can always rewrite my steps ( 4 ) and ( 5 ) so that they include the ceteris paribtts clause as part of the conclusion. Thus from our premises we would then have derived "Other things being equal Jones ought to pay Smith five dollars," and that would still be sufficient to refute the tradition, for we would still have a relation of entailment between descriptive and evaluative statements. It was not the fact that extenuating circum­ stances can void obligations that drove phi-

losophers to the naturalistic fallacy; it was ratlrnr a theory of language, as we shall see later on. \Ve have thus derived ( in as strict a sense of "derive" as natural languages will admit of) an "ought" from an "is." And the extra premises which were needed to make the derivation '\Vork were in no cause moral or evaluative in nature. They consisted of empirical assumptions, tautologies, and descriptions of word usage. It must be pointed out also that the "ought" is a "cate­ gorical" not a ''hyp othetical" ought. ( 5 ) does not say that Jones ought to pay up if he wants such and such. It says he ought to pay up, period. Note also that the steps of the derivation are carried on in the third person. We are not concluding "I ought" from "I said 'I promise,' " but ''he ought" from "he said 'I promise.' " The proof unfolds the connection bef:vireen the utterance of certain words and the speech act of promising and then in turn unfolds promising into obligation and moves from obligation to "ought." The step from ( 1 ) to ( 2 ) is radically different from the others and requires special comment. In ( 1 ) we construe "I hereby promise . . ." as an English phrase having a certain meaning. It is a consequence of that mean­ ing that the utterance of that phrase under certain conditions is the act of promising. Thus by presenting the quoted e,.-pressions in ( 1 ) and by describing their use in ( la) we have as it were already invoked the institution of promising. vVe might have started ,vith an even more groundfloor premise than ( 1 ) by saying: ( lb ) Jones uttered the phonetic se­ quence: /ai+hirbai+pramis+tapei+yu+smi0 +faiv+dalarz/

\i\Te would then have needed extra empirical premises stating that this pho­ netic sequence was associated in certain ways with certain meaningful units relative to certain dialects. The moves from ( 2 ) to ( 5 ) are rela­ tively easy. YVe rely on definitional connec-

HOW TO DERIVE "OUGHT'' FROM "IS" tions between "promise," "obligate," and "ought," and the only problem which arises is that obligations can be overridden br removed in a variety of ways and we need to take account of that fact. vVe solve our difficulty by adding further premises to the effect that there are no I contrary considera­ tions, that other things are equal.



tain words and that these words have the meaning they do surely objective facts. And if the statement of these two objective facts plus a description of the conditions of the utterance is sufficient to entail the statement ( 2 ) which the objector alleges to be an evaluative statement (Jones prom­ ised to pay Smith five dollars), then an evaluative conclusion is derived from des­ criptive premises without even going through steps ( 3 ) , ( 4 ) , and ( 5 ) .

In this section I intend to discuss three pos­ sible objections to the derivation.

Second Obiection

First Obiection Since the £rst premise is descriptive and the conclusion evaluative, there must be a concealed evaluative premise in the de­ scription of the conditions in ( lb ) . S o far, this argument merely begs the question by assuming the logical gulf between descriptive and evaluative v.rhich the derivation is designed to challenge. To make the objection stick, the defender of the distinction would have to show how exactly ( lb ) must contain an evaluative premise and what sort of premise it might be. Uttering certain words in certain condi­ tions just is promising and the description of these conditions needs no evaluative ele­ ment. The essential thing is that in the transition from ( 1 ) to ( 2 ) we move from the speci£cation of a certain utterance of words to the speci£cation of a certain speech act. The move is achieved because the speech act is a conventional act; and the utterance of the words, according to the conventions, constitutes the pedorm­ ance of just that speech act. A variant of this first objection is to say: all you have shmvn is that «promise" is an evaluative, not a descriptive, concept. But this objection again begs the question and in the end will prove disastrous to the original distinction between descriptive and evaluative. For that a man uttered cer-

Ultimately the derivation rests on the prin­ ciple that one ought to keep one's promises and that is a moral principle, hence evalua­ tive. I don't lmow whether •:one ought to > keep one s promises" is a «moral" principle, but whether or not it is, it is also tautologi­ cal; for it is nothing more than a derivation from the two tautologies: All promises are ( create, are undertak­ ings of, acceptances of) obligations, and One ought to keep ( fulfil l) one's obli­ gations. 'What needs to be e,._-plained is why so many philosophers have failed to see the tautological character of this principle. Three things I think have concealed its character from them. The first is a failure to distinguish external questions about the institution of promising from internal questions asked ,vithin the framework of the institution. The questions "Vi'hy do we have such an institution as promising?" and "Ought we to have such institutionalized forms of obli­ gation as promising?" are external ques­ tions asked about and not ,vithin the insti­ tution of promising. And the question "Ought one to keep one's promises?" can be confused with or can be taken as ( and I think has often been taken as ) an external question roughly expressible as "Ought one to accept the institution of promising?"



But taken literally, as an internal question, as a question about promises and not about the institution of promising, the question "Ought one to keep one's promises?" is as empty as the question "Are triangles three­ sided?" To recognize something as a prom­ ise is to grant that, other things being equal, it ought to be kept. A second fact which has clouded the issue is this. There are many situations, both real and imaginable, where one ought not to keep a promise, where the obligation to keep a promise is overridden by some further considerations, and it was for this reason that we needed those clumsy ceteris paribus clauses in our derivation. But the fact that obligations can be overridden does not show that there were no obliga­ tions in the first place. On the contrary. And these original obligations are all that is needed to make the proof work. Yet a third factor is the following. Many philosophers still fail to realize the full force of saying that "I hereby promise" is a perfonnative expression. In uttering it one performs but does not describe the act of promising. Once promising is seen as a speech act of a. kind different from describ­ ing, then it is easier to see that one of the features of the act is the undertaking of an obligation. But if one thinks the utterance of "I promise" or "I hereby promise" is a peculiar kind of description-for example, of one's mental state-then the relation between promising and obligation is going to seem very mysterious.

Third Objection

The derivation uses only a factual or inverted-commas sense of the evaluative terms employed. For example, an anthro­ pologist observing the behavior and atti­ tudes of the Anglo-Saxons might well go through these derivations, but nothing evaluative would be included. Thus step ( 2 ) is equivalent to "He did what they call promising" and step ( 5 ) to "According to

them he ought to pay Smith five dollars." But since all of the steps ( 2 ) to ( 5 ) are in oratio obliqua and hence disguised state­ ments of fact, the fac t -value distinction remains unaffected. This objection fails to damage the derivation, for what it says is only that the steps can be reconstrued as in oratio obli­ qua, that we can construe them as a series of external statements, that we can con­ struct a parallel ( or at any rate related) proof about reported speech. But what I am arguing is that, taken quite literally, without any oratio obliqua additions or interpretations, the derivation is valid. That one can construct a similar argument which would fail to refute the fact-value distinction does not show that this proof fails to refute it. Indeed it is irrelevant.

III So far I have presented a counterexample to the thesis that one cannot derive an "ought" from an "is" and considered three possible objections to it. Even supposing what I have said so far is trne, still one feels a certain uneasiness. One feels there must be some trick involved somewhere. VVe might state our uneasiness thus: How can my granting a mere fact about a man, such as the fact that he uttered certain words or that he made a promise, commit me to the view that he ought to do some­ thing? I now want briefly to discuss what broader philosophic significance my at­ tempted derivation may have, in such a way as to give us the outlines of an answer to this question. I shall begin by discussing the grounds for supposing that it cannot be ans,vered at all. The inclination to accept a rigid dis­ tinction between "is" and "ought," between descriptive and evaluative, rests on a cer­ tain picture of the way words relate to the world. It is a very attractive picture, so



attractive ( to me at least) that it is not mend, to advise, and so forth. Once ,ve see entirely clear to what extent the mere pres­ the different jobs the two perform, we see entation of counterexamples can challenge that there must be a logical gulf between it. \1/hat is needed is an explanation of how them. Evaluative statements must be and why this classical empiriCist picture different from descriptive statements in fails to deal with such counterexamples. order to do their job, for if they were 'Briefly, the picture is constructed some­ objective they could no longer function to thing like this: first we present examples of evaluate. Put metaphysically, values cannot so-called descriptive statements ( "my car lie in the world, for if they did they would goes eighty miles an hour," "Jones is six cease to be values and would just be feet tall," "Smith has brown hair" ) , and we another part of the world. Put in the for­ contrast them with so-called evaluative mal mode, one cannot define an evaluative statements ( "my car is ·a good car," "Jones word in terms of descriptive words, for if ought to pay Smith five dollars," "Smith is one did, one would no longer be able to a nasty man" ) . Anyone can see that they use the evaluative word to commend, but are different. '-''e articulate the difference only to describe. Put yet another way, any by pointing out that for the descriptive effort to derive an «ought" from an «is" statements the question of truth or falsity is must be a waste of time, for all it could objectively decidable, because to know the show even if it succeeded would be that meaning of the descriptive expressions is to the "is" ,vas not a real "is" · but only a dis­ lmow under \Vhat objectively ascertainable guised "ought" or, alternatively, that the conditions the statements \vhich contain "ought" was not a real "ought" but only a them are true or false. But in the case of disguised "is." evaluative statements the situation is quite This summary of the traditional empir­ different. To know the meaning of the eval­ ical view has been very brief, but I hope it uative expressions is not by itself sufficient conveys something of the power of this pic­ for l"Ilowing under what conditions the ture. In the hands of certain modern statements containing them are true or authors, especiaIIy Hare and NoweII-Smith, false, because the meaning of the expres­ the picture attains considerable subtlety sions is such that the statements are not and sophistication. capable of objective or factual truth or fal­ VVhat is wrong ,vith this picture? No sity at all. Any justification a speaker can doubt many things are vvrong ,vith it. In give of one of his evaluative statements the end I am going to say that one of the essentially involves some appeal to atti­ things wrong with it is that it fails to give tudes he holds, to criteria of assessment he us any coherent account of such notions as has adopted, or to moral principles by commitment, responsibility, and obligation. which he has chosen to live and judge In order to work toward this conclu­ other people. Descriptive statements are sion I can begin by saying that the picture thus objective, evaluative statements sub­ fails to account for the different types of jective, and the differerice is a consequence "'descriptive" statements. Its paradigms of of the different sorts of terms employed. descriptive statements are such utterances The underlying reason for these as «my car goes eighty miles an hour," differences is that evaluative statements "Jones is six feet taII," "Smith has brown perform a completely . different job from hair," and the like. But it is forced by its descriptive statements. . Their job is not to own rigidity to construe "Jones got mar­ describe any features of the world but to ried," hstitute either of these pkases for _dous difference of view as to the application 'morally good' in 'he is a morally good man' of tlie term �right.' Suppose, for instance, would obviously be not merely unidio­ that a man pays a particular debt simply matic, but absurd. It should be obvious, from fear of the legal consequences of not then, that 'right' and 'morally good' mean doing so, some people would say he had different things. But some one might say done what was right, and others would that while 'morally good' has a wider appli­ deny this: they would say t.1:ia t no moral cati.on than 'right,' in that it can be applied value attaches to such an act, and that to agents as ,veil as to acts: yet \Vhen since 'right' is meant to imply n1oral value, applied to acts they mean the same t.½ing. I the act cannot be right. They might gen­ should like therefore to convince him that eralize and say that no act is right unless 'right act' cannot mean the same as 'act it is done from a sense of duty, or if they that ought to be done' and also the sa.."1le as shrank from so rigorous a doctrine, they 'morally good act.' If I can convince him of might at least say that no act is right unless this, J think he v;,fl! see the propriety of not done from some good motive, such _as using 'right act' in the sense of 'morally good act.' either sense of dutv or benevolence. This differenc� of view may be due to But we ought Jirst to note a minor either of two causes. Both parties may be using 'righf in the same sense, the sense of cli:fference between the meaning of 'right' 'morally obligatory,' and differing as to the and the meaning of 'something that ought further character an act mnst have in order to be done' or 'that is my duty' or 'that is to have this quality. Or the first party may incumbent on me.' It may sometimes be using 'right' in this sense, and the happen that there is a set of two or more second in the sense of 'morally good.' It is acts one or other of which ought to be not clear to me which of these two things is done by me rather than any act not belong­ usually happening when this difference of ing to this set. In such a case any act of view . arises. But it seems probable that this set is right, but none is my duty; my both things really happen�that some duty is to do 'one or other' of them. Thus people fail to notice the distinction �righf has a son1ewhat wider possible between 'right' and 'morally good,' and that application than 'something that ought to others, while distinguishing the meaning of b e done' or any of its equivalents, But we these terms, think that on.Iv what is morallv \vant an adjective to express tJ10 same good is right. A discussi�n of the first ;£ meaning as 'something that ought to be these positions only is strictly in point here, done,' and though we have 'obligatory' at where we are discussing the meaning of our disposal, that also has its a.'llbigulty, 'right.' It seems to me clear that 'r ight' does since it sometimes means And e!sewhere 7 he says 'admitting ( what is not true) that the word right can have a mean­ ing v;,:ithout reference to utility.' Yet, as Sidgwick points out,' 'when Bentham explains ( Principles of Morals and Legisla­ tion, Chap. I, § I, note) that his funda­ mental principle "states the greatest happi­ ness of all those whose interest is in ques­ tion as being the right and proper end of human action," we cannot understand him really to mean by the word "right" "condu­ cive to the general happiness"; for the proposition that it is conducive to general happiness to take general happiness as an end of action, though not exactly a tautol­ ogy, can hardly serve as the fundamental principle of a moral system.' Bentham has evidently not made up his mind clearly whether he thinks foat 'right' means 'pro· ductive of the general happiness,' or that being productive of the general happiness is what makes right acts right; and would very likely have thought the difference unimportant. }.1iH does not so far as I know discuss the question whether right is definable. He states his creed in the form 'actions are right in proportion as they tend to promote happiness," where tlrn claim that is made is not that this is what 'right' means, but that this is the other character­ istic in virtue of which actions that are right are right. And Sidgw:iok says 1• that the meaning of "right' or ·oughf 'is too elemen­ taiy to admit of any formal definition,' and Principles of Morals and Legislatiori, Ch. I, § 10, ' ib. §14. 10. Methods of Ethics, ed. 7,, 26 n. UtiUtarianism, copyright eds., 9. 10 Methods of Ethics, ed. 7, 32. 6

8 9

ETHICAL NON-NATURALISM expressly repudiates" the view that 'right' means 'productive of any particular sort of resu1t.' The most deliberate claim that 'right' is definable as 'productive -of so and so' is made by Prof. G. E. Moore, who claims in Principia Ethica, that 'right' means 'produc­ tive of the greatest possible good: Now it has often heen pointed out against hedon­ isir1, and by no one more clearly than by Professor Moore, that the claim that 'good' just means 'pleasanf cannot seriously be maintained; that while it may or may not be true that the only things that are good are pleasant, the statement that the good is just the pleasant is a synthetic, not an ana­ lytic proposition; that the words 'good' and 'pleasant' stand for distinct qualitie.s, even if the things that possess the one are pre­ cisely the things that possess tl1e other. If this were not so, it would not be - intelligi­ ble that the proposition 'the good is just the pleasant' should have been maintained on the one hand, and denied on the other, ,vith so much fervour; for vi1e do not fight for or against analytic propositions; we take them for granted. Must not the same claim be made about the statement 'being right means being an act productive of the greatest good producible in the circum­ stances'? ls it not plain on reflection that this is not ,:vhat we mean by right) even if it be a true statement about what is right? It seems clear for instance that "Nhen an ordinary man says it is right to fulfil prom· ises he is not in the least thinking of the total consequences of such an act, about which he knows and cares little or nothing. 'Ideal ntilitarianisn{1-2 is, it would appear, plausible only when it is understood ·not - as an analysis or definition of the notion of 'right' but as a statement that all acts that are right, and only these, possess the fur­ ther characteristic of being prodllctive of ib. 2S--6. :u: I'use this as a well-known way of r0ferring to � Professor Moore's: view, �Agathistic utilitarianism' would indicate more distinctly the differenc-e bernreen it and hedonistic ut:Lltarianism. st

THE MEANING OF "RIGHT" the best possible consequences, and are right because they possess this other c:1ar­ acteristic. If I a.,"11 not mistaken) Professor lvloore has moved to this position, from the posi­ tion that 'right' is analysable into 'prod,;c­ tive of the greatest possible good.' In Prin­ cipia Ethica the latter position is adopted: e.g. 'This use of "'right/' as denoting what is good as a means, ,vhether or not it is also good as an end, is indeed the use to which I shall conflne the word.''" 'To assert that a certain line of conduct is, at a given time, absolutely right or obligatory, is obviously to assert that more good or less e,11 w:ill e:.tist in the world, if it be adopted, than if anything else be done instead,'" 'To ask what kind of actions one ought to perform, or what kind of conduct is right, is to ask what lin d of effects such action and con­ duct will produce . , . What I wish first to point out is that "right" does and can mean nothing but "cause of a good result," and is thus always identical with "useful" . . . That the assertion "I am morally bound to per­ fonn this action" is identical \\'1th the asser­ tion "this action will produce the greatest possib!e amount of good in the Universe" has already been briefiy shewn . . . ; but it is important to insist that this fundamental point is demo::,strably certain . . , . Our "duty/' therefore, can only be defined as that action, which w:i!l cause more good to exist in the Universe than any possible alternative. And what is ''righf' or "morally permissible" only differs from this, as what "ill not cause less good than any possible alternative.-'1� In his later book, Ethics, Professor 1vfoore seems to have to adopt the otl)er position, though perhaps not quite unequiv­ ocally. On page 8 he names as one of the 'more fundamental question.s' of ethics the question \vhat, after all; is it that we mean to say of an action when we say that it is right or ought to be done?' Here it is still u p, 18. l' p. 25. ii; pp, 146-8. Cf. also pp, 161, 1691 180-1.

111 suggested that 'right' is perhaps analysable or definable. But to this question Ethics nmvhere distinctly offers an answer, and on page 9 we find, 'Can we discover any single reason, applicable to all right actions equally1 which is 1 in every case, the reason why an action is right, vvhen it is right?' This is the question v,,hich Professor Moore in fact sets himself to answer. But the reason for an action being right is evi­ dently not the same thing as its rightness, and Professor �foore seems alreadv to have passed to the view that produ�thity of maximum good is not the definition of 'right' but another characteristic which underlies and accounts for the rightness of right acts. Again, he describes hedonistic utilitarianism as asking, �can we discover any charaeteristic1 over and above the mere fact that they are right, which belongs to absolutely all voluntary actions which are right, and which at the same time does not belong to any except those which are right?'1 8 This is the question which he describes hedonism as essentially answer­ ing, and since bis own view differs from hedonism not in logical form but just by the substitution of 'good' for 'pleasure,' his theory also seems to be essentially an ansv1er to this question, i.e. not to the ques­ tion what is rightness but to the question what is the universal accompaniment and, as he is careful to add,1,, the necessitating ground of rightness. Again, he describes hedonistic utilitarianisn1 as giving us �a cri­ terion, or test, or standard by -.vhlch we could discern with regard to any action whether it is right or \vtong:H And simi­ larly, I suppose, he regards his own theory as offering a different criterion of rightness. But obviously a criterion of rightness is not rightness itself. And, most plainly of all, he says, 'It is indeed quite 'plain, I think, that the meaning of the two words' ( 'duty and 'e.,.-pedicncy,' the latter being equivalent to m are irrelevant :o the position here taktm. Thns Mr. C. D. Broad attae�'\S the behaviorist who says "that all mental processes to An instructive ex.irnp1e of the futility of reduce without restdue to the fact thut the body is behaving in a certain specific way" ( The M-ir.d attempting to reduce Cesire to introspective terms and its Place in Nature, 1925, p, 616). I agree is uifor 'horesome/ 'tiresome,' nopeful,' which by their very formation point to a relation between a subject and an object, and the word 'good,' which equally clearly Faints to nothing of the kind but to a quality resident in the ob­ ject itself, independent of any subject's reaction to the object As regards 'beautiful' I am, as I shall point out later,"' inclined to agree faat the fact faat lies at the back of our predications of it is simply the power something has of producing a cer1

t$ H 20

Cf. pp. 121-2, A General Theory of Value, 3 L pp, 127-30.

tain kind of emotion in us; and the fre­ quent use of such words as 'channing,' 'de­ lightful,' almost as synonyms of oeautifuf may be held to lend this view some sup­ port. But it is surely a strange reversal of the natural order of thought to say that our admiring an action either is, or is what necessitates, its being good. \Ve think of its goodness as what we admire in it, and as something it would have even if no one ad­ mired it, somet.'1lng that it has in itself. We could suppose, for instance, an action of self- denial which neither the doer nor any one else had ever admired. If now some one were to become aware ,of it and admire it, he would surely pronounce faat it had been good even ,vhen no one had been ad­ miring it. Professor Perry makes the further ob­ jection that the $objective� theory derives all its plausibility from its exponei,ts' being preoccupied with 'the aesthetic and con­ templative values/ and t.'1at it precludes them from giving a comprehensive account of all values. The most serious defect of this type of theory is its failure to provide any systematic principle whatsoever. There are as many iI,definable values as there are feeling attitudes, and since these are to be regarded as objective qualities rather than as modes of feeling, there is nothix1g to unite them, not even the principle of feel� ing. If "good'� is a unique qcality, then so are "pleasa..'1.t/' "'bad," and �·ought'' There is no way of subsuming plea.,ant under good, or of denning the opposition of good and bad, or of subsuming bofo good and ought under a more genera� category such as value. If, on the other hand, value is defined :n terms of interest, then the varia­ bility of interest seems to account for both the unity and the diversity of values.' 21 His assumption, then, is that there must be some single sense of 'valuable' in which the word is ahvavs used, and his contention is that a subjective theory alone will serve to assign such a single meaning and. to show 1 �

A General Theory of Value, 34.


the relations betw�en the various speci.£.c kinds of value. And under the heading of 'valuable' he includes both things which would not naturally be described as being valuable at all, and things which we can surely recognize to have value only in fun­ damentally different senses. Does any one really think that obligatoriness is a special form of being valuable? 22 Is it not a hasty assumption to assume that it is an instance of the same kind of thing of which moral goodness or beauty is another instance? And is it not clear that what we call eco­ nomic values 23 are merely instrumental values, different in kind from the good­ ness of virtue or of pleasure? The assump­ tion that there must be 'a general theory of value' applicable to value in all the senses of that word seems to me to be unjustified. At the same time, I am inclined to agree with Professor Perry in one of his contentions, though not in what he ( if I understand him aright ) seeks to deduce from it. He is seeking to find a single thread of identity which unites all our ap­ plications of the word good, and to infer from this that the word 'good' has the sin­ gle meaning which he assigns to it. Now when I consider the variety of meanings of 'good' indicated in the preceding chap­ ter-the predicative and the attributive use, the meanings 'successful in his endeavour' and the 'useful,' the instrumental and the intrinsic sense-though I cannot agree that what we mean in all or any of these cases by 'X is good' is i being produc­ tive of pleasure or "ith being an object of desire, The point is not that the proposed definition is not seen at first sight to be true, or that it needs inquiry, but that it does not survive inquiry, Professor Pernls own criticism of Pro­ fessor Moore takes' the following fom1, Sup­ pose that 'good' be defined as 'desired by some one.' This definition is disproved, says Professor Moore, by the fact that even if war is desired by some one, it is still possible to inquire ,vhether ,var is good. Professor Perry seems to admit this as fatal to the proposed definition, for he proposes to sub­ stitute for it what he evidentlv thinks of as a different definition, 'go�d in some sense:::::::desired bv some one.1 And he en­ deavot:rs to tu,,:; the edge of Professor

Moore's object:on by saying faat the cor­ rectness of this definition is quite compati­ ble with our still being able to inquire ( as we evidently can ) whether ,var, if it is good in tbis sense, is also good in some other sense e.g. deslred by all men, or obli­ gatory, or beautiful. 2a TI1is seems to me strangely to miss the point No one would, I suppose, dream of objecting to the equating of 'good in some sense' vvith 'desired by some one' on the ground that war though desired by some persons is not dcisired by every o:ae, or not beautiful, or not obligatory. The objections are ( 1) that, even though desired by some persons, v:ar is not in any sense good ( though there may be elements in it that are good ), and ( 2) that, even if it were in some sense good, what would be meant by calling it good is most certain.'.y not that it is desired by some one. Professor Perry further tries to base an argument for the �elath1ty o: good to the interests of individuals: on, the fact �that the question may be submitted once again to each individual judge.' 'If when a given ob­ ject a is already acknowledged to be good the question of its goodness is nevertheless put to a subject M, the question is assumed to refer to the special sense of good which is relative to M.' 2' This would surely be a nonsensical procedure, If the goodness of a is already 'acknowledged,' i.e. admitted by both the persons involved, there is no sense in the one asking the other whether the ob­ ject is good; and anything that the other may say such as 'I desire it' or 1: don't de­ sire it' has no relevance to the question whether it is good, which has already ex h yp othesl been settled in the affirmative, The advo�.ates of the view he is criti­ cizing are, Professor Perry points out, anx­ ious to secure for ,.good� a mea:::ing \Vhich sha!! 'provide judgments of value w:ith a common object which will determine their truth or falsity.' And for this purpose, he in­ sists, �an interest is as good an object as any A General Theory of Value, 36-7, " lb, 31, 211


other. The fact that M takes an interest in a, consists iil a relation of a to M; but this fact itself is not relative to M's judgment about it, or to tbe judgment of any otber subject.' 30 There is, in fact, a great dif­ ference between a vie\V which makes the goodness of an object depend on a subject's judgement that it is good, and one which makes it depend on his interest in it. The former view is one that Vvill not stand a moment's examination; the latter is one that does provide judgements of value with some reality to judge about, and tbat there­ fore requires serious consideration, But while it provides for our judgements of value an object independent of our judge­ ments, it fails to do justice to what is also implied in our judgements of value, tbat when one person says an object is good and another says it is not, they are contra­ dicting one another. For if M only meant 'I take an interest in a' and N only meant 'I do not,' tbey would not be contradicting each otber. Professor Perry sometimes, for brevity, uses as equivalent to 'good' 'enjoyed by a subject,' and sometimes 'desired by a sub­ ject.' Neither of these phrases does full jus­ tice to his theory. His theory is that to be good is to be an object of interest, and in­ terest is thought of as covering both desire and enjoyment; i.e. the goodness of some things consists in their being enjoyed, that of others in their being desired, and that of otbers, perhaps, in their being both enjoyed and desired ( though this, as we shall see, is impossible). No\v so long as we say ( as he is apt to say) tbat 'the goodness of the primrose consists in its being desired,' 31 the theory seems at first sight attractive enough. But obviously it is only a rough and ready "c!escription of my desire to say I desire a primrose. YVhat I desire is to be seeing it or smelling it or possessing it. As soon as we describe defi­ nitely what it is that we desire, \Ve see that it is sometbing which does not yet 0 �


Ib. 38. Cf, A General Theory of Value, 133.

ETHICAL NATURALISM exist. There are no doubt cases in which we desire to go on doing what we are doing, or being in tbe same sort of state that we are in. But even if I desire, for in­ stance, to go on looking at a primrose, what I desire is not tbe looking which is taking place at present, but tbe looking which I wish to take place in the immediate future. The object of desire is always something non-existent. If it be said that it exists as a possibility, we must reply tbat tbat is an inexact way of saying that the possibility of it exists, which means that though it does not eA"ist, the nature of some or all of the things that do exist is not incompatible with its coming into existence. It is plain that in so far as goodness were either identical 1,,vith or dependent upon_ being desired, nothing could botb exist and be good. Now I suppose tbat we are all convinced both that some things that exist now are good and that tbings of certain kinds, which may come into exist­ ence in the future, will be good if and when they exist; and I suppose that apart from these convictions \Ve should have lit­ tle or no interest in the topic of 'good,' and etbics in particular would go by the board. Yet in so far as tbe theory identifies the good with the desired, it denies both these convictions. But it might be replied tbat the goodness of existent things consists in their being enjoyed, and the goodness of non-existent things in their being desired. I must take leave, ho\vever, to doubt whether we can say of a non-existent thing that it is good. However much one were convinced that conscientiousness, for exam­ ple, is good, and tbat A might become con­ scientious, no one would say ..A's conscien­ tiousness is good' if he were convinced that A is not in fact conscientious. But, our op­ ponent might reply, we can say of kind, of tbing tbat they are good even if we are not convinced that any instances of these kinds exist. 'We might say 'perfectly conscientious action is good,' even if ( as Kant suggests ) we are not convinced that there has ever been such an action. But that is only a short-hand way of saying that without


being sure that such an action ever has ex­ isted, we can be sure that if any existed it would be good. Hypothetical goodness pre­ supposes hypothetical existence just as ac­ tual goodness presupposes actual existence. And if so, being good can never be identi­ cal \Vith beiug desired1 or even compatible with it. The relation in which the prinwose stands to desire is not that of being desired but that of exciting desire. This is a rela� tion in \Vhich existing things can stand to desire, and the theory might be transmuted into the form, 'the good is that which ex­ cites desire.' But the excitants of desire fall into two classes, There are things our ex­ perience of v,rhich is such as to make us de­ site to remain in our existing relation to them, or to get into some closer relation to them, and to others like them; and there are things our experience of which is such as to make us desire to get away from thern. Things of the second class are just as decidedly excitauts of some desires as things of foe first c!ass are of others. And obviously one ma::n sub-class ( if not the whole ) · of the second class mnsists of things that cause pain. Thus if 'good' meant 'excitan: of desire' we should be led to the conclusion that things that cause pain are, as such> an important class of goods. This conclusion would evidently not be ac­ cented, and therefore the theory would ha";,e to be modifled into the form 'the good is that \Vhich excites the desire to maintain our relations with it, or to get into closer relations with it, and with others of its kind�- what we may, for shorty call �positive desire.> Now, on the face of it1 some of the thbgs that excite positive desire32 do so because they are judged to be good. Prima facie one would say that if the mnscious­ ness of a good disposition in oneseH or the contemplation of it in another leads me to \Vish to maintain and develop that sort of disposition, it is not because I feel it to be :1:: Desire, of course, not for them to es:ist but for us to be in some ne'"v re]ation to them or to continue to be in the same relation to tbem.


pleasant but because I judge it to be good. But this alternative is not open to Professor Perry, for, in basing our taking an interest in the thing on our thini!:ing it good 1 it would involve the giving up of his main thesis, that a thing's being good either is or is based upon our taking an interest in it. All that is left for him therefore is to iden­ ti£y what is good vvith that which by virtue of the pleasure it causes excites desire for a closer relation with it and with other things like it. '.·v'hat is good, then, for hin1 is that which excites and thereby excites such a desire, And though he includes both these elements in his formula, the fact of exciting pleasure is e·,idently the root fact of which the other is a mere consequence. Not only, however, is pleasantness the fundamental and tendency to excite desire only a consequential ele�ent in goodness, according to the thoory in the form i n which i t seems necessary to restate it 1 but it is far more plausible to put forward pleas­ antness, thaJ1 to put forward t.�is tendencyi as the essence of goodness. If we say 'that which produces so-and-so is, as doing so, good/ we are evidently :h:nplying that what is produced is inL-insically good, and what produces it instrumentally good. And it is plausible enough to say 'pleasure is intrinsi­ cally good, and what produces it inst:c"U­ mentally good'; there is a pretty general agreement that pleasure, whether it is the good or not, fa at least good. But there is !10 general agreement that desire, or even positive desire, is good. If we take the moral standpoint we must say that some desires are good and others b�d, and that when desires are good they are good not because they are desires but because they are the sort of desires they are. And ·if we take the hedonistic standpoint, we must say that desires are good or bad ( wh:ch will 1nean 'pleasant o:r unpleasant') not in virtue of being des'res but mainly ( I sup­ pose) in virtue of their being supposed to be likely or unlikelv to be folfllled. Desire ( even positive de.;ire) thus not being a thing necessarily good fr� itself, there is no reason \Vhy} in general, things that excite

166 desire ( or positive desire ) should be good. So long as we thought of things as objects of desire, it ,vas perhaps not unplausible to say that objects of desire are good even vvhen the desire is not; but if we recast the theory in the form in which we have found it necessary to recast it, and say the good is that which excites positive desire, i.e. which is to it as cause to effect, there is no reason ( obvious or alleged) why, positive desires not being always good, their exci­ tants should nevertheless always be so. Th� most favourable way, then, of pre­ senting the theory we are examining is to exclude from it the reference to desire and to reduce it to the form \vhat is good is that which produces pleasure.' 33 But no one would in fact say that everything which produces pleasure is good unless he thought pleasure itself good; 34 and the theory emerges in the final form 'pleasure, and pleasure alone, is good by its own na­ ture; and what produces pleasure, and only what produces pleasure, is good because it produces something good.' The heart of the theory, then, in spite of all it has said by way of attack on ordinary notions of intrin­ sic good, is that there is one thing, and one thing only, that is intrinsically good, viz. pleasure. The theory when reduced to its . simple terms seems to be our old friend, hedonism. After all the able refutations of hedonism that have been published in re­ cent years, it seems to me unnecessary to tread once more on this rather hackneyed ground, and I suppose that Professor Perry ·would agree that hedonism is untenable, and claim that his own theory is tenable This is ambiguous, since it may mean ' "good" means "productive of pleasure," ' or 'what is good is good because it produces pleasure'; i.e. the ambiguity involved in, the theory from the start ( cf. pp, 80-1 ) still remains. 3 + And inferred from this that what produces it is good. But it is surely plain that it does not follow from a thing's being good that what pro­ duces it is good, in the same sense of 'good.' It must be admitted that we often call 'good' things that are merely useful, but then 'good' is being used improperly. ·where I use the phrase 'instru­ mentally good,' I use it to indicate this common but loose sense of 'good.' 33

ETHICAL NATURALISM only in virtue of elements that distinguish it from hedonism. But these elements are, if I am not mistaken, among the least tenable elements in his theory. There is, however, one more point of view from which the theory may be exam­ ined. Professor Perry describes 'the most popular' objection to it as being that 'the fact of desire is not accepted as .final in most judgments of value. Objects of desire are held to be bad in spite of their being desired, and desires themselves are held to be bad whether or no they are satis­ fied.' 3 5 I need not consider ( a ) one form of this objection with which I have no sympathy-the view of Schopenhauer and others that all desire is bad; that is an ex­ travagance of quietism for which there is little to be said. ( b ) The first real difficulty to which the theory is exposed is that named next by Professor Perry, viz. the fact that 'the same object may be liked or de­ sired by one man, and disliked or avoided by another.' 36 This fact, taken with the identification of 'good' with 'object of inter­ est,' leads to the conclusion that the same thing may be both good and bad. On the face of it, this result is paradoxical, and all but self-contradictory; but he claims that 'a relational de.:B.nition, such as that here pro­ posed, is the only means of avoiding con­ tradiction.' 37 The claim is an odd one: by identifying good with object of interest we get into the paradox of calling the same thing good and bad ( a paradox which an absolute theory at least escapes, whatever be its other merits or demerits) ; and then we triumphantly get out of the difficulty by saying, 'Oh, but good only means good for one person, and bad only means bad for another person, so that there is no paradox.' Is it not clear that when we assert the goodness of anything we do assert some­ thing which we believe to be incompatible with the same thing's being bad? We may describe a thing as 'both good and bad,' A General Theory of Value, 134. lb, 135, 3 1 lb. 136.

a; 3



THE MEANI?-;G OF 'GOOD' but SU-Ch language is not strict. ( i ) We may mean that the thing contains some ele­ ments t.liat are good and some that are bad, but then that is the right way of putting the matter, and 'the thing is both good and bad'· is only a loose way of putting it. It is implied in our t:i.ought on the subject both that if we push our analysis far enough ,:ve shall find some elements that are simply good and others that are simply bad, and that the whole is not both good and had but is either on the whole good or on the whole bad. (ii ) It may be suggested that, without thinking 0£ a thing as consisting of good and bad elements, we may judge it to be good from one point of view and bad fron1 another-that a state of mind, say, may be morally good and intellectually bad. But this turns out to be reducible to the former case, in which analysis reveals a good and a bad element. If we take a tern· para! section of the history of a mind, how­ ever short be tlie section there will be ele­ ments in it of knowledge and opinion which have a certain value, and actions or dispositions to act which have a certain value ( positive or negative) . The whole state of mind, then, cannot be judged from the moral point of view} nor from the intel­ lectual, but some elements in it from the one and some from the other. And each such element will have a goodness that is incompatible with its being bad, or a bad­ ::1ess that is incompatihle with its being good; and the whole state of mind w:ill have a degree of goodness or else a degree of badness, which can be assessed only from a point of view in ,vhich we tra�­ scend both the moral and the intellectual point of view. ( c) 'The case which has most deeply affected popular habits of tl10ught, and which is mainly responsible for the preju­ dice against the present theo1y of value,' says Professor Perry, 'is the case in which an interest or its object is morally conde..'11ned.' sa It is certainly an obvious objection to the theory that all objects of interest are good, that :in point· of fact we :HI

Jb. 1.)6,

do judge to be bad many things in which nevertheless some one or other takes or has taken an interest. Professor Perry's answer to this objection is to urge that in such a case we are perlorming a moral judgement, and that 'moral judgments are not con­ cerned with value in the generic sense; but with a specific and complex aspect of it. . . . They do not deal w:ith interests per se, but with the relation of interests to the complex purposes in which they are in­ corporated.' 39 In answer to this it is important to point out that the ter.n 'moral judgement' contains a serious a,_-.n_biguity. There are three types of judgement which have by various writers been termed moral judge­ ments. These are ( i ) the judgements in which an act is pronounced to he right or ,:vrong; (ii) the judgements in which an action or disposition is judged to he mor­ ally good, or bad, or indifferent, i.e. to have ( or fail to have) the kind of goodness or badness that only dispositions and actions can have; (iii ) the judgements in ,vhich something is said to be good or. bad or indifferent sans phrase. The first hvo may be said to be departrnental judgements, in the sense that each of them is applicable only to one class of objects, the first to acts considered apart from their mo­ tives, the second to dispositions and ac:ions considered in respect of their motives. Judgemcnts of the third class are not in any way departmental; they may be made about anything in the whole world. It can be said of some things-I suggest, as at any rate an adumbration of the things of which it can be said, virtue, kno\.vledge and weI!­ grounded opinion, and pleasure40-that they are good; of others-vice, badly grounded opinion, and pain-that they are bad; and of other things that they are indifferent, i.e, considered in themselves, though many of them may be instrnmental to good or to evil. In making such judge:m Ib. •0

136-31. To avoid maki:-ig any statement too compli�

cated, I omit a further kind of good which mentioned later1 cf, p, :..38.

,vil: he


ments we are not adopting a narrowly ethi­ cal standpoint; we are saying for instance that wisdom and pleasure are good, though they are not morally good. We are taking the most commandi.c1 g point of ,1ew that can be taken vcith regard to the value of the thi. it ,vnuld be inconvenient to be precluded from using the term 'interest' when 'Ne wished to signify tl1at cognitive attentiveness which the terrr:. is commonly used to signify. 'Liking,' then, seems on the whole to


be the best term to express the basic ele­ ment in the consciousness of value-for-self. We must now take up our task of tracing the manner in which the concept of value-for-self acquires for the reflective consciousness a much more specific mean­ ing than mere 'objective of selfs liking,' owing to the recognition of certain distinc­ tions within objects of liking. I shall set out these distinctions in logical rather than his­ torical order, beginning with the most sim­ ple and working towards the more com­ plex. The first distinction is L'ie very elemen­ tary one between ob'ects liked rrwre and objects liked less. Thi; distinction need not detain us. It introduces, L'1 its crudest form, the distinction between major and minor values-for-self, corresponding to objects of major and minor liking respectively. It ought not, in stricb1ess, to have any effect in modifying the meaning of the concept 'value f- or-self; although I shall have to point out later that it is not altogether cer­ tain that it does not in fact some slight ·modifying influence. 111e next distinction has greater impor­ tance. There are some objects that ,:ve li1::e for themselves, others that we like only be­ cause they help in the attain::nent of objects liked for themselves. Recognition of this distinction leads us to make a distinction between end values-fm·�seif and i11-strwnen� tal values-for-seif, corresponding to objects of independent and dependent liking re­ spectively. And with the emergence of this distinction the meaning of 'value-for-self does begin to undergo definite modi­ ficatio;:;.. Sillce it is from end values that in· strumental values derive all the value that they possess for us, it will be natural to rec­ ognise that it is only end values that have a direct claim to the title 'value.' In so far as this distinction is active in the mind, therefore, there will be a tendency to iden­ tify value-for-self not with object of any !iking of the self, but with object of an in­ dependent liking of the self. Our next distinction, one which has

ETHICAL NATURALISM very i:mportant consequences, arises '-Vifuin the field of end values. It rests upon recog­ nition of the fact that some end values have also instrumental value, being condu­ cive to the attainment of certain other things liked for themselves, while other end value..:; have from the same point of view a definite disvalue. Thus ends like health and knowledge may be liked for the.'11selves, hut liked further because seen to contrib­ ute usefully to the attainment of many other liked things, whereas a good many things liked for themselves, e.g., idleness and gluttony, have quite obviously an op­ posite tendency. 'Ne may perhaps express the situation that arises by say5ng that end values fall roughlv ., into hvo classes, accord­ ing as their m�i::1 tendency is to co-operate with or to obs:ruct the end values of the self as a \Vhole. Now the 'co-operative' end values; if we may so christen them, ,vill certainly be accorded a much higher status as values­ for�self than the 'obstructive� end values. But the precise nature of this fligher status1 is sometl:ing which we must detennine with a good deal of care, for in truth we meet here with a very important develop­ ment in the mea."llng of the concept 'value-for-self.� There is no question1 in­ deed, of any modification in principle of the original equat'.on 'value-fo r -self = 'liked by self.' But a complexity now reveals itself within the meaning of 'liked by self which reacts profoundly upon the meani.'lg of 'value-fo r -self.' For it now appears that 'liked by self may mean either, on the one hand, to be an ob]ect of liking to the se!f as a whole, to the self as the unitary centre of its several likings-as in varying degrees is the case wifo what we have called the 'co�operative' end values-or it may mean; on the other hand, to be an object of liking to the self only in some very partial aspect of its being. and so iriimical to the self's other likings as to be more properly called an object of disliking to the self as a ,vhole--as in the case of the 'obstructive' end values. Now it can hardly be denied, I


think, that it is in the former of these two conceptions of itself, i.e., in its being as the unitary centre of manifold likings, that the self recognises its essential selfhood to con­ sist. Accordingly, it will be only those ob­ jects of liking which are harmonious, if not positively at least negatively, with the self's likings as a whole, objects which belong to the class of co-operative end values, or, at the least, of neutral end values, which will now be accepted by the self as genuinely representing what is 'liked by self'; and it is they alone which will now be recognised as 'values-for-self.' It is evident that when this distinction has become operative in the mind many things previously regarded as values-for-self will present themselves quite definitely in the light of disvalues: because, though in one sense still 'liked by self,' in a more profound sense of 'self' they are in antagonism to what is 'liked by self.' The distinction which is now engaging us concerns our ultimate purpose so closely that I may be excused if, in spite of the limited space at my disposal, I dwell for a little upon its general principle. The gen­ eral principle is, I think, neither obscure nor seriously debatable. The essence of the matter is just this, that as self-consciousness develops, and the self becomes conscious of itself as the unitary centre of manifold lik­ ings, the meaning of 'object of liking to self,' and consequently of 'value to self,' be­ comes deepened, and in a manner trans­ formed. v\lhatever is now regarded as good or valuable for the self has got to be some­ thing that respects the systematic manifold­ ness that belongs to the nature of a self. being in perfect ac­ cord with all that he says, I propose to waste even fewer by making no further reference to this type of theory. But we ::nust follow Dr. Ross into his sub-division of group A, to which our own theory would most naturally belong, and which Dr. Ros� allows to possess a much greater prima facie plausibility than the other group. The sub-division adopted will be best explained in its author's ovvn words. *Theories of this type," he says, ·are divisible into those which identify goodness with the presence of some fee:ing ( 1) in at least one person, no 1natter who he is, ( 2) i"'l the person ,vho judges fill object to be good, ( 3) in a ma­ jority of persons of some class or other-say persons belonging to a particular stage in the history of civilisation, ( 4) in a majority of mankind, or ( 5) in all mankind.' 11 Now of cou:rse the most L'Onspicuous feature of Dr. Rosls classification from our point of view is that it omits altogether the particular variety of type A which seems to ourselves to offer the true definition. For it need scarcely be pointed out that being an 11

The Right and the Good, pp. 82-83.



object of liking to human nature is by no whose summum bonum is sought? Not means identical with being an object of lik­ surely any particular man, but just 'man as ing either to all mankind or to a majority of such,' the exemplar of our common human all mankind-while much less is it identical nature. The Greek moralist works with a with any of the other suggested formulae. type man, constituted by the conative, Yet there is surely nothing unintelligible, or emotional, and intellectual proclivities be­ even strained, about the concept 'object of lieved to be common to human nature, sets liking to human nature.' There are ap­ him in a natural and social environment propriate objects of liking to human nature which, though inevitably relative to the just as there are appropriate objects to cat age and place of the moralist, is made as nature or dog nature. Cat nature is so con­ little specific as possible, and seeks to de­ stituted as to like stalking its prey and to termine what mode of life will afford the dislike immersion in water. \i\1hat is the fullest satisfaction to a being so constituted difficulty about saying that human nature is and so conditioned. It is not anything es­ so constituted as to like and dislike certain sentially dissimilar, in my judgment, that specific things also? Indeed, aren't we say­ mass opinion has been doing in the long ing that kind of thing ahnost every day of process of constructing its list of 'goods for our lives? And aren't our psychologists bus­ man.' The chief difference is due to the ily engaged at this very time in trying to simple fact that in the one case the process ascertain just what the basic likes and dis­ is undertaken with scientific thoroughness likes of human nature are? And wasn't it and method, and in the other case not. the chief aim of the Greek moralist to de­ That is why mass opinion is content with a termine what mode of life human nature set of pre-eminent goods, and does not con­ was so constituted as in the end, and on cern itself with the deeper, and to ethical science vitally important, question of the the whole, to like best? I hope, then, that no one will retort relation of these goods to one ariother against me that while Tom, Dick or Harry within the unity of the good. But so far as can have likings, it is not possible to assign the concept of 'human nature' is concerned, likings to what is not a person but an ab­ the procedure seems fundamentally the straction, viz., human nature. If ,ve are to same; and there seems no more difficulty in take that view in earnest, then we ought applying the concept in one case than in like\vise to insist, I presume, that instincts the other. I must claim, then, that the subjectivist cannot intelligibly be assigned to human nature either, since, strictly speaking, it is definition of good which I have placed be­ only an actual living creature that can have fore you is not to be put out of court on the an instinct. But I fancy that the critic score of being unintelligible, and I must in­ would wish neither to forbid other people sist that Dr. Boss's criticism of subjectivist to speak, nor himself to refrain from speak­ theories is not exhaustive of the type so jng of 'the instincts of human nahlre.' In long as the classification upon which it is neither case is there any real difficulty based ignores this particular variety. It about the meaning that is intended. Just as might be the case, indeed, that the exclu­ there are instincts which men have in sion of it was only formal. That is, it might virtue of their common human nature, so be the case that some of Dr. Ross's criti­ too there are likings ,vhich men have, or cisms of the theories on his list are capable tend to come to have, in virtue of their of being adapted, with more or less trifling modifications, to the destruction of our common human nature. But perhaps the best analogy for our theory too. It is therefore of first- r ate im­ usage is provided by the usage of the portance to observe that this is not even re­ Greek philosophers in their search after motely the case-as anyone may assure man's summum bonum. YVho is the How, one might ask, vmuld Dr. Ross s argument require to be supplemented in order to become relevant to the question of the obiectivity of the value of knowledge? It would, I think, be formally satisfactory if, in being invited to appraise the relative value of these t\vo states of the universe, we were at the same time instructed to rule out from our minds all considerations aris­ ing from our familiarity ,vith human likes and dislikes. Thus we should be obliged to suppose, for the sake of the argument, that there is no native impulse of curiosity in man \vhich makes him come to like know­ ing for its own sake. If, fulfilling these con­ ditions, we ,vere to proceed to put to our­ > selves Dr. Ross s question, an affirmative answer would, I think, imply the recogni­ tion of a strictly 'intrinsic' value in lmowl­ edge. But should \Ve then be able to return an affirmative answer? It is a matter which each must decide for himself by personal experiment, but I feel convinced that, if the experiment be performed with due observ­ ance of the conditions, only a negative answer will be found possible. No legiti­ mate ground remains, I believe, for judging the first state of the universe to be the bet­ ter. Old emotional habits, like old cognitive


habits, die hard, and there is undoubtedly a great mass of prepossessions to be broken through before one can hope to return a fair answer. But I must leave the experiment with the reader, pausing only to draw a t ­ tention to t\vo pitfalls which seem espe­ cially liable to engulf the unwary. ( 1 ) We know that 1.,-iowledge, even if it were not itself liked, would still be instrumental to a host of things that are liked, and this makes it difficult for us not to think of the state of the universe with lmowledge as the 'better' state. But it is clear that the terms of the hypothesis make this consideration irrele­ vant. Reference to other goods, whether as 'objects of liking' or in any other sense, is definitely ruled out. ( 2 ) A good many of us are more deeply influenced than we are apt to realise by an inherited religious tradition which leads us to think of our faculties as given to us by God for use and develop­ ment, so that it is in accord with the Will of God, and so far 'good,' that they should be exercised to the full. On this ground alone there is a powerful disposition in most of us to regard a state of the universe in which the faculty of knowledge finds ac­ tive expression as better than a state in which it does not. Evidently, however, we are not, in making this judgment, recognis­ ing knowledge to be something 'good in i t ­ self.' '''e are merely recognising respect for the gifts of God, or obedience to the Di­ vine Will, to be a duty.14 It appears to me, then, that Dr. Ross totally fails to demonstrate the objective must be admitted that it is extremely H It difficult, in an ideal experiment such as that which we are here called upon to perlorm, to prevent one's value-judgments from being affected by the cross-currents of moral and religious duty, Many, perhaps most, educated persons, even apart from religious considerations, believe it to be their duty to develop their capacity for kno-wledge, On that account the conception of 'knowledge' is closely associated in their minds with the conception of 'goodness.' But clearly the effort has got to be made to abstract from the influence of this connec­ tion, if we are seeking to discover, by the device of experiment upon our value-responses, whether knowledge is good simply as such, in and by itself, as the Objectivist claims that it is.


goodness of his 'intrinsic goods,' and I am not able to conceive any other method likely to yield a different result. Dr. Moore's attempt fails for exactly the same reason as Dr. Ross's. His argument in Chapter VI of Principia Ethica depends es­ sentially upon a prior supposed refutation of relational theories of the meaning of good in Chapter I; and, as we noted ear­ lier, Dr. Moore is himself the latest recruit to the ranks of those who find the reasoning of Chapter I fallacious. I do not find, there­ fore, in Dr. Moore's attempted demonstra­ tion of intrinsic goods anything which places in serious jeopardy the central con­ tention of this paper- the contention that there is nothing whatsoever, '\\rith the single exception of moral virtue, which does not derive its goodness from a relationship to subjective liking. 6 It is more than time that this paper drew to its belated close. Nevertheless, I must add just a few brief words upon a matter al­ luded to at an early stage of the paper, if I am to round off my theory with any pre­ tence of completeness at all. It has been an implication of my argument that when we predicate value of morally virtuous conduct we mean something rather radically dif­ ferent by the term value from what ,ve mean when we predicate value of anything else whatsoever. In the latter applications the meaning always involves an essential relation to human liking. In the former ap­ plication no such reference is involved. Yet there clearly must be some common factor in the two meanings. Othenvise, why use the same term 'good' or 'valuable' in both cases? VVhat is this common factor? VI/hat is the analogy between the usages which justifies us in employing common terms? The correct answer, I think, is that all usages of the term 'good' signify at least this common feature in that to which good-


ness is attributed, viz., that it is the object of what may perhaps least misleadingly be called a pro-attitude: just as that to which badness is attributed is always the object of a contra-attitude. An object of liking is quite obviously the object of a pro-attitude. But it is equally obvious, when we reflect upon it, that morally virtuous conduct is likewise the object of a pro-attitude. The latter pro-attitude is certainly not the same kind of pro-attitude as is entertained to­ wards an object of mere liking. But a pro-attitude it undoubtedly is, The identity and the difference can probably be made most plainly apparent by reflecting upon the value-judgments ingredient in any sim­ ple case of so-called 'moral temptation,' in which the course we believe that we ought to follow is recognised to be incompatible with the course that we like best. Our men­ tal attitudes tov,.rards the two courses are con­ spicuously different, but both of them are be­ yond question pro-attitudes. Indeed, if they were not, there could be no consciousness of inner conflict. It is equally certain, on the one hand, that the morally right course does not appeal to our mere 1iking,' and, on the other hand, that it does appeal to us. Its appeal is such, indeed, that it may be made by us the motive of our act, and thus be adopted as our 'end' in preference to that which we 1ike best.' If this is true, it appears that what we mean ultimately when we predicate good­ ness of moral virtue is that it is an object of approval or favour to the moral conscious­ ness. Does this then imply, it may perhaps be asked, that even the goodness of moral virtue is in the last resort "subjective,' con­ sisting in a certain relationship to a state of consciousness? It is, I think, partly a ques­ tion of the use of words. Most people, how­ ever, would probably agree that the moral consciousness, though it must be in a sub­ ject, is yet not 'subjective' in the same sense as desires and likings are 'subjective.' But I cannot now embark upon the long and ar­ duous task of defining the true status within the self- much less within the whole

MORAL AND NON-MORAL VALUES scheme of things- of the moral conscious­ ness. A fully adequate theory of value could not, I am sure, be dispensed from this obli­ gation. But the pretensions of the present paper are more modest. I am content if I


have established a prima facie case for the view that the meaning of value as applied to all the so-called intrinsic values with the exception of moral virtue involves an essen­ tial relationship to human liking.

Method in Ethics PA UL HENLE Pragmatic philosophers have on the whole tended to treat problems of ethics as paral­ lel to problems of lmowledge. The object of this paper is, first, to make the parallel per­ fectly explicit and, second, to show that it is not peculiar to a pragmatic view but is compatible with any one of a wide range of epistemological and ethical positions. VVithin this range and for a considerable number of problems, at least, the aim is to set up a one-to-one correspondence be­ tween ethics and epistemology with the re­ sult that any conclusion from one discipline · may be carried over to the other. The method of reaching this conclusion is to consider the relation betvveen pleasure and value. Sharp definitions of neither of these terms can be given, since in part the aim is to provide a method of determining the nature of these entities. The usual injunctions concerning the term ''pleasure'' are, however, necessary. It is not intended to refer exclusively, or even pri­ marily, to physical pleasures, and there is no assumption that all pleasures, or even any of them, can be ranked on a single lin­ ear scale of magnitude. Pursuant of this plan, too facile rela­ tionships of pleasure and value ,vill be dis­ carded in Section I, the parallel to episte­ mology will be established in Section II, some consideration of the ontology of ethReprinted from the journal Ethics by Paul Henle by permission of The University of Chicago Press. Copyright 1943.


ics will be noted in Section III, and the problem of ethical choice will be examined in Section IV.

I In general, this paper is in sympathy with naturalistic conceptions of value, holding that felt pleasures must be ethically impor­ tant and that any attempt to disregard them leaves an ethical theory perhaps con­ sistent in itself but utterly useless, since it creates an insoluble problem of ethical mo­ tivation. To build a whole ethical super­ structure vvithout answering the question, "v'/hat motive is there for doing good?" is to deprive ethical systems of any efficacy they might have and to reduce them to the status of an empty play of logical def­ initions and theorerns1 devoid of applica­ tion, and not complex enough as logic to make them even interesting. And the only way of avoiding this result is to incorporate the problem of motivation within the theory of good itself in the manner of the ethical naturalists. This paper, then, begins from a naturalistic position, though I am not at all sure it ends with one, for theories which equate good with a totality of pleasures or satisfactions, such as utilitarianism, are un­ workable from two points of view. They


answer problems neither of personal nor of social morality. To consider first the case of personal morality. The answer to an ethical dilemma must be: "Do what will create most pleas­ ure," either individual pleasure in the case of egoistic hedonism or pleasure of society in the case of utilitarianism. But this ans,ver, in the one case as in the other, is likely to leave the problem unsolved, for the alternative which will create the greater pleasure depends on which can be shown to be more worth ,vhile. This is to say that pleasure is not an intrinsic charac­ teristic of an activity, nor are pleasures rel­ ative merely to the actor, but they seem relative as well to the ideal they further or with which they are in conflict., To take a concrete case, consider the now familiar sit­ uation of a stndent faced with the alterna­ tives of continuing his studies or enlisting in the army. To advise him to do what will cause the most pleasure is only to invite the reply: "If you can show that the paramount necessity is the defense of the country, then enlisting will bring most pleasµre; but if you can sho,v that this necessity is not im­ perative, or if you can show that the intel­ lectual tradition should be continued at any cost, then continuing stndies will bring the greater pleasure-but, until you show one or the other, your advice is useless." Such a case, as well as the foregoing considerations, would seem to indicate that there are at least two components in pleas­ ure, the one which might be termed immediate pleasure, resulting from an ex­ perience inherently satisfactory to the per­ son experiencing it. Predominantly of this sort are the pleasures of corning upon an unexpectedly beautiful landscape or of smoking a cigarette. Another. aspect of pleasure lies in accomplishing or furthering some purpose considered valuable for its own sake. Satisfactions of writing a book or taking an active part in some civic organi­ zation are of this latter sort. Often the two components of satisfaction may be evenly


balanced, as in the satisfaction of viewing a landscape for the sake of which one has climbed a mountain or in the satisfaction of smoking after a full day's work. And in no case do we "1sh to suggest that the two as­ pects are experienced separately; rather they interpenetrate, fusing into a unitary e1.-perience whose dual character is re­ vealed only in analysis. This is true even in the case in which the two elements have opposed values, as in the case in which the pleasure of a cigarette is tempered by the feeling that time is being wasted or in the case that the dreariness of writing is miti­ gated by the feeling of the importance of what is being written. The point of these illustrations, then, is that, while immediate pleasures might con­ ceivably be reckoned by some hedonistic scheme, the second group of pleasures are dependent on the acceptance of some un­ derlying purpose, so that the problem can­ not be decided in terms of pleasures alone. For the sake of clarity and in a manner which follows current usage, ,ve shall refer to the second class of pleasures as satisfac­ tions. Again, on the side of social applica­ tions, a naturalistic ethics requires the em­ ployment of some sort of hedonistic calcu­ lus. Now it is generally recognized that such a calculus is very difficult to construct, too difficult for any precise applications; but what is less frequently recognized is that the calculus, even in theory, is impos­ sible. The objection is not merely based on applications-it is no proper objection to contemporary physics, for example, that it would have great difficulty gi,�ng an accu­ rate estimate of the number of electrons in this building-but rather that even under ideal conditions the calculus could not be applied. This may be seen from the most cursory attempt to sketch its workings. A unit of pleasure would be required for this calculus, just as a unit of length is required for any calculus of distances. In the first case, as in the second, the unit


must be arbitrary, and we might tentatively choose as our unit the amount of pleasure obtained from eating a bar of chocolate of specified size and composition. Of course, this pleasure varies with one's hunger, so, just as the meter bar is defined relative to conditions of temperature and pressure, the unit of pleasure must be defined relative to more complex conditions of health, pre­ vious taking of nourishment, and exertion since eating last. VVith the necessary re­ strictions of this sort we might, appropri­ ately, name our unit of pleasure the "bent­ ham." So far the difficulties have been merely technical, but, in the application of the standard, insuperable difficulties arise. V\Te might list desires as being greater or less than a bentham, but there seems to be no way of giving meaning to the addition of pleasures. Is a pleasure of one bentham, combined with a pleasure of one bent­ ham, a pleasure of two b8ntharns? And, in any case, what might be meant by a pleasure of two benthams? Thus the scale would not be additive, and quantitative determinations become impossible. Again, there might be people for whom the assimilation of the chocolate bar under the given conditions would provide neither pleasure nor displeasure, only pro­ foundest apathy. Any pleasure measured by this scale would then appear to be in­ finite, and no comparison of values could be obtained. Finally, if, in addition to the bentham, some other unit of pleasure were proposed, say the '\vimpy," defined as the pleasure of eating a hamburger of specified dimensions and consistency under specified conditions, no transformation formula could be worked out to go from one method of measurement to the other. The conversion formulas would vary from person to person and from moment to moment. Of two courses of ac­ tion, one might produce more pleasure measured in one sort of unit and the other measured by the other, so that which course of action were better would depend on which unit \Vere chosen-surely an


undesirable condition and a reductio ad absurdum of any projected calculus of the sort required.2 Both, then, because of the failure to provide any means of reaching individual decisions and because of the theoretical difficulties inherent in any attempt to for­ mulate the hedonistic calculus, the straight­ forward attempt to identify good with any sum of pleasures must be abandoned, and ethics would seem to be in the position of being able neither to avoid the notion of pleasure nor to dispense ,vith it. In perhaps a different sense than Plato intended, the ethicist is in danger of being overcome by pleasure.

II A workable theory of ethics would seem to require some sort of relationship be'hveen pleasures and purposes, and the difficulty in ethics lies in determining the nature of this relationship. Faced vvith this situation, it may be worth while to adopt a proce­ dure similar to that of nineteenth-century physicists who were wont to construct me­ chanical models to serve as explanations in a variety of fields. In this case, of course, a mechanical model of ethical phenomena \Vill not do, but an epistemological one may be of service. Epistemology is by no means a settled field, but it seems to me, at least, to be a field in which solutions have been niore clearly thought out than in eth­ ics, in which more :ingenuity has been ex­ pended, and in which more complex struce.g., suppose A and B are the only people involved in alternative courses of action, I and II. Suppose the pleasures involved to be the follow­ ing; Course I: +3b for A, -2b for B; Course II: -lb for A, +3b for B; Therefore, Course II is preferable. 2

Suppose for A, lb, = 2w; for B, lb = 1/2w. Course I: for A, -%w for B; Course II: -2w for A, lw for B ; Therefore, Course I is preferable.




tures have been proposed than are to be found in ethics. Hence an analogy between theory of knowledge and ethics may be useful, though I do not think it will solve all problems. This is not to say that ethics is deriva­ tive from epistemology or that theory of knowledge is logically prior to ·ethics but merely that a technique successful in the one field may advantageously be applied to the other. And this proposal is nothing new. After all, the method of attack in Kant's second Critique is a conscious imita­ tion of that employed in the first. The suggestion to be advanced here is that the relation of pleasures to ethical values is analogous to the relation of sense data to physical objects. It will . be objected at once that there is no more disputed ground in epistemology than this relation­ ship. This, of course, must be admitted, but it still may be the case that there is enough basic agreement on many topics to provide a fruitful analogy. That this is the case, we proceed to show by an enumer!3,tion of the points of agreement in epistemo]ogy. 1�'e may begin by specifying the sense of the term "sense datum" required. The word is sometimes used with the signif­ icance of an isolated sound or patch of color; with the significance that Russell a t ­ tached to the term "particulars" about thirty years ago.3 It may also be used to in­ dicate a complex of such data having a Gestalt character and interpreted as indi­ cating some object. In this latter sense the term "appearance" might be more apt, ex­ cept for its connotation of being "merely appearance." The latter sense is the one we shall employ. At least six relations and distinctions may be indicated as involving sense data and physical objects. l. The data are given in a sense in \Vhich the object itself is not. Depending on one's epistemological theory, the data may be parts of the object, effects of the object, or e.g., Mysticism and Logic ( London, 1918), chap. vii. 3


signs of the object, but in any case they will be given with a kind of immediacy the object lacks. This is not to rule out the pos­ sibility of a direct, Bergsonian intuition of objects but merely to insist that this intui­ tion, if it exists, is qualitatively different from the immediacy of sense data. 2. Because of this immediacy, the dah1m is subjective in a sense in which the object is not. Communication of the sheerly sensuous character of a datum is impossi­ ble, and there can be discourse only con­ cerning the relations be"b,veen data. An ex­ ample of this is the well-known fact that if green appears to one man as red appears to another, they will agree as to what things are green and will rl.ever discover the dif­ ference. 3. The datum symbolizes the object. It may symbolize in the way in which an effect symbolizes its cause, or in which a part symbolizes the whole, depending on one's theory; but in any case the reference is there. 4. The datum may be modified by the hypotheses concerning the existence of a given object. Thus a person looking for a friend in a crowd is likely to mistake sev­ eral people for him before the friend ac­ tually arrives. These false recognitions i n ­ volve data, modified by the hypotheses. 5. By and large data are of little inter­ est on their own account, but only because they are symbolic of, and related to, ob­ jects. 6. Finally, the data verify hypotheses concerning the mdstence of objects. The exact theory of this verification awaits satis­ factory explanation, but at least the follow­ ing factors would seem to be involved: ( a ) some sort of conception of the object, pos­ sibly vague; ( b ) e,q,ectations of the rela­ tion of this object to other objects; ( c) ex­ pectations of the sort of sense data which would occur as a result of ( b ) ; ( d ) the final occurrence or failure of occurrence of these data. Thus, if I would verify the pres­ ence of a cat in the next room, I must know (a) something at least of what a cat is in



order that ( b ) I may conjecture where the cat is likely to be discovered among the ar­ ticles of furniture of the room. Otherwise I may fail to discover the cat because I sought it on the ceiling rather than on the floor. ( c ) I must form some expectation of what a cat would look like under these conditions, otherwise I may stare at the cat without recognizing it. Finally, ( d ) the re­ quired data may either actually appear or fail to appear. It should be noted also that the data which are predicted of stage c need not be a literal imagination of those ,vhich ac­ tually appear at stage d. As a rule, the ex­ pectation is vague and permits of a variety of actual situations. The verillcation con­ sists not in discovering a qua�tative iden­ tity between expectations and presentations but rather in the fact that the data fit into an anticipated schema or pattern. The verification then consists in a fulnlment or disappointment of these fairly general ex­ pectations. Other relations between data and ob­ jects might be suggested, but they are con­ troversial and would depend upon one's epistemological views. Enough has been suggested, however, to introduce our paral­ lel which would claim analogous relations bet\veen the following sets of entities: In Epistemology

sense data

objects hypotheses con­ cerning existence of objects verification of hypotheses

In Ethics

pleasures and dis­ pleasures values aims or intentions realization of aims

In order to make the intended parallel clear, some explanation is required of the sense of the terms t operate through sticks and stones alone; words play a great part. People praise one another, to encourage certain inclinations, and blame one another, to discourage

ETHICAL NON-COCNITIVISM ot.l:iers. Those of forceful personalities issue comrnands which weaker people, for com­ plicated instinctive reasons, find it diffkult to disobey, quite apart from fears of conse­ quences. Further influence is brought to bear by ,vriters and orators. Thus social influence is exerted, to an enormous extent, by means that have not.¾ing to do with physical force or material reward. The ethi­ cal terms facilitate such influence. Being suited for use in suggest-ion, they are a means by which men's attitudes may be led thfa way or that. The reason, then, that we find a greater similarity in the moral atti­ tudes of one community than in those of different communities is largely this : ethi­ cal judgments propagate themselves. One man says "' This is a nonemotive definition; .and 1 being made to order, as it were, to fit my conception of a personal decision, it is free from the above objec� tion. But since it introduces the pronoun, "l," it does not make clear how two specta­ tors can disagree: when one says "X is good" an2 the other says "X is not good" each is talking about himself and each may be telling the truth. An emotive conception, on the other hand, can easilv avoid L¾is dif­ ficulty, as I have shown ei;ewhere in con­ trasting disagreement in helief about atti­ tudes with disagreement in attitude.' This point \Vould lead us a\v""ay from the per­ sonal to the interpersonal aspects of a:::i eth­ ical problem, hmvever, so I shall keep \Vithin my prescribed limits and say no more about it. The view of John Dewey, who has been so sensitive to the cognitive complex­ ity of ethics, raises a son1ewhat rlifferent question. I am greatly indebted to Dmveyi as this essay readily indicates. And yet I cannot believe that he has been successful in analyzing the ethical terms. He is con­ tent to say that they affect condnct and satisfaction by being predictive. But, since all predictive statements tend to affect con-' duct and satisfaction) and since not all of them, pre-�umably� are ethical, we must ask what sort of predictions am in question, And to this Dev,,ey gives iiO p::-ecise ans,ver. Nor do I see how Dewey could suc­ ceed-apart from introducing emotive meaning in the way I shall presently dis­ cuss-,in repairing his analysis. The cogni­ tive elernents that are relevant to a conflict are no less varied than the attitudes hetween which they mediate. I should sup­ p'ose, moreover, that they are different for different indh1duals; and I should suppose that1 even for a given individual; they would vary with different-problems. N"ow Dewey wants to pack all these elements into the very meaning of an ethical term: � Essays I and II, pp, 1 ff. arn! 26 f. and Ethics and Lm1guage1 chs. l and 8. 1

272 he wants them to be relevant to an ethical judgment by dc'finition. But they are so complicated that he is unable to specify what they are. So he can give only the genus of a defi:.'!ition, without the needed -differentiae, V Let me no,v turn to the more constructive part of this essay. I hope to show that emo­ tive meaning is likely to succeed where cognitive meaning is likely to fail, that it will restore the thoughtful and reflective elements of ethics to their rightful place. The precise definition of "emotive meaning" is itself a complicated matter; but the various details will not, I think, greatly affect the simple point I am ahout to make. So I shall assume that "emotive n1eaning/' whatever else1 refers to a tend­ ency of certain words to express or evoke attitudes; and I sha11 assurne that :it is one thing to express or evoke attitudes and another thing to designate them. That is to say, the interjection, "alas," which expresses or evokes sorrow, functions rather differ­ ently from the noun "sorrow" itself, which designates sorrO\v. It will be unnecessary for me to show1 I trust, that the ethical terrns have an emo­ tive mean:ng....._so long1 that is, as I do not insist that it is their only sort of meaning. The controversy has been concerned not with this point ·b1.0t rather with the impor­ tance of their emotive meaning. It is to be mentioned only to be put to one side so that it \vill not distract us from what is reallv essential: or is it itself an essentiai faet�r? \Vhen we limit attention to prob!ems of the sort I have been emphasizing- ­ evaluative decisions that a man makes in private rather than in discussions with other people-the emotive meaning of the ethical te:ms may at .first seem trivial. It

ETHICAL NON-COGNITIVISM may remind us merely that ethical deci­ sions are sometimes attended bv self-ex­ hortation. Although sel£-exhortation is in­ teresting enough, it is scarcely a matter to be dwelt c:pon. T.:len; is another respect1 hovvever, in wI'ich attention to emotive mean:llg is more rewarding. It helps us, in cases where a man is making a decision1 to see how his language reflects his problem-how it reflects his effort to make his attitudes speak '\vith one voice. It does so in this simple way: Suppose that the man first withholds such terms as «good" and "bad"; that he next uses them some..-vhat tentatively, or else alternates between the one tenn and the other; and that finally he uses one of them onlv. and with conviction. If we take his ethic;i terms as emotive, and hence as expressing his attitudes, we can easily explain the fact that they are verbal clues to the nature of his prob'.em; for at first he has no unimpeded attitude to express, being .in a state of conilict; and, as his atti­ tudes speak more and more ,,vith one voice, he expresses them 111ore anC more freely. Let me here emphasize a point that I feel to be of central imnortance. If we take the �an's ethical term; as expressing hls attitudes, we can become sensitive to the nature of his problem without difficulty, But if we take them as mere'.y designating his attitudes1 we are likely to miss the very aspect of his problem that makes it an eval­ uative one. For suppose we were to insist that his ethical judgment was no more than atti­ tude-designat'ng, like the statement, "Care­ ful introspection assures me that I approve of this." That would immediately suggest to us that the man's nroblem was one of describing his own stale of mind and hence a problem in psychology. \Vh.ereas we have seen that it is something else. The man. is trying to resolve a conflict, and the process of resolving it is much more complicated tllan the introspective process of describing

THE EMOTIVE CONCEPTION OF ETHICS it, In other words the attitude-designating tenns would be t\vice removed from his problem; they would formulate beliefs that were about it. And by emphasizing these beliefs, instead of the many others that he is really concerned \Vith, they would sug­ gest that he is simply looking at his con­ flict. But in fact he is living through it .and a!l the activities that attend its resolution, the task of looking at it being compara­ tivelv inessential. To restore the correct emphasis, then, we must take the ethical terms not as atti­ tude-designating but as attitude-expressing and hence as emotive. For in the latter capacity the tenns are only once removed from the man's attitudes; thev are related to his attitudes by a direct route and not by the indirect route of expressing beliefs about them. By causing us to look to the attitudes themselves, rather than to beliefs that do no more than describe them, emo­ tive meaning frees us from the ter.dency of supposing that an evaluative decision is somehow a::J exercise in introspective psy­ chology. It reminds us that the man's efforts throughout his decision are to cllange his very attitudes. He must actually make this change and not merely describe it as a self-conscious spectator1 as if all the work were being done for him by some­ body else. Thus emotive meaning, once it is taken into account, mai..-presses a pose that \Ve are straining ordinary lan­ special sort of approval, moral approval, so guage when we do so. But how is Steven­ "valid" might have a special sort of emotive son to explain this application of appar­ meaning because, for example, it expresses ently logical categories when he maintains only epistemic approval. Although Steven­ that in most ethical arguments the p::·e::n­ son nowhere develops such an analysis, it is isses are related to the conclusion psycho­ suggested by the way in which he marks logically r::,ther than logically? In many off «interest in knowledge" from the many passages he sea'TIS to say that the relevance other kinds of interests which are involved or irrelevance of some consideration, and in ethieal discussions ( EL, pp. 284-286). � C. ·wellinan, The Language of Ethics ( Cam­ The other direction in which Stevenson might move is to incorporate some descripbridge, Mass., 1961 ).

284 tive meaning into judgments of validity.

This would be in line with his version of

emotivism and is even suggested by his dis­

cussion of the reasons \vhy rational meth­

ods are usually to be preferred to irrational ones ( EL, pp. 156---157 ) . Thus he might say that "valid" means «in accordance with method of arguing M," where M is spelled out by specifying such things as avoiding logical inconsistency, not falsifying the facts of the case, using all available infor­ mation. Just what descriptive meaning a

speaker would build into his use of the term "valid" would, presumably, depend upon which methods he happens to ap­ prove. If Stevenson were to follow out the two directions I have indicated, the result would be something like this: "valid" means "in accordance vvith method of ar­ guing M" and m,.-presses a positive epis­ temic attitude. \%at is one to say of such an emotive theory of logical terms? ( 1 ) Probably it could be worked out in detail and, if it were, would have much to recommend it.

( 2 ) Some such theory is essential to Ste­

venson�s position unless he is prepared ei­

ther to deny that we in fact apply logical terms to ethical arguments or abandon his view that the conclusions of these argu­ ments lack objectivity. Therefore, until he works out his analysis of logical terms more fully, his theory is essentially incomplete. ( 3 ) To interpret logical terms like "valid" and "invalid" emotively would seem to in­ fect logic with the same subjectivity that emotivism insinuates into ethics. "'\1/ould

Stevenson be willing to accept this conse­ quence? Everything he says about logic, and science too, indicates that he regards these as very different from ethics in their

genuine claim to objectivity. ( 4 ) Would

S tevenson be willing to extend an emotive analysis to other epistemic terms like "true," "reasonable,» or «correct"? If so, what becomes of objective truth in factual matters? If not; what are his criteria for distinguishing the tvm species of epistemic terms, the emotive and the non-emotive?


( 5 ) We do not argue for or against judg­ ments of validity in at all the way this analysis would imply. On this theory the valid argument is the one in accordance with the approved method, and when Ste­ venson gives examples of reasons for ap­

proving a method he always gives pragmat­

ic considerations, such as, that using the

method will not lead to disastrous, conse­ quences, that its use will create social har­

mony, etc. It does not seem to me that

such pragmatic matters enter into our

judgments of validity at all.

But this criticism leads to another basic problem in Stevenson's position. Do rational methods of argument have any

privileged position in ethical discussion? It would ahnost seem that Stevenson would be forced to answer this question in

the negative because all sorts of nonration­ al methods of persuasion were possible in

ethics, many of these are rhetorically ef­

fective, and in most cases logical criticism

of such methods is out of place. Moreover,

he points out that in certain cases, as when

dealing with the feeble-minded or emotion­

ally upset or vvhen there is no time for

lengthy discussion, nonrational methods of persuasion are fully justified. Nevertheless, Stevenson insists that rational methods do have priority in ethical discussion because under normal circumstances these are the

best methods. He supports this normative judgment with many reasons. If people's attitudes are formed in ignorance of the

facts, their action is likely to be disorga­ nized and blundering; and there is a con­ siderable value in building up the habit of inquiry in most men (EL, pp. 156---157 ) . Rational methods give a more per­ manent agreement and a more stable per­ sonal conviction than "rhapsody or exhor­

tation" ( FV, pp. 7--8 ). If one person dis­

covers that another person has given him

a one-sided presentation of the facts, he will resent this treatment ( FY, p. 195 ). For the individual, it is better to consider both

sides of the case both to avoid disastrous consequences \Vhen the resulting attitude

EJylOTIVISM AND ETHICAL OBJECTIVITY is put into practice and to gain tl1e intrinsic value paying attention to all the facts ( FV1 p. 196 ). vVhat disturbs me about this case for rational methods is that it is framed in purely pragmatic terms. One chooses ra­ tional methods because they are better, and they can be seen to be better by consider­ ing the value consequences of using or not using them. It is my contention that the privileged position of rational methods in e'!:hical discussion fa not shnnlv ' • a matter of their greater utility. Any method is a method of doing some­ thing, and this relational aspect must not be overlooked in choosing a method. Shouting at a man may be a very good method of waking him up but a very poor method of putting him to sleep. Accord­ ingly, we must ask what we are trying to do when we discuss ethical issues. If we are simply trying to persuade someone, then we do indeed have a choice between rational and no:ll'ational methods, Looking at et.'iical arguing as a purely psychological matter of creating and reinforcing atti­ tudes, Stevenson is probably right that the choice of methods is the pragmatic one to be determined by relative rhetorical effec­ tiveness and various value consequences. But suppose that ethical arguing is thought of as establishing the truth of an ethical conclusion. Then the entire picture changes. One cannot ask svhethe::: rational or nonratio:ial methods are the best m.eans of proving a conclusio2 true, for nonra­ tional methods do not really prove or dis­ prove anything, L'1e man v,rho chooses ra­ tional over nonrationa1 methods of ethical arguing is not simply choosing a hetter method to achieve the same end ( persua­ sion ) but the only possible method to achieve a very, dilferent end ( knowledge of the truth) . VVhy is this so? It is because to make a statement is to make an implicit claim to truth, a claim that the statement is supported by the ,veight of the evidence. And the only possible test of this claim is reasoning. Thus rational methods of ethical


arguing have a privileged pos.ition ,vhich is denied by Stevenson's emotive theory of ethics. 11y contention rests, to be sure, upon my viev.r that there is a claim to rationality built into ethical statements. Is there any real evidence that this is so? \:Yell, one bit of evidence is the existence of ethical dis­ agreement. Just because Stevenson de­ nudes ethical statements of their claim to rationality he is unable to explain ethical disag:::eement. His suggestion is that two speakers disagree ethically when ( a ) their utterances exp:ess opposed attitudes to the same thing, and ( b ) at least one speakes has a motive to alter the attitude of the other (EL, p. 3). VVhy is this second condition neces­ sary? Probably because Stevenson recog­ nizes that there are differences whieh do not involve any disagreement. Hence he says that where two people have opposed attitudes but neither has a sufficient motive to ehange the attitude of the other, they merely d:iffer iu attitude ( EL, pp. 4-5 ) . But this does not seem to b e true. It often happens that two people argue strenuously about some ethical issue onlv to find that neither can convince the ;ther. At this point both may give up and "agree to dis­ agree." Alfaough neither speaker any long­ er has a sufficient motive to try to change the attitude of the other, it does not seem to me that they have ceased to disagree. Or suppose that two speakers e"'I'ress opposed attitudes to something, one saying •tperts. But this i,; not so for two reasons. ( a ) U:1 some cases there are tests of competence which are purely objective and empirical. Some men, for example, have perfect pitch, can detect minute musical intervals, can recognize and accurately reproduce long and complicated tunes and so on, while others cannot; and these are matters of fan!:, In judging their expertise v,re m�t, of course, rely on the ability of other experts to assess their competence; but the judge­ ments of these experts is 'objective' be­ cause they fulfil the requirements for ob­ jective language discussed in chapter 4. It is possible that one man might have a finer ear than all other men) so that in a case in which he said that two notes were slightly different when everyone else said they were the same he would be right and they wrong. But if there were no indfrect tests, such as the appeal to readings of scales and meters, for deciding ,vhether he can really detect these differences or is only bluffing, and if those who horn,stly claimed to be ahle to make fine discri..TUinations did not on the ,vhole agree with each other, ,ve could not call their judgements 'objec­ tive'. Now from the fact that a man is able to make these fine discriminations or tO perforn1 better than others in these objec· tively testable ways, it does not, of course, follow that he is a good judge. For to say that he is a good judge is either to state that he is good at appl;1ng :he accepted


criteria for what is good ( which is different from being good at passing the objective tests ) or to express approval of his judge­ ments, to praise him, to encourage others to accept his judgements, and so on; and in most cases it is to do all these things at once. But, once again, the reason why ,ve allow that, in a general ,vay, the most tech­ nically competent people are the best judges lies in the facts. A man who is tone-deaf is unlikely to be able even to dis­ tinguish one piece of music from another and his value-judgements ( if he makes them ) , are not likely to be consistent with each other; so that his value-judgements would be useless as a guide to others. A man who knows little Greek could not be a good judge of a piece of Greek prose. Con­ sistency and fine discrimination are not sufficient conditions of good taste or moral insight, but they are necessary conditions if criteria are to be used for the purposes for which they are used, ( b ) Secondly, the person who rejects the criteria usually employed or the verdict of the acknowledged experts may do so in two ways. He may simply refuse to be guided by them on the grounds that he happens not to like what is usually called good. But, if he goes further and says that the usual criteria are not good criteria, he is not just rejecting them; he is himself using criterion-applying language and he implies that he has second-order criteria for judging ( and condemning) the usual fust-order criteria. To the questioning of criteria there is no end; but if ,ve ask whether the criteria for judging Xs are good criteria we must, at ,vhatever level we have reached, use criteria for deciding whether they are good or not, It is logically absurd to ask a ques­ tion v?ithout laio\ving how the answers to it are to be judged to be good or bad answers. The appeal to criteria accepted by experts is not circular, but regressive; and the regress is not a vicious one since, al­ though we can ahvays question the criteria,


there is no practical or logical necessity to do so. The self-guaranteeing criteria so vainly sought by some moralists are neither possible nor necessary.

V Non-Practical Appraisals

We often make appraisals in contexts where there is clearly no question of choos­ ing or advising, for example moral judge­ ments about historical or fictional charac­ ters. And this seems to involve a difficulty for theories which make appraisals logical­ ly dependent on pro-attitudes. Hutcheson and Hume, for example, tried to reduce moral judgements to expressions of feel­ ing. They were not guilty of the Natural­ istic Fallacy, since they were prepared to allow that moral approval and sympathy are special, moral feelings distinct from other types of feeling. But even this con­ cession to the peculiarity of the moral use of language does not save them from an important objection that seems at first sight fatal to their case. Sentiments, as Hume noticed, seem to vary in rough pro­ portion to the propinquity of their objects. We are not moved by the iniquity of re­ mote historical characters as we are by those closer to us; and we feel more ap­ proval for and sympathy ,vith those near to us than with those who are more remote. Yet our moral judgements do not vary in the same way. '"'\iVe read Cicero now v:.rith­ out emotion, yet we can still judge Verres to be a villain. According to Hume's theory our judgement must change as do our feel­ ings. I do not feel indignation as· strongly now about the German invasion of Czecho­ slovakia as I did at the time it happened; yet I do not judge the action to be less wrong than I did then, or the agents less criminal. . . . It is but a weak subterfuge to say we transport ourselves by the force of


imagination into distant ages and countries, and consider the passions which we should have felt on contemplating these characters had we been contemporaries and had com­ merce with the persons. . . . I now feel completely indifferent to Verres, and know it. Yet, Hume tells-me, when I judge Verres to have been a villain, I am so deceived by my imagination that I talk as if I felt a strong feeling of anger."" Dr. Raphael's criticism is fatal to the theory that a man who makes a moral ap­ praisal is always expressing a feeling; and a similar criticism could also be made of any theory which says that to appraise is always to praise, advise, commend, etc. On some occasions a man may be simply ap­ plying the criteria that he and others cus­ tomarily use for these purposes. To call Verres a villain is to pass a verdict on him, to condemn him. Now the Moral Sense School were, I think, mistaken in con­ struing moral approval and disapproval as feelings, since this suggests too strongly the analogy with itches, aches, and tickles. But they were right to connect moral appraisals and verdicts with approval and disap­

proval. For although a man ,vho passes a

verdict need not be expressing a pro- or

con-attitude, we have seen that the criteria he uses are directly or indirectly linked with these attitudes; and in the case of moral judgements they must be linked in a special way that may be absent in other cases. I said earlier that, although in other cases 'good' might lose its gerundive force, it cannot wholly do so when used to make moral appraisals. The reason is that, v,rhat­ ever may be the case with other types of appraisal, moral appraisals must be univer­ sal. Anyone ,vho makes a moral appraisal even of a remote character must be ,villing to apply the same criteria universally. And it follows from this that he must be willing to apply them in practical conte,d:s. If I am j,


D. D. Raphael: The 1"\1oral Sense, pp. 88 and


not prepared to condemn anyone whose behaviour is like that of Verres in all rele­ vant respects, then, in calling Verres a vil­ lain, I am not making a genuine moral judgement; and the relevant respects are all of an empirical, objective kind. It would, of course, be trivial to include among them an objective property of vil­ lainy or moral turpitude; all that is neces­ sary is that I should be prepared to con­ demn anyone who did the sort of thing that Verres is called a villain for having done, anyone who oppressed the poor, robbed the rich, took bribes, and cheated the treas­ ury, and all for his own personal profit. J\ Ioral appraisals are therefore con­ nected with choosing and advising in a way that non-moral appraisals need not be. It is not logically odd to say "This is the better wine, but I prefer that"; but it is logically odd to say "This is the ( morally ) better course; but I shall do that."' And a man cannot be making a genuine moral judgement about Verres if he would him­ self be prepared to act on the same princi­ ples on which Verres acted and prepared to exhort others to do so. In condemning Verres he is not expressing any emotion; but he is affirming his own moral princi­ ples. 1

VI 0bjective-Subjective In chapter 6 I said that the distinction be­ tween "For what job is the word '. . .' used?" and "Under what conditions is it proper to use that word for that job?" throws light on the objective-subjective dispute. As we should expect, both parties are n This may sound surprising. V{e all know what it is to take what we know to be the morally worse course. I shall try to remove the air of para­ dox in chapter 18.


right. Just as the subjectivists are right in denying that A-words stand for special pro­ perties and explaining them in terms of people's reactions, so they are also right in connecting 'good' and 'bad' with people's desires, tastes,. interests, approvals, and dis­ approvals. There is a logical absurdity about calling a play 'amusing' if the speaker believes that it never has amused anyone and never will; and there is the same logical absurdity in calling something 'good' without any direct or indirect refer­ ence to a pro-attitude. If the connexion be­ tween 'good' and the pro-attitude that is contextually relevant were not a logical one, a gap would emerge between calling something good on the one hand and de­ ciding to choose it, choosing it or advising others to choose it on the other which ,vould make these activiti"es unintelligible. 1oreover, the subjectivists are also right in connecting 'good' with the pro-attitudes of the speaker, at least in moral cases. But the objectivists are also right. They are mistaken in denying the points made by the subjectivists above and in thinking that goodness must be a unique, non-natural property. It is sometimes argued that if there were no such property we could not account for the fact that we use the impersonal form 'this is good' rather than the personal form 'I approve of this', and those who use this argument are inclined to forget that we have an imper­ sonal form 'this is nice' as ,veil as the per­ sonal form 'I like if, so that niceness would have to be an objective property too. It would indeed be puzzling to under­ stand why we use these impersonal forms if ,ve were just talking about or expressing our own approvals; but this argument does not shmv that ,ve are talking about some­ thing else, still less that this must be a unique property. We can account for the objective formula, as we did in the case of 'nice', by saying ( a ) that 'X is good' is not only used in the context of choice and ( b ) that, ,vhen it is so used, it implies a great 1 ;\.


deal that is not implied by 'I approve of X' and is expressly denied by 'I happen to ap­ prove of X.' It implies that my approval is not an unusual one and that I could give reasons for it. It implies also-what is a matter of objective fact-that the object conforms to certain standards which are generally accepted. It is sometimes argued that 'this is good' cannot just mean 'I approve of this' on the ground that we can say "I approve of this because it is good". Approval must therefore be an intellectual emotion which arises in us only when we recognize some­ thing to have the objective property 'good­ ness'. But it has never been clear what the connexion between the approval and the recognition of the property is supposed to be. Is it logically necessary that anyone who recognizes the property should feel approval or is it just an empirical fact that people who notice the property, and only they, have the feeling? Each of these answers involves insuperable difficulties; but if neither is correct we must find some other way of explaining the 'because' in 'I approve of X because it is good'. The need for such an explanation van­ ishes when we see that this is not a reason-giving 'because' like that in 'I ap­ prove of Jones because he is kind to chil­ dren' but more like 'I like Jones because he is likeable'. It rebuts the suggestion that I just 'happen to' approve of X and it implies that X has certain properties which make it worthy of my approval and that it con­ forms to the known standards for Xs. The objectivist is right in drawing at­ tention to the factual background which makes impersonal appraisals possible; but the facts ,vhich it contains are ordinary, empirical facts, not special, non-natural facts. Unlike the subjectivist ( who tends to ignore the background altogether), he tries to include the background in the meaning of the word; and this, combined with the mistake of confusing practical and descrip­ tive discourse, leads him into the vain pur-


suit of a single ingredient to which we al­ ways refer when we call something good.

VII The Naturalistic Fallacy

V\Te are now in a position to see why the moral philosophers of the past subordin­ ated the critical or appraising uses of moral language to the practical uses. Each pre­ supposes the other, but in a different way. The practical uses presuppose the apprais­ ing use in that we could not use 'good' as we do for choosing, advising, and praising if we did not employ criteria or standards; since we only use 'good' for these purposes when vve are employing standards. Never­ theless people who did not know what standards were could do things recogniz­ ably like what we, who have standards, call choosing, advising, and praising. They would be very rudimentary performances, hardly deserving the names of choice, ad­ vice, and praise; but they could occur. vVe dra,v a distinction between 'good' and 'hap­ pen to like' which people without standards could not draw; and we, who have the dis­ tinction, would describe their activities in terms of what they 'happen to like', because they could not do anything that we would call 'choosing the best'. In this way the prac­ tical uses of 'good' imply the appraising use. But the practical uses are logically prior to the appraising use in a much more fundamental way. Unless men had pro­ attitudes, there could not be even rudimen­ tary analogues of what we ln10w as ap­ praising, judging, or passing a verdict. For these involve the use of standards; and without pro-attitudes we should neither have any use for standards nor even be able to understand what a 'standard' was. \Ve can imagine a world in which there was choosing, but no appraising and also a world in which there was classifying, sort­ ing, and ordering ( for example by size )

MIXED VIEWS but no choosing; but, in a world in which

there was no choosing, there could be no such thing as appraising or grading. Ethical Naturalism is the attempt to trace logical connexions behveen moral ap,· praisals and the actual pro- and con­ attitudes of men, their desires and aver­ sions, hopes and fears, joys and sorrows. One-track naturalistic theories always fail to do justice to the complexity both of the facts and of the logical connexions, since they suggest that there is only one thing to­ wards which men have a pro-attitude, pleasure, or that all pro-attitudes are de­ sires. And these theories are both psycho­ logically and logically misleading. Opponents of the Naturalistic Fallacy have pointed out the logical errors. It is true that gerundive and deontological words cannot be defined in terms of pleas­ ure, desire, or even purpose; and I shall try to show how they are connected with these teleological concepts later. It is also true that gerundive judgements and value judgements do not follow logically from descriptive statements about what men like, enjoy, and approve of. But the reason for this is not that gerundive words and value words refer to special entities or qualities, but that a person who uses them is not, except in certain secondary cases, describing anything at all. He is not doing what psychologists do, which is to describe, explain, and comment on what people likei enjoy, and approve of; and he. is not doing what moral philosophers do, which is to describe, eAJ)lain, and comment on the way in which people use moral words; he is himself using moral language, expressing approval, praising, advising, exhorting, commending, or appraising. The attack on the Naturalistic Fallacy is thus far justified. But the conclusion which is commonly drawn, that moral con­ cepts are a special sort of concept which must be purged of all association with the 'merely empirical or phenomenal' concepts of enjoying, wanting, and approval is not justified. Psychology is not as irrelevant to



ethics as some modern philosophers insist; that, human beings being what they are, for, although moral judgements do not fol­ there are certain types of activity that are low from psychological statements, we can­ in fact satisfactory to them and that it is not understand what the terms used in possible empirically to discover what these moral judgements mean unless we examine are. No doubt they often made mistakes of them in the context of their use; and they fact, for example that of supposing that are used either directly to express a pro- or what is satisfactory to one man would be con-attitude or to perform some other task satisfactory to another; and they made mis­ which beings who had no pro- or con­ takes of logic, for example that of suppos­ attitudes could not perform or even under­ ing that 'good' could be extracted from its stand. The various ways in which 'good' is context and be said to mean the same as used are unintelligible unless they are di­ 'satisfactory'. But they do not seem to have rectly or indirectly connected with choice; been mistaken in their basic assumptions and I shall try to show later that the same that the language of obligation is intelligi­ applies to 'ought'. ble only in connexion with the language of Moral philosophy does not, therefore, purpOse and choice, that men choose to do "rest on a mistake". For the great philoso­ what they do because they are what they phers were not primarily interested in the are, and that moral theories which attempt question whether deontological words to exclude all consideration of human na­ could be analysed in terms of 'merely em­ ture as it is do not even begin to be moral pirical' or 'natural' concepts. They believed theories.

What Is a Value Judgement ? R. M. HARE MEANING AND CRITERIA It is a characteristic of 'good' that it can be applied to any number of different classes of objects. We have good cricket­ bats, good chronometers, good flre­ extinguishers, good pictures, good sunsets, good men. The same is true of the word 'red'; all the objects I have just listed might be red. \Ve have to ask first whether, in ex­ plaining the meaning of the word 'good', it would be possible to explain its meaning in all of these eA-pressions at once, or whether it would be necessary to explain 'good cricket- bat' :first, and then go on to explain ·good chronometer' in the second lesson, 'good fire-extinguisher' in the third, and so on; and if the latter, whether in each lesson we should be teaching something entirely new-like teaching the meaning of 'fast dye' after we had in a previous lesson taught the meaning of 'fast motor-car'- or whether it would be just the same lesson over again, with a different example- like teaching 'red dye' after we had taught 'red motor-car'. Or there might be some third possibility. The view that 'good chronometer' would be a completely new lesson, even though the day before we had taught 'good cricket-bat', runs at once into difficulties. For it would mean that at any one. time our Reprinted from The Language of Aforals by R. i,r. Hare, 1952, by permission of the author and the Clarendon Press, Oxford. Pp, 95--97, 111- 119, 121-133, 148-150.


learner could only use the word 'good' in speaking of classes of objects which he had learnt so far. He would never be able to go straight up to a new class of objects and use the word 'good' of one of them. When he had learnt 'good cricket-bat' and 'good chronometer', he would not be able to manage 'good £re-extinguisher'; and when he had learnt the latter, he would still be unable to manage 'good motor-car'. But in fact one of the most noticeable things about the way we use 'good' is that we are able to use it for entirely new classes of ob­ jects that we have never called 'good' be­ fore. Suppose that someone starts collecting cacti for the first time and puts one on his mantel-piece- the only cactus in the coun­ try. Suppose then that a friend sees it, and says ·r must have one of those'; so he sends for one from wherever they grow, and puts it on his mantel-piece, and when his friend comes in, he says Tve got a better cactus than yours'. But how does he know how to apply the word in this way? He has never learnt to apply 'good' to cacti; he does not even know any criteria for telling a good cactus from a bad one ( for as yet there are none ) ; but he has learnt to use the word 'good', and having learnt that, he can apply it to any class of objects that he requires to place in order of merit. He and his friend may dispute about the criteria of good cacti; they may attempt to set up rival criteria; but they could not even do this unless they were from the start under no difficulty in using the word 'good'. Since,


therefore, it is possible to use the word 'good' for a new class of objects without further instruction, learning the use of the word for one class of objects cannot be a different lesson from learning it for another class of objects-though learning the crite­ ria of goodness in a new class of objects may be a new lesson each time. . . .


There are two sorts of things that we can say about strawberries; the first sort is usually called descriptive, the second sort evaluative. Examples of the first sort of re­ mark are, 'This strawberry is sweet' and 'This strawberry is large, red, and juicy'. Examples of the second sort of remark are 'This is a good strawberry' and 'This straw­ berry is just as strawberries ought to be'. The first sort of remark is often given as a reason for making the second sort of re­ mark; but the first sort does not by itself entail the second sort, nor vice versa. Yet there seems to be some close logical con­ nexion between them. Our problem is: 'What is this connexion?'; for no light is shed by saying that there is a connexion, unless we can say what it is. The problem may also be put in this way: if we knew all the descriptive proper­ ties which a particular strawberry had ( knmv, of every descriptive sentence relat­ ing to the strawberry, whether it was true or fals e ) , and if we knew also the meaning of the word 'good', then what else should we require to know, in order to be able to tell whether a stra\vberry was a good one? Once the question is put in this way, the answer should be apparent. We should re­ quire to know, what are the criteria in virtue of which a strawberry is to be called a good one, or what are the characteristics that make a strav,,berry a good one, or what is the standard of goodness in straw­ berries. We should require to be given the major premiss. VVe have already seen that


we can know the meaning of 'good straw­ berry' without knowing any of these latter things- though there is also a sense of the sentence 'vVhat does it mean to call a strawberry a good one?' in which we should not know the answer to it, unless we also knew the answer to these other questions. It is now time to elucidate and distinguish these two ways in \Vhich we can be said to knoW what it means to call an object a good member of its class. This will help us to see more clearly both the differences and the similarities between 'good' and words like 'red' and 'sweet'. Since we have been dwelling for some time on the differences, it will do no harm now to mention some of the similarities. For this purpose, let us consider the hvo sentences 'M is a red motor-car' and '}yf is a good motor-car' . . . The first similarity between 'M is a red motor-car' and '}.1 is a good motor-car' is that both can be, and often are, used for conveying information of a purely factual or descriptive character. If I say to some­ one 'J\1 is a good motor-ca/, and he himself has not seen, and knows nothing of M, but does on the other hand know what sorts of motor-car we are accustomed to call 'good' ( knows what is the accepted standard of goodness in motor- cars ) , he undoubtedly receives information from my remark about what sort of motor-car it is. He -will com­ plain that I have misled him, if he subse­ quently discovers that M will not go over 30 m.p.h., or uses as much oil as petrol, or is covered with rust, or has large holes in the roof. His reason for complaining will be the same as it would have been if I had said that the car was red and he subsequently discovered that it was black. I should have led him to expect the motor-car to be of a certain description when in fact it was of a quite different description. The second similarity between the two sentences is this. Sometimes we use them, not for actually conveying information, but for putting our hearer into a position subse­ quently to use the word 'good' or 'red' for


giving or getting information. Suppose, for example, that he is utterly unfamiliar with motor- cars in the same sort of way as most of us are unfamiliar ,vith horses nowadays, and knows no more about motor-cars than is necessary in order to distinguish a mo­ tor-car from a hansom cab. In that case, my saying to him 'M is a good motor-car' will not give him any information about M, be­ yond the information that it is a motor-car. But if he is able then or subsequently to examine M, be will have learnt something. He will have learnt that some of the char­ acteristics which M has, are characteristics which make people-or at any rate me-call it a good motor-car. This may not be to learn very much. But suppose that I make judgements of this sort about a great many motor-cars, calling some good and some not good, and he is able to examine all or most of the motor-cars about which I am speaking; he will in the end learn quite a lot, always presuming that I observe a con­ sistent standard in calling them good or not good. He will eventually, if he pays careful attention, get into the position in which he knows, after I have said that a motor-car is a good one, what sort of a motor- car he may expect it to be- for example fast, sta­ ble on the road, and so on. Now if we were dealing, not with 'good', but with 'red', we should call this process 'explaining the meaning of the vvord'- and we might indeed, in a sense, say that what I have been doing is explain­ ing what one means by ca good motor-car'. This is a sense of 'mean' about which, as we have seen, ,ve must be on our guard. The processes, . however, are very similar. I might explain the meaning of 'red' by con­ tinually saying of various motor-cars 'M is a red motor-car', 'N is not a red motor-car', and so on. If he were attentive enough, he would soon get into a position in which he was able to use the word 'red' for giving or getting information, at any rate about mo­ tor-cars. And so, both with 'good' and \Vith 'red', there is this process, which in the case


of 'red' we may call 'explaining the mean­ ing', but in the case of 'good' may only call it so loosely and in a secondary sense; to be clear we must call it something like 'ex­ plaining or conveying or setting forth the standard of goodness in motor-cars'. The standard of goodness, like the meaning of 'red', is normally something which is public and commonly accepted. When I explain to someone the meaning of 'red motor-car', he expects, unless I am lmown to be very eccentric, that he will find other people using it in the same way. And similarly, at any rate with objects like motor-cars where there is a commonly ac­ cepted standard, he will expect, having learnt from me what is the standard of goodness in motor-cars, to be able, by using the expression 'good motor-car', to give information to other people, and get it from them, without confusion. A third respect in which 'good motor-car' resembles 'red motor-car' is the following: both 'good' and 'red' can vary as regards the exactitude or vagueness of the information which they do or can convey. VVe normally use the ex­ pression 'red motor-car' very loosely. Any motor-car that lies somewhere between the unmistakably purple and the unmistakably orange could without abuse of language be called a red motor-car. And similarly, the standard for calling motor-cars good is commonly very loose. There are certain characteristics, such as inability to exceed 30 m.p.h., which to anyone but an eccentric would be sufficient conditions for refusing to call it a good motor-car; but there is no precise set of accepted criteria such that we can say 'If a motor-car satisfies these condi­ tions, it is a good one; if not, not'. And in both cases we could be precise if we wanted to. \Ve could, for certain purposes, agree not to say that a motor-car was 'really red' unless the redness of its paint reached a certain measurable degree of purity and saturation; and similarly, we might adopt a very exact standard of good-




ness in motor-cars. We might refuse the name 'good motor-car' to any car that would not go round. a certain race-track without mishap in a certain limited time, that did not conform to certain other rigid specifications as regards accommodation, &c. This sort of thing has not been done for the expression 'good motor-car'; but, as Mr. Urmson has pointed out, it has been done by the Ministry of Agriculture for the eA'Pression 'super apple'. 1 It is important to notice that the exact­ ness or looseness of their criteria does abso­ lutely nothing to distinguish words like 'good' from words like 'red'. ·words in both classes may be descriptively loose or exact, according to how rigidly the criteria have been laid do\vn by custom or convention. It certainly is not true that value-words are distinguished from descriptive words in that the former are looser, descriptively, than the latter. There are loose and rigid examples of both sorts of word. iVords like 'red' can be extremely loose, without be­ coming to the least degree evaluative; and expressions like 'good sewage effiuent' can be the subject of very rigid criteria, with­ out in the least ceasing to be evaluative. It is important to notice also, how easy it is, in view of these resemblances be­ tween 'good' and 'red', to think that there are no differences-to think that to set forth the standard of goodness in motor-cars is to set forth the meaning, in all senses that there are of that word, of the expression 'good motor-car'; to think that 'M is a good motor-car' means neither more nor less than 'M has certain characteristics of which "good'� is the name'. It is worth noticing here that the func­ tions of the word 'good' which are con­ cerned ·with information could be per­ formed equally well if 'good' had no com­ mendatory function at all. This can be made clear by substituting another word, made up for the purpose, which is to be 1

Mind, lh: ( 1950), 152· ( also in Logic and Language, ii, ed. Flew, 166 ) .


supposed to lack the commendatory force of 'good'. Let us use 'doog' as this new word. 'Doog', like 'good', can be used for conveying information only if the criteria for its application are known; but this makes it, unlike 'good', altogether meaning­ less until these criteria are made known. I make the criteria known by pointing out various motor-cars, and saying 'Nf is a doog motor-car', 'N is not a doog motor-car', and so on. vVe must imagine that, although 'doog' has no commendatory force, the criteria for doogness in motor-c�s which I am employing are the same as those which, in the previous example, I employed for goodness in motor-cars. And so, as in the previous example, the learner, if he is su­ fficiently attentive, becomes able to use the word 'doog' for giving or getting informa­ tion; when I say to him ·z is a doog mo­ tor-car', he knows what characteristics to expect it to have; and if he wants to con­ vey to someone else that a motor-car Y has those same characteristics, he can do so by saying 'Y is a doog motor-car'. Thus the word 'doog' does ( though only in connexion with motor- cars ) half the jobs that the word 'good' does-namely, all those jobs that are concerned ,vith the giv­ ing, or learning to give or get, information. It does not do those jobs which are con­ cerned with commendation, Thus we might say that 'doog' functions just like a descrip­ tive word. First my learner learns to use it by my giving him examples of its applica­ tion, and then he uses it by applying it to fresh examples. It would be quite natural to say that what I was doing was teaching my learner the meaning of 'doog·; and this shows us again how natural it is to say that, when we are learning a similar lesson for the expression 'good motor-car' ( i.e. learning the criteria of its application ), we are learning its meaning. But with the word 'good' it is misleading to say this; for the meaning of 'good motor-car' ( in an­ other sense of 'meaning' ) is something that might be known by someone who did not


know the criteria of its application; he would know, if someone said that a mo­ tor-car was a good one, that he· was com­ mending it; and to know that, would be to know the meaning of the expression. Fur­ ther, as we saw earlier (6.4), someone might know about 'good' all the things which my learner learnt about the word 'doog' ( namely, how to apply the word to the right objects, and use it for giving and getting information ) and yet be said not to know its meaning; for he might not know that to call a motor-car good \Vas to com­ mend it. It may be objected by some readers that to call the descriptive or informative job of 'good' its meaning in any sense is il­ legitimate. Such objectors might hold that the meaning of 'good' is adequately charac­ terized by saying that it is used for com­ mending, and that any information ,:ve get from its use is not a question of meaning at all. V\lhen I say 'M is a good motor-car', my meaning, on this view1 is to commend NI; if a hearer gets from my remark, together with his knowledge of the standard habitu­ ally used by me in assessing the merits of motor- cars, information about what descrip­ tion of motor-car it is, this is not part of my meaning; all my hearer has done is to make an inductive inference from 'Hare has usually in the past commended motor-cars of a certain description' and 'Hare has commended M' to 'M is of the same description'. I suspect that this objec­ tion is largely a verbal one, and I have no wish to take sides against it. On the one hand, we must insist that to know the crite­ ria for applying the word 'good' to motor­ cars is not to know- at any rate in the full or primary sense-the meaning of the e x ­ pression 'good motor-car'; to this extent the objection must be agreed witb. On the other hand, the relation of the expression 'good motor-car' to the criteria for its appli­ cation is very like the relation of a descrip­ tive expression to its defining characteris­ tics, and this likeness finds an echo in our language when we ask 'What do you mean,


good?', and get the answer 'I mean it'll do 80 and never breaks down'. In view of this undoubted fact of usage, I deem it best to adopt the term 'descriptive meaning'. Moreover, it is natural to say that a sen­ tence has descriptive meaning, if the speaker intends it primarily to convey in­ formation; and when a newspaper says that X opened the batting on a good wicket, its intention is not primarily to commend the wicket, but to inform its · readers what description of wicket it was. It is time now to justify my calling the descriptive meaning of 'good' secondary to the evaluative meaning. 1.{y reasons for doing so are hvo. First, the evaluative meaning is constant for every class of ob­ ject for which the word is used. 'When we call a motor-car or a chronometer or a cricket-bat or a picture good, we are com­ mending all of them. But because we are commending all of them for different rea­ sons, the descriptive meaning is different in all cases. V\le have knowledge of the eval­ uative meaning of 'good' from our earliest years; but we are constantly learning to use it in new descriptive meanings, as the classes of objects whose virtues we learn to distinguish grow more numerous. Some­ times \Ve learn to use 'good' in a new des­ criptive meaning through being taught it by an expert in a particular field- for exam­ ple, a horseman might teach me how to rec­ ognize a good hunter. Sometimes, on the other hand, we make up a new descriptive meaning for ourselves. This happens when we start having a standard for a class of objects, certain members of which we have started needing to place in order of merit, but for which there has hitherto been no standard. . . . The second reason for calling the eval­ uative meaning primary is, that we can use the evaluative force of the word in order to change the descriptive meaning for any class of objects. This is what the moral re­ former often does in morals; but the same process occurs outside morals. It may hap­ pen that motor-cars will in the near future



change considerably in design ( e.g. by our become conventional. It is, of course1 :im­ seeking economy at the expense of, size ) . It possible to say exactly when this has hap­ may be that then we shall cease giving the pened; it is a process like the coming of name �a good motor-car' to a car that now winter. Although the evaluative meaning of would rightly and with the concurrence of all be allowed that name. How, Hnguisti­ 'good' is primary, the secondary descriptive cally speaking, would this have happened? meaning is never wholly absent. Even At present, we are roughly agreed ( though when we are using the word ·good' evalua­ only roughly) on the necessary and su­ tively in order to set up a new standard, fficient criteria for calling a motor-car a the word still has a descriptive meaning, good one. If what I have described takes not in the sense that it is used to convey place, we may begin to say 'No cars of the information; but in the sense thar its use in nineteen-fifties were really good; there setting up the new standard is an essential weren't any good ones till 1960'. NO"'-'-'' here preliminary--like definition in the case of a we cannot be using 'good' with the same purely descriptive word-to its subsequent de,,scriptive meaning as it is now generally use with a nev;r descriptive meaning, It is used with; for some of the cars of 1050 do also to be noticed t:hat the relative promi­ indubitably have those characteristics nence of the descriptive _ and eva!.uative which entitle them to the name 'good mo­ meanings of 'good' varies according to the tor-car' in the 1950 descriptive sense of that class of objects '>ithin which commenda­ word. · what ls happening is that the eval­ tion is being given. vVe may illustrate this uative meaning of the ,vord is being used b y taking two extreme examples, If I talk in order to shift the descriptive meaning; of 'a good egg', it is at once known to what we are doing what would be called, if description of egg I am referring-namely, 'good' were a purely descriptive word, one that is· not decomposed. Here the de­ redeflning it. But we 0.annot call it t:hat, for scriptive meaning predominates, because the evaluative meaning remains constant; we have very fixed standards for assessing we are rather altering the standard. This is the goodness of eggs. On the other hand, if I similar to the process called by Professor say that a poem is a good one, very little Stevenson 'persuasive definition»; the information is given about what descrip­ process is not necessarily, hov:ever1 high!y tion of poem it is--for there ls no accepted standard of goodness in poem.a. But it must coloured ·with emotion. . . . Although with 'good· the evaluative not be thought that 'good egg' is exclu­ meaning is primary� there are other ,vords sively descriptive, or 'good poem� exclu­ in which the evaluative meaning is second­ siveiy evaluative. If, as the Chinese are al­ ary to the descriptive. Such words are 'tidy' leged to do, we chose to eat eggs that are and 'industrious'. Both are normally used to decomposed, we should call that kind of cornmend; but \Ve can say\ \vithout any egg good, just as, because we choose to eat hint of irony, 'too tidy' or 'too industrious'. game that is slightly decomposed, we call it It is the descriptive meaning of these �well-hung' ( compare also the expression words that ls most fumly attached to them; "good Stilton cheese') . And if I said that a and therefore, although: we nmst for certain poen1 was good, and was not a very eccen­ purposes class them as value-words ( for if tric person, my hearer \Vould be justiB.ed in we treat them as purely descriptive, logical assuming that t:he poem was not 'Happy errors result), they are so in a less full birthday to you!' sense t:han 'good'. If the evaluative mean­ In general, t:he more fixed and ac­ ing of a word, which 'tvas primary, comes cepted the standard, the more information to be secondary, t:hat is a sign that the is conveyed. But it must not be thought standard to which the word appeals has that the evaluative force of the ,vord varies


at all exactly in inverse proportion to the descriptive. The two vary independently: where a standard is firmly established and is as firmly believed in, a judgement con­ taining 'good' may be highly informative, without being any the less commendatory. Consider the following description of the Oxford Sewage Farm: The method employed is primitive but efficient. The farm is unsightly, obnoxious to people dwelling near it, and not very remuner­ ative, but the effiuent from it is, in the techni­ cal sense, good. 2 Now here, as may be seen by consulting

handbooks mi the subject, there are per­ fectly well-recognized tests for determining whether effiuent is good or bad. One manual3 gives a simple field test, and ano_ther4 gives a series of more comprehen­ sive tests which take up seventeen pages. This might tempt us to say that the word is used in a purely descriptive sense and has no evaluative force. But, although admit­ tedly in calling effiuent good in this techni­ cal sense we are commending it as effiuent and not as perfume, \Ve are nevertheless commending it; it is not a neutral chemical or biological fact about it that it is good; to say that it was bad would be to give a very good reason for sacking the sewage-farmer or taking other steps to see that it was good in future. The proper comment on such a lapse was made by a former Archbishop of York, speaking to the Congress of the Royal Sanitary Institute, 1912: There is now, I hope, no need of the tren­ chant eloquence of that noble-hearted pioneer of sanitary science, Charles Kingsley, to insist that it is not religion, but something more near­ ly approaching blasphemy, to say that an out­ �reak of disease is God's ,vill being done, 2

Social Services in the Oxford District, p. 322. s Kershaw, Sewage Purification and Disposal, pp. 213-14. ., Thresh, Beale, and Suckling, The Examinntion of ·waters and Water Supplies, 6th ed.

MIXED VIEWS v;rhen patently it is man's duty which is being left undone. 5

It is true that, if the word 'good' in a cer­ tain sentence has very little evaluative meaning, it is likely that it has a fair amount of descriptive meaning, and vice versa. That is because, if it had very little of either, it would have very little meaning at all, and would not be worth uttering. To this e1.'tent the meanings vary inversely. But this is only a tendency; we may do jus­ tice to the logical phenomena by saying that 'good' normally has at least some of both sorts of meaning; that it normally has sufficient of both sorts taken together to make it worth uttering; and that, provided that the first two conditions are satisfied, the amounts of the hvo sorts of meaning vary independently. There are, however, cases in which we use the word 'good' with no commendatory meaning at all. We must distinguish several kinds of such non-commendatory uses. The first has been called the inverted-commas use. If I were not accustomed to commend any but the most modem styles of architec­ ture, I might still say 'The new chamber of the House of Commons is very good Gothic revival'. I might mean this in several senses. The first is that in which it is equi­ valent to 'a good example to choose, if one is seeking to illustrate the typical features of Gothic revival' or 'a good specimen of Gothic revival'. This is a specialized eval­ uative sense, with which we are not here concerned. I might mean, on the other hand, 'genuinely preferable to most other examples of Gothic revival, and therefore to be commended within the class of Gothic revival buildings, though not with­ in the class of buildings in general'. With this sense, too, we are not now concerned; it is a commendatory use, with a limited class of comparison ( 8. 2 ) . Tbe sense with which we are concerned is that in which it means, roughly, 'the sort of Gothic revival building about which a certain sort of peoi;

Kershaw, op. cit., p. 4.


ple�you know who �would say «that is a good building".' It is a characteristic of this use of 'good' that in expanding it we often want to put the word 'good' inside inverted commas; hence the name. \Ve are, in this use, not making a value-judgement our­ selves, bi,t alluding to the value-judge­ ments of other people. This type of use is extremely important for the logic of moral judgements1 in which it has caused some confusion. It is to be noticed that it is easiest to use 'good' in an inverted-commas sense '\vhen a certain c1ass of people, who are sufficiently numerous and prominent for their value-judgements to be well known ( e.g. the 'besf people in any field), have a rigid standard of commendation for that class of object. In such cases, the inverted- commas use can verge into an ironic use, in which not only is no commen­ dation being given, but rather the reverse. lf I had a low opinion of Carlo Dold, I might say 1f you want to see a really "good" Carlo Dolci, go and look at the one in . . . '. There is another use in which the ab­ sence of evaluative content is nlJt su­ fficiently obvious to the speaker for us to call it either an inverted-commas or an ironic use. This is the corwentioruil use, Ill which the speaker is merely paying lipserv· ice to a convention, by commending, or saying commendatory things about, au ob­ ject just because everyone else does. I might, if I myself had no preference at all about the design of furniture, still say ''This piece of furniture is of good design', not because I wisheC to guide mv own or anv­ one else's choice of furnitur�, but simply because I had been taught characteristics which are generally held to be criteria of good design, and wished to show that I had < ho--;,vever: that the most useful value-judgements are those \vhich have reference to choices that we might very likely have to make. It should be pointed out that even judgements about past choices do not refer merely to the past. As we shall see, all value-judgements are covertly universal in character, which is the same as to say that they refer to, and eA-press acceptance of, a standard which has application to other similar instances. If I censure someone for having done something, I envisage the pos­ sibility of him, or someone else i or myself, having to n1ake a similar choice again; oth­ erwise there would be no point in ce:asur­ ing him. Thus, if I say to a man whom I am teachbi to drive 'You did that man­ oeuvre badly' this is a very typical piece of driving-instruction; and driving-instruction consist,; in teaching a man to drive not in the past but in the future; to this end we censure or commend past pieces of driving) in order to impa1t to him the standard which is to guide him in his subsequent conduct. \\1hen we commend an object, our judgement is not solely about that particu­ lar object, but is inescapably about objects


WHAT IS A VALUE JUDGEMENT? like it. Thus, if I say that a certain mo­ tor-car is a good one, I a1n merely saying something about that particular motor-car. To say something about that particular car, 1nerely, would not be to commend. To commend, as we have seen, is to guide choices. Now for guiding a particular choice we · have a linguistic insb·ument which is not that of commendation, namely, the singular imperative. If I wish merely to tell someone to choose a particu­ lar car, with no thought of th e kind of car to which it belongs, I can say 'Take that one'. If instead. of this I say 'That is a good one', I arh saying something more. I am im­ plying that if any motor-car were just like that one, it would be a good one too; whereas by saying 'Take that one', I do not imply that, if my hearer sees another car just like that one, he is to take it too. But further, the implication of the judgement That is a good motor-car' does not extend merely to motor-cars exactly like that one. If this were so, the implication would be for practical purposes useless; for nothing is exactly Hke anything else. It extends to every motor-car that is like that one in the relevant particulars; and the relevant parti­ culars are its virtues-those of its character­ istics for \Vhic-h I was commending i( or wbich I was calling good about it. vYhen­ ever we commend, we have in mind something about the object commended which is the reason for our commendation. It therefore always makes sense, after someone has said 'That is a good mo­ tor-car', to ask iV'nat is good about it?' or iVhy do you call it good?' or 'What fea­ tures of it are vou commendingF It may not ahvays be easJ to ansvter this question" pre­ cisely, but it is always a legitimate ques­ tion, If ,ve did not understand why it was alvvays a legitimate question� we should not understand the way in which the word 'good' functions. We may illustrate this point by com­ paring two dialogues: (1) X. Jones" motor-car is a good one.

327 Y. What makes you call it good? X. Oh, just that it's good. Y. But there must be some reason

for your calling it good, I mean some property that it has in vi.-1:ue of which you call it good. X. No; the p::operty in vi..."T',,1e of \Vhich I call it good is jl!st its goodness and nothing Y. But do you mean that its shape, speed, weight1 manoeuvrability &c., are irrelevant to whether you call it good or not? X. Yes, quite irrelevanti the only relevant property is that of good­ ness, just as, if I caaed it yellow, the only relevant property would be that of yellowness. · ith (2) The same dialogue, only w 'yellow' substituted for 'good' and 'yellowness' for 'goodness' through­ out1 and the last clause ("just as . , yellmvness') omitted. The reason vvhy X's position in the .Srst dia­ logue is eccentric is that since, as we have already remarked, 'good1 is a ·supervenienf or . . . . That the descriptive meaning of the that people who collected a lot of scalps ,vord �good' is fn morals; as elsewhere, sec­ were good ( cannibal) , they would not be ondary to the evaluative, may be seen in disagreeing, because in English ( at any the following example. Let us suppose that rate missionary English) , 'good' would a missionary, anned vtith a grammar bookJ meaa among other things 'doing no mur­ lands on a cannibal island. The vocabulary der', whereas in the cannibals' language of his grammar book gives him the equiva­ 'good' would mean something quite dif­ lent, in the cannibals' language, of the Eng­ ferent, among other things 'productive of lish 'good'. Let us suppose that, by a queer maximum scalps>. It is because in its pri­ coincidence, the \VOrd is 'good'. And let us mary evaluative meaning �good' means nei­ suppose, also, that it really is the equiva­ ther of these things, but is in both lan­ lent-that it as the Oxford English Dic­ guages the most general adjective of com­ tionary puts it, 'the most general adjective mendation, that the missionarv can use it of commendation' in their language. If the to teach the cannibals Chri;tian morals. Suppose, however, tha: the mission­ missionary has mastered his vocabulary, he can1 so long as he uses the word evalua� ary's mission is successful. Then, the former tively and not descriptively, communicate cannibals will come to commend the same with them about morals quite happily. qualities in people as the missionary, and Thev know that when he uses the word he the words 'good man' will come to have a is c�mmending the person or object that he more or less conunon descriptive rneaning. applies it to. The only thing they will find The danger will then be tha: the cannibals odd is that he applies it to such unexpected may> after a generation or two, thhik: that people, peop'.e who are meek and gentle that is the only sort of meaning they have. and do not collect large quantities of 'Good� will in that case mean for then1, sin1sca!ps; whereas they themselves are accus­ ply 'doing what it says in the Sermon o:i tomed to commend people who are bold the Mount'; and they may come to forget and burly and co!lect more scalps than the that it is a \VOrd of commenr!.ation; they average. But they and the missionary are will not realize that opinions about moral under no misapprehension about the mean­ goodness have a bearing on what they ing, in the evaluative sense, of the word themselves are to do. Their standards will 'good'; it is the word one uses for com­ then be in mortal danger. A Communist, mending. If they were under such a misap­ landing on the island to convert the people ptehension, moral communication between to his way of life, may even take advantage of the ossification of their standards. He them would be impossible. 'vVe have thus a situation which would may say 'All these "good" Christians­ appear paradoxical to someone who missionaries and colonial servants a..'1d the thought that 'good' ( either in English or in rest- are just deceiving you to their own the cannibals' language) was a quality­ profit.' ThLs would be to nse the word des­ word like 'red'. Even if the qualities in peo- criptively with a dash of irony; and he


could not do this plausibly unless the standards of the Christians had become considerably ossified. Some of the ploys of Thrasymachus in the first book of Plato's Republic are very similar to this . . . . Such vicissitudes of the word 'good' reflect accurately [this] sort of moral de­ velopment . . . Moral principles or standards are first established; then they get too rigid, and the words used in referring to them become too dominantly descriptive; their evaluative force has to be painfully revived before the standards are out of


danger. In the course of revival, the stand­ ards get adapted to changed circum­ stances; moral reform ta}ces place, and its instrument is the evaluative use of value­ language. The remedy, in fact, for moral stagnation and decay is to learn to use our value-language for the purpose for which it is designed; and this involves not merely a lesson in talking, but a lesson in doing that which we commend; for unless we are pre­ pared to do this we are doing no more than paying lip-service to a conventional stand­ ard.

A Quasi-.lVaturalist Definition RICHARD BRANDT It seems worthwhile to consider a novel form of naturalistic definition. 111is defi­ nition, however, is so markedly different from other naturalistic definitions that we call it ..quasi-natcralist.'' ',,Ve shall explain the dellnition first. Let us first agree to speak of a "corre­ sponding attitude.. for every ethical term. To ..preferable" the attitude of preference '\\1ll correspond; to "obligation" something like "feeling obligated" for the case of agents and "inclined to demand" for ob­ servers; and so on, Roughly what we mean by a "corresponding" attitude is the atti­ tude someone justifiably has if some ethical statement is properly asserted by him. There can be further discussion, obviously, of the ouestion of what attitude corre­ sponds, 1;1 this sense, to any ethical term in which one happens to be interested. The quasi-naturalist definition pro­ poses that "x is E" ( where E is some ethi­ cal term) means the same as '"ThG E-corresponding attitude [ which will be detemi.ined once we have specified the eth­ ical term] to x satisfies all the condition., that would be set, as a general p0Bcy1 for the endorsement of attitudes governi.'1g or appraising choices or ae,tions, by anyone who was intelligent and factually informed and had thought through the problems of the possible different general policies for Richard B. Bmndt, ETHICAL THEORY, The Problems of Nor.native and Critical Ethics, @ 1959, Reprinted by permission of Prentice-Hall� Inc., Englewco:i Cliffs� New Jersey. Pp. 26.S -269.

the endorsement of snch attitudes." For ex­ ample, "x is preferable to y" would mean "Preferring x to y satisfies all the conditions that would be set. . . .''1 This definition is properly classifie d as naturalist,. because whether an "E­ corresponding attitude" satisfies "all the conditions» can in the end be decided in principle by observation and the methods of science. In other words, the definition can be translated into the language of em­ pirical science (p. 155). But it is also prop­ erly distinguished as "quasi-naturalist" in order to mark two important points. Some naturalist definitions. like the proposal that ux is vvorthwhile in itself" means 'x is pleas­ ant,"" have the consequence that i."Ttportant snbstautive ethical principles are true by definition (in this case "Something is worthwhile in itself if and only if it is pleasant" ) , The quasi-naturalist definition has no such implication. Second, other nat­ uralist de.finitions, like the Ideal Observer theory, have Llie consequence that it is true by definition that a certain method 1nust be used if true or · valid ethical propositions are to be discovered. The qu�i-naturalist definition1 again, has no such implication> for in order to determine what ethical 4


There is a point o-£ vagueness in this defin i ­ tion, corresponding to one in the Qualified Atti� tude Method1 which we shall discuss in the follo w ­ ing chapter. There are also particular phrases in the definition that are vague. For example, when has one "thought through" the proble:ns men� t!oned? Bi:: such vagueness is not necessarily an objection,



method will lead to valid ethical state­ ments, it implies we must £rst ascertain what conditions would be set by an intelli­ gent, informed person for the endorsement of attitudes of a certain kind. It is no objection to this definition, as it is an objection to the Ideal Observer definition, that some people do not use the Qualified Attitude Method,• but, rather, a theological or utilitarian method. The rea­ son for this is that someone's use of a theo­ logical IIlethod, for instance, may simply show that he believes that "all the condi­ tions that would be set" include some con­ dition having to do with the will of God. These distinctive features of the qua­ si-naturalist definition are strong recom­ mendations of it. Everyone is under the impression that whatever he is saying, when he makes an ethical statement, it is something which is correct or incorrect, valid or invalid. The quasi-naturalist de­ finition has the desirable implication that this impression is correct. On the other hand, the definition is framed so as not to [Editors' Note. The Qualified Attitude Method is described in the immediately preceding section of the chapter. In it Professor Brandt suggests a method for resolving ethical disputes. Instead of simply asking "VVhat is my attitude toward X?" the user of the Qualified Attitude Method considers what his attitude toward X would be if certain conditions are fulfilled; thus, his atti­ tude toward X is discounted ( 1 ) if it is not im­ partial, (2) if it is not sufficiently informed as to the facts of the situation, ( 3) if it is a conse­ quence of an abnormal state of mind such as anger, grief, fatigue, depression, etc., and ( 4) if its prompting would be incompatible with hav­ ing a system of principles that are consistent with one another and general ( in the sense of applying to all members of the class in question, and containing no proper names ). The requirements of the Qualified Attitude 1fothod are much the same as the criteria invoked by the Ideal Observer theory; but the Qualified Attitude Method dif­ fers. from the Ideal Observer theory in that it makes no claim at all about the meaning of ethical terms ( it is not committed, as the Ideal Observer theory is, to a naturalistic definition of ethical terms) ; it is neutral with regard to these meta-ethical theories, but claims only that there is a rational method-the one outlined by the Qualified Attitude lvfethod-for resolving ethical disagreements.} 4


exclude, as necessarily irrelevant, any of the arguments actually used about particu­ lar ethical principles, or any of the debates about ethical method which strike us as se­ riously relevant. Of course this definition must be able to survive criticisms of the kind that have been leveled at traditional naturalist de­ £nitions, in order to be acceptable as a re­ portive definition of ethical terms. It must also be able to stand up against the reasons that have been offered for the emotive theory (pp. 214-20 ) . It is not clear, how­ ever, that any of these points is a serious objection to it. Of all the naturalistic de­ finitions it seems decidedly the least open to objection. A noncognitive theory. One may think, however, that the quasi-naturalist defini­ tion is too remote from our overt meanings to qualify as what we mean when we use ethical terms. It is, therefore, worthwhile to consider whether some form of noncog­ nitive analysis of ethical language might be more satisfactory. Let us explore this. First of all, however, vve should re­ mind ourselves that a noncognitivist need not deny that the Qualified Attitude Method is the "standard" method of eval­ uating ethical statements, or that good rea­ sons can be given for following it, of the kind we offered. This is obvious, when we remind ourselves of the definition of "non­ cognitivism" ( p. 205 ) ; this theory asserts only that ethical statements are best under­ stood by likening them to some speech form different from fact statements, and that ethical terms are not property­ referring at all- or at least that, if they do refer to properties, the fact is of secondary import for understanding ethical discourse. The noncognitivist's theory commits him in consistency only to saying that ethical statements certified by the Qualified Atti­ tude Method need not be viewed as con­ firmed by evidence in the way in which the statements of the empirical sciences are confumed, either in vie\V of the meaning and function of ethical terms or in view of

A QUASI-NATURALIST DEFINITION any other facts. Let us, for the sake of the argument, not contest this point There is a more questionable assertion that the noncognitiv:ist may make, although he need not. This is that, even if most peo� pie are prepared to make an ethical state­ ment only if they think it meets the condi­ tions of the Quali£ed Attitude Method, and even if there are convincing reasons in sup­ port of so doing, the fact is unimportant and can be ignored-because whether peo­ ple accept the Qualified Attitude Method as a standard for ethical reasoning and whether they are convinced by reasons that support its use is in the end a matter of their attitudes and not of sheer logic, This point, ho,vever} is mistaken. It leads the noncognitivist who accepts it to an insup­ portable neutrality in his own appraisal of methods in ethics-to the view that one is just as good as another. It leads hb:n to over­ look the fact that there are conclusive reasons for preferring one method of ethi­ cal. deliberation to another. It leads him to overlook the fact that ethical thinking and debate are very similar to inductive reason­ ing-reasoning according to a "standard" form which can be supported by good rea­ sons and to which there is no serious alter­ native. Let us now st:ppose that \:.'le noncogni­ th�st agrees that in the sense explained the Qualified Attitude .Method is the "stand­ ard" method of ethical reflection, that fol­ lowing it can he supported by strong rea­ son, and that on this accoc:nt some ethical judgments-those which satisfy the condi­ tions of the .Method-have a correspond­ ingly approved status. But will this affect his account of the meaning and function of ethica1 tern1s? Let us see how we can or must reconstruct his theory so as to allow for agreeing to these various things. First, the noncognitivist can continue, "1thout change, to say that the primary job of ethical language is noncognitive. He may say that the primary job is to advise, or to urge, or to express attitudes, or to ex­ press over-all moral or impartial attitudes


in the sense described in the preceding chapter� or some cornbination of these things. Perhaps he will say that there are different noncognitive jobs done by the same ethical "\vords on different occasions. So much for the noncognit:ve side of ethi­ cal language. ,1/e have seen that there is reason to recognize another aspect of ethical lan­ guage. Toward the end of the preceding chapter we sa-,v that many noncognitivists suggest that ethical language does not do merely a noncognitive job, or at least that it does such a job in a special way. It has been proposed that the use of ethical lan­ _guage 11as certain "contextual h:nplications'' or makes certain claims. \,Yllat is d-istincti-ve of ethical language, it has been held, is the making of certain claims or the having of contextual implications of a special sort, But there were differences of opi:iion among noncognitivists about just what these claim.s or implications are. On the basis of our argument so far, we axe now in a position to make a pro­ posal about an important one of these claims or implications. Ot:r suggestion is not that ethical statements claim that the attitude ( etc. ) they express satisfies the conditions of the Qualified Attitude Method; it is not this, · because we agree that some persons do not use this ivlethod1 and have not thought of the reasons which support it. Our suggestion is rather a weaker one; it is that ethical lal1guage claims, of the attitude ( etc. ) which it ex­ presses, that it satisfies all the conditions which wouJc. be set, as a general policy for the endorsement of attitudes governing or appraising choices or actions, by anyone who was intelligent and factually inforrned and bad thought through the problems of the possible different general policies for the endorsement of such attitu.des. In other words, we suggest a noncognitivist should say that the use of ethical language makes a claim about the attitude it expresses, which is identical with part of what the quasi-naturalist definition says ethical


terms assert. For instance, "It is desirable to do x" may be construed as ( 1 ) tbe ex­ pression of an over-all impartial preference ( or, perhaps, as equivalent to "I advise you to do x"), and ( 2 ) as claiming or at least implying tbat favoring x is justified in the sense that it satisfies the conditions which would be set by intelligent persons, and so forth. Ethical statements, then, may be construed as botb doing something and making a validity-claim of tbis sort.2 In order to distinguish tbe foregoing analysis clearly from tbe quasi-naturalist one, we must, of course, be clear about just what it is for a statement to make an asser­ tion, and what are the criteria for deciding the content asserted by a given statement. " \1/e should notice that a philosopher might propose theories, like the three just formulated, not of ethical terms themselves, but of the term "justified." He might refuse to analyze ethical terms themselves, saying that we understand them well enough, but might propose any one of them for the term "justified" as applied to ethical state­ ments. Such a view is not very satisfactory. For we shall still want some view about ethical terms themselves. Moreover, there is no reason to think that such analyses of "justified" are any better established than corresponding analyses of ethical language itself.


Similarly, we must be clear what it is for a statement to express something, or to make a claim; and we must know what the crite­ ria are for deciding what is expressed or claimed. Let us assume that tbe problems involved in clarifying all tbose things can be satisfactorily overcome. Is there anything to choose between tbe quasi-naturalist analysis of ethical terms, and this noncognitivist account? De­ spite some reservations ( particularly about whether such a noncognitivist can consist­ ently say tbe noncognition function is pri­ mary), we prefer to say there is nothing to choose. On tbe contrary, we do assert tbat there is no insuperable objection to either one, and that one or the other of them is correct. For our future purposes, however, it is simpler to adopt tbe quasi-naturalist definition. We shall, therefore, in tbe fu­ ture, talk as if tbis definition should be given preference. But, whenever we use it, it would be possible to make tbe same point by using the noncognitive tbeory just suggested. YVe believe tbere is no sound objection to the quasi-naturalist definition, but we have no quarrel with one who pre­ fers the above noncognitive analysis in­ stead.

F. Relativism and Justification

Ethical Relativism RICHARD BRANDT A Greek philosopher who lived in the filth century B.C., named Protagoras, seems to have believed two things : first, that moral principles cannot be shown to be valid for everybody; and second, that people ought to follow the conventions of their own group. 1 Something like this combination of propositions probably had been thought of before his time. Primitive people are well aware that different social groups have dif­ ferent standards, and at least sometimes doubt whether one set of standards can really be shown to be superior to others. Moreover, probably in many groups it has been thought that a person who conforms conscientiously to the standards of his own group deserves · respect. Vievvs roughly similar to those of Pro­ tagoras may be classi£ed as forms of ethi­ cal relativism. The term "ethical relativ­ ism," however, is used in different senses, and one should be ,vary when one comes across it. Sometimes one is said to be a rel­ ativist if he thinks that an action that is wrong in one place might not be in an­ other, so that one is declared a relativist if he -i:hinks it wrong for a group of Eskimos ta strip a man of his clothing twenty miles Richard B. Brandt, Ethical Theory: The Prob­ lems of Normative and Critical Ethics, © 1959. Reprinted by permission of Prentice-Hall, Inc., Engl�wood Cliffs, New Jersey. Pp. 271-284. 1 For Protagoras' view, see Plato, Theaetetus, pp. 166.ff.; and F. J. Copleston, A History of Philos­ ophy, I ( London: Bums Oates & V/ashboume Ltd., 1956 ) , pp. 87-90.

from home an January 1, but not wrong for a tribe at the equator. If "relativism" is used in this sense, then practically every­ one is a relativist, for practically everyone believes that particular circumstances make a difference to the morality of an act-that, for instance, it is right to lie in some cir­ cumstances but wrong in others. Again, one is sometimes said to be a relativist if he as­ serts a pair of causal propositions: that dif­ ferent social groups sometimes have dif­ ferent values ( ethical opinions ) as a result of historical developments; and that an in­ dividual's values are near-replicas of the tradition of his group, however strongly he may feel that they are "his own" or that they are "valid" and can be supported by convincing reasons. YVe shall not use "ethi­ cal relativism" for either of these vie\vs, but reserve it for a theory at least fairly close to that of Protagoras.2 � It is useful to compare Protagoras' relativism -..vith the special theory of relativity in physics. One implication of this theory is that measurements of certain physical quantities, like the temporal dis­ tance between two events, will come out differ­ ently for different frames of reference ( one "frame of reference" being the set of observers having the same relative rectilinear motion) . · All the careful observations in one frame will give the same result; and in this sense there is a "right" answer for this frame. But different frames will have di.ffen,nt "right" ans-..vers, and there is in principle no way of showing that one of these is the "really right" answer. However, certain quan­ tities (like the spatio-temporal distance between two events ) are absolutes, in the sense that care­ ful measurements will give one right answer for everybody.



1. THE Q UESTION: "ARE CONFLICTING ETHICAL OPINIONS EQUALLY VALID?" The position of Protagoras, however, is somewhat vague, and if \Ve are to assess it, we must sharpen it. It is also convenient to deal separately with its two parts. We shall begin with a restatement of the first part of his theory, and then assess it at some length; only then shall we consider the sec­ ond half of his position. As we go on, we shall see that the first part of his theory is theoretically more interesting and impor­ tant than the second. For this reason, we shall apply the term "ethical relativism" to any theory that agrees with our sharpened form of Protagoras· first point, irrespective of its attitude toward the second. It is clarifying to substitute, in place of our initial statement of Protagoras' view, the following, as a brief formulation of the relativist thesis in ethics : "There are con­

flicting ethical opinions that are , equally valid." But this formulation requires discus­ sion in order to be clear. The first thing to notice-although the fact will not b.e obvious until we have ex­ plained the phrase "equally valid"-is that the statement is about ethical opinions or statements, but is not an ethical statement itself. It is not like saying, "Nothing is right or wrong!" or "Some things are both right and wrong!" It is a metaethical theory. Next, the statement is cautious. It does not say that no ethical opinions are valid for everybody. It says only that some ethi­ cal opinions are not more valid than some other ethical opinions that conflict with them. Third, our relativist thesis is not merely the claim that different individuals sometimes in fact have conflicting ethical opini-orw, It does assert this, but it goes fur­ ther. It holds that the conflicting ethical opinions are equally valid. We do not es­ tablish this merely by showing that people


disagree. Nor do we establish it by show­ ing that individuals' ethical opinions are at least to some extent dependent on the cul­ tural stream within which they stand. Everyone must agree to this-although everyone must also admit that somehow so­ cieties often spa,vn their O\vn moral critics. Nor do we establish it by sho,ving that the standards of a given society have their causes. Of course they do; and so do the scientific opinions in a given society, al­ though we hardly think this necessarily im­ pugns their universal validity. Fourth, what do we mean by "con­ flicting ethical opinions"? \Ve mean, of course, by an «opinion" the readiness to make a sincere statement. Thus, a person has an "ethical opinion" to the effect that a particular thing is right or wrong if he could, when asked, make without decep­ tion an ethical statement to the effect that that thing is right or wrong. ( We explained how to identify an "ethical" statement in Chapter 1. ) Now, suppose Mr. A makes an ethical statement, and Nfr. B makes a dif­ ferent ethical statement, How shall we tell whether the two statements "conflict"? A sufficient condition of conflict is this: that both statements are about the same subiect ( we explain this in a moment ), and the one applies to this subject an ethical predi­ cate P, and the other applies to it the same ethical predicate prefaced by the English "not" or something that means or entails the same. For instance, one may say "is morally right" and the other may say "is not morally right," of the very same sub­ ject. But now, when do two ethical state­ ments have the same subiect? This is a more awkward question. YVe cannot test this just by observing the verbal forms. For instance, Thomas Jefferson said> approxi­ mately, "A revolution every few years is a fine thing." But suppose Karl Marx also said, "A revolution every few years is a fine thing." Could we assume that these two men were necessarily saying the same thing? Of course not. Or again, suppose

ETHICAL RELATIVISM Mr. A, a resident of the South Pacific, says it is right to bury one's father alive on his sixtieth birthday, irrespective of his state of health; and suppose I say this is not right Are we talking about the same thing? Not necessarily. The kind of situation Mr. A has in mind is likely to be very different from the kind of situation I have in mind. Per­ haps he is assuming that the body one will have in the next world will be exactly like the kind one has just before departing this life ( and hence, may think it advisable to depart before feebleness sets in ) ; whereas I ma v think o:::1e has no further existence at all after one's earthly demise. He is talking about burying a!ive a !ather who \vill exist in the next world in a certain kind of body; and I am not. In this situation} it is only confusing to say that our ethical opinions "conilict." Let us say that two people are talking about the same subject only in the following situation. Let us suppose A and B make con:licting eti1ical predications abot:t something or some kind of thing, ostensibly the same for both. But suppose further there is some property P that A more or less consciously believes this thing or kind of thing '1as, whereas B does not believe this, Further, let us suppose that if A ceased to believe this, he ·would cease to have the same eGcal opinion about it but agree with B; and let us suppose that if B began to believe this ( other things being equal), he would change his ethical opin­ ion and agree with A. In this case, let us say that A and B are not appraising the same subject. But if the::.-e is no n1ore-or-less conscious belief ha,�ng the status de­ scribed, then we shall say that they are talking about the same subject, and that their ethical opinions are conflicting. But now, :S.na1Iy, what is the meaning of the phrase "equally valid"? In order to damy this, let us draw a parallel w:ith lan­ guage v,re use in appraising scienti£.c theo­ ries. Suppose we have two conflicting theo­ ries about natural phenomena. Each of these theories might explain a part of

337 the known £acts, but not all of them, at least not very well. \Ve might then say, "In the !igbt of presently known facts, the t\vo theories are equally plausible." On the other hand, we might make a more radical supposition. Suppose, when thL'lking about these theories, we make the daring forecast about future evidence, that ·when scientific investigation has been indefinitely pro­ longed and all possible experimental data are in, both of these theories ,vrn explain all the facts, and there will be no grountl for a rational preference of one to the ofaer, although parts of the two theories do contradict eacl1 other. In this case, vve might say, although this sounds startling, "Although these theories are mutually con­ tradictory in some respects, they are both valid." What a person who made such a statement would b e saying is that the t:se of a refined inductive logic, on a complete set of experimental data, would .support as strongly confirmed both of two conflict:ng theories. \11e need not argue whether in fact this case ever does or even could arise, but we can understand the possibilit)', and the important thing is the parallel wifo eth­ ics. Nowi the ethical relativist is not rnerely maki.">1g the uninteresting claim, ·when he says two conflicting ethical state-me::its are equally valid, that the :wo statements are equally plausible in the ligk of the facts known at present. He is saying something much more radical, about what would hap­ pen if one were testing these statements by the best possible ethical methodology, and in the light of a complete system of factual or nonethical knowledge. In other words, he is saying that the application of a ·�ra­ tionar' method in ethics would S'.;.pport, equally, t\,;,o con!licting ethical statements even if there were available a complete sys­ tem of factual knowledge--0r else that there is no "rational" rr1ethod in ethics com­ parable to an ideal inductive method for empirical science. I have used the phrase "rational method in ethics" as designating something


roughly parallel in ethics to inductive logic in empirical science. This idea will be fa­ miliar to us from the preceding chapter, where we argued that the Quali£ed Atti­ tude Method has this status. We can novv explain exactly what it means to say that two conflicting ethical statements are "equally valid." What it means to say this, is that either there is no unique rational or justified method in ethics, or that the use of the unique ra­ tional method in ethics, in the presence of an ideally complete system of factual !mowl­ edge, would still not enable us to make a distinction between the ethical statements being considered. The ethical relativist asserts that there are at least some instances of conflicting ethical opinions that are equally valid in this sense. There are more, and less, radical rela­ tivists. The more radical kind of relativist asserts that there are conflicting ethical opi­ nions and that there is no unique rational method in ethics. To mark this, let us call him a "methodological relativist" or an "ethical skeptic." The less radical relativist does not say there is no unique rational method, but says that there are still some instances of conflicting ethical opinions that are equally valid. Let us call him a "nonmethodological relativist.'� \¥e must look at the logic of, and the evidence sup­ porting, these two kinds of relativism sep­ arately.

2. METHODOLOGICAL RELATIVISM It would perhaps be better to call a "meth­ odological" relativist a "skeptic" and not a relativist at all.3 Nevertheless, it is esCertainly this position is different from "relativ­ ism" in physics. In relativity physics, there are correct judgements for each frame of reference; only, it is impossible to say that one of these judg­ ments ( correct for its frame of reference ) is really correct for everybody. 3


tablished usage to classify various writers, especially anthropologists, as "relativists," although they are methodological relativ­ ists in our sense. In order to avoid confu­ sion, we shall follow this terminology. In order to assess the truth of the theory, the Jlrst thing to decide is whether there are conflicting opinions about the same subject at all. It has been denied that there are. ·Karl Duncker, in an article in Mind, in 1939, questioned whether any anthropological evidence establishes that there are-and suggested that anthropolo­ gists had overlooked the fact that when different societies ostensibly advocate dif­ ferent moral principles ( for example, the U.S.A., monogamy; Moslems, polygamy), they actually have different situations in mind. However, we have already assessed the evidence on this point (pp. 99-103 ), and concluded that there are conflicting ethical judgments even when speakers have the same situation in mind. So far, then, methodological relativism stands up. But is the theory correct in its asser­ tion that there is no unique rational method in ethics? Obviously it is not, if the argument of the preceding chapter is sound. vVe need not go over this ground again. The reader ,viii by now have made up his mind whether or not this is the case. 4 If the reader found the preceding chapter convincing, he may be puzzled by the fact that there are methodological rela­ tivists among social scientists. The reason is simple: the kind of theory we are suggest­ ing is of a species that has only recently been proposed, and social scientists are un4

Notice, however, that the acceptance of the quasi-naturalist definition leaves open the question of methodological relativism, for it might be that intelligent ( and so on) persons would set rw condi­ tions for the endorsement of attitudes governing choices in community living, In this case, every attitude would pass, and there would be no ground for preferring one of two conflicting ethical statements to the other. The same is true for the analysis we discussed.


familiar with it.6 ( Neither, for the most part, are they familiar with sophisticated forms of naturalism.) \Vhen they say that one ethical statement cannot be shO\vn "objectively" to be more valid than another, what they mean, and all that they mean, is that one cannot show that ethical state­ ments are confirmed or refuted by obser­ vation in exactly the same way as are hypotheses in science. They rightly see that "is desirable" must be tested in a way different from ''is desired/' and they con­ clude that ethical statements cannot be evaluated at all- overlooking the fact that tests appropriate for assessing ethical judg­ ments may be somewhat distinctive, but none the less defensible, given their subject matter. .i\{oreover, many social scientists sim­ ply do not realize that their acceptance of inductive logic is no more "rational," in the sense of no more supportable by the canons of deductive logic, than is the "standard" method of ethical thinking. Yet, without qualms they make use of inductive log­ ic- but at the same time condemn as "sub­ jective" the appraisal of ethical statements, although in fact the "standard" method is warranted by reasons equally as good as those that can be adduced to support in­ ductive reasoning in science. Presumably, as time goes on ( and scientists become more familiar with the results of contempo­ rary thinking about inductive logic and ethics ) , social scientists will cease making this irrational distinction. The reader need not, incidentally, feel that he has to choose between what we 5

A contributing factor is that some social scien­ tists do not distinguish "relativism" in the sense of methodological or nonrnethodological relativism as defined above from other senses of "relativism" ( described on pp. 340 - 3 ) . Hence, they think that the truth of one can be inferred from the truth of the other. Since relativism in the senses described on pp, 340-3 is doubtless true, one who does not make the proper distinctions naturally concludes that relativism in one or the other of the two senses now under consideration is also true. Such an inference is, of course, entirely unwarranted.


have called the "standard" method, and methodological relativism. \Ve might be mistaken about what the "standard" method is. It might very well be that there is one and only one method that would be used to resolve ethical issues by intelligent ( and so on) people, but that it is some­ what different from the method we have described. The "methodological relativist" ( as we have defined this term) is making a strong statement: he is saying that there is no method that is a "rational" method in the sense of being the one unique method that would be used to resolve ethical issues by intelligent ( and so on) people. But there might well be such a method, even if it is not the one we have described. Of the theories we have been consider­ ing in previous chapters, which ones are forms of methodological relativism and which are not? Clearly naturalism is not, for naturalists so construe the meaning of ethical statements that ethical statements have the same capacity to be confirmed by observation as do the statements of empiri­ cal science. There is one "unique rational method" for assessing ethical statements 1 and it is simply the method of inductive logic. Some of the naturalists, on the other hand, are nonmethodological relativists, for instance, VVesterrnarck. 6 The emotive theory, on the other hand, as it is usually worked out, belongs to the methodological relativist species. The whole concept of validity in ethics is Different "unique rational methods" vary in the extent to which they can succeed in resolving ethical disputes. There could well be a ''method" -and it might well be the only method we could claim to be a "rational" method- that marked a few ethical judgments as definitely untenable, but gave us no help on the serious issues. In fact, this is the case with YVestermarck's view, The unique method is the method of empirical sci­ ence. But if ethical statements mean what Wester­ marck says they do (p. 166), the method can show that some ethical statements are unwarranted and false, but will not in all cases resolve ethical dis­ putes in the sense of showing that only one of two conflicting statements, made by different peo­ ple, is correct. 6


banned. The theory does not recognize any unique rational method of ethical delibera­ tion; on the contrary, anything is allowed that is effective, that wins harmony of atti­ tudes either interpersonally or intraperson­ ally. Indeed, ineffective reasoning is all right too, according to this theory; it merely i,s ineffective. As a result, there is no way in which any ethical conviction can be "objectively" criticized as being defective, incorrect, or erroneous. On the other hand, the emotive theory need not be a species of methodological rel­ ativism. For instance, if it is held that ethi­ cal statements are expressions of over-aII, impersonal attitudes, then an ethical state­ ment may be "mistaken" if the speaker does not have: the over-all, impersonal atti­ tude he purports to have. Much the same is true if it is supposed that the use of ethical language 9!�S certain "contextual implica­ tions" or makes certain "claims," for then we can say ·that ethical statements are at least «misleading" if not "incorrect"-if the «contextual implications" or "claims" dis­ tinctive of ethical language are not satis­ fied.7 If the argument of the preceding chap­ ter is correct, we have said, the methodo­ logical relativist is mistaken because there is a unique rational method in ethics. 'Norse still, it may be that he is contradict­ ing himself if he both affirms methodologi­ cal relativism and makes ethical statements ( which the relativist presumably, like other men, will often do ) . Whether he is contra­ dicting himself depends on what he means by his ethical statements. He certainly is contradicting himself if he means what the Ideal Observer form of natmalism ( in its absolutist form ) says ethical statements mean ( p. 173 ) . The relativist of this variety need not, hO\vever, renounce engaging in ethical de­ bate, or stop thinking that such debate is 1

The reader should examine R. M. Hare, The

Language of Morals ( Oxford: Clarendon Press, 1952 ) , p, 69, and P. H. Nowell-Smith, Ethics ( Baltimore: Penguin Books, 1954), p. 319,


fruitful. Indeed, he may think that the major questions of normative ethics can be answered-not in the sense of finding a "righf' answer for them, but in the sense that agreement is attainable. He can hold, as many social scientists do, that there are points of ethics about which there is uni­ versal agreement, common ground on the basis of which discussion and adjudication can fruitfully proceed. He may think there is a wide basis of agreement that can be extended further, by pointing out the im­ plications-in vie\V of kno,vn scientific facts-of commitments already made. For instance, agreement on a program of econ­ omic reform might be reached- starting from agreement that suffering is evil and to be avoided- by showing that these econ­ omic reforms are necessary means for avoiding suffering. Indeed, the methodo­ logical relativist can, in general, espouse ethical reasoning of the form licensed by "contextualism"; but no more than that.


The second and less radical form of relativ­ ism agrees with the first form that there are conflicting ethical judgments about the same subjects; but it differs by holding that there is a unique rational method for answering ethical questions. It then goes on to assert that when we apply this method even to an ideally complete set of data, it sometimes happens that it is impos­ sible to decide between conflicting ethical judgments. Practically, it is an important question hov.r often and where this «some­ times" is; but we postpone this question for the m·oment. Is methodological relativism true or tenable? We have already conceded the thesis it shares with the more radical view: that there are conflicting ethical judgments about the same subjects. But, where there are such conflicting judgments, are the judgments ever equally valid? This ques-


tion is a difficult one, Indeed, it may be in­ consistent to suppose that they ever are. \Ve must consider this. There are some metaethical theories that are consistent vsith nonmethod­ ological relativism, \\1th sa;�g that confilcting judgments are sometimes equally valid, Take, for instance, the \icw of VVestermarck. He is a naturalist and therefore thinks foat the rational method for answering the questions of ethics ls the method of science, On the other hand, he thinks that ,..x is \vrong" means "I have a tendency to feel impartial resentment to­ ward people who do things like x," Given his premises, is it consistent to be a relativ­ ist? That is, is it consistent to say :hat confilcting ethical statements are some­ t::mes both "valid" in the sense permitted by his theory? Yes, for we can describe conditions in which conB.icting statements would both be "valid." Suppose Jones and Smit.½ are debating whether it is right to pluck a chicken alive, l n order to secure a somewhat tastier dish, Jones says it is; Smith says it isn't. Now suppose Jones has been so conditioned that he really would tend to feel in1partial resentment toward anyone who did this sort of thing. Further, suppose Smith, who was reared in South America, would not; he simply cannot get excited about whether chickens are plucked before or after they are killed, and in any case he is very fond of the taste of succulent chicken. In this situation, accord­ ing to \Ves-tennarck"s analysis, it really would be true and correct for Jones to say, ''That is not wrong," and for Smith to say, "That is w:ong." Pechaps this outcome sug­ gests something 'l.vrong "\Vith VVestermarck's analysis; but this is ,vhat his analysis i.-.n­ pHes, and there is nothing inconsistent about the reasoning or conclusion. But it is inconsistent to - assert non­ methoC.ological relativism and also certain other metaethical theories: e.g., nonnatur� alism, Perry's theory, the absolutist form of Ideal Observer theorv. The intriguing, question for us is


whether the thesis developed in the preceding chapter leads to the same con­ clusion. The main burden of our previous chapter was this: ( l ) The Qualified Atti­ tt::de .Method is the "stand2rd" method for evaluating ethical statements and can be· defended by good reasons, ( 2) Ethical statement� either assert or claim O!' L"llply that a corresponding attitude (for in­ stance, preference in the case of judgments about what is preferable) meets all the conditions which would be set by informed, reflective people for the endors;ment of at­ titudes governing choices in community living. Is the thesis of nonmethodological re­ lativism inconsistent with either of these assertions? Let us consider them in order, ( 1 ) Nonmethodological relativism is consistent ·.vith our conclusions about the Qualified Altitude Method if and only if it is logically possible for two conflicting ethi­ cal judgments both to satisfy the conditions of this Method. Is this logically possible? When we think of it we can see that in one respect our description of the Qualified At­ titude Method was incomplete, It told us how each of us is to proceed in order to de­ cide whether a given ethical judgment is valid-that it must jibe with our corre­ sponding attitudes in so far as they need not be "discounted," and so on, But it did not make perfectly clear whether a person's ethical judgment is vacid if it satisfies the tests as made by him but not the tests as made by other persons. One could make it a part of the Qualified Attitude Method that a judgment is satisfactory only if it meets the prescribed tests as performed by everyone, Hi o n the other hand} it is enough for the validity of a judgment made by a particular person, that i t satisfy the tests as performed by him, then i t is logically possible that con.ilicting judgments both be valid, How shall we decide this matter? It is not easv to sav what is •rsta:ndard" practice. It is ,,'t least very infrequently that we think our judgment rnee�s the conditions


we have enumerated as parts of the Qual­ ified Attitude .\Jethod, and at the same time think that a conflicting judgment by someone else meets the same conditions, as tested by bim, So it is not easy to say whether, if we did think these things, we should feel free to assert our own ethical proposition, Nevertheless, we have ob­ served above (p. 175) that there are per­ sons who think there are great variations in moral beliefs, and who do not think it likely that everybody's judgments, however corrected or qualifled, will necessarily agree on many issues. Still they are quite prepared to go ahead and make moral as­ sertions as required by consonance \ViLl'l their own criticized attitudes. They think that a person can say, like lvlartin Luther, "Here I stand; I can do no other," irrespec­ tive of information about the attitudes of others. So a formulation o.f the Qualified Attitude Method consistent ·with relativism has some support in ordinary thinking, :Moreover, there are reasons for rec01n­ mending such a formulation-reasons to the effect that adherence to an absolutist for­ mulation would paralyze moral judgment by making i t impossible to claim either a pro- or a con- judg,..nent as justified i n far too many cases, Therefore we shall not re­ gard it as part of the Qualified Attitude Method that one's ethical judgment be found compatible with the discounted atti­ tndes of ecen;body; what is required is only that it be compatible with the judge's own discounted attitudes, So far, then, it is consistent for us to assert that the Qualified Attitude �fethod is the proper test of ethi­ cal judgments, and also to assert the thesis of nomnethodological relativism, (2) \'Ve stumble into a logical dif­ ficulty, however1 when with this conclusion in mind we consider the consistency of nonmethodological relativism with our pro­ posal that ethical statements assert or claim that a corresponding attitude meets all the conditions that would be set ( a..'Cld so on). The difficulty is as follows, Suppose Mr. A makes careful use of the Qualified Atiitnde



Method, and as a result says, "x is desirable." And suppose Mr , B, after the same process, says, ,...x is undesirable." The view of the nonmethodological relativist is that this situation can really arise, that both parties really can have applied whatever unique rational method is available. Bt:t how can one person say x fa desirable, and the other correctly deny this, if our pro­ posal about the meanings is correct? At least, how can it be i.f what Mr. A is saying is, "Desiring x, on the part of everybody, meets the conditions , , ,"? Si;rely this can­ not be asserted if Mr. B's desiring of x does not meet the appropriate conditions at least as tested by Mr. B. Evidently1 if we are to be consistent relativists, \Ve must not only have a specific understanding of the Qualified Attitude �fethod ( as suggested above ), but also a particular understandi,ig of the quasi­ naturalist definition. We must specify this definition in a relativist direction1 just as the Ideal Observer theory has a relativist form (p. 173 ) , 'We can say, to take "desira­ ble'� as an example, that "x is desirable" means ��Desire for x on my part satisfies all the conditions that would b e set . . ," ( and so on ), with the understanding that the "se t conditions" may he such that desire for x on the part of one person may meet them, and desire for non-x on the part of some other person may also meet them. With this emendation, the quasi-naturalist definition is brought even closer to the non­ cognitivist counterpart described at the end of Chapter 10---t.':te view that to say that "x is desirable" is ( l) to express a desire for x, perhaps an overall impartial one, and ( 2) to claim or imply that the desire expressed satisfie.s all the conditions ( and so on, as before) , If ,ve are not prepared to understand the quasi-naturalist definition in some such manner1 we must in consistency reject rela­ tivism, It makes little difference, for the topic-s we have to discuss beyond the present chaJ:>ter, whether we make these speci-


Hcations and adopt relativism, or do not make them and accept absolutism. The rea­ sons for relativism are fairly weighty, but we shall see that the issue hardly arises in later contexts. In particular, it will not be necessary to distinguish relativist from ab­ solutist forms of the quasi-naturalist def­ inition in later discussions. We shall feel free to ignore the difference partly because nothing ,vill turn on it, and because the rel­ ativistically-minded reader can supply the emendations without difficulty ( except per­ haps in Chapter 14, where the changes must be slightly more complex) . In order to continue the argument, let us assume that we are now agreed that it is consistent to adopt certain metaethical theories ( and in particular approximately the one outlined in Chapter 10), and at the same time to be a nonmethodological rela­ tivist. Nevertheless, it may still be that non-methodological relativism is just plain false. We must now consider this possibil­ ity. How shall we decide this? Again, it de­ pends on our metaethics. Take ·wester­ marck's view. If the attitudes of all impar­ tial persons were in agreement, then one person could never truly say, "I have an impartial tendency to feel resentment against the agent of acts like x," and at the same time someone else correctly say, of the same x, "I do not have an impartial tendency to feel resentment against the agent of acts like x." Then, according to Westermarck's definitions, one person could not truly say, "x is \\rrong," when an­ other one could truly say, "x is not wrong." Hence, conflicting ethical statements would in fact never be "equally valid." Relativism would be false. Similar reasoning must be used to de­ cide \Vhether relativism is true or false, if we adopt the view that the "rational" method in ethics is the Qualified Attitude Method. Essentially the issue is this : If one informed ( and so on) attitude in fact never clashes with another attitude that is equally qualified, both of course being di­ rected at the same act or thing, then one


person can never correctly claim, ··x is \VTOng," when someone else can correctly say, "x is not wrong." Valid ethical state­ ments would then never conflict; and rela­ tivism would be false. Relativism is right, then, ac�ording to our theory ( and Westermarck's ) , essen­ tially if "qualified" ( in the sense of not re­ quiring to be "discounted" ) attitudes to­ ward the same act or event can be con­ flicting. iVell, can they, or can they not? Or what should we believe? The simplest way to answer these questions, of course, is just to find two indi­ viduals, both qualified in the relevant ways, and observe whether in fact one wants, abhors, feels obligated to do, de­ mands from others, feels indignant or­ gusted at, admires, or prefers things, ac­ tions, or events to which the other individ­ ual takes an opposite attitude. It is difficult, though, ever to be certain that such indi­ viduals are before us. How can we be sure that all the relevant facts are believed by both, and that neither needs to be disa­ bused of false beliefs? How can we be sure that all the relevant considerations are pres­ ent to the minds of both, with requisite viv­ idness? Perhaps, of course, individuals on occasion may '\Vith reason be said to ap­ proximate to these conditions. It seems preferable, however, not to rest one's argu­ ment on such possible cases. There is an indirect method for answering our question. Consider a paral­ lel: that we feel free to make statements about how gases would behave at an abso­ lute-zero temperature, although we have not actually observed gases in this state. vVhy? \rVe draw inferences from relevant causal laws. The same is true in our case. If we have good reason to believe causal laws, to the effect that a person's attitudes are not a function solely of his information ( or its vividness ) and his state of personal needs or wishes ( at the time) and his nor­ malcy, then we have so much reason to think that "qualified" attitudes occasionally


vary. If we happen to know precisely the nature of these laws, we may be able to specify the conditions under which such variation will occur. Psychological theory and experiment, then, are the most obvious source for an answer to our question. Unfortunately, psychological theories do not provide a uniform answer to our question. Gestalt theory would lead us to believe that attitudes ( "ought" experi­ ences ) to a situation will be identical, if the situation is identically understood, and personal needs and interests do not play a distorting or blinding role. Psychoanalytic theory and Hullian learning theory, how­ ever, provide a different answer. Accord­ ing to these theories, two attitudes, equal­ ly «qualified" in the sense of occurring in minds with equal information ( and so on ) , can b e conflicting, depending on the his­ tory of the development of the persons: their past identifications, their past revmrds and punishments. The doctors, then, dis­ agree. But how does the currently available experimental evidence look? Does it fa­ vor the vie,v of either theory, on this par­ ticular point? To this our answer must be: There is no certainly correct reading of the evidence, but it appears to favor the relati­ vist answer to our question, for there is some reason to think that fundamental ori­ entations may be adopted from parents in early life, and that these may have a per­ manent influence on attitudes; that identi­ fications, emotional relations with impor­ tant figures in one's life, and feelings of security play a role in the development of one's values; that certain things or events maybe highly valued in compensation for the inaccessibility of other satisfactions at an earlier period, or as a result of depriva­ tions. Then, if thes� things are true, we can specify some occurrences in the life of an individual that would have the effect that his attitudes now, whatever the informa­ tion ( and so on) of his present state of mind, would be different from what they would have been had his earlier exper­ iences been different. Individuals with rele-


vantly different earlier experiences, then, may be expected to have different at­ titudes, despite identical qualifications with respect to knowledge, impartiality, and so forth. ' On the whole, then, the relativist is better able to claim the support of contem­ porary psychological theory and research than is his adversary. However, the issue is not closed. The facts of anthropology are also rel­ evant to our question, and in the following way. In the first place, we bave already no­ ticed ( p. 109) that studies of cultural change in primitive societies suggest that facts like personal conflicts and maladjust­ ments, the attitudes of one's close relatives ( for example, whether favorably oriented toward ,Vhite civilization ), and personal success in achieving status in one's group or outside one's group ( for example, \\�th White men) play an important role in the development of the values of adults. This finding is some support for our reading of the observational evidence of psychology. In the second place, there is the fact that various groups have different values. The mere fact that different ethical standards exist in different societies, of course, by it­ self proves nothing relevant to our present problem. Nevertheless, something impor­ tant is proved if the facts bear testimony that different standards can prevail even if different groups have the same beliefs 8 Vile should not, however, overlook the possibil­ ity that an individual might, if he knew that an attitude of his was a result of some type of early experience ( for example, a high valuation of knowledge being a result of the unsatisfactoriness of his personal relations at an earlier period of development ) , to some extent lose this attitude. In other words, perhaps self-understanding in the sense of understanding the genesis of one's own values is a fact relevant to what one's present attitudes will be: It is possible that any two individuals, otherwise equally "qualified," would in fact always have the same attitudes toward everything at the conclusion of a careful psycho­ logical treatment in which each acquired complete self-understanding. Is there evidence, from psycho­ analysts or other specialists in personality theory, that points in this direction? The writer does not know.


about the relevant event or act, and if there is no reaso::1 to suppose that the group standards reflect group differences in re· spect of other "qualifications." ( We must remember that attitudes common to a group cannot usually be discounted as be­ ing a result of personal interest or of an ab­ normal frame of mind.) The fact of varia­ tion of group standards, in these circum­ stances, would tend to show that attitudes are a function of such variables, that atti­ tudes could differ even if our "ideal quali­ fications" were all met. Is there such variation of group stand­ arC.S? YVe have seen that there is one area of ethical opinion where there is diversity in appraisal arid at the same ti"lle possible identity of belief about the action-that about the treatment of animals. On the whole, primitive groups show little feeling that it is wrong to cause pain to animals, · whereas the columns of The New York Times are testimonv to the fact that many persons in the U.S.A. take a vigorous inte;. est in what goes on in slaughterhouses. \Ve have alreadv mentioned some details about the attitud;s of primitive groups (p. 103 ) . Nevertheless, we cannot be sure that atti­ tudes of the groups here in question really do fulfill our "qualifications" equally well. Primitive peoples rarely make pets of the animals tbey maltreat. There is at least some question whether they have a vivid imagination of what the suffering of an ani­ mal is like, comparable to that of the au­ thors of letters to the Times. The writer has assured himself by personal investigation that there is no definite discrepancy be­ tween the Hopi beliefs, about the effects of maltreating animals, and those of what seems a representative sample of educated White Americans. Degrees of vividness of


belief, however, do not lend themselves to objective investigation, and it is not clear how we may definitely answer questions about them, either yvay. Perhaps :he sanest conclusion i s just to say that, as far as can be decided objectively, groups do some­ times make divergent appraisals when they have identical beliefs about the objects, but that the difficulties of investigation justify a healthy degree of skepticism about the conclusiveness of the inquiry. The fact that objective inquiry is dif. £cult naturally works both ways. It pre­ vents us from asserting confidently that, where there are differences of appraisaI1 there is still identity of factual beUef. But equally it prevents. us from den)ing con­ fidently that there is identity of belief, where appraisals differ. The anthropological e,idence, taken by itself, then, does not give a conclusine answer to our qr_;_estion. At the present time, foe anthropologist does not have two social groups of which he can say definite­ ly: "These groups have exactly the same beliefs about action A, on all points that could be seriously viewed as ethically relevant. But their views-attitudes-about the morality of the acts are vastly dif­ ferent," vVhether, everything considered, the relativist reading of the facts is not the more balanced judgment, is another ques­ tion. The \\Titer is inclined to think it is the better judgment. If we agree that the ethical standards of groups are not a function solely of their beliefs ( or the vividness of these ) , it is rea­ sonable to suppose that "ideally qualilied" attitudes may well conflict ,vith respect to the very same act or event. To say this is to say that foere is reason to suppose that nonmethodological relativism is correct.

The Justification of Value Judgments: Rational Choice PAUL TAYLOR A. THE CONCEPT OF A WAY OF LIFE I have defined a way of life as a hierarchy of value systems in ,vhich each system be­ longs to a different point of vievv. Since a value system is nothing but a set of stand­ ards and rules arranged according to their relative precedence, it follO\vs that a \vay of life is simply an organization of different sets of standards and rules. These sets ( value systems ) are in tum arranged ac­ cording to their relative precedence. How is their relative precedence determined? In order to answer this question ,ve must £rst consider what it means for a value system to be relevant to a situation and to be in confilct with another value system. It is only when two value systems are both rele­ vant to a situation and are in con£ict with each other that one can be said to take precedence over the other. In Chapter 5 I gave as an example a situation to which an aesthetic value sys­ tem and an etiquette value system are irrel� evant and to ,vhich a moral value system and a prudential value system are both rel­ evant. It was a situation in which one's own life and the lives of others are in dan­ ger and one is confronted ,vith the choice of whether to risk one's life to help others. Paul W. Taylor, NORMATIVE DISCOURSE, © 1961. Reprinted by permission of Prentice-Hall, Inc., Englewood Cliffs, New Jersey. 346

Now the fact that aesthetic considerations and considerations of etiquette are not rele­ vant to such a situation is a fact about a person's way of life. Another person with a different way of life might hold that they are. In the act of committing himself to a way of life, a person subscribed to the prin­ ciple that, if his own life and the lives of others were in danger, it would be irrele­ vant to use the standards and rules of aesthetics or of etiquette in deciding what to do. Another person, in committing him­ self to a way of life, may have subscribed to the opposite principle. We .cannot s�y ,vhether such value systems "really'' are rel­ evant or irrelevant to the situation. VVe can only decide the question on the basis of a given way of life, and different ways of life will yield different answers. What, then, does it mean to say that a value system is relevant to a situation? It is to say that, according to a certain way of life, the standards and rules of that system are to be used to guide the · choices and regulate the conduct of those in the situa­ tion. And this means simply that the stand­ ards and rules in question include the situ­ ation in their range of application. Accord­ ing to the given ,vay of life, it is legitimate and proper to judge the choices and con­ duct of people in the situation by the standards and rules of the value system. Conversely a value system is irrelevant when its standards and rules do not cover


the situation in their range of application, and so cannot be used to judge choices or conduct in the sitnation. It is possible for two vaiue systems, each belonging to a dilferent point of view, to be relevant to a situation but not to be in conflict. They do not conflict when it is possible for a person's choice and conduct to be in accordance with the standards and rules of both systems. Two relevant value systems are in con:8kt1 on the other hand, when a person's adopting one system in the situation prevents him from adopting the other, that isi when the standards or rules 0£ one system are in conflict with those of the other. From Chapter 3 we know that one standard conflicts with another when a feature of something which is good-making according to one will be bad-n1aking ac­ cording to the other. That in so far as an object fulfills one standard it fails to fulfill the other. And we know that one rule con­ flicts with another when acts which are right according to one are wrong according to the other. There are different degrees to which two value systems may. be in conflict, depending on how many of the standards and ruJes of one are in conflict with those of the other. That two dilferent value systems can be relevant to a situation and not in conflict may be illustrated as follows, Sup­ pose we are judging a pabting from two points of ,,iewJ the aesthetic and the econ­ omic, vVhen we judge it aesthetically, we apply the standards of a certain aesthetic value system we have adopted. When we judge it economically, we are interested in its worth as an investment and vve apply the standards of an economic value system we have adopted. The painting may be aesthetically good and a!so a good invest­ ment; it may be aesthetically bad but a good investment; it may be aesthetically good but a bad investment; or it may be bofa aesthetically bad and a bad invest­ ment, The two value systems are thus logi­ cally independent of each other. They may be said to be divergent, but not to be in


conflict. Even when they diverge ( i.e., when the painting is judged as aestheti­ cally good and economically bad, or vice versa) that which makes the painting aesthetically good is not that which makes it economically bad. In other words, the features of the painting which are good­ making characteristics from the aesthetic point of view are not the same features which are bad-making characteristics from the economic point of view. As a result; our judgment that the painting is heautiful does not entail that we buv it, and our judgment that i t is a bad in�estment does not entail that we find it aestheticallv dis­ pleasbg. Since the aesthetic system a�d the economic system are not logically con­ nected, to increase ( or decrease) the aes­ thetic value of something is not eo ipso to increase ( or decrease) its economic value, althoc:gh there may be a causal relation be­ tween the two. Therefore they cannot be in conflict. Fm· an example of two value sys­ tems in conflict, I would cite the ca?e dis� cussed above, where one's O\Vfi life and the lives of others are in danger. Here a moral valc.e system may indeed conflict with a value system of self.interest, since to act from self-interest would involve acting im­ morally ( i.e., violating a moral rule ) , and acting moraI!y would involve sacrificll1g one's ovvn interests. It should be noted that the decision as to whether two value systems conflict in a given situation does not depend on a way of life, but on the nature of the value sys­ tems themselves, They conflict when their constituent standards and rules conllict, re­ gardless of the way • of life that contains them. It is true tha-i: con:f:Uct does not a1ise unless the way of life allows the tv,o value systems to b e relevant to the same situa­ tion, But once this is so, then whether or not they conflict is not determined by a way of life. Let us no\v suppose that two value systems are in conflict. If we ask which sys­ tem takes precedence over the oilier, it is to the principles of a way of life that we



must refer for an answer. Different ways of tern take precedence over another when life will entail different answers, According they are in con:Sict Since each value sys­ to one, value system V will take prece­ tem that is part of a way of life belongs to dence over- value system V'; according to a different point of view, it may he thought another, V' will take precedence over V. that the commitment to a wav of life also There is nothing in the value systems them­ decides which point of oiew shall t11ke pre­ selves which renders them superior ( or in­ cedence over another. But the roost we can ferior) to others. We cannot show that a say is that a wav of life determines indi­ moral value system ahvays or necessarily re�tly the relativ� precedence of points of takes precedence over a prudentfal value vie;,v. ¥Ve can speak of one point of vie\.V system, for example, merely by examining ( say, the moral) taking precedence over the value svstems themselves. \Ve cannot another point of view ( say, the pruden­ even show this by analyzing the two points tial ) , hut only in the sense that, given a sit­ of view to which the value systems belong. uation in which a person must decide be­ The canons of reason:ng which define one tween acting morally and acting in his point of view do not stipulate that any self-interest, the person's way of life deter­ value system whicb is guided by them shall :nines that the moral system takes prece­ he superior to any value system guided by dence over the prudential system. In anoth­ another set of canons. Such a stipulation er situation, or with another way of lifet it can only be made outside all points of might be the case that self-in'terest will vie\\\ as a principle to which one sub­ take precedence over morality. Thc.s we . scribes as part of one's way of life. Thus we Cannot ask. "Does ori.e point of view in gen­ cannot say that the moral point of ,1ew eral take urecedence over another?.. We takes preeedence over the prudential point can only �sk "Does this particular value of view unless we have committed our­ system belonging to one point of view take selves to a way of life in which a moral precedence over that particular value sys­ value system tilkes precedence over a pru� tem belong' to another point of view?" dential value system '-Vhenever the two And this question can be answered only conflict. Uie mere fact that the one systen1 relatively to a way of life. Hmvever, under is moral does not make it take precedence special conditions it would be possible to over another svstem, Indeed, there is at make a generalization> albeit a somewhat least one way �f life, actually practiced by misleading one) concerning the relative a culture, which does not claim superiority precedence of points of view. Suppose, for on behalf of its moral system over the example, -that there is one way of life value system of prudence or self- i nterest. which is always preferred to every other This is the culture of the Navaho Indians. way of life on the basis of a rational choice. ( This has been argued by John Ladd in And suppose that, according to this way of The Structure of a Moral Code. Cam­ life 1 whenever a moral value system con­ bridget Mass.: Harvard University Press, flicts with a prudential value system the 1957, esp. pp. 212-213 and pp. 292-296. ) moral system takes precedence. We might However, for the Navaho there is perhaps then say, still somewhat misleadingly, that no conceivable conflict beh:veen 1norality the moral point of view takes precedence and prudence, in light of "the general Na­ over the point of view of self-interest. ( The vaho presu::uption that the \Ve!fare of oth­ relative precedence of points of view is dis­ ers is a necessary condition of one's O\VT1 cussed further in Section B of Chapter 11.) welfare." ( Ibid., 296. ) In summary, to commit oneself to a The commihnent to a way of life in� way of life is to subscribe to certain princi­ valves the decision to make one value s ys- ples. These principles are of two types:


THE JUSTIFICATION OF VALl'E Jl'DGMENTS principles of relevance and prineiples. of relative precedence. iVhen we subscribe to a princi;,le of the first type, we decide which vake systems shall be relevant to a certain kind of situation and which shall not. In choosing a way of life we nwke a given system relevant or not relevant to a given situation. Similarly, when we sub­ scribe to a principle of the second type, we decide that one value system shall take precedence over another in a situation where they conflict and to which they are both relevant. In c.'1oosing a way of life we stipulate the relative precedence of our value systems, Thus we cannot answer the question why a certain .value system is rele� vant or ,vhy it takes precedence over an­ ofaer. \Ve can only say that these simply are the principles to which we subscribe in virtue of the fact that we are committed to a particular way of life. In the very act of comrnitting oursel Ves, we make value sys­ tem V relevant to situation S and we make value system V take precedence over value system V'. We cannot give reasons for claiming that V is relevant to S or that V takes precedence over V'. We can only say we have chosen that way of life. Such a choice is our ttltirnal'e normative cmnmit­ ment. The only kind of reasons which can be given to justify the principles of a way of life are reasons which justify the way of life as a whole. As we shall see, such rea­ sons consist in showing that the way of life is rationally chosen. Variation in ways of life depends on the particular principles of relevance and relative precedence which define each. From this variation in wavs of life them­ selves; vie can distinguish 'another kind of variation-a variation in commitments to a way of life. Thus we 1nay classify varying commitments according to ( a ) their degree of coherence and stability, ( b ) their degree of depth, ( c) their degree of conventional­ ity or unconventionality, and ( d) their de­ gree of explicitness. ( a ) I have been speaking up to now as


if everyone commits himse;f to just one way of life, striving to lJve by its principles and to realize its idealE throughout his life. The fact is, hO\vever, that individualc; vary widely in the coherence and stability of their ultlntate commitments. In a society at large, there will always be a n10re or less definite way of life to which most members of the society have been committed by being brought up within it. They have :10t committed themselves as a matter of choice, but have been co�mitted bv others. If they remain uncritical and con,;entional in their outlook, they will have a coherent and stable of life. But there will be in­ dividuals who for one reason or another will come to doubt the way of life in which they have been brought up. They might then choose a new way of life which will be just as ' of . coherent and stable a,; the wav life of their society. On the other hand, they might become disillusioned \vith their new way of life. They might shortly .find themselves committed to another, which again might turn out to be unsatisfactory to them, To the extent that a person's commit­ ments are in this vvay ten1porary and con­ stantly changing, to that extent they lack coherence and stability. The exireme of in­ coherence and instabilitv is reached whe:1 an fodiv1dual has no way of life of his own. He lives without principles o: ideals. 'What­ ever standards and mies be does follow are not organized into unified value sys­ tems. He lacks second-order norms hv which to determine the relative precedenc� of other norms. And h e app:ies his norms inconsistently, sometimes having pro­ attitudes toward the very same things which at other times are the ohjeets of his con-attitudes. From this extreme there is a c_ontinuum of commitments of increasing coherence and stability until we arrive at 2 single:1 all-embracing1 permanent ,vay of life, in which the value sv.stems of all points of view are integrated in a consist­ ent hierarchy. ( b ) Commitments to ways of life may


vary not only in degree of coherence and stability, hut also in degree of depth or throughness. A strong-minded, deeply con­ vinced, thoroughly committed person will strive to live according to his way of life under all circumstances. He will exert every effort to fulfill its ideals, even at great cost to his own comfort or safety. A person who is not deeply committed to a way of life will he only weakly motivated to live by its principles and ideals, and will frequently fail to adhere to the standards and rules involved in it when it is not to his immediate advantage. ( I shall consider depth of commitment further in Chapter 12.) ( c ) Commitments to ways of life also vary in degree of conventionality or uncon­ ventionality, according to their agreement or disagreement with the way of life of the general culture or times. An individual's way of life is conventional to the extent that it agrees with the way of life of his family, of his religious background, of his economic and ,social class, and of the var­ ious groups to which he belongs. ("We shall see that the conventionality or unconven­ tionality of a way of life has nothing to do with whether it can be rationally chosen. ) ( d ) Another dimension in which com­ mitments to ways of life vary is in their de­ gree of explicitness. A person might live fully in accordance with the ideals and principles of a way of life and yet not be able to tell someone what they are. Such a person would be unable to make his way of life explicit; he could not articulate, either to himself or to others, bis basic beliefs and "values." V\Te might say of him that he lives as if he believed in certain ideals and prin­ ciples. His commitment is implicit, . not ex­ plicit. Another person might be able to state clearly and coherently what his way of life is. If he is of a certain bent and has certain abilities, he might even write out his "philosophy of life" in a book. He might also preach it to others and try to get them to become committed to it. He might ac­ cordingly be proclaimed a prophet or wise


man, or else ( depending on the social con­ ditions of bis time and on the nature of his way of life) a fanatic, a crank, a reformer, or a demagogue.

B . ABSOLUTISM AND RELATNISM Having considered what it means to be committed to a way of life, we are now prepared to continue our inquiry into the justification of value judgments. vVe have reached the fourth and final stage of such justification: the rational choice of a way of life. We have seen that the commitment to a way of life is an ultimate commitment, If we ask someone to justify his value judg­ ments and he appeals to standards or rules, and if we ask him for reasons for accepting his standards or rules and he validates them, and if we then ask him to justify his entire value system and he vindicates it, he must finally refer to his whole way of life. There he takes his final stand. The question now before us is: How can this ultimate commitment itself be justified? This ques­ tion, it seems to me, lies at the heart of the controversy between absolutism and rela­ tivism in values. No one has brought out more clearly the fact that commitment to a way of life is an ultimate commitment than R. NL Hare. In the following passage Mr. Hare gives his account of the justification of decisions and principles. ( The term "principles" covers not only what I have referred to as stand­ ards and rules, but also what I have called the ideals and principles of a way of life. ) . . . A complete justification of a decision would consist of a complete account of its effects, together with a complete account of the princi­ ples which it observed, and the effects of observing those principles . . . . Thus, if pressed to justify a decision completely, we have to give a complete specification of the way of life of which it is a part. This complete specifica­ tion it is impossible in practice to give; the nearest attempts are those given by the great


religions, especially those which can point to historical persons who carried out the way of life in practice. Suppose, however, that we can give it. If the inquirer still goes on asking 'But why should I live like that?' then there is no further answer to give him, because we have already, ex hypothesi, said everything that could be included in this further answer. \i\Te can only ask him to make up his o,vn mind which ,vay he ought to live; for in the end everything rests upon such a decision of prin­ ciple. He has to decide whether to accept that way of life or not; if he accepts it, then we can proceed to justify the decisions that are based upon it; if he does not accept it, then let him accept some other, and try to live by it. ( R. M. Hare, The Language of Morals, p. 69.)

\Ve have considered a sommvhat par­ allel situation with regard to the vindica­ tion of value systems. There, however, an appeal is made to "principles" in a wider context, for we found that a value system can be vindicated in terms of the ideals of a whale way of life. But there is no such wider context to refer to when we are asked to justify a way of life itself. Hare says that we can only ask the person to try to live by it. But is there no way to show that one person's choice of a way of life is more intelligent or enlightened than anoth­ er's? Hare himself argues that the choice of a way of life is not an arbitrary decision.

To describe such ultimate decisions as arbi­ trary, because ex hypothesi everything which could be used to justify them has already been included in the decision, would be like saying that a complete description of the universe was utterly unfounded, because no further fact could be called upon in corroboration of it. This is not how we use the ,vords 'arbitrary' and 'unfounded.' Far from being arbitrary, such a decision would be the most well­ founded of decisions, because it would be based upon a consideration of everything upon which it could possibly be founded. ( Ibid., p. 69 .) In light of these remarks, it would seem that there is at least one condition for


rationality in making a "decision of princi­ ple" to commit oneself to a whole way of life, namely the condition that one know what is involved in all the alternative ways of life among which one is choosing. I shall later specify what I consider to be the ne­ cessary and sufficient conditions for a ra­ tional choice among ways of life, and this condition will be included among them. Before doing this, however, it is important to clarify the general nature of a rational choice. The context in which I am dealing with such a c"p.oice is the context of justi­ fying a value judgment. My purpose is to make clear the logic of our reasoning when our value judgments are consistently chal­ lenged. In setting forth the conditions of a rational choice among ways of life, I shall not be trying to describe a situation of choice which would actually confront someone in everyday life. I shall be con­ structing a concept of a rational choice in the abstract. This concept is designed to answer the question, "\i\That sort of commit­ ment to a way of life would anyone on re­ flection be willing to call a justified one?" The question is not, "How do people ac­ tually come to commit themselves to a way of life?" Nor is it, "What would be the psy­ chological grounds ( causes ) for a person's choice if he actually were confronted with alternative ways of life among which he were asked to choose?" The philosophical question with which we are concerned arises only when we try to push back our defense or support of a value judgment as far as it will go- to its ultimate foundations, as it ,vere. The logic of our thinling moves step by step from the judgment to a standard or rule, from a standard or rule to higher standards or rules, from those to the highest standards or rules within the framework of valida­ tion, from there to the ·whole value system which sets that framework, and from there to the whole way of life in terms of which the value system is vindicated. Here we stop, until we notice that there are many


different and conflicting ways of life. We see that if we commit ourselves to one we will be able to justify our value judgment, but that if we commit ourselves to another we will be able to show that our value judgment is unjustified. It is then that we ask ourselves if all ways of life are equally justified. Is it not possible to give good rea­ son for accepting one way of life rather than another? The issue raised here is that which many philosophers _call the issue between relativism and absolutism. Are values rela­ tive or absolute? If we can trace the logical foundations of our value judgments back to our commitment to a certain way of life but cannot justify this commitment itself, then all values are said to be relative. They are relative to our way of life. ( :tv:Iore accu­ rately, they are relative to our value sys­ tems, and our value systems are themselves relative to our way of life.) This relativistic position holds that, if we find ( as we do ) that different societies and cultures have conflicting ways of life, then the struggle between them is a matter of brute force, unless they voluntarily decide to tolerate each other's differences. No rational choice can be made between them. If one has a moral code ( i.e., a moral value system ) which contradicts the moral code of the other-so that acts of a certain sort are right in the one culture and wrong in the other- we cannot talk about acts of that sort being "really" right or "really" wrong. They are right ( not merely believed to be right ) in one culture and ,vrong in the other. Good reasons can be given for doing the acts, if one accepts the supreme norms of the value systems as vindicated by the way of life of the first culture. Similarly, good reasons can be given against doing the same acts in the framework of the way of life of the second culture. Since good reasons cannot be given in support of one whole way of life rather than the other, any "good reasons" are relative to value systems and finally to ways of life. They are valid only in so far as one adopts a cer-


tain value system and with it a certain ,vay of life. vVe cannot escape this kind of relativ­ ism by arguing that the canons of reason­ ing which define normative points of view are equally as rational as the canons of rea­ soning which define the scientilic, the mathematical, and the historical points of view. It is true that just as we can explicate the canons of reasoning that govern the lat­ ter and thereby reveal the ideal of rational­ ity implicit in those ways of thinking, so we can explicate the canons of reasoning which govern the various nonnative points of view and thereby disclose the ideal of rationality implicit in them. The relativist can grant this and still claim that our ina­ bility to justify a whole way of life opens the way for skepticism, and destroys the ra­ tionality of normative thinking as a whole. His argument is based on the fact that the canons of reasoning which govern the scientific, the mathematical, and the histor­ ical points of view are independent of what­ ever way of life a scientist, a mathemati­ cian, or a historian may be committed to. A physicist will appeal to the same sort of evidence to ·confirm a hypothesis whether he be a Hindu or a Christian, a Communist or a Capitalist. But the reasons which he appeals to in trying to justify his value iudgments will vary according to the way of life he is committed to. Unless his way of life can itself be justified, the reasons which he gives will not make a universal claim on all men. The relativist would have to admit that, on the levels of verilication and validation, taking a certain point of view would entail the same rules of valid inference and relevance for everyone, no matter what his way of life may be. Hence to this extent a person can claim rationality on behalf of his way of justifying value judgments. It is only when we reach the level of vindicating a whole value system that the skeptical power of relativism is felt. At this level we can no longer appeal to the universality of our canons of reason­ ing as a sufficient ground for claiming that

THE JUSTIFICATION OF VALCE JUDGMENTS the justification of value judgments is a ra­ tional process. The universal canons at this level are the ru1es of valid inference which define the process of vindication itself. That Ls, they are the mles according to which we justify a value system by show­ ing that it has contributive and instrumen­ tal value to a way of life. But such rules allow for the sarn; value system to be both iustified and tmiustified, since a value sys­ tem ,vhich has contributive and instrwnen­ tal value to one way of life can have contri­ butive and instrumental disvalue to an­ other. Under the relativist's assumption that ways of life cannot rationally be justified, it seems vve must make an importan½ indeed a damaging., quali6cation in our claim to know what is good or bad, right or wrong. We must admit that such know;edge varies from way of life to way of life, whereas scientific, mathematical, and historical kno\vledge do not. This difference is suf­ ficient to make us doubtful about using the word "knowledge" at all in connection with value judgments. A similar doubt arises concerning our use of the words "tme" and "false." A value system may be vindlCated in one culture and an opposite or confiicting value system rnay be vindi­ cated in another culture, since each culture embodies a different way of life. This seems to imply that a given value judgment may be true in one culture and false in an­ other, Indeed, it is possible to define valua­ tional relativism as the vievv that no value judgment is simp!y tn1e or simply false, but is only true for someone { or for some group or culture ), and false for someone else ( o:- for some other group or culture ), Nobody would say this about scientific, mathematical, or historical state::nents. A scientific) mathematical, or historical state­ ment is simply ( absolutely, genuinely ) true or false, To say that it is true for someone means only that .someone believes it or thinks it is true. Value judgments may also be said to be true for someone in this sense. B�t this is not what the relath.rist is getting


at when he makes such a claim. He means that the truth of a value. judgment is rela­ tive to son1eone's ( or some group1s or cul� ture's ) way of life in such a ,vay that the same judgment will be false relative to an­ other person's ( or group's or culture's ) way of life. Valuational absolutism, on the other hand, claims that a value judgment is sim­ ply true or false, not trne or false for some­ one. It is true when it can be shovm to be justified, false when it can be sho,11n to be unjustified. It is shov,m to be justified when it is verified by appeal to a standard or rule which can be validated within a value svs­ tem, which in turn is vindicated bv refer­ ence to a way of life, and this way of life can rationally be prejerred to all others. If the way of life which vindicates the value system can ratioually be shown not to be preferable to some other way of life, then the value judgment is false, To say that the judgment is true or false is to say that it is really true or false, quite a;ide from whether or not people thL"lk so. Of course its truth or falsity does depencl on a given value system, but not all value systems are equally justifled. Some may he vindicated in terms of a non.rational way of life. Hence value systems do not fully provide a justi� fication for a value judgment. A value judg­ ment is completely justified ( i.e., it i s as justified as it can be ) when the value sys­ tem within the framework of it is verified and validated is itself vindicated by reference to a rational way of life. Only such a value judgment can be claimed to be really true. Its truth in some respects will be different from scientific and mathe­ matical truth, although> like the latter, it will be independent of cultural variation. A value judgment will be true even tl1ough its truth is not recognized by a whole cul­ ture ( just as "The earth is a globe" was true even when people thought that the earth was Hat) . For the vaJuational absoh::tist, what is the difference beD:veen the truth 0£ value ;udgments and the trutb of scientific or

mathernatical statements? The answer lies in how they are justified. Scientific and mathematical statements are completely justified in terms ofthe canons of valid rea­ soning set by the scientific and mathemati­ cal points of view. To decide to take such a point of view is to decide to reason in a certain ,vay, that :is, according to certain rules of inference. What is found to be in­ tellectually acceptable according to these rules will be designated as true; what is found, under the same rules, to be intellec­ tually unacceptable will be false. The can­ ons of reasoning ,vhich gov�rn the com­ plete justification of a value judgment, on the other hand, require the steps of vindi­ cation and rational choice. These latter do not correspond to any steps used in veri­ f)>ing scientific statements or ln pro,>ing mathematical statementt::. The verification and validation of value judgments are car­ ried out within the framework of a value system according to the canons of reason­ ing which define a normative point of view; they may be compared generally \Vith empirical verification and deductive reasoning in nonnormative points of view. But as ,ve have seen, the value system re­ ferred to in the verification and validation of value judt,tznents must itself be justified. The method used is that of a pragmatic test, ,v rhich in tum makes reference to a way of life. There is no similar test and no similar reference in scientific and mathe­ matical reasoning.

C. THE CONCEPT OF A RATIONAL CHOICE I shall now proceed to give the argument for valuational absolutism. To do so, I shall try to explicate the concept of a rational choice among ways of life, and thereby show that the preference for one way of life ra-U1er , than another i,; not arbitrar;. This is the fourth and final stage in the total justification of a value judgment. It can be accomplished by showing that when-


ever anyone is confronted with a situation ln which he is to choose between a given way of life and other ways of life, and whenever that situation satisfies the neces­ sary conditions '.for a rational choice, - the t,>ivcn "'"'Y of life will be chosen in prefer­ ence to any other. The necessary conditions for a rational choice among ways of life must be spe· cified. I suggest that these conditions may be grouped under three general headings: conditions of freedom, conditions of enlight­ enment, and conditions of impartiality. I shall say that a choice is rational to the ex­ tent that it is free, enlightened, and irnpar­ tial. Each of these conditions sets up an ideal. No actual choice can ever be com­ pletely free, completely enlightened, or completely impartial. Hence no choice ac­ tually made among alternative ways of life can be fully rational. As I have already pointed out, the concept of a rational choice is the concept of an ideal. In de­ scribing it, I am trying to explicate one of the - canons of rationality-to state what would be considered by anyone to be a ra­ tional choice, i'f such a choice were ever to occur under ideal conditions. I am not trying to describe any actual choice made · by someone. All that is necessary for ulti­ mately justifying a value judgment is that a meaning be given to the concept of rational choice that will make explicit the assump­ tions underlying the way a fully rational person would think in the given context. My claim here is that to the extent that any actual choice fulfills the conditions of ra­ tionality which I shall state, to that extent it justiiles the way of life chosen and conse­ quently can be used to justify a value judg­ ment. I sha.ll now specify the conditions of rationality under the three headings men­ tioned. l. Conditions of freedom. A choice is free to the extent that: ( a ) The choice is not decisively deter­ mined by unconscious motives. That is, if unconscious u10tives do have a role as psy­ chological determinants of the person's


choice 1 their role is not decisive. 'i\lhat -is decisive will be given in tbe fourth condi­ tion of freedom, stated below under ( d ) . ( b ) The choice is not at all deter­ mined by internal coristmint. That is, tbe person who makes the choice is under no element of compulsion, wheL½.er of irresisti­ ble impulse or extreme desire. He is calm and collected, in complete control of him­ self, ( c) The choice is not at all deter­ m.inecl by external constraint. That is, there is nothing in the physical or social environ­ ment of the person to compel him to make a choice. So far as the physical environ­ ment is concerned, the person is not in any immediate pbysical danger and does not have to suffer any immediate physical harm when he makes his choice. ( Some of the \Vays of life among which he is Ghoos:ing, however, might entail more physical dan­ ger and suffering than otbers.) So far as the social environment is concerned, no so­ cial pressure is being brought to bear on hi.'11 to choose one alternative rather than another. No one is goading him, threatenhim, or trying to intimidate him. He is not under any form of coercion or ( d ) The choice is decisively deter­ mined by the person's own preference. That is, his choice follows upon his prefer­ ence, though in order that his choice be r a ­ tional his preference must be enlightened and impartial ( as spelled out below) . To say that a choice is free is not to say that it is uncaused or undetermined. It is rather to say that the choice is the result of, and hence determined bv, the individual's mak­ ing up his own mind about the matter. I do not call this process deliberation, however , Deliberation is evaluation, and all evalua­ tion is made according to standards or rules. VVe are concerned here with a prefer­ ence, not an evaluation. There are no standards or rules to appeal to, since \Ve are dealing with a choice a111ong whole ways of life and all standards and rules are included in these. l f we were to appeal to a standard or rule, we would be presuppos-


ing a way of life in making the choice, and therefore our choice would not be a. choice among ways of life. 2. Conditions of enlightenment. A choice is enlightened to the extent that: ( a) The nature af each way of life is fr,lly k'tlown. ( b) The probable effects of lic'ing each way of Ufa are fully knoum. ( c) The means necessary to bring about each way of life (i.e., what is required to enable a person to live each way of life) are fully h'IlOWn,

There are three kinds of knowledge in, valved in each of these conditions: intellec­ tual knowledge 1 imaginative knowledge, and practical knowledge. Intellectual knowledge of a way of life includes all the empirical knmvledge necessary for a com­ plete and accurate des&-riptio,, of the way of life itself, of the probable effects of liv­ ing according to i� a.."'Jd of the necessary means for bringing it about. Such knowl­ edge must provide us ,vith anssvers to a whole series of questions. What is it like for a person ( or group) or culture ) to live the way of life? What are the ideals of the 1-vay of life; what vision of the summum bonum does it embody? 'vVhat value sys­ tems are relevant to different sorts of situa­ tions and what value systems take preced­ ence over others in those situaiions? Wbat wocld be the psychological and social consequences of a person's { or groupfs, or culture's ) living that way of lifo? What physical, social, and psychological condi­ iions must be realized before a person ( or group, or c,tlture) would he able to live the way of life? In addition to the knowledge that the empirical sciences must provide, inte11ec� tual knowledge of the nature of a wav of life must also include philosophical kn�wl­ edge. \Ve must know the canons of reason­ ing that constitute the point of view to ,vhich any value system in the way of life belongs. ·what are the rules of relevance and tl1e rules of valid inference which gov-


em the justiJlcation of judgments, stand­ ards and rules within the framework of each value system? That is, according to what criteria is a reason a relevant reason or a good reason in such justification? Intellectual knowledge of a way of life, in short, consists of all the scientific and philosophical knowledge that can pos­ sibly enlighten us concerning the value sys­ tems and the points of view which consti­ tute it. A person does not make a rational choice among ways of life, however, if he merely has intellectual knowledge of them. Imaginative knowledge is also necessary. He must be able to envisage what it is like to live each way of life. He must be able, by imagination, to convey himself into each way of life and experience it vicariously. Short of actually having lived a way of life, there are four particularly effective means for developing this imaginative knowledge: through personal contact with people who live the way of life, through the reading of history, biography, and to a lesser extent anthropology and sociology, through the study of religion, and through appreciation of the fine arts. vVhen one has personal friends or ac­ quaintances who live a certain way of life, or better, lives among people who follow it, one's imaginative insight into it is in­ creased, even if one does not share that way of life with them. Constant contact with people makes us subtly aware of their interests, attitudes, points of view, and as­ pirations, so that we can sense intuitively the way they look upon the world. The closer we get to people and the better we come to know them, the deeper becomes our understanding of their way of life. We can also increase our ability to en­ visage a way of life by reading the history of peoples who have embodied it in their culture, or by reading a biography of an in­ dividual whose life exemplifies it to a high degree. This is especially helpful when the historian or biographer is sympathetic to the way of life. If an individual's letters, es­ says, and speeches have been published,


reading these will also help to convey to us his way of life. Anthropological and socio­ logical studies of whole cultures or of sub­ cultural groups sometimes can be used to the same effect. Thus Margaret Mead's writings on Samoa can help us to imagine the Samoan way of life, and sociological studies of New York City's juvenile gangs can make us vividly aware of their way of life. Each of the established religions of a society is itself a total way of life and a sensitive reading of its sacred texts and other scriptures will to some extent enable a person who does not practice it to imag­ ine what it is like to practice it. The de­ tailed investigations of scholars in compar­ ative religion can also help a nonbeliever envisage what it is like to have a particular religion and to experience life from its point of view. Of course one can have full knowledge of a religion only if he himself practices it, that is actually lives the way of life which is the religion. This is what I have called "practical knowledge," and I shall consider it more fully below. My only claim here is that the sensitive reading of the literature of a religion can aid us in coming to an imaginative grasp of its meaning, even if we do not practice it or believe in it. It is necessary in addition to purely scientific and philosophical knowl­ edge about the practices and beliefs of the religion if a rational preference for the reli­ gion over another way of life ( or for an­ other way of life over the religion) is to be made. The critical analysis and appreciation of art has long been recognized as a way to deepen our understanding not only of an artist's personal outlook but also of the whole spirit of an age or the general world view of a culture. The music, the painting and sculpture, the dance, the architecture, the drama, and the literature of a culture all present to us the way of life of the cul­ ture. A thorough understanding of works of art in these various forms brings us to an imaginative awareness of a way of life


which no scientific or philosophical lmowl­ edge, however complete, could yield. One of the most interesting aspects of a great novel, poem, or drama, for example, is the way its author creates a world in which certain fundamental attitudes, points of view, and ways of life are expressed. A novelist, poet, or dramatist does not neces­ sarily attempt to persuade us to accept his world outlook or way of life. He confronts us with one, or sometimes several, for our imaginative contemplation. And his work often reflects the . entire world view which underlies his social milieu and cultural background. Reading a novel or poem, watching a drama or ballet, listening to music, looking at painting, sculpture, and architecture all can give us a direct insight, an intuitive grasp, of a way of life. Thus we gain an envisagement of what it is like to live a way of life which we ourselves may never have lived. The third kind of knowledge which we must have if our choice among ways of life is to be ideally enlightened is knowledge by acquaintance, or what I have desig­ nated "practical knowledge." A person has this kind of knowledge of a way of life when he actually has lived it. This means that he has been inspired by the ideals of the way of life and that he has adopted the appropriate value systems relevant to given situations. We recall that adopting a value system involves both reasoning in a certain way and living in a certain way. To have practical knowledge of a way of life is to know what it is like to live it because one has conducted his thinking and his behav­ ior in accordance with the value systems of which it is comprised. In the case of a reli­ gion, one has practical knowledge of it vvhen one has been a believer, has prac­ ticed it, has actively and sincerely partici­ pated in its form of worship. Although such knowledge by acquaintance can be one of the best means of enlightening ourselves about ways of life, it should be noted that a person who lives a way of life might not be able intellectually or imaginatively to


see it as a unified ,vhole. He may be too in­ volved in it to have the kind of detached understanding whiCh can come from an outsider's intellectual or imaginative knowl­ edge of it. A second limitation of practical knowledge is that we can come to know very few ways of life by means of it. One cannot just decide to "try out" a way of life as an experiment and live it for a few months or even a few years. A person must be educated in a certain way, and some­ times must receive special training, to come to the point where he can accept a way of life and commit himself wholly to it. It should further be noted that, at the time of the rational choice itself, · the person who makes the choice must not be committed to any of the ways of life among which he is choosing. But this brings us to the third set of conditions for a rational choice. 3. Conditions of impartiality. A choice is impartial to the extent that: ( a ) The choice is disinterested. That is, the choice is not at all determined by bribes, by exercise of favoritism, by desire to protect one's privileges ( or those of one's family, friends, or class), or by any emo­ tional prejudices on the part of the person who makes the choice. For example, if a person is influenced by anti-Semitism he cannot possibly make a rational choiCe be­ tween the way of life of Judaism and some other way of life. And this holds regardless of how much intellectual or imaginative knowledge he might possess of Judaism. There is one condition that can guar­ antee the complete disinterestedness of the choice, and I therefore include it among the conditions of a rational choice. ( It was suggested to me by Professor John Hos­ pers. ) vVe can eliminate entirely the ele­ ment of self-interest in a choice by stipulat­ ing that the person who makes the choice not know what position he himself would have in any chosen way of life, if it were to be realized on earth. Like the souls in Book X of Plato's Republic, his future destiny would be decided by lot. Thus there would be no possibility that the person was in-


f!uenced by desire for personal acvantage or protection of special pr.ivi1eges in mak­ ing his choice. For he would have no idea which way of life would he more in his self-interest to choose. ( b ) The choice is cktached or obiec­ tioe, By this I mean that it is a choice among ways of life other than that in which the person ,vho makes the choice was brought up and other than that to which he is committed at the thne of choice. The latter condition must be in­ ck:ded for the obvious reason that we are asking a person to state his preference for one way of life over others. If he is al­ ready committed to a way of H£e and yet ( under the second set of conditions) knmvs about other ,:vays of life, his preference is set in advance. He will prefer ills own way of life to the others. Such a person is not in a position to make a choice at all, to say nothing of a rational choice. The first quali:6.cation, however, Ce­ serves to be examined at greater length. A rational choice among certain ways of life can b e made only by those who were not brought up within the framework of a.._'1y one of those ways of life. }"or it is hnpossi­ ble entirely to escape the influence of early childhood, when we \Vere given rules to follow and standards to fulfill and all the pro-attitudes and con-attitudes that go with success or failure in doing this. Alt:hough, as we shall see when we come to the third condition of impartiality, a pers�n need not be biased as a result of his being brought up in one way of life, he will ahvays be under its influ ence. Consequently his choic� among ways of life which include that one will never he truly impartial. In order to insure the maximum degree of imr partia:ity, then, the following conditions of ''detachment'' or "objectivity" must ho!cl. Let us assume that a person \VtlS brought up in a ,vay of life, A, so that he is not qualified to rnakc a rational choice behveen way of life A and way of life B, or between way of life A and way of life C. But ( if all other conditions of a rational choice are sa-

RELATIVISM AND JUSTIFICATION tisfied) he is qualIBed to make a rational choice between B and C. Similarly, a p er ­ son brought up in B cannot make a rational choice between A and B or between B and C, hut he can between A. and C. And one brought up in C can choose between A and B, but not between A and C or B and C, Now suppose all persons who were brought up in ways of life C1 D, E, F, G . . . etc-., were to make a rational choice beDNeen A and B, and suppose that they all preferred -'� to B. Then under the assumption that in each case all the conditinns of rationality were satisfied, we may conclude that way of life A is ,:_1 ore justified than way of life B. This judgment is strengthened to the ex­ tent that ways of life C, D, E, F, G, etc., are verv different from one another and from l;oth A and B. The judgment is strengthened even more if persons brought up in B, upon satisfying all the other condi­ tions of a rational choice but this one, were to prefer A to B. It might be objected that, no matter 10\v free and enlightened the choice is7 and no matter hO\v disi:::iterested are the persons n1aking the choice, it is still not a truly :im­ partial choice. For a person brought up in C will prefer A to B because A is more sim­ ilar to C than is B. And the same for those brought up in D, E, F, G, etc. In each case A is more similar and B is less similar to the way of life in which the person was brought up, and for that reason the person always has a slight bias toward A and away from B. The influence of early childhooc is ever-present and therefore L"llpartiality is never attained. Indeed, it is unattainable. In order to face this objection, I set one ad­ ditional condition of impartiality. ( c) The choice is ,mbiased. In order to minimize the impact of early childhood environment upon a rational �hoice, I stip­ ulate not only that the person making the choice must not have been brought up in one of the ways of life among which he is choosing ( the condition of detachment ), but I also stipulate that the person ,vas not incloctrimrted into or conditioned blindly to


accept the way of life in which he was brought up. To put it positively, a choice is unbiased to the extent that ( i ) the person's upbringing was nonauthoritarian, (ii) the person's education was liberal, and (iii) the person's experience of life up to the time of choice ,vas of considerable variety, richness, and depth. All children are brought up within the frame\vork of some standards and rules. But there is a great difference between being conditioned to an unquestioning ac­ ceptance of standards and rules and being brought to see the reasons behind such standards and rules. When the standards are imposed and the rules laid down in an authoritarian manner by the parents, the child learns blind obedience and rigid con­ formity. When the parents impose stand­ ards and lay do,vn ndes i11 a nonauthorita­ rian manner ( and they must impose some standards, lay down some rules, or else the child will have no guidance), the parents encourage the child to question the stand­ ards and rules and to ask that they be jus­ tified. The parents' answers will at first, of course, be given in a relatively simple way, for instance, in terms of the usefulness of the standards and rules in accomplishing this or that specific purpose. Nonauthorita­ rian parents yvill also try to develop in the child, as he grows older, an ability to make ''decisions of principle," that is, to make up his own mind whether to follow a rule or standard or to make an exception to it in a given situation, or to decide to reject the rule or standard entirely. A nonauthorita­ rian upbringing enables a child as he grmvs older to justify his value judgments and his standards and rules, and .finaIIy to choose his own way of life ( which may or may not be the same as that of his parents ) . As a mature person he will be able to change his way of life when social, economic, political, domestic and other conditions of his life demand such a change. A similar contrast can be found in the difference between indoctrination and edu­ cation. To be educated within the frame-


work of a society's way of life is not neces­ sarily to be indoctrinated into that way of life. Indoctrination is a deliberate manipu­ lation of the mind of a child, an attempt to produce unquestioned belief in one way of life and a blind rejection of all others. Edu­ cation, on the other hand, is a process of giving the child tools of criticism as ,vell as of adjustment. As the child matures, his education becomes more liberal, presenting him with value systems and ways of life which are foreign to his own. At the same time his mind is trained to think critically about his own way of life, so that he is forced to make up his mvn mind about is­ sues on which he finds he must take a stand, or about controversies on which he finds he must take sides. The society which not only permits, but encourages, its mem­ bers to criticize its own foundations ( value systems ) is in no sense a society which in­ doctrinates its members. Finally, we may contrast the life of a person who has been brought up in a uni­ form culture, where he meets only people who have the same outlook and opinions as he does and whose value systems are the same as his, ,vhere he remains in one occu­ pation all his life, never travels either to foreign countries or to other areas of his native land, reads no books, remains pro­ tected from any great suffering, and never goes through any deeply emotional experi­ ences- we may contrast such a life with that of a person v.rho meets and gets to know people of varied backgrounds and from all walks of life, who travels widely, reads a great deal, and has a generally var­ ied and rich experience of life. Although both of these persons are, in some very general sense, "children of their culture," the latter is much less a child of his culture than the former. In the same way, an edu­ cated person is much less a child of his cul­ ture than an indoctrinated person, and a person with a nonauthoritarian upbringing than one with an authoritarian upbringing. To the extent that a person in these various ways is not a child of his culture, to



that extent he is better able to make an im­ and more people who make such choices partial choice among ways of life which ex­ tend to prefer a way of life A to a way of clude the one he was brought up in. And to life B, then to that extent it is reasonable that extent he will not automatically prefer for anyone to accept A as preferable to B. the way of life which is most similar to that Such acceptance must remain tentative of his childhood. Thus, although total ab­ only. It must be open to revision in the sence of bias cannot be guaranteed, we can light of further choices under conditions which more closely approximate those of in this way minimize it. These, then, are the conditions which an ideal rational choice. But it provides the define an ideal rational choice among ways best available way of knowing whether A of life. It is a choice which is totally free, really is preferable to B. It might be objected that no such totally enlightened, and totally impartial. No one is ever in a situation where he can agreement could ever be reached, even if actually make such a choice. We are never all the conditions of a rational choice were confronted with alternative ways of life fully realized. For it is always possible that under these ideal conditions. But that is variations in temperament among people not to the point. The real question is this. will result in variations in their preference What would make a reason for committing for ways of life. Thus even choices made ourselves to a way of life a good reason? under ideal conditions of rationality will be My answer, in sum, is that such a reason is subject to the disagreements among roman­ to be found in the situation I have de­ ticists and classicists, doers and thinkers, scribed-where we find that more and more rationalists and mystics, dogmatists and people brought up in a variety of ways of skeptics, optimists and pessimists, conserva­ life tend more and mOre to prefer one par­ tives and radicals, seekers of happiness and ticular way of life to all others, when their doers of duty. This possibility must be ad­ preference results from a free, enlightened, mitted, but it does not destroy the concept and impartial choice. \iVhat better reason of a rational choice . as such. It merely could there be for committing ourselves to leaves open the question-which would have to be left open even without this con­ a way of life? I have tried only to show that it makes sideration-whether increase in rationality sense to talk about a rational choice among of choice leads to agreement among the ways of life, and that therefore ihe relativ­ choosers. All that we can say is that, if ist's position is not tenable. In justifying under the conditions of a rational choice our value systems we can go beyond vindi­ there was a tendency among choosers to cating them in terms of a way of life. We agree that way of life A is preferable to can ask that the way of life itself be jus­ way of life B (. whether or not the choosers tified. It will always be impossible in prac­ were of the same temperament) , then we tice to know with certainty which way of would have just that much reason to con­ life is more justified than any other, since clude that A is more justified than B. That the conditions of an ideal rational choice such a tendency would be made manifest are such that it is difficult even to approxi­ as ideal conditions of rational choice were mate them. But it is theoretically possible more and more closely approximated must to do so, and therefore meaningful to speak remain an open question. of a rationally chosen way of life. Although we can never be certain v.rhether one way D . WHY BE RATIONAL? of life is rationally preferable to another, we can reach a probable knowledge of this in the following way. \Ve can say that, to There is one difficulty involved in this at­ the extent that choices become more and tempt to define a rational choice among more rational and to the extent that more ways of life which, if not satisfactorily met,


undercuts the entire project. Have I not imposed my own way of life upon the con­ cept of a rationally preferable way of life by stipulating just these conditions of ra­ tionality and not others? In other words, am I not begging the question by giving conditions for a rational choice which are themselves part of a way of life? Am I not merely presupposing a way of life and working within its framework, rather than taking a standpoint outside all ways of life? And in that case how do I know that the way of life I am presupposing is itself preferable to all others? Until I have shown that it is so, I cannot claim that the condi­ tions which · 1 specify for a rational choice really do justify one way of life rather than another. On the other hand, if I try to es­ tablish the preferability of this presup­ posed way of life by appeal to a rational choice, I am arguing in a circle, for the conditions of a rational choice are part of that way of life itself. My answer is to deny that the condi­ tions of a rational choice are part of a vvay of life. For they are the conditions which I presume anyone, in any ·way of life, would accept as defining a rational choice, in the ordinary sense of the word "rational." If people on reflection would not be willing to accept these conditions, I would not say they were making a mistake. Nor would I continue to impose my conditions upon them. I vmuld ask them what conditions they would give for defining a rational choice. If they offered some which I had not thought of and which did seem ( to me and to them) to elucidate further our ordi­ nary meaning of being rational in making choices, then these new conditions would go into the definition. \1/ould they not then be imposing their way of life upon the con­ cept of rational choice? The answer is no, because the concept of a rational choice is independent of all ways of life. Even if two persons were commitled to very different ways of life, both would have to admit that, if a rational choice were to be made between their two ways of life, it would have to be a free, enlightened, and impar-


tial choice. Or else it would have to be a choice under other conditions which better elucidate what a rational choice means. But whatever the conditions, they can­ not change when the way of life being judged changes. This follows from the very meaning of the word "rational." If the choice is to be a rational one, it must at least have properties of rationality which do not vary with the ways of life among which the choice is being made. Thus it must always be possible for a person to admit that a rational choice was made and yet another way of life was preferred to his own. To test the preferability of one's own way of life by the method of rational choice always presupposes the possibility that the test ,vill tum out negative. For otherwise it is no test at all. One would have set up the conditions of a "rational choice" ( the test conditions ) in such a way that one's own way of life would always come out on top. To vary the conditions so that they always bring about this result is a sign that the choice is not rational. For it is an essential part of the meaning of ration­ ality that its conditions not vary vvith what is being judged. In specifying the conditions of a ra­ tional choice, then, no particular way of life is involved. All that is involved is an at­ tempt to make explicit the idea which we all have ( no matter what may he our way of life) when we reflect about what an ideally rational choice among ways of life would be. I have spelled out what I think it would be. It is for others to challenge my account and to improve upon it. The condi­ tions I have speci£ed are, I think, an accu­ rate explication of the meaning of the word ''rational" in this context. Consequently I believe that any person would have to admit that this is what a rational choice consists in, even if making such a choice would result in other ways of life being preferred to his O\Vll. Furthermore, if my explication of ra­ tional choice is correct, then any person would have to agree with it, even though his own way of life denied freedom of


choice to people, prevented enlightened choices from being made, and did not de­ velop impartiality in people. It does not follow, however, that his way of life would never be preferred to others when a ra­ tional choice among them was made. \Ve . do not know what actual ways of life would be preferred when rational choices were made, and there is no necessity that

the preferred ways of life hove the same characteristics as the rational choice itself. The concept of a rational choice is not logi­ cally connected with the content of any particular way of life. I am not imposing the content of my own way of life upon others when I explicate the conditions of rationality in terms of freedom, enlighten­ ment, and impartiality. If one should ask me, "Why ought I to accept your condi­ tions of rationality?», my answer would be, «They are not my conditions of rationality, but yours too. Is this not what you would mean by an ideally rational choice?" If the reply is negative, then the way is open for further explication of a rational ideal com­ mon to both of us, with both of us trying to make our explication correct. Perhaps behind the foregoing objec­ tion is a deeper (but more confused) one. Suppose a person grants that, as far as he is concerned, my explication of a rational choice is correct. He admits that I have shown what being ideally rational means in the contei,,i: of choosing among ways of life. But he then raises the following objection. "Let us assume way of life A is rationally preferred to way of life B, that is, A is ra­ tionally chosen- in the sense you have spe­ cified-over B. vVhy does it follow that I should live according to A rather than B? I grant that A is rationally preferred to B. But why is the way of life which is pre­ ferred the way of life which ought to be preferred? Do not answer that it is because the preference is based on a rational choice. I already know it is based on a r a ­ tional choice. I am asking why I ought to follow a rational choice. Jv1y question con­ cerns the claim that a rational choice has


upon me. Why ought I to live the way of life that is rationally chosen? To put it all in a nutshell) ,vhy be rational?" This is a huge muddle. The confusion becomes apparent \Vhen we ask ourselves what sort of answer the person wants . What reply could possibly satisfy him? I shall try to disentangle this confusion by discussing four points: 1. The distinction between (a) giving a correct explication of a rational choice among ways of life and ( b ) giving reasons for trying to be rational in explicating a ra­ tional choice. 2. The distinction between ( a ) and ( b ) on the one hand, and, on the other, ( c ) giving reasons for the validity of the argument that "A ought to be preferred to B" follows from "A is preferred to B on the basis of a rational choice." 3. The distinction between ( c ) and ( d) giving reasons for the validity of the argu­ ment that "One ought to live according to way of life A" follows from 'Way of life A is rationally chosen over all others." 4. The distinction between ( c ) and ( d ) on the one hand, and, on the other, ( e ) giving reasons for committing oneself to try to discover ,vhich way of life is ra­ tionally chosen over all others. 1. If someone asks why he should live the way of life that is rationally chosen, one possible reply is that the question is beside the point. In defining ( explicating) the conditions of a rational choice I am not trying to argue that people ought to live the rationally chosen way of life. I am only trying to state what a rational choice means, regardless of whether or not people actually want to live a rationally chosen way of life. I am not saying they ought to want to do this. If people do want to live according to the most justified way of life, they must first know which way of life is most justified. In my explication of a ra­ tional choice I am trying to state how we find out what that way of life is. My expli­ cation ,vill be of little interest to those who do not want to discover the most justified


way of life. But if my explication is a cor, rect explication, it w11l be of help to those who have such an aim. The aim is to find out how to discover \Vhat it'means to live rationally, that is, according to the most justified way of life. But iu carrying out my explication I an1 not imposing this airn or this further purpose upon anyone. Nor am 1 claiming that anyone ought to have such an aim or purpose. So the question "vVhy be rational?" is simply irrelevant to the a t ­ tempt to explicate the method for finding the most justified way of life. But this reply might not the objector. He might make the following rejoin­ der. "My question 'Why be rational?' is not entirely irrelevant, for I can ask it about your explication itself. If you are trying to give a rational account of the method for finding the most jus!i!led way of life, as you say you a!'e, then ,vhat reasons are there for accepting your account? Suppose I do not accept it. You may claim that ill that case I would not be rational. But why be rational here?" There is a confusion in this rejoinder which stems from a failure to distinguish betyveen the demand for reasons for prefer­ ring a way of life and the demand for rea­ sons for accepting an account of such rea� sons. The first demand is expressed in the question: ( a ) What are good reasons for preferring one way of life to another? The answer to this question lies in the attempt to e;rplicate what it means to be ideally ra­ tional in preferring one way of life to an­ other. Let us suppose an answer is offered, such as the answer I have proposed in the concept of a rdtional choice. Then the sec­ ond question is asked: ( b ) \'\!hy should the canons of reasoning governing this answer be accepted? Let us see how this question arises. It may be that the concept of a ra­ tio�l choice as I have deflned it does not provide a correct or accurate explication of justifying a way of life. I may have failed in my attempt to answer question ( a ) and I am perfectly willing to be criticized on


that account. But any such m:itieism must itself be governed by the canons of reason­ ing that govern correct explication. The a person who makes the criticism is seekino 0 correct explication and therefore must accept the canons of reasoning which de£ne the philosophical point of view ( as distinct from the scientific, the mathematical; and the various normative points . :Of view ) . · within that point of view he may carry on arguments concerning the acceptability of answers to question ( a ) . But the question �vvhy be rational?" does not arise within such a rational framev;rork. Hm,\ then, does it arise? The ques­ tioner may explain that he means as fol­ lows: "When I ask 'Why be rational?' I am demanding reasons for anyone's placing himself "'lthin a rational framework, like that of the philosophical point of view. I am not asking for reasons within a rational framework. I am outside a point of ;1ew and am asking why I should take the point of viev,., and so have· my thinking governed by its canons of reasoning. vVby be r a ­ tional?' means Why take the philosophical point of view ( or any point of view, for that matter) ?' " Here the questioner want'> to be given reasons for accepting the canons of reason­ ing which govern a proposed explication. These are th e canons of the philosophical point of vimv. \¥hat sort of reasons could we give? I submit that only one sort of rea­ son can be given: If we want to live the most justified way of llfo, or if we want to answer question ( a ) > then takir1g the philo­ sophical point of view is indispensable. If a person did not care to live the most jus­ tified way 0£ life� or if he were not inter� ested in trying to find out what that way of life is, or if he did not want a valid proce­ dure for finding out what that way of life is, then i t would be pointless ( in the pres­ ent context) to take the philosophical point oc view. Taking that point of view is a necessary means to the three enGS ;ust mentioned. Unless a person takes that point of view he cannot achieve any of them. But


if he does not seek them, no reasons can be has very good reasons for taking the philo­ given for his taking the point of view. vVe sophical point of view. For only by doing cannot argue that he ought to seek those so ,vill he be able to discover which way of ends without assuming the canons of rea­ life ( if any) is the most justified. We might soning of some point of view which the ob­ call this a pragmatic justification ( vindica­ jector is also ,villing to assume. Since he tion) of the philosophical point of view, al­ has stated his unwillingness to be placed though this must not be confused with ,�n­ within any point of view, no argument can dicating value systems which belong to be given that "�II make a claim to his as­ normative points of view. 2. But perhaps the foregoing discus­ sent. In short, if his question "VVhy be ra­ tional?" is a demand for reasons for taking sion has missed the principal point of the any point of view, then it is logically im­ objection. V\lhat is being objected to is our possible to answer his question. In order to saying that a person ought to prefer a way answer the question "Vlhy?" we must give of life because it has been shown to be ra­ reasons, and giving reasons is a process of tionally chosen over other ways of life. thought governed by canons of reasoning Would we not be falling into the naturalis­ that define some point of view. Since no tic fallacy? ( A full account of the naturalis­ such canons are accepted, no reasons can tic fallacy will be given in Chapter 9. ) From the fact that people rationally prefer be given. Further reflection on this situation one way of life to anotlier we infer that we makes us doubtful about just what the dif­ ought to prefer it. Is this not going from ficulty is. If it is logically impossible to "is" to "ought"? We might reply that if a give an answer to a question, is there ac­ person does not prefer what is rationally tually any question to be answered? In the preferred, he is not rational. The question present case an answer to the question is then pops up, "Why be rational?" Here the logically impossible because the questioner challenge means, "Why ought one to prefer has deliberately refused to accept the con­ a way of life which, under conditions of a ditions required for giving an answer. It is rational choice, actually is preferred over the questioner himself who has made his other ways of life?" question unans-,verable. It is hardly surpris­ A reply would have to provide good ing, then, to discover that we cannot give reasons for going from "A is rationally pre­ any answer that will satisfy him. Indeed, ferred to B" to "A ought to be preferred to must not conclude that no genuine B." Is there not a fallacy in this inference? question is being asked? For the words I do not think so, for the following reason. "vVhy be rational?" can have meaning only iVhen we say that A is rationally preferred if reasons can be given for being rational. to B, we mean that whenever the condi­ But no such reasons are allowed under the tions of a rational choice hold, A is pre­ conditions set by the questioner. Thus in ferred to B. Let us assume that the condi­ asking the question he is demanding reasons tions as I have specified them do correctly and at the same time making it impossible explicate the justification of a way of life to satisfy the demand. The conclusion we ( and this assumption is not now in ques­ al'e forced to draw is that he does not know tion ) . Then it would follow that A is more what he is saying. He is merely pronounc­ justified than B. This means that there are ing vvords in an interrogative form outside better reasons for choosing A than for of any possible context for their use. choosing B. If a person, knowing this, were On the other hand, suppose the ques­ still to prefer B to A, his preference would tioner does want to live the most justi£ed be a paradigm of an irrational choice. The way of life, or wants to find out what such person might want to persist in his choice a life would be. In these circumstances he nevertheless, claiming that he honestly 364

THE JUSTIFICATION OF VALUE JGDGMEKTS does prefer B to A. There is no logical error in his doing this, so Tong as he does not claim that he has good reasons, or is being rational,. in doing it. He .is saying in effect that he does not care to be rational about this matter. But then he must not ask "\Vhy be rational?" For as soon as h e asks 'Why?" he is demanding reasons and thus presup­ posing rationality. To ask such a question is to speak as if only a rational ans\ver will be acceptable. But in the present situation he already knows what a rational prefer­ ence is ( namely, the preference of A over B ) and that his own preference runs coun­ ter to the rational one. What more does he need to know? Again his question appears to be outside any contexts for its possible use, 3. We may further distinguish be­ tween the question ( e) "What are good reasons for going from 'A is rationally pre­ ferred to B' to 'A ought to be preferred to B'?'' and a very similar question ( d) "vVhat are good reasons for going from '1\ is ra­ tiona,:y preferred to all other ways of life' to 'Everyone ought to live in accordance with A'?" Question ( d ) is, I think, the basic question that is in the back of many peo­ ple's minds when they ask ''\Vhy be ra­ tional?" In asking this they want to know why anyone ought to do what he already knows to be rationally justIBed. vVhy should one's action be motivated by one's knowledge of the good? I shall now try to sho,v that this is an ernpty question. To sav that someone alreadv knows that an act, X, is rationally justified is to say that he knows there are good reasons for doing X. Thus he already bas good rea­ sons for doing X. Why, then, should he ask for rnch reasons? He might say that he knows X is the rational thing to do but he wants to know why he ought to do the ra­ tional thing. He wants to know, in other words, by what rule of logic we can go . from (+X is the rational thing to do" to "X ought to be done." The ao.swer is that both statements ::::::iean the same thing, namely that there are good reasons for doing X. As


I shall point out in Part II, the word "ought" is here being used prescriptively, and this contextually implies that there are good reasons for doing the act prescribed. To prescribe act X by saying "X ought to be done'' is contextually to imply that there are good reasons for doing X. �ow if a per­ son wants to know why he ought to do X, he is askbg for good reasons for - doing X. But in the case at hand he acknowledges that he has good reasons for doing X. Hence his question is empty. It can only be "answered" by uttering a tautology: "There are good reasons for doing an act which you have good reasons for doing." Or: "It is rational to do what is rationally justified." Or: "It is rational to be rational.'' The same considerations applv if ·we demand good reasons for living way of life whb':i is acknowledged to be rationally justified. If a person hows that a way of life is rationally chosen and is therefore as justified as it can be, then he knows why he ought to live it. Indeed, he already bas the very best reasons for living it. This renders the demand for good reasons otiose. No further reasons can be given for him to live the way of life, s:ince ex hypothesi he al­ ready has the best reasons for living it. But then he does not need any further reasons. 4. It \Vas mentioned under point 1 that when a person wants to know how a way of life can be justified, he must assume that taking a rational point of view toward the problem is itself justified. Canons of rea­ soning ( in th:is case canons of philosophical reasoning) must h e accepted in any at­ tempt to solve the problem. So if he seeks an ailS\ver to his problem, he cannot de� mand reasons for taking the philosophical point of view. He cannot ask '�\Vhy be ra­ tional?'" since he already presupposes the justification of being rational in seeking a:, answer to his problem. K ow he might seek an answer to his problem purely from intel­ lectual curiosity. He might not be inter­ ested in justifying his own ,vay of life. He might simply want to know whether any­ one's way of life can be justified and if so,



how. It would then be perfectly consistent for him to find out how a way of life can be justified and then not try to justify his o,vn nor try to live in accordance with a justified one. Suppose, on the other hand, that a per­ son wants . to lmow how a way of life can be justified in order to discover whether his O\vn is justified and in order to live in ac­ cordance with a justified one. And let us suppose that such a person, on learning of the method of rational choice, accepts it as a correct explication of how a ,vay of life can be justified. Then suppose he asks 'Why ought I to try to live in accordance with the rationally chosen way of life?" We must now be puzzled about ,vhat his ques­ tion can mean. Can he be serious in asking it? For he is already committed to trying to live in accordance with a justified way of life. And his question cannot mean "VVhat makes a way of life justified?", since he ac­ cepts the concept of a rational choice as providing a correct answer to this. His question is rather, «Granted that a ration­ ally chosen way of life is a justified one, why ought I to try to live in accordance with it?" But if he is interested in finding out how a way of life can be justified in order to live in accordance ,vith a justified one, then he knows he is committed to liv­ ing a rationally chosen one, since this makes it justified. Hence he cannot be ser­ ious in demanding reasons for living such a way of life. He is already trying to do so. The person might then make the fol­ lowing move, in explaining his question. He might say, "It is true that I am already committed to living a rationally chosen way of life because I want to live in ac­ cordance with a justi£ed one. But it seems to me that my commitment is without rea­ son. I am asking if there are any reasons for committing myself in the way that I have. It is true that I am now try�pg to live rationally, but I want to know \yhat reasons can be given for my ( or anyone's ) decision to try to live rationally." This question is quite different from


the one discussed under point 1. For the person is not demanding reasons for taking the philosophical point of view. He already knows that taking this point of view is ne­ cessary if he is to ?-nd out how a way of life can be justified and so find out which way of life to try to live in accordance ,vith. His question is rather about his end of trying to live in accordance with a jus­ tified way of life. He wants to know if any reasons can be given for seeking such an end. Nforeover, his question is not the question discussed under point 3. He is not asking why he should live a rationally cho­ sen way of life. He knows why-because it is rationally chosen. He sees that this is an empty demand, since he acknowledges that a person who knows that his way of life is rationally chosen already has all the rea­ sons he can possibly have for living that way of life. The question he is concerned ,vith is not why a person should live a jus­ tified way of life, but why a person should try to find out what way of life is justified in the first place, and why he should seek a justified way of life in order to try to live in accordance with it. In other ,vords, why not simply disregard the problem, or seek some other end? How can such a question be ansvvered? It ,vould seem that if a person did not care about finding a rational way of life or did not want to try to live in accord­ ance with one, nothing could be said to show that he was unjusti£ed in his atti­ udes. It would seem that no reasons could be given against his lack of commitment to seeking a rational way of life, and also that no reasons could be given for such a com­ mitment. Either a person cares about such things or he does not, and that is all there is to be said on the matter. Yet something more can be said. In the first place, we can point out that a person who is not interested in finding a rational way of life or in trying to live in accord­ ance with it cannot give reasons for his lack of commibnent ( or for his commitment to other ends ) . Nor can he give reasons

THE JUSTlFICATION OF VALUE JUDGMENTS against the commitment of a man who does want to find out which way of life is most justified and who does care about living in accordance with it In the second place1 it is possible to interpret the question, "Why try to find a rational way of life, and why have the purpose of trying to live in ac­ cordance with it?" as a way of asking to be convinced or persuaded that one is right in making these commitments. Like the per­ son who asks "Whv be moral?" and wants as an answer to be' inspired to fulfill moral standards and to follow moral rules ( i.e., to adopt a moral value system ), the question­ er here wants to be moved to be more deep­ ly committed to his basic goals. He wants to have his attitudes strengthened, to be en­ couraged, to be given support in his endea­ vors. Can such a demand be satisfied? I think there ace two general methods which can be used. \1/e might try to convince him intellectually or we might try to persuade him emotionallv. It must be understood that the flrst ,;;ethod is not a matter of proof. \i\Te cannot glve reasons from vvhich it follows that a person is. right in being committed to seeking a rational way of life. VVe can, ho,vever� ask him to review his c01nn1itment in light of alternative commit� ments. We can ask him to think what a jus­ tified wav of life is--that it is the way of life which a person would have the hest reasons for living. If . he says he al­ ready kn0\V5 this ( as ,ve acknov.rledge that he does ), we can only invite him to think about it more deeply, to pay more attention to what it is he is seeking when he seeks a rational way of life, and to be folly aware of what it means not to care about seeking a rational way of life. Finally"' however, \Ve must leave it t:p to him. He must choose whether to seek a rational way of life or not. He makes no logical error in choosing not to seek it, but if he chooses not toi he n1ust not expect to Bnd a v,,ay of Hfe v\..hich he can justify when it is challenged. If, on the other hand, he does w�nt to flnd that way of life which can best be defended ra­ tionally against attack, then in the very fact


that he wants to do this lies his commit­ ment. He has made his choice, for to try to find such a way of file_ is already to seek the end he now' wishes to be enco�aged to seek. This is about all we can do in an intel­ lectual way to answer his question. What can we do in the wav of emotiona! an· peals? In order to p�rsuade the pers;n emotionally that he is right in seeking his end we must use techniques ( such as praise, rewards, pointing at inspiring exa1n­ ples, and so on ) which would be effective in strengthening his motivation to seek the end. This is not, strictly speaking, to answer his question. But it is to respond to his question in such a ,vay ( if ,ve are suc­ cessful) as to satisfy him. The outcome of this process is that he no longer asks the question. And this is not because we have silenced him, Our procedure C.oes not :in­ volve preventing hin1 from uttering a ques­ tion that is still in his mind. Rather, we put to rest the inner doubt. He no longer asks the question to himself. He has come to have a strong, stable disposition to try to Bud a rational way of life in order to live in accordance with it. He no longer demands that he be justified in having this disposi­ tion. This completes my account of the jus­ tification of value judgments, Throughout the discussion I have taken the philosophi­ cal point of view and so committed myself to approaching foe question as rationally as I could. But this commitment to rationality \Vas not my ultimate commitment; for one can always give reasons for taking the phil­ osophical point of view. These reasons would constitute a pragmatic justification (vindication) of taking that point of view. They would point out that it is a necessary means to a certain end we have chosen. If our end is to learn as much as '-Ve can about wh.t�J ,it means to be rational in justi­ fying value:, judgments, then taking the philosophi�a1 point of view is indispensable for achievitjg our end. The ulti-mate com­ mitment is our deciding to seek this end.


Throughout my discussion of the justi­ fication of value judgments, I have as­ sumed that the reader shares with me not only this commitment, but also the accept­ ance of the philosophical point of view which such a commitment requires. VVithin the framework set by the canons of reason­ ing of that point of view, the reader may \Vish to criticize what I have to sav about the justification of value judgments: But he cannot criticize me for taking that point of view on the grounds that when I state what it is to be rational in justifying value judgments I am assmning canons of ration­ ality and therefore arguing in a circle. For the canons of the philosophical point of vie,v only govern the correctness of an ex­ plication of the mles of reasoning used in justifying value judgments. Those canons are not themselves used in justifying va�ue judgments. ( The canons of normative points of view, however, are so used, ) Yiy explication of the justification of value judgments may well be incorrect at many points and I am open to criticism on that account. But whatever criticism is made 1 it must be made from within the framework of the philosophical point of ,1ew. And the one who offers such criticism is ultimately committed to the same ideal as that to \Vhich I a m committed- to 1ean1 as much as possible about what it means to he rational in justifying value judgments. We must distinguish this philosophical ideal from the practical ideal of actually trying to live a rational life. The philosoph­ ical idea] requires only that we be rational i n our intellectual inquiry into what it means to live a rational life ( including what it means to be rational in the justi­ fication of our value judgments ). The prac­ tical ideal requires that we be rational in all of life. The philosophical ideal is to find out whether value judgments can be jus­ tified and, if so, how. The practical ideal is to find out which value judgments are jus­ tified and to Hve in accordance with them. This means trying to fulfill the standards


and follow the rules that verifv our value judgments and that constitut� the value systems of a rationally chosen way of life. Is the commitment to live 1},is way of life an ultirnate coinm:itinent? Yes, if we mean by "ultimate" that no further justificatory reasons can be given for making the com­ mitment. Someone might now triumphantly eon­ clude, "You see, it is all uUimately absurd. ·we finally come to the point where we must commit ourselves ,vithout reason, reason far enough and you will arrive at unreason. All our ultin1ate commitments, being ultimate, are arbitrar;. Thus we might as well toss a coin to decide whetl1er to live a rational or a nonrational way of life. In the end they are equally n�nra­ tional. No reasons can be given for deciding in favor of a rational life rather than a non­ rational life. It is impossible to answer the question 'vVhy not live a nonrational lifer So both rational and nonrational ways of life are on an equal footing. The choice be­ tween them is arbitrary, unfounded, and absurd," The reply to this objection should now be obvious. It is the very reply I have made ( under point 3 above) to the person who demands that reasons be given for liv­ ing a rational life. No reasons can be given) it is true. But no reasons need be given. For knov;1ng that a certain way of life is :rational is knowing that one is wholly jus­ tified in committing oneself to it. To know that it is rational is already to have all the reasons one could possibly have for living it As Mr. Hare pointed out in a passage I have quoted 1 the decision to commit one� self to a way of life which is rationally cho­ sen over other ways of life ( each of which must be fully 1.'Il�wn for the choice to be enlightened and hence rational) is the most reasonable, least arbitrar)\ and best founded decision of all. It is the decisfon to live th e way of life one is most justified in living, all things comidered.

Aren 't Moral Judgments ''Factual''? MARTIN E. LEAN In this paper I am concerned with the ques­ tion of how moral terms and judgments may be regarded in terms of the notions of factual matter or judgment, value matter or judgment, objective, subjective, and oth­ er such kindred terms and notions. What I have done is to see how strong a case I can make for holding, in the face of some of the well-known argumentative moves for the contrary view, that moral terms are gen­ uine factual predicates and moral judg­ ments are factual in character. This task I first set for myself some ten years ago when, as my contribution to a program at Columbia University memorializing G. E. Moore ( under whom I had been privileged to study there some years before ) I elected to see what could be made of his position about this. His position-to which he stead­ fastly adhered despite the most cogent and persuasive arguments of the relativists, logical positivists, linguistic analysts and emotivists- was, it will be recalled, pre- · cisely that moral judgments are factual and that moral disagreements between people can and do involve moral judgments that stand to each other in the relation of gen­ uine logical contradiction. It was not a position with which I then agreed, and it would have been easier and more comfortable to have flourished a new cadenza on the more fashionable tune of the time, to play picador to, e.g., Prof. C. L. Reprinted from The Perscmalist, Vol. 51. A shorter version of this paper was presented at a philosophical meeting held at Columbia Univer­ sity in 1959 in commemora1?-on of G. E. Moore.

Stevenson's matador, than to essay the role of champion in behalf of a position so ob­ viously declasse. But Moore was dead and the vividness of his courageously steadfast honesty and carefulness was especially large before me, and indeed before all who had known him and were gathering to honor his memory. And so I dared to raise a lance and try a round in behalf of an un­ popular cause. Now i\1oore himself, it will be remem­ bered, did very little more than to assert the thesis that moral judgments are factual, and to reiterate it insistently. It did not to him, apparently, seem to need argument. His main concern was to descry and decry what he named the naturalistic fallacy, and to argue the indefinability, unanalysability and logical simplicity of the quality "good." I am not here primarily concerned with these issues. My concern rather is with what case can be made for the view that moral judgments are factual. The fact is that what began for me as no more than a dedicated exercise became a posture of conviction. The more I sought to see what moves could be made to and from i\1oore's position, my realization grew that he had a game with vastly more potential than I had previously been in the frame of mind to ap­ preciate. I had been persuaded by the fash­ ion of the times, by arguments more in the nature of la\V}'ers· briefs than of accurate and justly weighed presentations of the facts and issues pro and con, and perhaps by psychological considerations too messy and irrelevant to be dredged here. 369



What I have produced is, I dare say, the persuasive-emotivists. £\1oral expres­ but another lawyer's brief, arguing the sions and judgments, however sophisti­ other side. I trust that it "�ll be understood cated, and however disguised they may be that I do not claim it is the way Moore by their occurrence in the grammatically would have argued it. Indeed I have rea­ indicative mood, are nonetheless of the log­ son to believe that he would have either re­ ical and ontological status of primitive jected or eschewed or at least have been tribal taboos. They are inter-culturally rela­ extremely dubious about some of my argu­ tive, and intra-culturally evaluative and at­ ments. But two things perhaps may be said titudinal, hortatory and imperative, emo­ for what I have done. First it is not a mere tive and exclamatory, subjective in their reiteration of the old thesis but an argu­ application and in no sense factual. i'Vhat­ ment; and an argument that does not ap­ ever is irreducibly factual about moral peal to the old-fashioned gambits of ques­ evaluations, over and above the plain fact tion-begging intuition, divine revelation, or that people make them, resides in or con­ the like. Rather it attempts, successfully or sists of the extraneous other- than-moral not, to make the thesis respectable in con­ factual features of the situations in which temporary terms. And the second thing that or about which the moral evaluations are may be said for doing what I have done is made. The argument for this general view this. As Aristotle reminded Nichomachus, his son, in advising him about aiming at seems to rest mainly on t\vo premisses: the Golden Mean, the carpenter in straight­ First, that individuals and indeed whole so­ ening the warped board exerts pressure op­ cieties can and to a considerable extent do positely, to the other extreme, in seeking to disagree and even differ radically in their bring the board to true. So I, in wrestling moral assessments; and second, that there vvith the warp which seems to me now to is no operational definition or specifiable be clearly present in the currently fashion­ criterion of application for moral terms, able and uncritically received view about and hence no effective rational procedure moral judgments, have exerted pressure to for settling disagreements or differences in the opposite extreme. Perhaps your oppos­ their use. In this paper I want to examine eacb ing pressure now to my argument will con­ tribute to our getting the matter straight of these two premisses for truth and rele­ and true. vance. What is the extent and nature of moral disagreement and difference? Does such disagreement and differ­ ence, so far as it does exist, unambi­ I guously and necessarily imply either rela­ tivism or emotivism? Is there no rational It has become the dominant and prevailing technique for resolving or explaining such view in meta-ethics that moral terms are disagreements and differences? Is it essen­ not really factual predicates, and that tial to a term's being deemed a factual moral judgments are not factual. This is the predicate that there be precise, specifiably view, more or less, of Thrasymachus in the discernible criteria for its application? Is it Republic and of the Sophist Protagoras, for essential to a judgment's being deemed fac­ example, in ancient times; of David Hume tual that there be an effective rational deci­ in more modern times; and it is plainly the sion procedure for resolving disagreement argument of the ethical relativists from Ed­ and establishing truth? ward ·vvestermarck on, and of their more Now, to begin mth, it can hardly be fashionable contemporary cousins : the logi­ maintained that there is not widespread cal positivists, the linguistic analysts and disagreement and difference in moral eval-



uations, or that, on the other hand, there agreement in moral judgment, rather than are definitional criteria for moral terms and of disagreement ( which I shall come to decision methods for moral judgments. But Iater ) . this being granted, we are also obliged to There are two obvious ways in which aclmo1,,vledge that there is Jike\v:ise an ap­ the fact of "1despread agreement among preciable amount of agreement among peo­ people in their particular moral assess­ ple in the:r moral evaluations, that people ments might be accounted for. One expla­ do learn to use mora! terms vvith recogniz­ nation is that the moral terms are genuine able . consistency, and that there are ra­ factual predicates, and that moral judg­ tional moves by which people argue their ments are thereby factual in a relevant and moral evaluations. And if this also be significant respect. The other is that our granted ( as I think it must) then I can set moral terms are predicates in a merely the theme of my argument by pointing to grammatical respect, pseudoApredicates on­ these two sets of facts. For it seems to me tologically, but our judgments nonetheless logically ob,fous that if the former set of coincide because we learn the111 rote­ facts constitutes an argument ( however in­ fashion, from childhood on, as a sort of conclusive ) against t.oe thesis that moral moral catecl:tism�just as the members of a terms are proper factual predicates and primitive tribe learn the rituals and taboos, that moral judgments are factual, then by and thereafter heed and chant them ln uni­ the same token the latter set of facts should son, and react v;rith emotion whenever and constitute at least a prima- facie argllillent to whatever they are applied. Now it seems to me patent that the for that thesis. So is the battle joined. Those who would maintain the thesis that complexity of the phenomena of moral moral judgments are factual must account judgment discourse simply cannot be ac­ for the facts of moral disagreement and counted for on this latter view. 1 do not divergence; and they must also take ac­ deny, of course, that there is this process of count of the lack- nay, more, the seeming societal inculcation of n10ral concen1 in its impossibility-of effective definition and members. Nor do I deny that particular decision criteria. Those who ,yould deny precepts may be and often are taught in the thesis, on the other hand, must accouilt uncritical rote fashion1 commonly in a reli­ in some other way for the amount and kind gious context) and that the whole may ac� of agreement, and the reconciliation tech, quire for us complex affective associations of the most varied sort. 'What I challenge is niques, that do exist. It seems to he a feature of the histori­ only that this account is adequate to all the cal dialectic of the problem that relatively facts. But before turning to this, I want to little has been clone with the latter obliga­ consider for a moment the logical relevance tion in recent years, even by-what shall of the whole account in terms of rote­ we call them?-the absolutists or objectiv­ learning to the issue of wheth. er or not ists, whose cause it would serve. The best moral judgments are factual. It is surprising hov;r readily philoso­ that the latter seem to have been able to muster, in answering the case of the rela­ phers fall into the trap of suppos:ing that tivists or emotivists and pressing their O\'Vn, the "tribal taboo" account of moral agree­ have been metaphysically and logically du­ ment implies ( or entitles one to say) that bious appeals to theological revelation, or moral judgments are not factual. Let us seemingly question-begging intuition, or suppose ( -.,,yhat is surely not the case either, ( though I would say that this is not wholly for that matter) that all the particular ta­ boos in a given set have no other property irrelevant) some form of "naturalism," common1 are bound together by no other in I propose therefore to begin with a consideration of the fact of widespread principle than that they have been singled


out and called "taboo" by the tribal priests. So that to say that a particular form of be­ havior is "taboo" would be to say nothing more than that it is one of the proscribed forms of behavior on this arbitrary list. The term "taboo" would then function some­ what as do proper names, ,vhich designate or denote particulars without describing or characterizing them- except in certain de� rivative and incidental ,:vays. (E.g.) in terms of acquired emotional associations, or in terms of developing convention, or the like--as the proper name "Mary" is a fe­ male's name, or as we all seem to be able to understand what it means to say of a po­ litical leader that "He's no Winston Churchill/' or uHe�s a veritable Hitler.") The point is this. Surely the mere fact that an arbitrarily and whimsically selected set of particulars has arbitrarily been desig­ nated by an otherwise :neaningless sound or mark in no way negates that anyone who gras11s and accepts the arbitrary con­ vention, and subsequently identifies the particular or set by this terrr•0 thereby makes a factual judgment-and if the iden­ tification be correct, then a true judgment. ( I am not prepared to go all the way "�th tl1ose linguistic philosophers who have said that knowing what something is is simply knowing its correct name. But I think it is undeni;ble that the converse-so far as it goes, which is admittedly not very far-is certainly true.) Thus, even ii moral judgments were no more than tabOo-identiflcations, or proper- name identifications, they would still be factual. "That is Tom Jo�1es is a factual judgment and is either true or false. It could even be a true judgment if Tom were an illegitimate foundling child, having neither biological nor adoptive relation.. ship to any family named Jones, and had been assigned this name quite arbitrarily. Moreover, to say "That is Tom Jones" is not to make the still more trivial statement '1f that is the person I think it is, then it is ( Le., his name is ) Tom Jones" ( a state­ ment, that is1 which ventures no more than

RELATIVISM AND JUSTIFICATION a memory-claim judgment about a proper name). To say "Tnat is Tom Jones" is surely to claim at least both that the indi­ vidu�l is the one that the speaker takes hi:n to be, and that that is his name. Similarly, to say that "Taking the tribal god's name in vain is taboo" is to express a relatively complex factual judgment which is either true or false, regardless of the origin of this particular taboo for the particular tribe, and regardless of the origin or the word­ concept "taboo." And to say of a specific action in a specific instance that "In doing that he violated the taboo," would be to make a factual judgment of a still more complex type. By the same token, even if the "tribal-taboo" or "catechism" account of n1oral judgments were the who1e story, this would not deprive them of their rightful classification as factual judgments, judg­ ments which not only may call for the re� sponses "I agree/7 or ''I disagreei " but which also properly take the semantic ad­ jectives "true" and "false." Thus, the state­ ments "Wanton killing is wrong" and "It was wrong to kill him when he was plainly surrendering" clearly express factual judg­ ments. They are either false or ( as I submit all of us who have learned to use moral terms and sentences recognize in, this in­ stance) true. To be sure" some really inter­ esting and very difficult questions remain: vV}iat mo:e, if anything ( other than that it is a moral fact) can we say about the nature of the fact that we are rec-ognizing v;rhen we correctly judge and identify something as being wrong, or, as the case may be, right? How, if at all, can the word-concepts in whinh we make our n1oral judgments be analysed-and if they cannot be, why not? \ social or other ernpirical factors generate and affect our moral discern­ ments? But wherever these quests might lead us, it is a sheer mistake to suppose that the taboo-catechism account implies that moral judgr::ients are not factual. Failure to recog­ nize this obvious logical point seems to me


a howler on a par with the one introduced by David Hume ( and alas since perpe­ tuated) when he made the distinction be­ tween "matters-of-fact'' and "relations­ of-ideas"-as though it were not a fact that two and two make fouri or true that every debtor has a creditor. ( The admirable Hume, it may be noted in passing, also contributed his part to the vulgar confu­ sions which he decried in discussions of ethics-but that is another story. ) The question of what kind of facts or truths log­ ical facts or truths and moral facts or truths are, and how they are discovered or deter­ mined, is a different question. That they are facts or truths is simply not in question. Thus far I have been discussing the ta­ boo-catechism type of explanation for ·the degree of consensus that does obtain in moral discourse, and the mistake in the common supposition that such an account implies that moral judgments are not fac­ tual in character. I turn now to the alterna­ tive explanation: that moral terms are gen­ uine factual predicates. In this connec­ tion I shall in due course take up the ques­ tion of the moral disagreement that exists. And I shall also consider ( what it may be noted I have stated but not yet discussed) the inadequacy of the taboo-catechism ac­ count as an explanation for all the phe­ nomena of moral discourse. On the view that moral terms are gen­ uine factual predicates, moral judgments are factual in more than the respect in which they would be, as I have shown, even if the taboo-catechism account were the whole story. That story, so far as it goes, need not be discarded on the present view, it should be appreciated. It could hardly be denied that, whatever else \Ve may say of them, moral terms are predi­ cates of incalculably great social impor­ tance- like such undisputeclly genuin::pect him to discern in common among the diverse moral precepts, or among the varied examples in_ terms of which a given one is taught. How would this be possible, and especially how would it be possible that he could by himself dis­ cern the n10ral features in new and increas­ ingly more complex types of circumstances, and initiate new and subtle judgments of his own? Let us consider what it is for a term to have semantic reft?,rence and for a judg­ ment to be factual in character. In a certain perhaps Pickwickian but nonetheless lllu­ rninating sense� every factual assertion ex� presses what, to mark the Pickwicban re­ spect, and to distin'guish it from what is noimaUy understood by the words ''Judg­ men t of comparison," I shall here term an extrinsic judgm ent of comparison." W'hat is normally meant by the phrase ajudgment of carnparisod' are those which have some such explicit form as ''this is like 1 or dif� fere,:it from, or larger than that," These we might describe by way of contrast as intri-n� sically comparative judgments, for i t is their essence and distinction that thev refer to one thing in comparison with a..'1other . . Moreover, they e>.'Press the · comparison ex­ plicitly, in contrast with those which; while analysis reveals them to be intrinsicallv . comparative, express a more obscure or in{_plicit sort of comparison. I h2ve in 1nind here such judgments as "'He i"llproved the situation," and especially the 1nore subtle cases such as "That weighs ten pounds/' or "The specific heat of that compound ls 2.38"'-at the botton1 of ,vhich, of course, are comparisons or ratios. Now the respect in which all judg­ rnents may be regarded as re.:8.ecting a com­ parison is perhaps not wholly unlike the re­ spect in which these last, more s,1btle kinds of cases do so-weight, specific heat; mass, atomic number, hard, soft, and the like. Yet a


obviously it must be a different respect. For it is a respect, I am saying, in which all judgments reflect comparisons. What is this respect? In order to judge in a present instance that something is or is not the case, it is clear that \Ye must previously have expe­ rienced ( or in some way based on our ex­ perience have been infonned about) the kind of thing that we are judging the pres­ ent something to be, or not to be,. a case of. Correspondingly, to judge that a partic­ ular expression is sernantically appropriate to describe a present case, we must pre� viously have experienced ( or have been in­ formed about) the kind of case to which it is semantically appropriate. Thus in judg­ ing that the situation before us is of a spec­ Hied character and semantically warrants a given verbal expression, we are in effect judging-recognizing, that is-that 1J1e situ­ ation is, in the relevant respects, like the s:t­ uations in which we previously encoun­ tered that property and learned to use that semanticaIIy appropriate expression. And if the situation is indeed recognizably of that character, then what we have judged and said to be the case is the case, is so, is true, is a fact. It should of course be understood that in presenting this perspective about all judgment and all assertive utterance I need not hold any such naive view as that we learn all our words and concepts in tenns of a specifiable set of unarguably paradigm cases at speciliable times in our lives. Nor am I suggesting, what is also patently con­ trary to the facts, that whenever we judge something or employ an expression we must do so self-consciously, or go through any conscious or even unconscio�s steps, or ho.ld up in our heads some paradigm image, The point I am making about all jud!,rment and assertive c:tterance is a logi­ cal not a psychological�one. To judge that the man on the left is taller is bdubita!:ily a comparative judg­ ment of the normal sort, even though I

RELATIVISM AND JUSTIFICATION need not attend to anything but the two in­ dividuals when I make it. I need not be self-conscious nor e:1tertain any images or other thoughts about standards of linear measurement or of verbal propriety. My judging, here, except in well-k!!own special kinds of circumstances. normallv does not consist in anything mo�e than shnply look­ ing, noticing, and perhaps saying. Yet ob­ vim:sly the logic of what has occurred \Vhen someone asserts that the man on the left is taller involves what in an undeniable respect we must acknowledge to be a judg­ ment of compa::ative height. In the same way, then, I vvould argue that considera­ tion reveals that the logic of the whole en­ terprise of j,:dging and asserting, including the intrinsically and explicitly comparative judgments normally identified as such, rests on what in an '-!ndeniable further and spe­ cial respect we I!lUSt acknowledge to be implicit comparison. Consider now moral judgments. We lean, in chi'.dhood and subsequently to use rnorai terms by experiencing their use in concrete situations. As \Vith al! simple terms learned early in our linguistic life, we learn them. ostensively�-i.e,, not by their being analysed or verbally defined for us, and not1 mo!'eover, as self-conscious exer­ cise in vocabulary building, bc:t rather hy experiencing their use in practical connec­ tion with the very kinds of things, proper­ ties or situations to which they pertain. And to put it succinctly, 2lbeit admittedly crudely, this experience of their nse is what they mean to us. Our early use of such terms, at least, is of course a kind of conditioned response, a rote use. By the very irnport of "';:ote," even consistent rote use of ru1 e;,.,,-pression is ob­ viously not a guarantee. of meaning iillder­ stood; nor, to be sure, does it establish t!" an expression is not being used as a proper na1ne, but designates fo.r the user an identifiable characterizing property or kind of situation predicable in a factual judg­ :nent. AIi this is established only when the



with rigor or certainty in many or even most cases-·though quite easy to apply in some. ( Consider the simple-and I might add "non-natural"-property of entaiinoent, or of contradiction, which is easy to per­ ceive in some propositional sets, and ex­ tremely difficult to decide in others. ) Are we really prepared to classify as non-factual all issues for which there is no acknowledged method of conclusively es­ tablishing truth and resolving dispute? Surely if there were no pre-judgment of the case; no other reasons-or causes­ persuading us that moral terms are not II genuine predicates and moral judgments not factual, this lack of a decision method Lest my argument seem to be no more than would not of itself lead us lo that conclu­ a tour de force resting solely on the admit­ sion. Ho·w many1 reaJly, of our predicates tedly technical ( though I insist sound) and judgments, outside of mathematics point of what it means to be a factual predi­ and, less certainly, the quantitative natural cate and a factual judgm ent, I need to say sciences� are determinable by UJ1iversally something more about fue troublesome acknowledged operational definitions and business of rnoral disagreement and diver­ dec'.sion techniques? Many, most, perhaps gence, and about our :inability to resolve even all of our everyday concepts are to this by established definitions and decision some degree vague and ill-defined; and hence so must our jud1,tments be. Few of procedures. It is, also, true that moral clifferences them permit of unequivocal test even mav and freauentlv do persist after the dis­ ,vhere there is agreement in application, putants are talkea' out. The disputants dis­ nor of resolution \Vhere there is not. Though we tend not to foc1:s on this, engage, either moral1y indignant or sadlyi realize when vve do that \Vith may ,ve but in either case not knowing what more to do or say- and perhaps even admitting many of the predicates and jud;,tments that that there is nothing more that can be done we do not hesitate to cla£Sify as factual, it or said. In this respect moral disputes are is by no means a foregone conclusion that like disputes of taste-and de gustibw: non the class of borderline cases must be es'/; disputandum we all know. Or do we? smaller than the class of clear- cut ones. Perhaps so about de gustibus, and perhaps Consider such notoriously vague and loose, not. That is another pickle. The point I but nonetheless admittedly factual terms want to make here is that the facts of as; "useful/' "'plausible," "'justifiable;' "'ra­ moral differences are not at all difficult to tional/' '"intelligent," �(witty/ "cleari" reconcile with the thesis I am advancing. ute the proposition that intrinsic value is always io proportion to quantity 0£ pleasure we must dispute this argument But the argument may seem to be ab10st indisputable. It has, in fact, been used as an argument in favour of the proposition that intrinsic value is always in proportion to qua.'1tity of pleasure, and I think it has probably had muc.¾ influence in inducing people to adopt that view, even if they have not expressly put it in this fonn, How, then, can \\'e dispute this argu­ ment? \Ve rnight, of course, do so, by re­ jectlng th e proposition t.liat no whole can ever be intrinsically good) unless it con­ tains some pleasure; but, for n1y pa.rt:, though I don't feel that this propo­ sition is true, I also don't feel at all certain that it is not true. TI1e pa.'i: of the argument which lt seems to me certainly can and ought to be dispt:ted is another part­ namely, the assumption that, where a.



whole contains two factors, A and B, and cases )-that these two facts together must one of these, B, has no intrinsic goodness at have a certain a1notu1t of intl"',..nsic value) all, the intrinsic value of the whole cannot that is to say must be either intrinsically b e greater than that of the other fac'i:or, A. good, or intrinsically bad, o, intrinsica.lly This assumption, I think, obviously rests on indifferent, and that the amount by which a still more general assumption, of which it this value exceeds the value which the ex­ is only a special case, The general assump­ istence of A would havP½ if A existed quite tion is: That where a whole consists of two alone, need not be equal to the value which factors A and B, the amount by which its the existence of B would have, if B existed intrinsic value exceeds that of one of these quite alone, This is all that we are saying. two factors must always be equa.l to that of And can any one pretend that such a view the other factor. Our special case will fol­ necessarily contradicts the laws of aritlh,ne­ low from tbis general assmnption: because tic? Or that it is self-evident that it cannot it will follow that if B be intrinsically indif­ be true? I cam1ot see any ground for sayfrig ferent, t.1:iat is to say, if its intrinsic so; and if there is no ground, then the ar­ value - 0, then the amount by which the gument which sought to show we can value of the whole A+B exceeds the value never add to the value of any whole except of A 1nust also = 0, that is to say,, the value by adding pleasure to it, is entirely base­ of the whole must be precisely equal to less. If, therefore, we reject the theory that that of A; while if B be intrinsically bad, that is to say, if its intrinsic value is less intrinsic value is always in proportion to than 0, then the amount by which the quantity of pleasure, it does seem as if we vahie of A+ B w'.'.; exceed that of A will onay be compelled to accept the prindple also be less than 0, that is to say, the value that the amount by which the value of a of the whole will be less than that of A. whole exceed$ that of one of its factors i, Our special case does then follow from the not necessarily equal to that of the remain­ general assumption; and nobody, I think, ing factor-a principle which, if true, is \Vould maintain that the special case \Vas very important in many other cases. But, true \Vithout maintaining that the general though at first sight this principle may assumption was also true. The general as­ seem. paradoxical; there seems to be no rea­ sumption may, indeed, very naturally seem son ·whv we should not aceen'!: it; while to be self-evident: it has, 1 think, been gen­ there a;e other independent r'"easons why erally assumed that it is so: and it may we should _accept it. And, in any case, it seem to be a mere deduction from the laws seems quite clear that the degree of intrin­ of arithmetic. Buti so far as I can see, it is sic value of a whole is 11ot always in pro­ not a mere deduction from the laws of portion to the quantity of pleasure it con­ arithmetic, and, so far from being self­ tains. But, if we do reject this theory, what, evident, is certainly untrue. it may be asked, can we subs:itute for it? Let us see exactly what \Ve are saying) if we deny it. 'l'Ve are saying that the fact How can we answer the question, vvhat that A and B both exist togethe,, together kinds of consequences are intrinsically bet­ with the fact that thev have to one another ter or worse than others? vVe may, I think, say, first of all, that any relation which th�y do happen to have ( when they exist together, they always for the same reason for which 'WC have re­ mus� have some relation to one another; jected the view that intrinsic value is a l ­ and the pcecise nature of the relation cer­ ways in proportion to quantity of pleasure, tainly may in some cases 111ake a great \Ve must aJso reject the vie,.v that it is al� difference to the value of the whole state of ways in proportion to the quantity of any things, though, perhaps, it need not in all other single factor whatever. \.Vhatever sin-


gle kind of thing may be proposed as a measure of intrinsic value, instead of pleas­ ure-whether knowledge, or virtue, or wis­ dom, or love- it is, I think, quite plain that it is not such a measure; because it is quite plain that, however valuable any one of these things may be, we may always add to the value of a whole which contains any one of them, not only by adding more of that one, but also by adding something el.,e instead. Indeed, so far as I can see,, there is no characteristic whatever which always distinguishes every whole which has greater intrinsic value from every \vhole which has less, except the fundamental one that it would always be the duty of every agent to prefer the better to the worse, if he had to choose behveen a pair of actions, of which they would be the sole effects. And similarly, so far as I can see, there is no characteristic whatever which belongs to all things that are intrinsically good and only to them-except simply the one that they all are intrinsically good and ought al­ ways to be preferred to nothing at all, if we had to choose between an action whose sole effect \vould be one of them and one ,vhich vmuld have no effects whatever. The fact is that the view which seems to me to be true is the one which, apart from theo­ ries, I think every one would naturally take, namely, that there are an immense variety of different things, all of which are intrinsically good; and that though all these things may perhaps have some char­ acteristic in common, their variety is so great that they have none, which, besides being common to them all, is also peculiar to them- that is to say, which never be­ longs to anything which is intrinsically bad or indifferent. All that can, I think, be done by way of making plain what kinds of things are intrinsically good or bad, and what are better or worse than others, is to classify some of the chief kinds of each, pointing out what the factors are upon which their goodness or badness depends. And I think this is one of the most profita­ ble things which can be done in Ethics, and


one which has been too much neglected hitherto. But I have not space to attempt it here. I have only space for two final re­ marks. The first is that there do seem to be two important characteristics, \Vhich are common to absolutely all intrinsic goods, though not peculiar to them. Namely ( 1 ) it does seem as if nothing can be an intrinsic good unless it contains both some feeling and also some other form of consciousness; and, as we have said before, it seems possi­ ble that amongst the feelings contained must always be some amount of pleasure. And ( 2) it does also seem as if every in­ trinsic good must be a complex whole con­ taining a considerable variety of different factors- as if,. for instance, nothing so sim­ ple as pleasure by itself, however intense, could ever be any good. But it is important to insist ( though it is obvious) that neither of these characteristics is peculiar to intrin­ sic goods: they may obviously al.,o belong to things bad and indifferent. Indeed, as re­ gards the first, it is not only true that many wholes which contain both feeling and some other form of consciousness are in­ trinsically bad; but it seems also to be true that nothing can be intrinsically bad, unless it contains some fe8ling. The other £nal remark is that we must be very careful to distinguish the two ques­ tions ( 1 ) whether, and in what degree, a thing is intrinsically good and bad, and ( 2 ) whether, and in what degree, it i s capable of adding to or subtracting from the intrin­ sic value of a whole of which it forms a part, from a third, entirely different question, namely ( 3 ) whether, and in what degree, a thing is useful and has good effects, or harmful and has bad effects. All three questions are very liable to be con­ fused, because, in common life, we apply the names 'good' and 'bad' to things of all three kinds indifferently: when we say that a thing is 'good' we may mean either ( 1 ) that it is intrinsically good or ( 2 ) that it adds to the value of many intrinsically good wholes or ( 3 ) that it is useful or has


good effects; and similarly when we say that a thing is bad we may mean any one of the three corresponding things. And such confusion is very liable to lead to mis­ talces, of which the following are, I think, the commonest. In the first place, people are apt to assume fvith regard to things, which really are very good indeed in senses ( 1 ) or ( 2 ) , that they are scarcely any good at all, simply because they do not seem to be of much use-that is to say, to lead to further good effects; and similarly, ,,�th re­ gard to things which really are very bad in senses ( 1 ) or ( 2 ) , it is very commonly as­ sumed that there cannot be much, if· any, harm in them, simply because they do not seem to lead to further bad results. Noth­ ing is commoner than to find people asking of a good thing: vVhat use is it? and con­ cluding that, if it is no use1 it cannot be any good; or asking of a bad thing: vVhat harm does it do? and concluding that if it does no


harm, there cannot be any harm in it. or, again, by a converse mistake, of things which really are very useful, but are not good at all in senses ( 1 ) and ( 2 ) , it is very commonly assumed that they must be good in one or both of these two senses. Or again, of things, which really are very good in senses ( 1 ) and ( 2 ) , it is assumed that, because they are good, they cannot possi­ bly do harm. Or finally, of things, which are neither intrinsically good nor useful, it is assumed that they cannot be any good at all, although in fact they are very good in sense ( 2 ) . All these mistakes are liable to occur, because, in fact, the degree of good­ ness- or badness of a thing in any one of these three senses is by no means always in proportion to the degree of its goodness or badness in either of the other two; but if we are careful to distinguish the three different questions, they can, I think, all be avoided.

Why Not Hedonism? A Protest RALPH MASON BLAKE In current discussions of ethical theory it has become the tradition to treat hedonism on the more or less definite assumption that its rejection is a foregone conclusion. Its falsity is so thoroughly taken for granted that it is thought possible to dispose of its claims in very short order indeed. It is usually treated with ill-disguised contempt as an antiquated heresy that has so long since been definitely refuted that its truth can scarcely be contemplated as a genuine possibility at all. Its "fallacies" are summa­ rily pointed out, in a few brief pages, and we are hurried on to a consideration of theories more worthy of the attention of a mature mind. I find this state of affairs extremely u n ­ satisfactory. For my own part, I may as well say at once, I accept the hedonistic position. If there are any serious difficulties in the way of such acceptance, I very much wish to know what they are. The current "refutations" seem to me for the most part simply puerile-so much so, indeed, that I should feel a sense of shame in proposing to discuss them seriously were it not for the scandalous fact that they are still solemnly repeated and piously deferred to. Perhaps it is some such sense of shame that keeps Mr. S antayana, who, ahnost alone among contemporary writers, adopts a fully he­ donistic position, from dealing with the current objections. Mr. Santayana, how­ ever, prefers in general to expound rather than argue his philosophy, and the truth of Reprinted from the journal Ethics, 1926, by R. lvf. Blake by permission of The University of Chicago Press.


hedonism in particular doubtless seems to him so evident that he cannot bring himself to any careful consideration of the alleged difficulties. The occasion therefore seems ripe for such a discussion, and it is soine­ thing of the sort which, in summary fash­ ion, I propose to undertake. Let it first be clearly understood, how­ ever, that in attempting to defend hedon­ ism from its critics I am by no means con­ cerned with the integral defense of any his­ torical hedonistic system. By hedonism I do not mean Epicureanism, or Benthamism, or the doctrines of J. S. Mill. There is not the slightest difficulty in showing, and it has in fact been demonstrated ad nauseam, that these historic theories are one and all in­ fected with serious fallacies and gross er­ rors. I believe, however, that there is a set of fundamental principles which, ,vhether or not it has ever been held in this precise form by any of the classical proponents of hedonism, at any rate seems to have been more or less approximated by each of them; that this set, moreover, will be rec­ ognized as undoubtedly constituting a he­ donistic system; and, finally, that this he­ donistic system is by no means to be dis­ posed of by the simple device of showing that it has usually been inadequately stated, defended by fallacious arguments, and combined with inconsistent or erro­ neous principles. This central core of hedonistic doc­ trine has been most clearly and completely disengaged from its various historical ac­ companiments not by any advocate of the theory, but, oddly enough, precisely by its


acutest critic, Mr. G. E. Moore ( in his Eth­ ics, Home University Library). Ignoring all complications and refinements of interpre­ tation1 the bare essentials of this view can be stated in a highly compressed fonn in seven propositions. Of these the 6rst two are simple prnliminary definitions of tenns. They are as follows: ( 1 ) To say of a thing that it is intrin­ sically good means that it would be good even if it existed quite alone, without any accompaniments or effects '\Vhatever. ( 2 ) To say of a tlring that it is ulti­ mately good, oc good for its e>um sake, ::neans ( a ) that it is intdnsically good, and ( b) that it contains no part which is not in­ trinsically good. The next tbree propositions are deflni­ tions of moral concepts. They are not pecu­ liar to hedonism as such, but characterize it as teleological, rather than as a formalistic system of etbics. (For this distinction, cf. Paulsen's System of Eth£cs, Book II, Chap­ ter l. ) These propositions are as follows: ( 3 ) To call a voluntary act wrong means that the total con.sequences of some other action possible to the agent under the circumstances form a whole which is intrin­ sically better than the whole formed by the total consequences of the act in question. ( 4) To call a voluntary act right 111eans that it is not wrong. ( 5 ) To call a voluntary act a duty for a given agent, or to say that the agent ought to perform it, means that1 among t11e acts possible to the agent under the cir­ cumstances, the total consequences of the act in question fonn a whole which is in­ trinsicatv better than the whole formed bv the totaf" consequences of any of the oth;r possible acts. The next proposition also is not neces­ sarily peculiar to hedonism as such. It amounts to a denial of Mr. Moore's pdnci­ ple of "organic unity" ( cf. his Principia Ethica) , ,rnd might fom1 a part of a non­ hedonistic system. This principle is as fol­ lows: ( 6 ) The intrinsic value of a whole is al-


ways in proportion to the amount of ulti­ mate value which it contains� The last proposition is the characteris­ tic and peculiar thesis of hedonism: ( 7 ) Pleasurable consciousness is a l ­ ways ultimately good ( o r good for its own sake); and nothing else is ever ult::mately good. And now for the current objections. Most of these, I think it will be readily seen, simply do not touch at all the theory previously stated. So far as it is concen1edi they are completely beside the mark. For example, much ink is still expended on the refutation of psychological hedonwn, i.e., the once fashionable theory that the sole human motive is the desire for pleasure. ( Cf., e.g., Dewey's constant recurrence to tlris subject in his Human Nature and Co"' duct. ) I n view of the fact that no sedous thinker, so far as I know, now holds any such theory, this ink ls simply wasted. The doctrine \Vas,. of course, accepted by most of the hedonists of t:.'1c oast and was i n ­ deed frequently put fo�ara" by them a s a proof of the truth of their ethical theory. Its falsity is, however, now generally recog­ nized, and its uselessness as a basis for demonstration of the ethical theory fully admitted. It evidently forms no part of the doctrine previously state� and is in no ,vay implied thereby. Discussion of it is there­ fore whollv irrelevant to anv living issue. But e;:,en those who re�ognize the irrel­ evance of psychological hedonism some­ times formulate and criticize ethical hedon­ ism i n a way which is ab1ost equally irrele­ vant. Hedonists hold-so the matter ls fre­ quently put-that even though the desire for pleasure is not actually the sole human motive, yet nevertheless it ought to be. Thus Miss Calkins tells us ( The Good Man and the Good, p. 73) that "according to this the017 the proper, though not the in­ vadable, object of the morally willing self is pleasure." And according to �1r. J oad ( Commo� Ethics, p. 13; cf. p. 15 ) "although other things besides pleasure may be desired, pleasure is the only thing


that ought to be desired." Now, whether or not such a doctrine has actually formed a part of the hedonistic systems in the past, it certainly ought to be plain that it is by no means an essential feature of such a sys­ tem. Not only is it not in any way implied in the foregoing propositions, but on any reasonable view of things would seem even to be excluded thereby. For according to these principles to say that no man ought ever to desire anything other than pleasurn means that the total consequences of a de­ sire for something other than pleasure never form a whole which contains more pleasurable experience than does the whole formed by the total consequences of any other possible desire; and to most observ­ ers of human life this statement seems to be plainly false. Hence, indeed, the famil­ iar "hedonistic paradox" to the effect that "pleasure to be got must be forgot"-a par­ adox which may certainly be quite consist ­ ently accepted by hedonists. The fact that hedonists judge the value of acts by refer­ ence to their consequences in pleasure by no means commits them to ·the view that such consequences are best attained by making them directly the sole motive and the sole object of human desire. Nobody makes any scruple of admitting that such a valuable end as health, for example, is not best attained by making it a direct and constant object of conscious concern. It should therefore not surprise us to find that the like is true also of other valuable ends. In fact, I am not familiar \vith any system of ethics in which it is held that desire for the ultimate end proposed by the system should be made the sole human desire. vVhy then should it be supposed that he­ donists alone are bound to maintain such a doctrine? As for the connection of hedonism with egoism� it might seem superfluous at this late date to insist that hedonists need in no way be adherents of egoism. Yet this confusion still to some extent persists, even among ,vriters who ought to know very much better. Thus Professor Muncterberg


seems to have supposed that a hedonist can

consistently regard any given act as consti­ tuting for him a duty only provided he can view it as resulting in a preponderance of pleasure for himself; for he argues against hedonism as follows ( Eternal Values, p. 39 ) : "When we will the morally good, we do indeed wish that the good also give us joy, but we know that it is not good simply because it gives us pleasure. . . . E'ven if we acknowledge the pleasure in the minds of other human beings as goal for our moral action, the moral self is not therefore based on pleasure. . . . We feel it our duty to serve the pleasure of others, but this duty cannot itself come into question as a pleas­ ure. \Ve may submit to it with pleasure, but we do not submit to it because it gives us pleasure." But need a hedonist maintain that our duty must needs be a pleasure in the doing? Mr. Joad also thinks it an inconsis­ tency . in a hedonist to admit "that the indi­ vidual can, and ought to, desire something which may have no relation to his mvn pleasure, namely, the good of the commu­ nity" ( op. cit., p. 15 ) ; for he tells us not only that Mill's implicit admission of this doctrine involved him in inconsistencies, but also that c'these inconsistencies in Nfill are important, and I have dwelt on them at some length because they demonstrate the impracticability of maintaining, even with the best "�11 in the world, that pleasure is the only thing of value . . . They reveal themselves most completely in lvli!Ys work, but they are implied in any form of utilita­ rian hedonism." ( Op. cit., p. 16.) I should very much like to have it pointed out just how the rejection of egoism implies any in­ consistency in such a form of utilitarian he­ donism as that outlined above. Another classical line of attack upon hedonism consists in elaborate criticism of the so-called hedonistic calculus. If the rightness or wrongness of actions depends upon the degree to which a greater or less "quality of pleasure" is realized in their consequences, then, in order definitely to



determine upon the rightness or wrongness This seems to n.1e so1nething of a caricature of anv action, we must be able somehow to of the hedonistic view, I venture to think predi�t the consequences of various actions that even Bentham and lviill, who no doubt and to estiinate the relative "'quantities of entertained somewhat exaggerated notions pleasuren involved. Now no one would with regard to the applicability of the cal­ deny, I suppose1 that it is no easy matter to culus, never went to anything like such forecast the future, especially io such a lengths. And what ls Professor Dewey's co:::plicated sphere as that of human con­ O\v:!l conclusi�n with regard to the calcu­ duct and its effects; nor will anyone be dis­ lu.s? He tells 'us that posed to doubt that there are grave diffi­ culties involved in the determination and the problem of deliberation is not to calculate comparison of quantities of pleasure and future happeningst but to appraise present proposed action. VVe judge present desires and displeasure. I cannot here enter into the habits by their tendency to produce certain details of th:s question, but it ought to be eo::isequences , , , The future outcome is not clearly unde,stood that precisely similar certain . . • But its tendency (i.e., the tendency difficulties affect any teleological system of of t.¾e fire vvhic-h Professor Dewey uses as an ethics whatever. Every such system makes example) is a kno,vable matter) what it will do i:he value of actions depend upon the quan­ under certafrl circumstances. And so we know tity of good which they succeed in realiz­ \vhat is the tendency of malice, charity� con­ iog; every such system holds faat some ac­ ceit� patience, \;\,1e know by observing their tions are better t..\an others; that some real� consequences, by recollecting what we have ize more and some less good. In every such observed" by usfag that recolJectfon in con­ system, therefore, some smt of ''calculus" is structive imaginative forecasts of the future, by using the thought of future consequences to necessary, and I flnd it very dL'licu!t to un­ tell the quality of !he act now proposed. derstand how it can be any easier to deter­ (Human Nature and Conduct, pp. 206-7.) mine such quantitative questions as are here involved in terms of "satisfaction of Precisely. And I believe that this is very desire" or ,•nony" or ·(self-realization" much the sort of thing that the hedonist than it is in terms of pleasure and displeas­ means by his despised "calculus." But, it wm perhaps be saicl, we cannot ure. Hedonists have at :ea5't made a reso­ lute attempt to deal with this aspect of the trust men to guide their actions by any matter. I fail to see how any teleological such calculus. The vast majority of man­ system can view such an attempt as kind are incapable of judging with any a c ­ superfluous; and I an1 not aware that any curacy concen1ing the consequences of system has made the attempt with more their acts; they can neither foresee them all earnestness or \\1th greater success than he­ nor correctly estimate their value. So far as this is true, ho;,vever, it is true on amt teleo­ donism. Professor Dewey is one of those who logical syste111. And in any such sys7;m it is make much of the defects of t-lie "'r"'1Jou]ul' O:lly .so far as some such calculus can more as an objection to hedonjsm. B::-;t I cannot or less roughly be carried out that any help t.¾inking that he interprets the doc­ knmvledge of the rightness or. wrongness of trine in a manner that is highly artifici,,l actions can be attaioed. vVhere the esti- · and unreal. He seems to suppose that in mate fails we must be content to remain jn the ,iew of hedonists the calculus can be ignor2nce. Those who are incapable of applied with perfect mathematical preci­ gc.iding themselves must, as in other mat­ s:on to fue determination of results of indi� ters, be gc,:ced and taught by those pos­ vi.dual acts severallyi and that, too, vvith sessed by more adequate vision. "But in the crisis of tleG'lSion a man absolute accuracv and certaintv of result. (Human Nature'and Conduct, 'pp. 50-51.) cannot be expected always to ·defoy action


until he has completed such a calculation of cons.equences." In truth, no one expects him to do so. As well expect a man on every practical problem of mathematics to calculate his logarithms afresh. In ordinary practice the mathematician relies upon his logarithmic tables. Just so in practical ac­ tion are we guided by more or less expli­ citly formulated codes of action, systems of general rules which we take to be the re­ sult of past e>.'])erience, and to embody a measure of wisdom and foresight. V{e do not question the validity or the value of a method of calculating logarithms because not every Tom, Dick, and Harry is capable of performing it, nor because in practice we follow the results of past calculations. Similar considerations should apply to the calculus of pleasures. "But a calculus of pleasures can give no definite result. \'Vhat is one man's meat is another man's poison. VVhat is happiness to one man is unhappiness to another." This, of course, simply means that what gives pleasure to one man fails to give it to another. This is certainly a fact which must be taken into account by any theory, but if the sources of pleasure thus vary from man to man, is it not so also with the sources of any ultimate good which one may choose? Are the sources of «satisfactions of desire" or of "self-realization" any less diverse and conflicting? Hedonism here again appears to be no worse off than any other teleologi­ cal system. Professor Albee ( History of English Utilitarianism, p. 274 ) puts the argument a little differently: A direct computation of the consequences of actions, in terms of happiness or unhappiness, can never afford the foundation for a scientific Ethics, not merely, or principally, because experience shows that individuals derive pleas­ ure and pain, as the case may be, from very different things; but because it is absolutely certain, on general principles, that every advance in morality involves a shifting of the scale of hedonistic values. Otherwise eA-pressed, individuals and nations are con-


stantly, if generally slowly, discarding one scale of hedonistic values for another, pre­ viously assumed to be ultimate, and this in proportion to the development of moral char­ acter. Reduced to its lowest terms, this means that hedonistic values vary as moral character varies.

It is difficult to see that this way of stating the matter makes things really any harder for the hedonist. It is true that men's judg­ ments regarding the sources or causes of pleasure are subject to frequent change, but so are the judgements with regard to the source of any sort of ultimate value. It is true that as men change they derive pleasure from different things and in diffe­ rent degrees than formerly. But whatever theory of ultimate value is adopted, it seems likely that changes in human thought and character or in other condi­ tions of life will bring with them similar al­ terations in the sources of value. The mere fact that changes in moral character consti­ tute one cause of such alterations in the sources of value seems to introduce no es­ sentially novel difficulty into the argument; and in any case the fact would remain the same and the difficulty equal on any theo­ logical theory whatever. But arguments based on the difficulties of the calculus do not exhaust the case against hedonism. There is no stopping the chorus of objections. "A state of unbroken pleasure would not really be pleasant. A continuous heaven of constant enjoyment would be intolerable boredom." Such an objection surely represents mere confusion of thought. How can pleasure be unpleas­ ant? How can enjoyment be boredom? "But uninterrupted pleasure is an ideal which is impossible of attainment in any actual human life. It is a mere chimera.'' Or, as Professor Rogers puts the point: "That at \Vhich a sensible human being aims is no unimaginable state of the intens­ est possible pleasure unaccompanied by pain. -. . . Rational satisfaction is no dream of an undisturbed and impossibly complete state of felicity." ( Theory of Ethics, pp.


49-50.) But what hedonist has asserted the attainability of such an ideal? Are hedon­ ists, then, wholly ignorant of the conditions of human life? In truth they are not so fool­ ish as to maintain the attainability of any such pedect -consummation, or to counsel attempts to realize the impossible. \Vhat they hold is simply that the intrinsic value of any state of affairs is in proportion to the amount of pleasurable experience it con­ tains, and that human effort should be so directed as to make this amount as great as possible. No overstrained idealism is im­ plied. A hedonist may, indeed, consistently also be a pessimist. vVe should not forget that the truth of pessimism has often been argued from hedonistic premises. But then, says Professor Rogers, "in practice the only clear meaning, therefore, that a 'sum of pleasures' carries is this, that I want my life to be a continuous series of satisfied moments lasting as long as possi­ ble. But this is pretty much an empty plati­ tude, which throws ahnost no light at all on what constitutes satisfaction at any given moment." ( Op. cit., p. 48.) But is it to be expected of any theory of the nature of the ultimate end that it should automatically reveal the particular means to that end? For Professor Rogers "that at which a sen­ sible human being aims is . . . the realiza­ tion that he is making the most of life that is possible for him, with his particular in­ terests and limitations, to make, consider­ ing the means at his disposal." ( p. 50 ) How much light does this way of formula t ­ ing the end throw upon "what constitutes satisfaction at any given moment"? The interesting feature of Professor Rogers' case is that de facto he comes very near, despite his protests, to being a hedon­ ist himself-if only he were not so desper­ ately afraid of the name. "Only when we can point to pleasure," he writes, "is the judgment of value felt to be justified." (p. 31) ''No aim will be called reflectively a good aim unless it tends to result in pleas­ ure." ( p. 38 ) But still he will not be called a hedonist. He does not "intend to say that


mere pleasurableness by itself is a good. Pleasantness as such is not good because pleasantness does not exist by itself; a good is concrete, and pleasantness merely an ab­ stract quality." But, may we ask, what he­ donist attaches any value to mere "pleas­ antness" as an abstract universal? It is, of course, only concrete pleasurable experi­ ence which he values. Professor Dewey finds it a fundamen­ tal defect in utilitarianism that it thinks "of the good to which intelligence is pertinent as consisting in future pleasures and pains." This, in his view, involves a "catastrophe."' He emphasizes the "contrast between such conceptions of good and of intelligence and the facts of human nature according to which good, happiness, is found in the pres­ ent meaning of activity, depending upon the proportion, order, and freedom intro­ duced into it by thought as it discovers ob­ jects which realize and unify othenvise contending elements." (p. 212) Professor Dewey's thought, here and elsewhere, seems to be that utilitarian hedonism un­ duly emphasizes satisfactions to be realized in a distant and none too certain future at the expense of the happiness to be gained in the present through the solution of immediate conflict; that it unduly exalts re­ mote and more or less doubtful advantages at the expense of present fruition ( cf. p. 265 ) . That such a tendency to some extent may have operated in the history of utilita­ rianism I will not deny. But I see not the slightest reason to suppose that such an emphasis is the only one which can be con­ sistently adopted by one who accepts the hedonistic position. Professor Calkins rejects hedonism be­ cause of the "narrowness" of its conception of the good. " 'vVhy then,' the non-hedonist protests to the hedonist, 'why do you ex­ clude activity and thought from your con­ ception of the good, or the ultimate end, why do you limit this chief good to happi­ ness alone?"' ( The Good Man and the Good, p. 76. ) The very asking of this ques­ tion, it appears, is sufficient to refute he-



donism, for without more ado Professor pleasure than a wrong ad?" To the ques­ Calkins continues on the next page, "Now tion so stated no hedonist would dream of that the hedonistic answer to this question giving an affirmative answer. Professor is discredited, we tum, naturally enough, to Fullerton also gives a criticism of hedonism anti-hedonistic doctrines." And what do we based on this same misunderstanding. After find? "It is already clear that only one qual­ quoting Bentham to the effect that all itative theory of the good can escape chal­ pleasure is in itself good, even the pleasure lenge for its narrowness. This is the doc­ a malicious man "takes at the thought of trine which describes the good not in terms the pain which he sees, or e;,..-pects to see, of any . one kind of consciousness, as pityl his adversary undergo," he asks, "Can the loyalty, wisdom, or happiness, but as inclu­ pleasure of a malignant act properly be sive of all these experiences and of all oth­ called morally good at all?" ( A Handbook ers which people wish or will for them­ of Ethical Theory, p. 224; cf. p. 240. ) This selves" (p. 78 ) . But is it really so clear that question being answered in the negative, there is no "narrowness" here? :Might we Bentham's position is considered to be re­ not ask, ,vith a pertinence equal to that of futed. But, as we have seen, no hedonist the non-hedonist's former inquiry, "Why do dreams �f maintaining that pleasant experi­ you exclude things not wished or willed ence is in itself a moral good. In fact, for from your conception of the good? vVhy do hedonism nothing whatever is in itself, i.e., you limit the ultimate end to things wished intrinsically, a moral good. Another favorite procedure of the crit­ or ,villed alone?" It is, of course, evident that any theory of the end must distinguish ics is based upon the fact that hedonists between what is ultimately good and what have often professed to give demonstrative is not, and must therefore inevitably suffer proofs of the truth of their theory. These attempted demonstrations are examined from this form of «narrowness." tvfany criticisms of hedonism reduce to and found to be inconclusive, or positively most elementary misunderstandings of the fallacious. Hedonism, it is thus discovered, hedonistic distinction between pleasurable has "failed to prove its case," and we pass experience as intrinsic and ultimate good, on to consider the alternative theories. and moral value as a species of extrinsic or Now let us admit at once that the funda­ instrumental good. Thus A. E. Taylor mental principles of hedonism are incapa­ writes ( Problem of Conduct, p. 327 ) : "A ble of demonstration. So far as I can see, man is not morally good because his career the critics are quite right in rejecting all has been marked by extraordinary cases of the alleged proofs that have ever been good luck, nor is the life of one of the offered, and I know of nothing that can be lower animals to be reckoned morally good set in their place. But how stands the case because it may contain a vast number of with the alternative . theories? Is it possible pleasant moments." But hedonists, of to give a demonstration of their fundamen­ course, do not say that moral goodness con­ tal principles? Their advocates scarcely sists in enjoyment of pleasure; for them en­ pretend that it is. These alternatives are ac­ joyment is an extrinsic value derived from cepted, not because of any rigorous proof the fact that certain actions and disposi­ of their truth, but on quite other grounds. tions result in consequences ,vhich are The fact that hedonism 'Tails to prove its more pleasurable than the results of other case" in itself certainly constitutes no proof acts and dispositions. The same confusion of the truth of any other theory. The truth is apparent also in Taylor's treatment of the of the matter seems to be that no theory of question, Is the good always pleasant? He ethics, in so far a.s it is a question of ulti­ interprets this to mean "Is the morally good mate ends, is susceptible of "proof' in the or right act ahvays accompanied by more strict sense. In fact, even Bentham and Mill

WHY NOT HEDONISM? · A PROTEST were on occasion prepared to admit as much. Thus Bentham says of the funda­ mental principle of hedonism, "Is it suscep� tible of any direct proof? It should seem not; fo r that which is used to prove every­ thing else, cannot itself be proved." ( Selby­ Blgge, British Moralists, Vol. I, Sec. 364. ) And Mill also remarks: "To be inca­ pable of proof by reasoning is common to all Brst principles." ( Utilitarianism, Every­ man Library edition, p. 32. ) Unfortunately, however1 l\ifill nevertheless attempted, in his fourt.\.i chapter, precisely such a proof as he had he,e stated to be impossible. His . lack of success is certainly not sarprising. Fundamental ethical principles, in fact, as Mr. Moore so properly insists, are ac­ cepted or rejected on intuitive grounds. The most that any adherent of any ethical system can do by way of persuading an­ other to accept his theory is to state its fun­ damental principles as clearly and ade­ quately as possible, to take care that these are properly interpreted, and that the is­ sues are not obscured by any confusion \Vith irrelevant or inconsistent doctrines) to exhibit the implications of these principles and their consistency or inconsistency with other human beliefs, and then simply ap­ peal to the reflective judgement of his hearer. If the latter, h:n1ng once clearly understood the principles and their impli­ cations, thereupon rejects them, then cad-it quaestio. For my own part, when I subject to such a test the fundamental principles of hedonistic ethics, they appear to me to ring true. Indeed, it seems to me to be actually self-evident that all pleasurable experience is cltimately good. It does not seerr1 to rn-e self-evident that nothing but pleasurable e;.-perience is ever ultimately good; but much careful reflection has hitherto failed to reveal anything else which does seem to me u,timately good. Again, it is not self­ evide::i:t to me that the intrinsic value of a whole is ahvays in proportion to the amo=t of ultimate good which it con­ tains; but in every instance \Vhich I have

399 ever considered it has always seemed to me that this is actually the case. Consequently 1 am forced to adopt a hedonistic position. lf other men judge these matters differently I know of no way of ''refuting" them; but, on the other hand, I have never been able to see that any of the considerations ad­ vanced in opposition to hedonism consti­ tute a refutation of it. From what has been said, however, it is obvious that there may be perfectly legitimate criticisms of hedonism-those, namely, yvhich consist simply in presenting for judgment ''hard cases" concerning ,,>hich it is thought that the only conclusion consistent with hedonistic principles will nevertheless, on careful reflection, be re­ jected. But such criticisms, however legiti­ mate in method, have never actually seemed to me in the least conclusive. Such force as they at Brst sight sometimes ap­ pear to have always turns out to arise, so far as I can see1 from some confusion of thought which still c:ouds the issue. Once these confusions are cleared a\vav, I never seem to £nd i n these '"hard casei anything incompatible with the truth of hedonism. Any adequate consideration of this phase of the matter would lead us too far afield to allow of our undertaking it on the present occasion with any degree of full­ ness. I shall thecefore simply illustrate the 'way in which it seems to me possible to dispose of such hard cases by the examina­ tion of a few upon which Mr. Moore chiefly depends, and which r hope will be more or less typical. The following ls an instance which he believes ""11 persuade us that eve!1 wholes containing no pleasure may be intrinsically valuable: "'Let t:s imagine one world exceedingly beautiful. Imagine it as beautiful as you can . . . and then imagine the ugliest world you can possibly con­ ceive. Imagine it simply one heap of filth�' ( Principia Ethica, p. 8 ) , and then suppose that no one ever can or does receive pleas­ ure or displeasure from eit�er \Vorld in any respect or degree whatever. '\iVould i t not be vvell to do ¥,diat we could to produce


the beautiful world rather than the other?" ·would not the former be intrinsically bet­ ter than the latter? Now I ask myself

\.vhether this case does not derive most of its apparent force from the circumstance that the reader who makes the imaginative comparison very naturally revolts from the image of the ugly world and at the same time takes pleasure in the thought of the beautiful world, and that he neglects expli­ citly to notice and discount this fact. I also ask myself whether the reader is not in­ fluenced, and his judgement unconsciously perverted, by the fact that we can scarce­ ly compare these two imaginary worlds without the thought that the beautiful world obviously possesses greater pleas­ ure-producing potentialities than the ugly one; by the fact that it is difficult to com­ pare these Dvo imaginary worlds \Vithout reference to the consideration that the one world provides, for any conscious being that might sometime be introduced upon the scene, a better basis for enjoyment than does the other. Once I carefully notice and discount such sources of bias, I entirely fail, for my O\Vll part, to see the superior value in the beautiful world. Another of Mr. Moore's examples-one of those which to his mind "constitute a re­ ductio ad absurdum of the view that intrin­ sic va_lue is always in proportion to quan­ tity of pleasure," is as follows. If this he­ donistic principle is true, it «involves our saying . . . that a world in which absolutely nothing except pleasure existed-no knowl­ edge,. no love, no enjoyment of beauty, no moral qualities-must yet be intrinsically better- better worth creating- provided only the total quantity of pleasure in it were the least bit greater, than one in which all those things existed as well as pleasure." (Ethics, pp. 237, 238.) This in­ stance seems almost deliberately framed to confuse the issue; for it is very difficult in considering the matter to remember that, if vve are not illegitimately to introduce into our second . \Vorld an additional increment of pleasure, by "enjoyment of beauty" we


n1ust here distinctly mean merely contem­ plation of beauty, wholly divorced from any element of pleasure. Moreover, it is difficult to keep our minds wholly free from the thought of the greater hedonic potentialities of a world possessing so many elements \vhich in our experience are fruit­ ful sources of enjoyment, as compared with a world from which these sources are elimi­ nated. Once I clear my mind from such confusing associations, hO\vever, I feel no further difficulty in reaching the hedonistic conclusion. Mr. Moore also points out that the he­ donistic theory compels us to assert that "the state of mind of a drunkard, when he is intensely pleased with breaking crockery, is just as valuable, in itself-just as well worth having, as that of a man who is fully realizing all that is exquisite in the tragedy of King Lear, provided only the mere quantity of pleasure in both cases is the same" ( p. 238 ) . Here again, once I care­ fully abstract from all tacit reference to the differing promise and potentiality of these two states of mind, from all larger thought of their vastly differing significance for the total lives of these men and their fellows, I find myself quite clearly committed to the hedonistic vie\V of the matter. I thus do quite clearly embrace the conclusion which Mr. Moore thinks self­ evidently mistaken, "that if we could get as much pleasure in the world, without need­ ing to have any knowledge, or any moral qualities, or any sense of beauty, as we can get with them, then all these things would be entirely superfluous" (p. 238 ) . But I also quite as heartily agree ,vith Mr. Moore that "the question is quite incapable of proof either way" ( p. 238 ) , and that "if any­ body, after clearly considering the issue, does come" to the contrary conclusion, �'there is no way of proving that he is wrong" (p. 238 ). My point simply is that there is no short and easy way with hedon­ ism, and that the cavalier way in which it is commonly treated is wholly unreasonable and unjust.

Intrinsic Value 1110NROE C. BEARDSLEY I lvlany ph:losophers apparently still accept the proposition that there is such a thing as intdnsic value, i.e., that some pait of the value of some things ( objects, events, or states of affairs) is intrinsic value. John Dewey's attack seems not to have dis­ lodged this proposition, for today it is sel­ dom questioned. I propose to press the at­ tack again, in terms that owe a great deal to Dewey, as I understand him. The predicates ( 1 ) ". . . has intrinsic value," ( 2) ". . . is intrinsically valuable," and ( 3 ) ". . . is intrinsically desirable,r will be used interchangeably-not for the sake of elegant variation,. but because each per­ mits idiomatic constructions that bring out different features of what I take to be the same concept. ''Desirable�' associates with "to desire/' which vvill be convenient to have available at a later stage of the argu­ ment. ( No doubt "valuable" and �to value" are similarly related, but the latter is not free from the suggestion of reflective ap­ praisal ) On the other hand, the noun "value_,,, is useful because we can speak of "a value" and aa kind of value." '"\talue" lends itself more readily than "desirability" to such adjectival qualifications as "cogni­ tive value" and ,tmoral value." Two phrases are the most often used in defining "intrinsic value": '"for its own sake" �d "in itself." Their m,,anLin1:;s are .Reprinted from Philosophy and Phenomenclogi� cal Research, XX.VI, Sept. 1905, pp. 1--17'� by per­ mission of the author and the publisher.

close1 but not identical, and the seco!ld seems more satisfactory than the first. We might say that something is intrin­ sically valuable, in some degree, if it is val­ uable for its own sake, and that if it has value for the sake of something else, then its intrinsic value, if any, ls that which would remain if that other-regarding value were subtracted. One inconvenienc-e of this definition can be brought out as follows; A sheet of postage stamps has been misprint­ ed-the central Bgure, say, is i:!lvertetl. The stamps derive part of their value from their rarity. Is one of these stamps valuable, ,n part,. for its own sake? ,Veil, its value is not for the sake of anything else-···if we speak of its philatelic value, not its. market value. But is this value then intrinsic? It seems strange to say this ,vhen it can be taken away, without altering the stamp at all, simply by having the Post Office Depart­ ment print a few hundred million more copies. Since its rarity is a relational prop­ erty, there is a sense in which the rare stamp is valuable not for the sake of any­ thing else, either. It might be replied that, even if the issue becomes plentiful, the phi­ latelic value of each individual stamp is not destroyed, but only reduced; rarity cannot transform an object with no value into one with value; it can only increase certain sorts of value in things that already have some degree of it. Still, that part of the stamp's philatelic value that is st:pplied by rarity seems to be neither intrinsic nor e x ­ trinsic} i f these are defined as ""for its own sake" and "for the sake of something else" respectively. •



The second definition of "intrinsic value" is that proposed hy G. E. Moore in his paper on "The Conception of Intrinsic Value." Suppose we can distinguish be­ tween the internal and external properties of a thing, that is between ( 1 ) its qualities and inner relations, and ( 2J its relations to other things. Then the value that depends upon a thing's internal properties alone is its intrinsic value; the value that depends ( wholly or partly ) upon a thing's external properties is its e;,,_'trinsic value. The intrin­ sically good thing is "good in itself." Moore states the definition this way:

To say that a kind of value is "intrinsic" means merely that the question whether a thing pos­ sesses it, and in what degree it possesses it, depends solely on the intrinsic nature of the thing in question. 1

It is this definition of intrinsic value that leads to Moore's thought- experiment in Principia Ethica. To decide the question "vVhat things have intrinsic value"?, he says, it is necessary to consider what things are such that, if they existed by themselves, in absolute isolation, we should yet judge their existence to be good. 2

If the intrinsic value of a thing is inde­ pendent of its relationship to anything else, it cannot be destroyed by the removal of everything else. Moore holds that only by applying this test can we sort out intrinsic from extrinsic value with clarity and con£­ dence. Some puzzles in 1 Ioore's definition of "intrinsic nature" have been further dis­ cussed by him and by others. 3 One diffi­ culty is to explain how intrinsic goodness can be a property dependent solely on the intrinsic properties of things, without itself Philosophical Studies, London 1922, p, 260, :, Principia Ethica, Cambridge University 1903, p. 187; cf. Ethics, London 1947, pp, 42, 101. See C. D. Broad, "G. E. Moore's Latest Published Views on Ethics," Mind LXX ( 1961), pp, 435-57. 1




being an intrinsic property-this explana­ tion is needed to fix the status of goodness as a "nonnatural property." Another is, to decide whether dispositional properties are to be called internal or external, and to dis­ pose of a certain element of arbitrariness or conventionality in the classification of some properties as dispositional. In the case of the most self-contained and self-sufficient of the valuable objects we are acquainted with- that is, works of art-the internal­ external distinction has an immediate ap­ peal. Thus we tell people to listen to the music "itself' and pay no attention to any­ thing outside it, such as objects it might suggest or the biography of the composer. And the so-called "Formalist" has been known to assert that the ( aesthetic) value of a painting depends on internal proper­ ties alone ( lines, shapes, and colors ) and owes nothing to its representational rela­ tionship to the world outside it. But even if these notions are regarded as acceptable, serious questions can be raised about the sharpness and decisiveness of the internal­ external distinction. And it is no doubt for this reason, as well as for others, that most contemporary value-theorists have concluded that if any­ thing is intrinsically valuable it is not an extem,al object, but an experience or psy­ chological state. At best, the work of art could only be said to have "inherent value" (in C. I. Lewis's terminology ) , if exposure to it can result in an aesthetically enjoyable experience. The distinction between inter­ nal and external properties seems clearer when applied to experiences. Of course, in� tentionality-a reference to other states of mind or to the external \VOrld- rnust be ta�en as internal to the experience itself; but neither an ostensible memory of a past pleasure nor the expectation of a future one is causally dependent upon the occur­ rence of those pleasures, and so, theoreti­ cally at least,. ,ve can conceive any short stretch of conscious life apart from its ante­ cedents and consequents. And apparently we can ask ,:vhether it has intrinsic value.


I take it that to say that something is valuable is to say that it deserves to be valued; and to say that something is desira­ ble is to say that it is worthy of being de­ sired. Now when we add "intrinsically" to, say, "desirable," how does it fit into the de­ finiens? Does "X is intrinsically desirable" mean ( 1) "X Ls intrinsically w01thy of being desired" (that is, by de£.nition, «x is wor­ thy-of-being-desired on account of its internal properties alone")?

Or does it mean

(2) "X Ls worthy of being intrinsically desired" ( that is, by definition, "X is worthy of being-desired:..on-account- of ­ its-internal-properties-alone") ?

I have puzzled over the relationship be­ tween these two expressions, and find that I can only understand the former in terms of the latter. For if Xs desirability depends on its internal properties alone, then these must be the properties that ought to be, or deserve to be, desired; and what ought to be the case is that X is desired on account of these internal properties. Thus in order to attach a sense to "in­ trinsically desirable" we must first attach a sense to "intrinsically desired." And there is, I think, no trouble about this. For there is an evident psychological distinction be­ tween desiring something on account of its internal properties alone,. and desiring it on account of its relationships to other things. The distinction is not easy to apply, be­ cause in most of our desires, care for the thing itself and concern for what \vill come of it are thoroughly mixed. But we can pretty well fix the extremes: at the one end, the candy that the child in the grocery store ·wants, and screams for, and evidently would count the world well lost for; at the other end, the pieces of string and bits of ceIIophane that \Ve instantly discard, once we have secured the goods they serve to wrap. And we can use :tvfoore's test \Vith some success to make the psychological


distinction, by asking a person, for exam­ ple, to think what he would choose to do if he had only a short time on earth, with lim­ itless. resources and no obligations to oth­ ers. The child no doubt would gorge him­ self on candy, but nobody would pore over chewing-gum wrappers. So I would like to allow, and exploit, the term "to desire intrinsically." But of course it does not follow automatically that because we can attach "intrinsically" to "desire," we can therefore attach it to '(de­ sirable." There are obviously other adverbs that go ,vith "desired," but not ,vith "de­ sirable" ( for example, "eagerly," "strongly," and "widely" ) . An argument must be made out to show that things can be intrinsically desirable as well as intrinsically desired.

II The question,. then, is this: What good rea­ sons are there, if any, for believing that there is such a thing as intrinsic value? Since this proposition is seldom considered to be in need of elaborate proof, formal ar­ guments are difficult to collect. I can only discover three such arguments: ( 1 ) an a r ­ gument from definition, ( 2 ) a dialectical demonstration, and ( 3 ) an attempt at em­ pirical confirmation. ( l ) The Argument from Definition. In the vievi-' of some thinkers, the existence of intrinsic value can be simply shown in this way: Some extrinsic value is instrumental value, which is defined as follows : "X has instrumental value" means "X is condu­ cive to something that has intrinsic value." ( Call this Definition A.)

I have selected the loose term "conducive to/' in order to avoid some distinctions that \Ve do not need at present. Thus if Y is an end to which X is a means, or a whole of which X is a necessary part,. then X is con­ ducive to such intrinsic value as Y pas-

404 sesses. Obviously, if we accept Definition A, we are as committed to the existence of intrinsic value as we are to the existence of instrumental value-however long the chain of conduciveness may be. But must we accept this definition? In order for Y to confer its value on X, when X is conducive to Y, it is certain that Y must have some value to confer, but whether that value is intrinsic or instru­ mental does not matter as far as X is con­ cerned. So the following alternative defini­ tion should be acceptable: '°X has instrumental value" means "X is condu­ cive to something that has value." ( Call this Definition B. ) The Arguer from Definition rejects Definition B. It is all right if the key term in the definiens, "value," can be defined by itself, without reference to intrinsic value. But the word "value," he might contend, is necessarily an ellipsis; it cannot stand by itself. Up to this point I have been speak­ ing as though value is a genus with two species, so that value can be defined first and then divided by Moore's test. And it is true that in describing that test, Moore speaks as though we could first know that an object has a certain total value, before going on to discover how much of that total is intrinsic, and how much extrinsic. But this, according to the Arguer from D ef­ inition,. is all misleading. The terms "extrin­ sic value'' and "intrinsic value," despite the noun they share, do not name coordinate species of a genus,. but designate two very different concepts, one of which is deriva­ tive from the other. 4 And Moore's own way of speaking agrees with this interpretation, on occasion. For example, he, speaks5 as ' "Intrinsic" and "extrinsic," in the convenient terminology of Austin Duncan-Jones, would then be "sense-discriminating," rather than "concept­ modifying," adjectives (see "Intrinsic Value: Some Comments on the ·work of G. E. 1v1oore/' Philoso­ phy XXXIII ( 1958 ) , pp. 240-73, esp. pp. 261-62 ) . r; Principia Ethica, p . 24.

WHAT THINGS HAVE VALUE? though the phrase "good as a means" (i.e., having instrumental value ) is synonymous with the phrase "a means to good" ( i.e., conducing to intrinsic value) . Being good as a means is not a way of being good- the instrumentally valuable thing is not a valu­ able instrument, strictly speaking, but an instrument of value. And similarly, To have value merely as a part is equivalent to having no value at all, but merely being a part of that which has it. 6 This line of thought would issue in the re­ jection of my Definition B. For it seems to show that instrumental value can only be defined in terms of intrinsic value; so that the existence of the former automatically entails the existence of the latter. Nevertheless, it seems to me that the word "value" does have a meaning by it­ self, and does mark out a genus. \1/e can sensibly ascribe value as such to things, and this is even more clearly true of "de­ sirable" and '\vorth having." 7 The terms "good as a means" and "good as an end in itself," if they may be taken as synonymous with "instrumental good" and "intrinsic good," suggest that the distinction is be­ t\veen two grounds of goodness, not two senses. This is the position assumed by Moore in his Ethics when he asserts that saying a thing is intrinsically good . . . means it would b e a good thing that the thing in ques­ tion should exist, even if it existed quite alone, without any further accompaniments or effects whatever. 8 For here "intrinsically good" is defined by means of "a good thing." If this is a correct procedure, Definition B is acceptable, and the Argument from D efinition fails. ( 2 ) The Dialectical Demonstration is closely connected with the Argument from Definition, but deserves exhibition on its own. I think of it as logically parallel to the Ibid., p. 35. See Duncan-Jones, pp, 257-58. a Ethics, p. 42.




First Cause argument for the existence of God. "Instrumentally valuable" is a rela­ tional concept- X borrows its value from Y, or Y confers its value upon X. If the value Y confers is itself instrumental, so that it is merely passed along from Z, then where does Z get its value? In the last analysis, something must ( according to this argu­ ment) possess its value in itself, or nothing can get any value. 9 So the existence of any instrumental value proves the existence of some intrinsic value, just as the occurrence cif any event is said to prove the existence of a First Cause. To align the Demonstration of intrinsic value with the venerable First Cause argu­ ment may lend it prestige, but may also suggest its faults. As Kant showed, the First Cause argument projects a certain kind of ideal explanation that cannot be completed if the causal series has no begin­ ning term. That is, if to explain an event, X, requires not only that we assign cause, W, but that we assign an expla-ined cause, then the ideal explanation of X would involve the explanation of all its causal antecedents; and if these have no first term, no such explanation can be given. Similarly, the Dialectical Demon­ stration of intrinsic value projects a certain kind of ideal justi£cation that cannot be completed if the series of means and ends has no last term. That is, if to justify ascrib­ ing a value to X requires not only that we show it is a means to Y, but also that \Ve justify ascribing a value to Y, and if there is no stopping point, no such ideal justifica­ tion can be given. But ordinary justi£ca­ tion, like ordinary causal explanation, in­ volves no such in£nite regress. The Dialectical Demonstration cannot be a pure formal demonstration, for, unless the Argument from Definition is valid, it cannot be proved strictly self-contradictory to assert the existence of instrumental value a Cf. Hume's argument that "something must be desirable on its own account," Inquiry Concerning the Pri-nciples of Morals, N,Y,: Liberal Arts Press, 1957, p. lll.


but deny the existence of intrinsic value. The Demonstration must rather be thought of as applying to our knowledge of value. Premise 1 : \1/e 1..'Tiow, or have good reason to believe, that some things are instrumen­ tally valuable. Premise 2 : We could not know this unless we knew some things to be intrinsically valuable. Conclusion: We know some things to be intrinsically valu­ able. Now, Premise 1 seems to me clearly true. But the conclusion seems to me quite clearly false. The paradoxical feature of our value-knowledge is just that we have a good deal of sound knowledge about in­ strumental values, but are in considerable doubt about intrinsic values. Philosophers have disputed, and still dispute, about whether pleasure is an intrinsic value, and, if so, whether it is the only intrinsic value; and it is signi£cant that as ordinary people we have not had to wait upon the settle­ ment of these issues before discovering a great many valuable things nearer to hand. \Ve must have some way of knowing that in many concrete situations it is better for a person to be healthy than sick, without knowing whether that is because it is in­ trinsically best for him to maximize his net positive hedonic quality, or realize his po­ tentialities, or cultivate a good ,vill, or whatever. The apparent hopelessness of resolving problems about instrumental value ,vithout ln1owing antecedently what, if anything, in­ trinsic value may be, can be cleared a\'vay, I believe, if we go back to the concrete contexts of value- problems. 10 Y\71len we are in the position of having to decide what is valuable, or more valuable, we are in Dew­ ey's "problematic situation," and such a sit­ uation is one in which certain ends are in grave doubt and others are ( on that occa­ sion) taken as temporarily fixed. If the value of everything in the situation were in 10

Here I want to acknowledge my debt to Sidney Hook's essay on "The Desirable and Emo­ tive in Dewey's Ethics," in his collection1 John Dewey: Philosopher of Science and Freedom, N.Y. 1950.


406 question at once, nothing could be decided at all, and indeed no problem could even be conceived; only in terms of certain ten­ tatively-held values, can we decide, or even ask, whetber other tbings are valuable or not. There must be a basis on which it is reasonable for us to pick out salient ele­ ments in the situation, and assign them probable values, witbout transforming the task of assigmnent into another proble­ matic situation, thus endlessly postponing a decision. A state of affairs such as good healtb, for example, has retained its eligi­ bility through earlier problems and expe­ riences; its value has survived them, in the sense tbat up to this point health does not seem to have interfered with our pursuits, but a lack of it from time to time has not only contributed to tbe rise of difficult problems, but has limited our capacity to resolve them. We need not suppose that health is an intrinsic good, or that it is al­ ways good, or even that we are necessarily right in taking it to be good on tbe present occasion- but we do have rational justifica­ tion for supposing that it has positive value that ought to be taken into consideration now. This is tbe merest sketch of a way of looking at tbe problem of evaluation. The gist of it is that reasonable decisions about instrumental values do not presuppose, or wait upon, previous reasonable decisions about intrinsic values ( even if such deci­ sions were possible, which I shall argue-in tbe next section- tbey are not ) . So tbere is no infinite regress in a purely instrumental­ ist theory of value. ( 3 ) The Empirical Confirmation. Some writers on the theory of value have not relied on eotber of the a priori forms of argument that I have just considered1 but have held that the existence of intrinsic value is attested by direct experience. One of the most carefully considered tbeories is tbat of C. I. Lewis. He distinguishes tbree kinds of value-predication, of which tbe basic kind is the �·expressive statement," having a form like "This is good," where

"this" refers to immediate presentations in experience. 1 1 E xpressive statements report "value-apprehensions" or "direct findings of value-quality in what is presented,"12 and are therefore incorrigible. 13 . . . It will hardly be denied that there is what may be called "apparent value" or "felt good­ ness," as there is seen redness or heard shrill­ ness. And \vhile the intent to formulate just this apparent value- quality of what is given, \vithout implication of anything further, encounters linguistic difficulties, surely it -will not be denied that there are such immediate experiences of good and bad to be formulated.14 It is this immediate value-quality tbat Lewis calls intrinsic value, and that imports intrinsic value into those experiences that possess it. To it all forms of value-judgment are ultimately anchored. And because it is only in exp eriences tbat we are directly ac­ quainted with such qualities, only experi­ ences can, strictly speaking, have intrinsic value. VVhen an object enables us to "real­ ize" intrinsic value in experience, by being directly presented to us, tbe object has "in­ herent value," which is a form of extrinsic value. 15 "The goodness of a good object is a potentiality for the realization of good­ ness in e1.-perience."1 6 VVe have suggested-and intend to abide by-distinction between intrinsic and extrinsic values by reference to the question, "Is that which is valued, valued for its own sake or for the sake of something else?"17 VVhat is most surprising in Lewis's characterization of tbis value-quality, is tbe way he employs alternately, and indiffer­ ently,, the two e,q,ressions that other phi11

Theory of Knowledge and Valuation, LaSalle,

Ill., 1946, p. 374. 1 .2 Ibid., p. 365. 18 Ibid., p. 375. " Ibid., pp, 374-75. " Ibid., pp. 386-87, 391. 1 6 Ibid., p. 389. 11 Ibid., p. 385.



losophers have been at such pains to keep distinct: "valuable in itself" and "valued in itself." There are passages in which Lewis seems to suggest that being valuable and being valued may not be the same thing, since one of them may be evidence of the other. Here being liked is an index of being

Such appellations as these-1 do this onlv as an intermediate end of his action. The' idea has an appar­ ent plausibility, but is nevertheless a mis­ take. Tiie truth seems to be that a man can pursue tile happiness of others either as in­ termediate or as ultimate end. The delight of a king can be the hap­ piness of his subjects. He gives all his ener­ gies and ,vork to the promotion of this end. Maybe he sacrifices his so-called 'personal happiness' for the good of those over whom he is set to rule. Yet, if this is 'What he likes to do, it is also that in vvhich his happiness consists, To gay this is not to C!stort facts logically. But to say that the king sacrifices himsel£ for the sake of becoming happy and not for the sake of making others happy, would be a distortion. It would be a distortion similar to that of which psycho­ logical hedonism is guilty, v1hen it main­ tains that everything is done for the sake of pleasure, on the ground t.11at all satisfaction of desire may be thought intrinsical�y pleas� ant Can a man's welfare be an end of his own a for example: a man whom we call happy and wlio is hit by a sud­ den blow of bad luck, say, the loss of a child in an accident. He \V:ill experience painful agonies and a·dreme sadness. 'News of the disaster made him dreacifullv un­ happy,' we might say, thinking of ' these emotional effects on him. I( however, we were to say that the ne,vs made him an t-m� happy man, we should be thinking not only, or maybe even not at all, on those emotional effects, but on effects of a less immediate showing and of a longer lasting. If we can say of him some such things as, 'For years after he was as paralysed; none of the things, which used to delight him, gave him pleasure any longer/ or �Life seemed to have lost meaning for him,-for

420 a time he even contemplated suicide,' then the accident made him unl,appy as distinct from merel.v. sad. But whether things, bear­ ing on the distinction, can be truly said of the man, is not to be seen in an inkling. AnaJogous things can be said about changes in the reverse direction. A piece of news; say of an unexpected in'!:ieritance� can make a man jump wlth joy. But whether it makes him happy as distinct from merely glad can only be seen from fects of a longer lasting and less obvious showing on his subsequent life. Should we say 'the whole of his subse­ quent life'? I think not. Happiness is nei­ ther a 1nomentary state nor 15 it a sum tota1 to be found out when ,ve close our life's ac­ count. A man can become happy, be happy, and cTwnge from happy to un­ happy. Thus, in the course of his life, a man can be both happy and unhappy. And he can · be happy and unhappy more than once. ( See section 1.) 'vVe could make a distinction between a happy man and a happy life ·and regard the second as a thing of wider scope. This would make it possible to say of somebody that he had a happy life although, for some time, he was a most unhappy man, 6. A judgment to the effect that some being is happy or is not happy or is un­ happy we shall call an eudaimonic judg­ ment. I think it is illuminating to compare the logic of the eudaimonic judgment to the logic of the statement 'This is pleasant'. Of the sentence "This is pleasant' · \Ve said that it conceals a logical form. ( See C:,. IV, sect. 6. ) It suggests t.'iat pleasantness is a property whi ch we attribute to some object or state, whereas in fact to judge something pleasant is to verbalize a relationship in which the judging subject stands to this L'iing. To judge something hedonically good is to manifest an attitude, one could also say, to certain things ( activities, sensa­ tions, the causes of sensations) . The logi­ cally most adequate fom1 of the verbal:za­ tion is therefore, it seems 7 the relational

WHAT THINGS HAVE VALUE? form 'I like this' or some similar relational form. In an analogous sense the sentence 'He is happy' may be said, I think, to conceal a logical fom1. It suggests a view of happi­ ness as a property which the happy indi­ vidual exhibits-- which shines forth from him. Whereas, in fact, to be happy is to be in a certain relationship. A relationship to what? it may be asked. A relationship to one�s cir,cumstances of Ufe, I ,vould answer. To say 'He is happy' is similar to sa11ng 'He llkes it', the 'it' not meaning this or that partic'Ular thing or activity but, so to speak, cthe whole thing. One could also say, 'He likes his life as it is.� On this vie\v, if a 1nan says of himself 'I am happy,' he manifests in words an atti­ tude which he takes, or a relationship, in which he stands, to his circumstances of life. Happiness is not in the circumstances -as it were awaiting the judgment-but springs into being with the relationship. (Just as hedonic goodness does not reside in the taste of an apple, but in somebody's liking the taste of an apple. ) To judge one­ self happy is to pass judgment on or value one's C'..rcumstances of life. To say 'He is happy' can mean two different things. It ean mean that the man, o! vvhom ,ve are talking, iE in the relation­ ship to his circumstances which, if he were to verbalize his attitude, he could express in the 1.vords 'I am happy'. T}1en ese consequences make that other man un� happy. If our fool accepts the conse­ quences v,dth a cheerful heart1 the man cannot insist that he must be right. He cannot do so on the ground) say, that those same consequences would have made hhn, or most people, miserable, Nor can he pre­ tend that the lightsome fellmv is 'really' un­ happy, though unaware of his own misery. But cannot a man be mistaken in thinking that he is happy? In a sense he can not, but in another he can. 'He says he is happy, but in fact he is not' can express a true proposition, But does not the truth of this proposition entail that the person who professes to be happy is lying? And is this not uninteresting? The answer is that, be­ side uninteresting lies1 there exist pro­ foundly interesting lies in the matters, \vhich we are now discussing. First person judgments of happiness cari be insincere, and insincerity may be regarded as a spe­ cies of lying. The same, incidentally, holds good for £rst person hedonic Judgments too. A youngster may profess to like the taste of tohaeco, which in fact he detests, just for the sake of showing off. He may even make himself believe this, in some involved and twisted sense of 'believe'. A polite man may say he likes the taste of a wine merely to please his host. The insincerity of such first person judgments may be relatiYely easy to unmask. In the case of first person judgments of happiness and misery, the problem of sin­ cerity is most dif!leult----both psychologi­ cally and conceptually. I shall not here try to penetrate its logical aspects, which I find verv bewildering, ( I am not aware of any satisfactory dise�ssion of the topic in th� literature. ) I shall make a shortcut through


the difficulties and only say this much in conclusion: However thoroughly a man may cheat himself with regard to his own happiness, the criterion of cheating or insincerity must be that he admits the fraud. A judgment is insincere when the subject 'in his innermost self' admits that it is not as he says it is. If his lips say 'I am happy' and he is not, then in his heart he n1ust already be saying to himself 'I am not happy'. He, as it were, does not hear the voice of his heart. These are similes, and I am aware of the tempta­ tion to misuse them. ( They are the same sort of similes that are used and misused in psychoanalysis- the similes of the subcon­ scious, the super-ego, etc. ) vVhat I mean by them could perhaps be said most plainly as follows: The fact that first person judg­ ments of happiness can be insincere must not be allowed to conflict logically with the fact that whether a person is happy or not depends upon hw own attitude to his cir­ cumstances of life. The supreme judge of the case must be the subject himself. To think that it could be otherwise is false ob­ jectivism. 8. Judgments of the beneficial and the harmful, i.e. of that which is good or bad for a man, involve two components. e have called them the causal and the axio­ logical component. ( See Ch. III, sect. 5.) ·vve must now say some ,vords about each of them. vVhen something happens, i.e. the world changes in a certain respect, there will usually also be a number of subse­ quent changes, which are bound (by so­ called 'natural necessity' ) to come about, once the first change took place. These sub­ sequent changes ,:ve here call the conse­ quences of the first change. If the first change is of that peculiar kind which we call a human act, then the subsequent changes are consequences of action. The change or changes upon which a certain further change is consequent ( i.e. the con­ sequence of which this further change is) we shall call the cause(s) of this further change.



J\fost things which happen, perhaps all, would not have happened, unless certain antecedent changes had taken place in the world. These antecedent changes we shall call the causal prerequwites or requirements of the subsequent change. They are some­ times also called 'necessary causes'. The necessary causes may be, ·but need not be, 'causes' in the sense defined above. These explanations are very summary. Not least of all considering the importance to ethics of the notion of consequences of action, it is an urgent desideratum that the logic of causal relationships be better elaborated than it is. vVe shall not, how­ ever, attempt this here. Only a few obser­ vations will be added to the above. The notions both of consequences and of prerequisites and of causes of a change are relative to the further notion of a state of the world. Thus, e.g., a change which is required in order to effect a certain change in the world as it is to-day, may not be re­ quired in order to effect this same change in the ,vorld as it is to-morrow. It is sometimes said that every event ( change) 'strictly speaking' has an infinite number of consequences throughout the whole of subsequent time, and that for this reason we can never know for certain which all the consequences of a given event are. These statements, if true at all, hold good for some different notion of consequence, but not for the notion ,vith which we are here dealing. Exactly what could be meant by them is not clear. Yet we need not dismiss them as nonsense. vVhen, for example, something which hap­ pens to-day is said to be a consequence of something which took place hundreds of years ago, what is meant is perhaps that, if we traced the 'causal history' of this event of to-day we should find among its 'causal ancestry' that event of hundreds of years ago. Here the notions of causal ancestry and causal history could be defined in terms of our notions of cause, consequence, and prerequisite and yet it need not follow that, if an event belongs to the causal an­ cestry of another event, the first must be a

THE GOOD OF MAN cause or prerequisite of the second or the second a consequence 0£ the first, For ex­ a.."'llple: Let event b be a consequence ( in our sense ) of event a and a causal prere­ quisite ( in our sense) of event c. It would then be reasonable to say that event a is a • causal a.."1cestor� of event c, or that tracing the 'causal history' of o takes us to a. In some loose sense of the words, a may be said to be a 'cause' of c and o a �conse­ quence' of a. But in the more precise sense;, in which we are here employing the tenns, a is not ( necessarily) a cause of c, nor c ( necessarily) a consequence of a. The causes and consequences of things which happen, are often insufficiently known and therefore largely a matter of belief and conjecture. Sometimes� however, they are known to us. T11e statement, sho�u ld it be made, that they cannot ( 'in principle') be known either is false or ap­ plies to some different notions of cause and consequence from ours. By knowledge of the causes and conse­ quences of things which happen, I here mean knowledge relating to particulars. An example would be knowledge that the death of N. N. was due to a dose of arsenic, which had been mixed into his food. Such knowledge of particulars is usually grounded on knowledge of general proposi­ tions-as for example that a dose of arsenic of a certain s+uength will ( unless certain counter2cting causes in!:ervene) 'inevitably' kill a man. vVhcther all such knowledge of particulars ls grounded on general knowl­ edge, we shall not discuss. Vi'he n in the sequel we speak of knowledge of the causes and consequences of things, or of known ca...:ses ai1d conse­ quences, 'knowledge' is short for 'knowl­ edge or belief' and �known' for 'known or believed'. Tbe consequences which are known ( i.e. known or believed ) at the time when the thing happens, we shall also call foreseen consequences. So much for the causal component in­ volved in judgments of t.½e beneficial and the harmful. \Ve now turn to the a.-:io1ogi­ cal component. A preliminary task will here


he to clarify the notions of a wanted and an umcanted thing. 9. The notion of a wanted thing, which I shall now try to e,-plain, is not the same as that oE a,-., end of action. I shall call it the notion of being wanted in itself. How things vvhich are \Vanted in themselves, are related to things which are wanted as ends of action, \vill be discussed presently. Cor­ relative with the notion of being wanted in itself fa the notion of being unwanted in i t ­ self. 'Between' the two falls a notion, which we shall call the notion of being indifferent in itself. The notion of being wanted in itself is the nearest equivalent in my treatment here to the notion of lntrinsic value in fVfoore and some other writers. P:foore1 when discussing the notion of ir!binsic worth 1 often resorts to a logical fiction which 1 mutatis - mutandi.s, mav be resorted to also for explaining the ;ieanlng of a thing being wanted, unwanted, or indiffer­ ent 'in itself'. This fiction i.s that of a preferential choice between two alternatives. A major difficulty is to formulate the tem1s of the choice correct!y for the purpose of defining the axiological notions under discussion. ( :Vloore's explanation of intrinsic value in terms of bettemess of alternatives cannot be regarded as logically satisfactory-apart from questions of the meaningEulness of the very notion. 8 ) Our proposal here of a solution to the problem is tentative only. Assume you were offered a thing X which vou did not alreadv uossess. \Vould you th�n rather take it th;n"leave it, rather have it than ( continue to ) be without it? The offer must be considered apart from questions of caGsal requirements and of consequences. That is: considerations of things which you y,;rill have to do in order to get X, and of things which "111 happen to you as a consequence of your having got the thing X must not influence your choice. If then you would rather take X than leave 8 See Ethics, pp. 42-44 nnd fa particulRr, 1 Moore's rep:y to his critics in The Phil.osophy of G. E. Moore,. pp. 554-5$7.


it, X is wanted in itself. If you have the op­ posite preference, X is unwanted in itself. If you have no preference, X is indifferent


10. Anything which is an- intermediate or ultimate-end of action, can be called a good ( for the' subject in pursuit of the end ) . in itself. ( Cf. above Ch. I, sect. 5 and Ch. III, sect. As readily noted, the ideas of the in i t ­ 1. ) Anything which is an end of action, can self wanted and unwanted, which we have also be said to be a wanted thing. thus tried to explain in terms of a fictitious Also every thing, which is wanted in preferential choice, are necessarily relatlve itself, can be called a good ( for the subject to a subject. Nothing is wanted or un­ to whom it is wanted ) . And every thing, wanted cin itself, if the words 'in itself' are which is unwanted in itself, can be called a supposed to mean 'apart from any rating or bad ( for the subject who shuns it) . Ends of action and things wanted in valuing subject'. The words 'in itself mean 'causal prerequisites and consequences themselves thus both fall under the cate­ apart', A thing, which for one subject is a gory 'goods'. Ends of action also fall under wanted thing, may be regarded as un­ the category 'things wanted'. The question may be raised, how ends wanted by another subject. A thing, fnr­ thennore, which is wanted now, may be of action and things v.ranted in themselves un\vanted at another time-the subject are mutually related. The question is com­ being the same. The notion of being plicated and I shall not discuss it in detail. wanted or unwanted in itself is thus rela­ It is reasonable to think that only things, tive, not only to a subject, but also to a par­ which are attainable through action, can be ends of action. e only thing which is ulti­ ma;tely good or desirable, and pain the only thing which is ultimately bad or undesira­ ble; or as asserting that pleasure is the only thing which is good for its own sake, and pain the only thing which is bad for its own sake. And there is, I think, a sense in which it does assert this. But these expres­ sions are not commonly carefully defined; and it is worth noticing that, if our theory does assert these propositions, the eh.'Pres­ sions ·ultimately good' or 'good for its own sake' must be understood in a different sense from t½at \Vhich has been assigned above to the expression intrinsically good.� 'We must not take 'ultimately good' or 'good for its own sake' to be synonyms for For our tl1e01y ::nost does not assert that pleasure is the only thing intr!nsically good1 and pain the only thing intrinsically evil. On the contrary1 it asserts that any whole vvhich contains an excess of pleasure over pain is intrinsically good, no matter how 1nuch else it may contain besides; and simi­ larly that any whole which contains an ex­ cess of paln over pleasure is intrinsically bad. This distinction between the concep· tion e,tpressed by 'ultimately good' or 'good for its own sake,' on the one hand, and that expressed by 'intrinsically good; on the other, is not commonly made; and yet ob­ ,iouslv we must make it, if we are to say that o·ur theory does assert that pleasure is the only ultimate good, and pain the only ultimate evil. 'The tvvo conceptfons, if used in this \Vay, have one important point in common, namely, that both of them will only apply to tirings whose existence 1oould be good> even if they existed quite alone. \�! we assert that a thing is 'ulti­ mate}y good' or 'good for its own sake' or 'intrinsically good/ \Ve are always asserting 1

that it would be good; even if it existed quite alone. But the two conceptions differ in respect of the fact that; whereas a whole which is 'intrinsically good� rnay contain parts which are not intrinsically good, i.e. woulil not be good, if they existed quite alone; anything which is 'ultimately good' or �good for its 0\\'11 sake' can contain no such parts. This, I think, is the meaning svhich we must assign to the e:-..J)ressions ·ultimately good' or �good fo{ its o,v.n sake,' if we are to say that our theory asserts pleasure to be the only thing 'ultimately good' or �good for its own sake: \Ve may, in short, divide intrinsically good thi:1gs into two classes : namely (1) those which, while as wholes they are intrinsically good, nevertheless contain some parts which are not intrinsically good; and ( 2 ) those, \vhich either have no parts at all, or1 if L'½ey have any, have none but what are them­ selves intrinsically good, And we may thus, if we please, confine the terms 'ultimately good' or 'good for their own sakes' to things which belong to the second of these two classes. \Ve may, of eourse, make a pre­ cisely si:milar distinction between two classes of iomnsically bad things. And it is only if we do this that our theory can be truly said to assert that nothing is 'ulti­ mately good' or 'good for its own sake,' ex­ cept pleasure; and nothing 'ultimately bad' or 'bad for its U\VTI sake/ except pain. Such is the ethical theory which I have chosen to state, because jt seems to me par­ ticularly simple, and hence to bring out particularly clearly some of the main ques­ tions which have formed the subject of eth­ ical discussion. \:Vhat is specially important is to dis­ tinguish the question, which it professes to answer in its flrst part, from the much more radical questions> which it professes to answer in its second. In its first part, it only professes to answer the question: vVhat c:1aracteristic is there which does actually, as a matter of fact, belong to all right vol­ untary actions ) which ever have been or will be done in this world? Vv11ile, in its


second part, it professes to answer the much more fundamental question: What characteristic is there which would belong to ·absolutely any voluntary action, which was right, in any conceivable Universe, and under any conceivable circumstances? These two questions are obviously ex­ tremely different, and by the theory I have stated I mean a theory which does profess to give an answer to both. \Vhether this theory has ever been held in exactly the form in which I have stated it, I should not like to say. But many people have certainly held something very like it, and it seems to be what is o�en meant by the familiar name 'Utilitarianism,' which is the reason why I have chosen this name as the title of these two chapters. It must not, however, be assumed that any-


body who talks about 'Utilitarianism' al­ ways means precisely this theory in all its details. On the contrary, many even of those who call themselves Utilitarians would object to some of its most funda­ mental propositions. One of the difficulties which occurs in ethical discussions is that no single name, which has ever been pro­ posed as the name of an ethical theory, has any absolutely fixed significance. On the contrary, every name may be, and often is, used as a name for several different theo­ ries, which may differ from one another in very important respects. Hence, whenever anybody uses such a name, you can never trust to the name alone, but must always look carefully to see exactly what he means by it.

Utilitarian Generalizatio,i D.4VID LYONS Sometimes an act is criticized just because the results of everyone's acting similarly would be bad. The generalization test, 'What would happen if everyone did the same?' is often used in raising such criti­ dsms; and a principle \Va:-!'anting t..¾:e criti­ cism is of the following kind: ( Gl) If the consequences of everyone's doing a certab sort of tlling would he undesirable, then it would be wrong for anyolle to do such a thing. This principle is clearly teleological ( utilitarian) since in appealing to it, in de­ termining whether acts are wrong, we con­ sider onlv desirable and undesirable effects-their utility. It is also a generaliza­ tion principle: the consequences of a gen� era! practice ( everyone·s doing the same) are considered; a uarticular act is assessed as an act of that kind; and thus the verdict applies to all such acts. Such a principle may therefore be called a fonn of utilita­ rian generalizat-ion. Challenges employing the generaliza­ tion test are not t:ncommon in everyday n1oral a:gument. The significa�ce of the test has, however, puzzled philosophers. Thus the subject is of some interest in its own right. But the generalization test has philosophical importance today primarily because it has been associated vdth the forms o!: utilitarian generalization,. and be­ cause these princip!es, in tum> have seemed ·,o accomplish certain moral tasks Reprinted from Forms and Limits of Uti'litarian­ ism by David Lyons by permission of the author and the Clare:idon Press, Oxford. Co;;,yright 1964.

on strictly utilitarian grounds which other forms of utilitarianism fail to do. Our main subject then is this family of principles-their logic and their substantive import: the members of the family; how they co::npare with the more traditioaal kmd of ut:litarian principle; how they may properly be applied . :Vforeover, I shall use this inquiry-the methods of analysis which I shall adopt and the conclusions that are reached--as a basis for examining a much-heralded recent theory in this trndi­ tion 1 rule-utilitarianism. I shall relate all these principles to cor.siderations of justice and fairness-to show that the apparent force of the generalization test requires ap-. peal to more than utility. In this first chapter I shall identify and characterize utilitarian generalization, sug­ gest difficulties and issues, sketch an histor­ ical frarnevvork, and indicate the dimen­ sions within this class of principles.

A. TW'O KINDS OF UTILITARIANISM 'Oh look!' she said, pointing off to the right. 'The apples are ripe in that orchard. Let's stop and pick some.� 'No . . . .' He d:nve on, more slow;y, 'I do:1'!: think we should. Suppose everyone ' did that!' 'Don't be silly-not everyone ,vill. And. the few we·d take woulddt be missed/ 'But that's beside the point. If '.Ve can do it then so can anyone else. And if everyone did the same . . . • 451


And if everyone did the same, if every pas. ser-by picked as he chose, this grower ( or perhaps all growers) would suffer irretriev­ able losses. ?1foreover, he m:ight ask him­ self: 'Does it pay to take such care of my orchards if others are to pick them bare?' Thus, his incentive could be undermined a!'ld. future production could t..\ereby be damaged. Or he might he obliged, at con· siderable cost, to post guards and erect fences that would mar the now pleasant landscape. If such contingencies were the ground for our ::noralizer,s objection, the� he VJas emp!oying the generalization test. He was appealing to a form of utilitarian generaJi. zation, suc.'1 as ( GI). Notice how our moralizer did not argue. H e did not cfoim that the grower would sufEer hardship or loss as a result of the small expropriation proposed by his companim:i. :'.:;or did he say that such hard· ship or loss would indirectly flow from the act� as a result of their example inciting others to do likewise, sparking a chain re­ action leading to a devastation of the or· chard. Nor did he maintain that in doing such ·a thing he and hi.s companion ,vere disposing themselves to act in ft:.ture in ways wl'.'Jch ultimately would have bad consequence.,;. Finally, our moralizer did not mention tJ1e contingency, the outside chance that others would in fact do the same and that, under the circumstances, this act might contribute to a bad state of affairs. That is to say, the moralizer did not argue that the over�all effects of the one act would be undesirable ( or worse than those of some alternative) and that this was the reason against taking some apples. He might have argued in this way while still appealing to utility. But such an argument r,sts upon applying the test of utility in a radically different way-in what I shall call a simple utilitarian way. Simple utilitarian considerations are those that concern all the effects of the par­ ticular act in question ( or the effects of


that act as compared with those of the alter­ native acts) . If the moralizer had appealed to such considerations he would have asked, 'What "·ill happen if this act is per. formed?' and not vVhat would happen if everyone did the same?' In contrast, general utilitarian consid­ erations concern the total effects that could be produced i£ all acts similar to the one in question, which could be performed> actually were performed. That is, in ap­ plying a for:n of utilitarian generalization, we describe the p2rticular act in some "vay1 thus marking o:'F a class of acts, whfoh could he performed, that are similar in the respects specified. Vie do not assume that others will do the same. We are only to suppose that the kind of act specified is generally practised and to evaluate the effects of this hypothesized practice. These two kinds of utilitarian­ ism- simple and general-are &tinguish� able in two respects: ( 1) the manner in which value-crlteria1 fi1e tests for utility, are applied to acts, and ( 2, ) the generality of the judgements derivable. In the case of simple utilitarianism, ( l) value-criteria are applied to the elfects of particular acts bfren separately, and ( 2) judgements co:i­ cem only particu:ar acts. The rightness or \vrongness of a particular act depends upon the value of its effects, i.e. upon its simple utility; or alternatively upon the value of its effects as compared with the values of the effects of the alternative acts, i.e. upon its relative simple utility. In the case of util­ itarian generalization,. on the othe� handi ( 1) value-c.riteria are applied only to what I shall call the tendency of an act, i.e. to the efl:ects of everyone's doing the same sort of thing; and ( 2) the judgements di. rectly derivable concern a class of acts that are similar in the specified way, each one detennined as right or wrong or obligatory, or prirna facie so) as the case may be. The rightness or ,,rrongness of a particular act here depends upon the value of its tend­ ency1 i.e. upon its generalized utility; or al­ tematively upon the value of its tendency

UTILITARIAN GENERALIZATION as compared with the values of the tenden­ cies of the alternatives, i.e. upon its relative generalized utility. The generalization test occurs in var­ ious familiar linguistic shapes and often in­ corporates the substance of the matter at hand. Thus we may have: vVhat if everyone dodged the draft? Suppose everyone lied just to suit his own convenience? But suppose everyone failed to pay his taxes! What would happen if no one both­ ered to vote? vVhen such objections are made, many kinds of disagreement can arise. For exam­ ple: ( 1 ) ''Nell, what would happen? The question you pose is more complex than it may appear. The total effects of everyone's doing what I propose to do may be quite different, qualitatively different, from the effects of this one act. If people haven't generally acted in this way before, we may not know, we may have no reliable idea of, what consequences would result.' The particular details of the factual problems suggested here will not be con­ sidered in this study. We are concerned with the general nature of these principles, and this will lead us to examine some em­ pirical ( causal) phenomena. But we shall not be concerned "�th the specific applica­ tions of the principles, and thus not gener­ ally with the practical problems of getting the required information and correctly in­ ferring judgements from the principles on the basis of that information. The practical problems here are akin to, though more complex than, a set of difficulties faced in applying simple utilita­ rianism. In that case the implications of a given principle depend upon all the effects ( all the utilities and disutilities) of individ­ ual acts, no matter how remote or indirect they may he. Such practical obstacles to success in discovering what a given princi­ ple actually implies are compounded in the

453 case of utilitarian generalization, for there one is concerned, not \vith all the effects of one act, but with all the effects of every one of a class of similar acts, supposing that all are performed. ( 2 ) 'But would the results be as had as you suggest? Would they be bad at all? How do you judge so? Why in that way?' This value-theoretic set of problems, in practice linked with ( l ) , will not concern us either. I am distinguishing two features of a teleological or utilitarian theory and dealing with one only. vVe are leaving value-theory aside and shall concentrate upon the structure of utilitarianism-how the value-criteria are to be applied. Thus we shall not ask 'What are the criteria of intrinsic goodness?' or 'What things are de­ sirable ( undesirable ) ?' or 'How can we de­ cide what is a desirable goal?' We shall consider only questions related to differ­ ences in utilitarian theories such as the differences between simple and general uti­ litarian considerations. VVe are doing so be­ cause some have thought that a mere difference in structure along these lines re­ sults in a substantive difference in the im­ plications of utilitarian principl�s. But if we do not concern ourselves with value-criteria, and therefore set no restrictions upon them at all, this will allow us to call certain theories 'utilitarian' even though they might not ordinarily be so called. For example, 'self-realizationist' te­ leological theories might be counted as uti­ litarian; and is this not a confusion to be avoided? The answer is, that we need not be concerned with such distinctions. It is merely a terminological- and partly histori­ cal-point, which principles we choose to call 'utilitarian. The forms of utilitarian ge­ neralization and also the species of rule­ utilitarianism that we shall examine are in fact, usually supposed to be applied in don­ junction with universalistic value-criteria ( where the interests of each person count equally) , and these theories may therefore be counted as 'utilitarian' in one restricted sense. Bue we are not assuming that a utili-


tarian theory is necessarily hedonistic, for example ( i.e. based upon a pleasure princi­ ple), and we need impose no other evalua­ tive restrictions. The reason some have been concerned to restrict value-criteria used in conjunction with 'utilitarian' principles is that by adopt­ ing certain ad hoc valuations the utilitarian seems to escape at least some of the tradi­ tional criticisms of his theory. Thus, as we shall see, the 'ideal' utilitarian can claim that just distributions are intrinsically good ( and unjust distributions intrinsically evil ) and thereby attempt to assimilate justice to utility and in that way accommodate utili­ tarianism to a class of criticisms based on appeals to justice. But, as I shall argue in the last chapter, even this move will take the utilitarian only so far and not far en­ ough. For what the utilitarian cannot allow is that some value related to the rightness or wrongness of acts is characteristic of acts of certain kinds, e.g. unfair acts, inde­ pendently of their effects. The only condition we must impose is that, when principles are compared, the value-criteria employed in conjunction with them must of course be ( whatever else they are ) identical. This will tacitly be as­ sumed- for our arguments,. as opposed to il­ lustrative examples, will be strictly sche­ matic, requiring no specification of value­ criteria. The following issues will receive atten­ tion in the immediately succeeding chap­ ters: ( 3 ) 'What is the force of "everyone" in your objection? Who is to count? Surely not everyone, for not everyone will have occasion to do this kind of act. Shall we consider merely those who will pass by this orchard, or all those who will pass by all similar orchards? Or shall we consider only those who will notice the apples? Or per­ haps only those who will be strongly tempted to take some? How do we decide which class to consider? How does one show that a particular method of selection is not arbitrary?'


Similarly: ( 4) 'What we to counrt as the same sort of action? And how do we de­ cide? Shall we consider "picking apples" or "stealing apples"? Shall we mention that no one is looking or that there will be many left when we have taken some?' The latter two sets of problems are more fundamental than ( 1 ) and ( 2 ) , for any defensible application of utilitarian generalizatiQn presupposes answers to the questions raised. ( 5 ) 'But of course not everyone will do the same.. To suppose that they will is to suppose fal,;ely. And to act upon such a false supposition intentionally is to mislead oneself regarding the circumstances-and therefore the effects-of one's act - a very unutilitarian thing to do.' ( 6 ) 'But of course few others will do the same. Therefore the evil will not be produced anyway, regardless of what I do, so my act cannot be wrong.' ( 7) 'But of course most others ,vill do the same. Therefore the evil will be pro­ duced anyway, regardless of what I do, so my act cannot be wrong.' ( 8 ) 'My act itself will not have bad effects. And I am responsible for my acts alone, not for what others will or might do. Thus there is no utilitarian ground against my acting this way- regardless of what oth­ ers do.' These objections involve a set of re­ lated misunderstandings regarding utilita­ rian generalization which none the less suggest real problems as to the relevance of the behaviour of others. We shall deal with the relevance of others' behaviour in some detail. ( 9 ) 'Granted that this is an act of the kind you specify; but there are also impor­ tant differences. This is a special case which deserves special consideration ( or in­ dulgence).' And £nally: ( 10) '\,V'hat does it matter? Why should I consider such an objection at all?'



bly provides a ground o,· criterion for their ascription. The test is for the universaliza­ bility of a maxim, i.e. whether the maxim of one's act could possibly become a 'uni­ Until recently, the notion of generali­ versal law' and whether an ideal rational zation in ethics was not normally asso­ agent could consistently will that lt become ciated with utility. Generalization has had a universal law. Clearly, neither notion of generaliza­ two primary associations: the principle of tion is at all related to utility. It would generality and Kant's ethical doctrines. The principle of generality-otherwise therefore be misleading to speak of utilita­ called, e.g. the principle of impartiality or rian generalization as 'Kantian' simply be­ equity-merely asserts that moral consider­ cause it involves a notion of generalizatio:i. The earliest, pioneering study of a fonn ations have a universal character or 'bind­ ingness'. A common formula for this notion of utilitarian generalization was made by is �Treat like cases alike�-and, as we under­ C. D. Broad two generations ago. {'On the star1d, Treat relevantly different cases Function of False Hypotheses in Ethics', differently.' More particularly, we may say: International Journal of Ethics xxvi (April if it is right ( or wrong) for someone to do a 1916), 377-97. ) Broad considered the more certain klnd of thing, then it is likewise common negative form, that \vhich concerns right ( or wrong) for anyone to do a similar only the undesirable effects of general thing. Sometimes, the principle is under­ practices. He pointed out that such a prin­ stood as requiring that moral criticism and ciple is nonnally applied when it is as c e r ­ justification turn upon rules and princi­ tain as possible that not everyone will do ples-or at least tum upon general reasons. the same-that the general practice will not The principle of generality thus has a actually occur. The supposition (hypothe­ minimal content. It says nothing about sis) that everyone ,,ill do the same is nor­ which acts are right or wrong, nor why mally counterlactual. Thus Broad called ar­ some are right and others wrong; nothing guments based on the generalization test about which are to be regarded as similar 'the method of false hypothesis' or of 'false and vvhich as different, nor why they may universalisation in ethics. He argued that be so regarded. In this sense, it is a fonnal such a principle when viewed as st-ictly principle ( thus, sometimes called the for­ utilitarian was paradoxically most unutili� mal principle of justice) : it te1ls us about tarian, since it required acting upon a 'false morality, about the generality of moral account of t'ie circumstances'. This led him considerations, but not about their content, towards an 'ideal' utilitarian position. vVe nothing about the rightness or wrongness shall examine these arguments in their of acts as such. One might say that it con­ turn. Broad, s study neither represented nOr cerns the correctness or soundness of moral reasoning as distinct from the direct assess­ occasioned a n10vement in moral philoso­ ment of acts ( or of other moral subjects ) . phy. It developed a.s a critique of one com­ The Kantian notion of generalization mon method of moral reasoning which pre­ ( or of universalizability) is not, on the sented obvious difficulties. It is understand ­ other hand, strictly formal. Kant's theory able, then, that for two decades after directly concerns the assessment of moral Broad's unfavourable review utilitarian ge­ subjects ( in this case, L½e maxims of ac� neralization was largely ignored by aca� tions, e.g. 'Do such and such' or �n such demie philosophers. ivfeanwhile, utilitarianism came under and such circumstances� do so and so'); it concerns the application of first-order severe attack. Actually, crlticis� was di­ moral tenns such as "good\ and it presuma- rected at simple utilitarianism-or at the


456 predominant form of it, Act-Utilitarianism. For at that time the distinction between simple and general utilitarianism svas un­ noticed, and thus Act-Utilitarianism was taken as the paradigm theory. Roughly speaking ( as I shall e:q,lain), Act-Utilitarianism is the theory that one should always perform acts the effects of which would be at least as good as those of any altemative. These are right actions; ail others are wrong. It is one's duty, or over-all obligation, to perform right acts only; and thus if one act has the best consequences, that act is the thing to be done. In our terminology, this grounds the moral assessment of acts upon their relative simple utilities. This theory has otherwise been called 'crude'� 'extreme', or 'direct' utilitarianism--although these terms have also been used rather generally with regard to simple utilitarianism. Admittedlv, the classical utilitarian theories might not properly be character­ ized as purely simple utilitarian. None Uie less, partly through the influence of G. E. Moore ( Principia Ethica and Ethics;, in this century the traditional variety had corne to be viewed as simple utilitarian, and Act-Utilitarianism as a coherent fonnula­ tion of the predominant traditional theory. Admittedly, there were differences among utilitarians during the first three or four decades of this century-concerning value-theory ( e.g. hedonistic versus 'idear uti!itariuns ) ; concerning the scope of moral considerations (positive versus negatve utilitarians ) ; concerning responsihility ( whetlier actual or probable consequences should be considered); and so on- but these differences developed within the con­ fines of simple utilitarianism. Thus, while outside cri3,cisms have rn.ainly been directed against Act­ Utilitariauism, their point has been that utility is not the sole ( or perhaps not at all a ) detem1inant of right action. These criti­ cisms have had two related aspects. First, counter-examples v;,rere offered) examples of purportedly strong o hligations, the exJst-

WHAT ACTS ARE RIGHT? ence or strength of whieh could supposedly not he accounted for bv Act-Utilitarians. Criticisms have, for exa,;ple, turned upon purported 'prima facie obligations' such as those of fidelity, obligations resting more upon past acts or circun1stances than upon the effects of present and future acts. It has been claimed that Act-Utilitarianism can­ not adequately account for our obligations to keep our promises, to repay our debts, to tell the truth, to punish the guilty and pro­ tect the innocent. In particular, it has been held that a really wrong act can appear right on Act-Utilitarian grounds, just be­ cause a condition of secrecy shrouds the act. And it is supposed that 'a condition of secrecy should not Y'/ea."-

457 of moral reasoning have stratifled it into several 'levels': £rst, the justiflcation of particular judgements about the rightness or \.vrongness of an act by reference to a good reason pro or con kinds of acts, or by reference to a moral rule; secondly, the validation of such reasons or roles by refer­ ence to higher-order rules or principles or criteria; and perhaps third ( there are var­ iations here ), the vindication or ultimate defence of these higher rules, principles, IYf cdteria. ( See, e.g., H. Feig), 'Validation and Vindication', in W. Sellars and J. Hos­ pers, Readings in Ethical Theory [New York: Appleton-Century-Crofts, Inc., 1952], pp. 667-80; K. Baier, The Moral Point of View [Ithaca, New York: Corne!'. Uni­ versity Press, Hl58]; P. VV. Taylor, Nor­ mative Discourse [Englewood Cliffs, New Jersey; Prentice-Hall, 1961]. ) Now when the first-order rules ( or reasons ) are grounded upon a second-order criterion of utflity1 we have rule-utilitarianism. To complicate matters, however, the most notable early theories, such as Toul­ min'.s, had impure second-order criteri a., not strictly utilitarian, Toulmin placed a special premium upon the social accept­ ance of rules as opposed to their utilities. ( S. E. Toulmin, .A.n Examination af the Place of Reason in Ethics [Cambridge, England : Cambridge University Press, 1950] . ) This impurity in the rule-utilitarian tradition is one reason £or coining the new label, 'utilitaria n generalization'. But, as I have said, the latter also involves no direct reference to rules. Another mode of argument intending to lead towards rule�utilitarianism or to� wards utilitarian generalization is based upon the time-honoured method of appeal­ ing to examp!e, By showing that the new type of principle does not fall prey to the traditional criticism, the counter-examples originally offered against util:.tarianism in general are vitiated. There is a dialectic .in� volved here which I shall comment on presently. The following examples illustrate one frequent aspect of the argument, that



of the revisionistic against the traditional reject utilitarianism. He would argue that most voters face a similar predicament, utilitarian: ( 1 ) Should one hother to vote when it each finding it inconvenient to cast his sin­ is inconvenient to do so? One knows, gen­ gly indecisive hallot. If each reasoned in erally, that his single ballot will not be the Act-Utilitarian way he would decide especially signi£cant; therefore, the direct against voting. But what would happen if effects of voting and of not voting will everyone who found it inconvenient to vote hardly be different, if at all. And, regarding failed to vote? Since, as vre are supposing, indirect effects, while one's absence from most vvill find it inconvenient to vote, very the polls will ( let us assume) not be no­ few would then vote; consequently, the ticed by others, and therefore will not in­ wrong man could be elected; or worse still, fluence their behaviour, it would on the the mass abstentions might seriously hann other hand be more convenient not to vote. the electoral system which is, in the long If we tally up the score, it appears that run, of great importance to all. And this an Act-Utilitarian must hold that it would evil would far outweigh the total in­ be wrong to vote under such circum­ convenience which could be avoided by stances, since the over-all effects of voting abstaining. Thus, if one abstained, the bal­ are worse than those of abstaining. Thus it ance of utility vmuld be positive; no harm seems that mere inconvenience provides a would be done, and inconvenience would reason overriding whatever good reason \Ve be avoided. But if everyone who found it ordinarily have to vote. And this, many inconvenient to vote abstained, the over-all effects would be bad, much worse than if would hold, is simply not so. An Act-Utilitarian can, of course, ob­ all those bad voted and suffered the in­ ject that our facts are mistaken, that we conveniences . Only when the general prac­ have weighed the utilities incorrectly, that tice is considered-and not always \vhen \Ve have overlooked certain pernicious indi­ the individual act is considered separ­ rect effects. Here innumerable argumenta­ ately-can we take into account certain tive complications can arise which, for our undesirable consequences. Only by appeal­ present purposes, vve need not consider. ing to utilitarian generalization can certain Let us assume, for the sake of the immedi­ undesirable consequences be avoided. ate argument, that there can be such a case But it might be objected that this is of voting that, because of inconvenience, illusory. For if others do not act in the con­ would be wrong 011 the Act-Utilitarian ac­ demned way, e.g. if others do vote even count. when it is inconvenient, the undesirable Now some \vould hold this as a case consequences in question will not mate­ against Act-Utilitarianism. It is supposed rialize whatever one does. And conversely, that one has a good reason for voting that no matter what one does, if others do not is stronger than Act-Utilitarianism sug­ vote the evil will be produced. The propo­ gests: mere inconvenience ( as opposed to nent of utilitarian generalization is then serious suffering or hardship ) does not pro­ pressed to different reasons for urging that vide a sufficient countervailing reason. utilitarian generalization is none the less a And, considering the generality of the Act­ tenable moral principle while Act­ Utilitarian theory, if the theory is wrong in Utilitarianism is not. He then argues that a one case it cannot be accepted; it is not an valid moral principle is one - that everyone adequate account of the rightness or can hold and act upon. If everyone fol­ wrongness of actions. lowed Act-Utilitarianism, certain bad con­ The rule-utilitarian- or better, the sequences would result that could be proponent of utilitarian generalization­ avoided if everyone in fact acted according might accept these criticisms and yet not to utilitarian generalization. VVe shall put a

UTILITARIAN GE;,;ERALIZATIO.s' fuller perspective upon arguments like these later on, in the third and fourth chapters. (2) Suppose now that one lives in a tmv:n in which racial segregation is the bru­ tally enforced rule, and in which the pros­ pects of changing the oppressive system are quite slim, How should one take these facts into account? One who contravenes the segregation rules endangers his family and friends, jeopardizes his home and live­ lihood, perhaps removes himself from fur­ ther activity in the community. For ex� ample, fairly certain dangers face a racially mixed group even if they gather at one's home, and face someone who vigorously agitates for reform, The good results of such unorthodox beha,1our will ( let us suppose) be far ou:weighed by the venge­ ful harm done those who refuse to ac­ quiesce in the rules. It may therefore be argued that Act­ Utilitarianism counsels inactivity in such a case-or at least that one refrain from activities which expose one or others to clanger. A pro?onent of utilitarian generaliza­ tion might argue, however, that if everyone were to continue to acquiesce, if everyone failed to take the risks entailed by defying and seeking to change the rules, then the sdering imposed by the system would continue unabated. But if1 on the other hand, everyone were to run the risks, very 2.esirable consequences would result1 far outweighing indhodua: sacrifices which might be made, since the joint effort wonld be sufficient to change tbe system, Again, we have an example suggesting a peculiar divergence between the two kinds of utilitarianism. But the main inter� est in this case is that personal sacf'lice is involved, and that the proponent of utilita­ rian generalization wodd hold that1 if such sacrifice ,vere cailed for by his principle, then it must be suffered. Notice, hmvever, tha!: if one or only a fe·.v take the risks and suffer the burdens, that will not be suffi­ cient to produce the good that could he produced by ( or to eliminate the evil that

459 could be e:irninatec! by) everyone's doing the same. None the less, it appears that no part of the generalization test concerns what others actually are doing or will do, Thus, these applications of utilitarian generalization make that sort of principle appear guite unutilitarian: For the generalizer appears to surrender that hard-headed praeticality which has been the hall-mark of utilitarianism. He is. not supposed to be concerned with the actual effects of his acts, but only with the conjec­ tured effects of an hypothesized general practice, He may have to accept the sac­ rifice, in accordance with his principle, on tbe false supposition that others will do the same. How should one act? Should one re­ fuse to doff one's hat before the tyrant's statue, knowing that if one acts alone one will suffer as a result and that no appreci­ able good w�ll come of it, and knowing also that others will not refuse to doff their hats? Are there social situations, political regimes, in which one simply ought not to acquiesce? relevance of present cases to these questions is that utilitarian generalization appears to provide a utilita­ rian ground for disobedience, a utilitarian ground for seemingly unutilitarian but­ shall we say?� -morally i.-nperative acts, "\Ve shall have to examine whether ou::: ex­ amples mislead us, whether this is really a utilitarian argument after all. ( These issues will arise in the last chapters.) Let us now consider such examples in the light of the relevant developments in moral philosophy. A three-sided conflict has accompanied the rise of the nevv utilitarianism. In the first place, critics of utilitarianism have claimed that it cannot account for certain reliable moral beliefs or data. ( These arguments, as \Ve have noted, have mainly been directed against si:::nple utilitarianism, although they have also been extended to apply against general utilitarianism. And there have been attacks on the latter by simple utilitarians, How­ ever, criticisms of general utilitarianism


have either been based on peculiarities of particular theories or presentations of them, or rest upon an inadequate grasp of the nature of utilitarian generalization. YVe can, therefore, ignore this aspect of the de­ bate.) Secondly, simple utilitarians have attempted to reject or accommodate the criticisms. They may claim that the charges are based upon factual error or evaluative oversight; they may go so far as to claim that the charges are based upon moral error; or they may qualify simple utilitarianism, patching it up to meet the objections. Finally, the proponents of utilitarian generalization ( or of rule­ utilitarianism ) generally accept the tra­ ditional criticisms of simple utilitarianism while claiming that these are ineffective against their new theories. Thus, examples such as those I have outlined are supposed to establish the su­ periority of the new to the old utilitarian­ ism. But what is our method and what are our criteria for criticizing and comparing alternative moral theories? I shall not go into this general question extensively, but I shall deal (in Chapter IV) with a characteristic argument purporting to show the superiority of the new utilitarianism. It should be obvious, however, that any attempt to displace simple by general ( or rule- ) utilitarianism presupposes a pos­ itive answer to the following questions : Are there in fact any substantive differences be­ tween these theories? If so, between which ones? And are these differences in the requisite directions? Utilitarians have generally failed to ex­ amine these questions. Instead their argu­ ments develop along the lines already suggested. A proponent of the new utilitarianism accepts, as ( most probably ) correct, judgements or generalizations, about particular acts or kinds of acts, that seem inconsistent with simple utilitarian­ ism. Or he accepts, as ( most probably ) correct, rules or reasons that are used in criticizing or justifying acts but which are


prima facie unutilitarian. For example, he may agree with the critics that the strength of one's obligations to keep one's promises ( or to tell the truth ) is greater than the rel­ ative simple utility of promise-keeping ( or veracity) would make it appear. In this re­ spect, some hard moral data are more or less assumed- are held less vulnerable to criticism than simple utilitarianism. But it must be observed that the supposed inconsistency of such data with simple utilitarianism is not rigorously substantiat­ ed. For any such argument requires that we pin le think that individuals should go ahe,;d ,;ith a good example and not wait until the mle-making powers of the group are used. Others argue L'1at this is putting loo great a burden on the public-spirited, Thus, compulsory military service vdth exemptions granted to those engaged in important national industries is said by some to be fairer, volunteering for is said by others to be mor­ ally preferable. I can see no reason for the latter view. It may indeed seem preferable from the military point of view} for it may be argued that volunteers are better so1diers. But there is no reason why if keen­ ness is \vanted volunteers should not have preferential rights to serve in the army rather than in industrv. On the other hand, there is no reason ;hy the sacrifices in� volved in the defense of their country should be borne only by those who are tak­ ing their moral responsibilities seriously, and no reason why those who are not should benefit gratuitously. In the absence of argument showing that the method of individual initiative vields a ::nore efficient army, the other se�ms to me preferable and, in any case) obviously fairer. IIesita­ tion to use the 1awmaking force of the com� munity· is understandable, for such use may endanger individual freedom, but often this hesitation is supported on the grounds of the moral preferability 0£ individual sacri­ fice and initiative. Such arguments seem to me unsound.

What Makes Right Acts Right? SIR DA FID ROSS The real point at issue between hedonism and utilitarianism on the one hand and their opponents on the other is not whether �right' means 'productive of so and so'; for it cannot with any plausibHity be main­ tained that it does. The point at issue is that to which we now pass, viz. whether there is any general character which makes right acts righ� and if so, what it is. Among the main historical attempts to state a sin­ gle characteristic of all right actions which is the foundation of their rightness are those made by egoism and utilitarianism. But I do not propose to discuss these, not because the subject is unimportant, but be­ cause it has been dealt with so often and so well alreadv, and because there has come to be so m,;ch agreement among moral phi­ losophers that neither of these theories is satisfactory. A much more attractive theory has been put forward by Professor Moore: that what makes actions right is that they are productive of more good than could have been produced hy any other action open to the agent1 This themy is in fact the cuL'Ilination of all the attempts to base rightness on productivity of some sort of result. TI1e first form this attempt takes is tbe attempt to base rightness on conduciveness to the ad­ vantage or pleasure of the agent. · This Reprinted from The Right and the Good by Sir David Ross by permission of the Clarendon Press, Oxford. Copyrigh! 1930. 1 I take the theory which, as I have tried to fo:nvard in Ethics rat.'ler shovt, seems to be than the earlier plausible theory put forward in Priri.;ipia For the difference, cf. my pp. 8-11.


theory comes to grief over the fact, which stares us in the face, that a great part of duty consists in an observance of the rights and a furtherance of the interests of others, whatever the cost to ourselves may be. Plato and others may be right in holding that a regard for the rights of others never in the long run involves a loss of happiness for the agent, that 'the just life profits a man/ But this, even if true, is irrelevant to the rightness of the act. As soon as a man does an action because he thinks he will promote his O'\VD. interests thereby, he is acting not from a sense of its rightness but from self-interest. To the egoistic theory hedonistic utili­ tarianism supplies a much-needed a:nend­ ment. It points out correctly that the fact that a certaL'1 pleasnre will be enjoyed by the agent is no reason why he ought to bring it into being rather than an equal or greater pleasure to be enjoyed by another, though, human nature being what it is, it makes it not unlikely that he will try to bring it into being. But hedonistic utilitar­ ianism in its tum needs a correction. On reflection it seems clear that pleasure is not the only thfrtg in life that we thin.le good in itself, that for instance we think the posses­ sion of a good character, or an intelligent understanding of the world, as good or bet­ ter. A great advance is made by the substi­ tution of 'productive of the greatest good' for 'productive of the greatest pleasore.' Not only is th's theory more attractive than hedonistic utilitarianism, but its logi­ cal relation to that theory is such that the latter could not be true unless it were true,



while it might be true though hedonistic utilitarianism were not It is in £act one of the logical bases of hedonistic utilitarian­ ism . For the view that what produces the maximum pleasure is right has for its bases the views ( l) that what produces the max­ imum good is right, and ( 2) that pleasure is the only thing good in itself. If they were not assuming that what produces the maxi­ mum good is right, the utilitarians' attempt to show chat pleasure is the only thing good in itself, which !s in fact the point they take most pains to establish, would have been quite irrelevant to their attempt to prove that only what produces the maxi­ mum pleasure is right. If, therefore, it can be shown that productivity of the maxi­ mum good is not what makes all right ac­ tions right, we shall a fortiori have refuted hedonistic utilitarianism. Wlien a plain man ful£ls a promise be­ cause he thinks he ought to do so, it seems clear that he does so with no thought of its total consequences, still less with any opin­ ion that these are likely to be the best pos­ sible. He thinks in fact much more of the past than of the future. What makes him think it right to act ,n a certai.'l way is the fact that he has promised to do so--that and, usually, nothing more. That ms act will produce the best possible consequences is not his reason for calling it right. vVhat lends colour to the theory we are examin­ ing, then, is not the actions ( which form probably a great majority of our actions) in which some such reflection as 'I have promi.Sed' fa the only reason \Ve give our� selves for thinking a certain action right, but the exceptional cases in which the consequences of fulfilling a promise ( for instance) would be so disastrous to others that we judge it right not to do so. It must of course be admitted that such cases exist. If I have promised to meet a friend at a particnlar time for some trivial purpose, I should certainly think myself justined in breaking my engagement if by doing so I could prevent a serious accident or bring relief to the victims of one. And the sup-

porters of the view we are examining hold that my thinking so is due to my thinking that I shall bring more good iota existence by the one action than bv the other. A different account mayi ho\� ever, be given of the matter, an account which will, I be­ lieve, show itself to be the true one. It may be said that besides the duty of fulfilling promises I have and recognize a duty of re­ lieving distress,2 and that when I think it right to do the latter at the cost of not doing the former, it is not because I trunk I shall produce more good thereby but be­ cause I think i t the duty which is in the cir­ cumstances more of a duty. T'nis account surely corresponds much ,nore closely with what we really think in such a situation. If, so far as I can see, I could bring equa: amounts of good into being by fulfilling my promise and by helping some one to whom I had made no promise, I should not hesitate to regard the fonner as my duty. Yet on the viev;;r that what is right is right �ecause it is productive of the most good I should not so regard it. There are Wo theories, each in its wav simple, that offer a solution of such cases ;£ conscience. One is the view of Kant} that there are certain duties of perfect obliga­ tion, such as those of fulfllling promises, of paying dehts, of telling the truth, which admit of no exception whatever in favour of duties of imperfect obligation, such as that of relieving distress. The other is the view 0£1 for instance 7 Professor Nfoore and Dr. Rashdall, that there is only the duty of producing good, and that all 'conflicts of duties' should be resolved by asking 'by whlch action will most good be produced?' .But i t is more important that our theory fit the facts than that it be simple, and the ac­ count we have given above corresponds ( it seems to me) better tI,an either of the sim­ pler theories with what we really think, viz. that nmmally promise-keeping, for exam­ ple, should come before benevolence, but 1

but t1 These are not strictly s eaidng duties, f. things t._'-:ta.t tend to be our outy, or prirr.a fade du:ies, Cf. pp. 19-20,

484 that when and only when the good to be produced by the benevolent act is very grea: and the promise comparatively triv­ ia� the act of benevolence becomes our duty. In fact the theory of 'ideal utilitarian­ ism/ if I may for brevity refer so to the theory of Professor lvfoore, seems to sim� pH-'}, unduly our relations to our fellows. It says, in effect, that the only morally signifi­ cant relation in which my neighbours stand to me is that of being possible beneF.ciaries by my action.' They do stand in this rela­ tion to me: and this relation is :norally sig­ ni!icant. But they may also stand to me in the relation of promisee to promiser1 of creditor to debtor, of wife to husband, of child to parent, of friend to friend, of fellow countryman to fellow countryman, and the like; and each of these relations is the foc,ndation of a prima facie duty, which is rnore or less incumbent on me according to the circumstances of the case. 'When I am in a situation) as perhaps I always am1 in which more than one of these prima facie duties is incumbent on me, ,vhat I have to do is to studv the situation as fullv as I can until I form, the coasldered opm'ion (it is never more ) that in the circumstances one of them is more incumbent than anv other; then I am hound to think that to' do this prima facie duty is my duty sans phrase in the situation. I suggest 'prim� facie duty' or 'condi­ tional duly' as a brief way of referring :o the characteristic ( quite distinct from that of being a duty proper) which an act has, in virtue of being of a certain kind ( e.g. the keeping of a promise ) , of being an act which would he a duty proper if it were n ot at the same time of a...,other kind ,vhich is morally significant. ' Whether an act is a dt:ty proper or actual duty depends on all the morally .signincant kinds it is an in3 Some will think it, a_part from other considera­ tio:1s, a sufficient refutation of this vi:ew !o point out that I aiso stand in that relation to myself, so that for this view the distinction of oneself from others is morally insignificant.

WHAT ACTS ARE RIGHT? stance of. The phrase 'prima facie duty' must be apologized for, since ( 1) it sug­ gests tliat what we are speaking 0£ is a cer­ tain kind of du:y, whereas it is in fact not a duty, but something related in a special way to duty. Strictly speaking, we want not a phrase in which duty is qualified by an adjective, hut a separate noun. ( 2) 'Pr!md fac!e suggests that one is speaking only of an appearance which a moral situation p:e� sents at first sight, and which may turn out to be illusory; whereas what I am speaking of is an objective fact involved in the na­ ture of the situation, or more strictly in an element of its nahtre, though not, as duty proper does, arising from its whole nature. { can, however, think of no term which folly meets the case. 'Claim' has been sug­ gested by Professor Pricha.:d. The word 'claim' has the advantage of being quite a familiar one in this connexion, and it seems to cover much of the ground. It would be quite natural to say, 'a person to whom I have made a promise has a c1aim on me,' a."!d also> ; the name is wide enough if we are willing to include under it implicit promises, i.e. modes of behav­ iour in which '\vithout explicit verbal prom­ ise \Ve lnten::ionally create an expectation that ,ve can be counted on to bellave in a certain way in the inte!"est of another per­ son. These seem to be, in principle, all the ways in which prima facie dcties arise. In actual experience they are compounded to­ gether in highly complex ways. Thus, for example, the duty of obeying the laws of one's country arises partly ( as Socrates


contends in the Crito ) from the duty of gratitude for the benefits one has received from it; partly from the implicit promise to obey which seems to be involved in perma­ nent residence in a country whose lmvs we know we are expected to obey, and still more clearly involved when we ourselves invoke the protection of its laws ( this is the truth underlying the doctrine of the social contract ) ; and partly ( if we are fortunate in our country ) from the fact that its laws are potent instruments for the general good. Or again, the sense of a general obliga­ tion to bring about ( so far as we can) a just apportionment of happiness to merit is often greatly reinforced by the fact that many of the existing injustices are due to a social and economic system which we have, not indeed created, but taken part in and assented to; the duty of justice is then rein­ forced by the duty of reparation. It is necessary to say something by way of clearing up the relation between prima facie duties and the actual or abso­ lute duty to do one particular act in parti­ cular circumstances. If, as almost all moral­ ists except Kant are agreed, and as most plain men think, it is sometimes right to tell a lie or to break a promise, it must be maintained that there is a difference be­ tween prima facie duty and actual or abso­ lute duty. vVhen we think ourselves justi­ fied in breaking, and indeed morally obliged to break, a promise in order to relieve someone's distress, we do not for a moment cease to recognize a prima facie duty to keep our promise, and this leads us to feel, not indeed shame or repentance, but certainly compunction, for behaving as \Ve do; \Ve recognize, further, that it is our duty to make up somehow to the promisee for the breaking of the promise. We have to distinguish from the characteristic of being our duty that of tending to be our duty. Any act that we do contains various ele­ ments in virtue of which it falls under var­ ious categories. In virtue of being the breaking of a promise, for instance, it


tends to be wrong; in virtue 9f being an instance of relieving distress it tends to be right. Tendency to be one's duty may be called a parti-resultant attribute, i.e. one which belongs to an act in virtue of some one component in its nature. Being one's duty is a toti-resultant attribute, one which belongs to an act in virtue of its whole nature and of nothing less than this.7 This distinction between parti-resultant and toti­ resultant attributes is one which \Ve shall meet in another context also. 8 Another instance of the same distinc­ tion may be found in the operation of natu­ ral laws. Qua subject to the force of gravi­ tation towards some other body, each body tends to move in a particular direction \vith a particular velocity; but its actual move­ ment depends on all the forces to which it is subject. It is only by recognizing this dis­ tinction that we can preserve the absolute­ ness of Iavvs of nature, and only by recog­ nizing a corresponding distinction that we can preserve the absolutness of the general principles of morality. But an important difference between the tvm cases must be pointed out. vVhen we say that in virtue of gravitation a body tends to move in a cer­ tain way, we are referring to a causal in£u­ ence actually exercised on it by another body or other bodies. When we say that in virtue of being deliberately untrue a cer­ tain remark tends to be wrong, we are re­ ferring to no causal relation, to no relation that involves succession in time, but to such a relation as connects the various at­ tributes of a mathematical figure. And if the word 'tendency' is thought to suggest too much a causal relation, it · is better to talk of certain types of act as being prima facie right or wrong ( or of different per­ sons as having different and possibly con­ flicting claims upon us ) , than of their tend­ ing to be right or wrong. Something should be said of the rela­ tion between our apprehension of the 1 8

But cf. the qualification in p, 33, n: 2. Cf. pp. 122-3.


prima fade rightness of certain types of act and our mental attitude towards particular acts. It is proper to use the word 'appre­ hension' in the former case and not in the latter. That an act, qua fuIBlling a promise, or qua effecting a just distribution of good, or qua returning services rendered, or qua promoting the good of others, or qua pro­ moting the virtue or insight of the agent, is prima facie right, is self-evident; not in the sense that it is evident from the beginning of our lives, or as soon as \Ve attend to the proposition for the first time, but in the sense that when we have reached sufficient mental maturity and have given sufficient attention to the proposition it is evident without any need of proof, or of evidence beyond itself. It is self-evident just as a mathematical axiom,. or the validity of a form of inference, is evident. The moral order expressed in these propositions is just as much part of the fundamental nature of the universe ( and, we may add, of any pos­ sible universe in which there were moral agents at all) as is the spatial or numerical structure expressed in the axioms of geome­ try or arithmetic. In our confidence that these propositions are true there is involved the same trust in our reason that is in­ volved in our con£dence in mathematics; and we should have no justification for trnsting it in the latter sphere and distrust ­ ing it in the former. In both cases we are dealing with propositions that cannot be proved, but that just as certainly need no proof. Some of these general principles of prima facie duty may appear to be open to criticism. It may be thought, for example, that the principle of returning good for good is a falling off from the Christian prin­ ciple, generally and rightly recognized as e,q,ressing the highest morality, of return­ ing good for evil. To this it may be replied that I do not suggest that there is a princi­ ple commanding us to return good for good and forbidding us to return good for evil, and that I do suggest that there is a posi­ tive duty to seek the good of all men. What


I maintain is that an act in which good is returned for good is recognized as specially binding on us just because it is of that character1 and that ceteris parihu� any one would think it his duty to help his benefac­ tors rather than his enemies,. if he could not do both; just as it is generally recognized that ceteris paribus we should pay our debts rather than give our money in char­ ity, ,vhen we cannot do both. A benefactor is not only a man, calling for our effort on his behalf on that ground, but also our ben­ efactor, calling for our special effort on that ground. Our judgements about our actual duty in concrete situations have none of the cer­ tainty that attaches to our recognition of the general principles of duty. A statement is certain, i.e. as an expression of knowl­ edge, only in one or other of two cases: when it is either self-evident, or a valid conclusion from self-evident premisses. And our judgements about our particular duties have neither of these characters. ( 1 ) They are not self-evident. vVliere a possible act is seen to have two characteristics, in virtue of one of which it is prima facie right, and in virtue of the other prima facie wrong, we are ( I think) well aware that we are not certain whether we ought or ought not to do it; that whether we do it or not, we are taking a moral risk. -VVe come in the long run, after consideration, to think one duty more pressing than the other, but ,ve do not feel certain that it is so. And though we do not always recognize that a possible act has two such character­ istics, and though there may be· cases in which it has not, we are never certain that any particular possible act has not, and therefore never certain that it is right, nor certain that it is ,vrong. For, to go no fur­ ther in the analysis,. it is enough to point out that any particular act will in all proba­ bility in the course of time contribute to the bringing about of good or of evil for many human beings, and thus have a prima facie rightness or wrongness of which we know nothing, ( 2 ) Again, our


judgements about our particular duties are not logical conclusions from self-evident premisses. · The only possible premisses would be the general principles stating their prima facie rightness or \Vrongness qua having the different character.sties they do bave; and even if ,ve could ( as we cannot) apprehend fae extent to which an act will tend on the one hand, for example, to bring about advantages for our benefac­ tors, and on the other hand to bring about disadvantages for fellow men who are not our benefactors, there is no principle by which we can dra\-v the conclusion that it is on tbe whole right or on the whole wrong. In this respect the judgement as to the rightness of a particular act is just like the judgement as to the beauty of a particular natural object or ,vork of art. A poem .is, for instance, in respect of ce1tain qualities beaut:ful and in respect of certain others not beautiful; and our judge,nent as to the deg,ce of beau:y it possesses o;:i the whole is never reached by logical reasoning from the apprehension of its particular beauties or particular defects. Both in this and in the moral case we have more or less prob­ able opinions which are not logically justi­ fied conch.:sions from the general principles that are recognized as self-evident. TLlere is therefore much trulh in the description of the right act as a fortunate act. If we cannot be certain that it is right, it is our good fortune if the act we do is the right act. This consideration does not, hoyv� ever, make the doing of our duly a mere matter of chance. There is a parallel here between t!ie doing of dnty and the doing of what will be to our personal advantage. \Ve never know 1-vhat act will in the long run be to our advantage. Yet is is certain that we are more likely in general to secure our advantage if vve estimate to the best of our ability the probable tendencies of our actions in this respect, than if \Ve act on ca;nice. �A..nd similarly we are more likely to do-our duty if we reflect to the best of our ability on the prima facie rightness or wrongness of various possible acts in virtue


of the characteristics we nerceive them to .have, than if we act without reflection. With this greater likelihood we must be content. Many people •.vould be inclined to say that the right act for m e is not that whose general nature I have been de.scribing, viz. that ·which if I were omniscient I should see to be my duty, but that which on all the evidence available to me I should think to be my duty. But suppose thatf rom the state of partial knowledge in which I think act A to be my duty, I could pass to a state of perfect knowledge in which I saw act B to be my duty, sboulc: I not say 'act B was the right act for me to do'? I should no doubt add 'though I am not to be blamed for doing act A' But in adding this, am I not passing from the question 'what is right' to the question "what :is morally good'? At the same time I am not mak;ng the full passage :from the o:ie notion to the other; for in order that the act should be :norally good, or an act I am not to be blamed for doing, it must not merely be the act which .it is reasonable for me to think my duty; it must also be done for that reason, or from some other mora1lv good motive. Thus the conception of th� 1ight act as the act which ,it is reasonable for me to foink my duty is an unsatisfac­ torv compromise betvveen the true notion of the right act and the notion of the mor­ ally good action. The general principles of duty are ob­ viously not self-evident from the beginning of our lives. How 'do they come to be so? The answer is, that they come to be self­ evident to us just as mathematical axioms do. We find by experience that this cm,ple of matches and that couple make four matches, that this couple of balls on a wire and that couple make four balls; and by re­ fl.ection on these and similar discoveries we come to see that it is of the nature of two and t\vo to I!1ake four. In a precisely simi­ lar way, we see the prima facie rightness of an act which would be the folElment of a particular promise, and of another which


would be the fulfilment of another promise, and when we have reached sufficient ma­ turity to think in general terms, we appre­ hend prima faoie rightness to belong to the nature of any fulfilment of promise. ,:Vhat · comes first in time is the apprehension of the self-evident prirna facie rightness of an individual act of a particular type. From this we come by reflection to apprehend the self-evident general principle of prima facie duty. From this, too, perhaps along with the apprehension of the self-evident prirna facie rightness of the same act in virtue of its hav::ng anofaer characteristic as well,· and perhaps in spite of the appre­ hension of frs prima facie wrongness in virtue of its having some third characteris­ tic) \Ve corne to believe something not self-evident at all, but an object of prob­ able opinion, viz. that this particular act is ( not prima 'facie but) actually right. In this respect there is an important difference between rightness and mathe­ matical properties. A triangle which is isos­ celes necessarily has two of its angles equal, whatever other characteristics the triangle may have-whatever, for :instance; be its area, or the size of its third angle. The equality of the two angles is a parti­ resultant attribute.' And the same is true of all mathematical attributes. It is true, I may add, of prima facie rightness. But no act is ever, in virtue of falling under some general description, necessarily actually right; its rightness depends on its whole nature10 and not on any element in it. The reason is that no mathematical object ( no £gure, for instance, or angle) ever has two characteristics that tend to give it opposite resultant characteristics, while moral acts 11

Cf. pp, 28, 122-3 [in The Right and the Good]. To avoid complicating unduly the statement of the general view I am putting fonvar (10.1 ). The object of such a mancemTe might be to convince us t.'iat all moral words have their descriptive meanL>1g irremovably attached to t..liem; but, fortunately for the usefulness of moral language in expressing cha.,-,ging standards, this is not so. To take this line would be to give an account of moral lan­ guage which is, so far as :it goes, true, but

503 not sufficiently general ( in the sense in which Newtonian mechanics is not suffi­ ciently general ) . The account would suffice for the moral language of an irrevocably clos?d society, in which a change in moral standards was unthinkable; but it does not do justice to the moral language of a so­ ciety like our o,vn1 in ,;,vhich some people sometimes think about ultimate moral questions, and in which, t.lierefore, morality changes. Orwell's Newspeak in 1984 was a language so designed that in it dangerous thoughts could not he expressed, Much of Oldspeak is like this too-if we want, in the Southern States, to speak to a negro as an equal, we cannot do so by addressing him as a nigger; the word "nigger' incapsulates the standards of tbe society, and, if we ,vere confined to it, we could not break free of those standards. But fortunate!y we are not so confined; our language, as ,ve have it, can be a vehicle for nev,.r ideas. 2.8 It must be noticed that the mere fact that the descriptive meanings of ::noral words can alter does not distinguish them from ordinary descriptive words. All words can alter their; dictionaries are full of sub-headings which begin 'Obs.', And even in the case of words in cuuent use their meanings vary from occasio:::1 to occas:on within at tin1es quite \ Hmits. And there is 'family resemblance' and 'open te.�ture' and all Lliat Some people have been misled into thinking that, since de­ scriptive words have these features, and since what has caused a lot of the trouble with value-words is their shifting descrip­ tive meanings, the trouble can he. cleared up ·without distinguishing betw'een the hvo classes of ,vo:-ds. The premisses of this in­ ference are perfectly true, but tbe conclu­ sion misses the point. ValGe-words are i!)­ deed like descriptive words1 both in that they have descriptive meanings} and in that the descriptive meanings of both are alter­ able, flexible, and so on. So, if we cared to concentrate on the resemblances between the two classes of words, and ignore their differences, \Ve could call them all 'descrip-


tive \VOrds', meaning by tl1is �words having descriptive meaning'. But to do this would be to neglect an important ( indeed essen­ tial) part of the meaning of moral and other value-words; and the philosopher who wishes to do justice to this will !:lave to be more careful in choosing his tem1inol­ ogy. The tenninology to which I have my­ self cried to be consistent is the following. An exuressfon which, in a certain context, has d�scriptive meaning and no other, I call a descriptive tenn1 word� or expression, as used in that context; one which has prescriptive meaning ( whether or not it also has descriptive meaning) I call a pre­ scriptive terrn; and one which has both kinds of meaning I call an evaluative term. A value-judgement or evaluative judge­ ment is a judgement b1 which such a te1m is used; on the other hand the ::nention of an evaluative term inside quotation !!larks1 or similarly 'insulated', does not :::nake a judgement evaluative. Not all moral judge­ ments are value-judgements ( LM 11.3 ). In The Language of Morals I used the words �evaluative meaning' for �he prescriptive meaning of evaluative expressions. This had some advantages, as being a less ques­ tion-begging expression which did not pre­ suppose t..;at ,vhat gave these terms their e•,,-aluative meaning was their prescriptiv� ity; but in the end it turned out to be in the interests of clarity to make this, in effect, true by definition ( LM 11.2; 5.7 ) ; and so in the present context I feel at li::,­ erty to use the words 'prescriptive meaning' which do carry this presupposition, and are sornewhat clearer in that they avov>? it. To give examples of the use of t½ese terms: 'red�: i n most contexts, is a descriptive term ( though not when used of communists by conservatives) ; 'good� is� as typically used, an evaiuative term, and so are tright' and 'ought.3 These terms are primarily evalua2 Some ,vriters use the words 'evaluative' and 'vaiue-judgement' in a narro1-ver sense than this. They cal: judgements containing the word 'good' and some similar '-VOrds 'evaluative' or 'value­ judgernents', and distinguish these from judge­ ments cor.ta:ning the words 'right', \vrong',

WHAT ACTS ARE RIGHT? tive; words like ·industrious', 'honest', and 'courageous' are, as explained above, secon­ darily evaluative. All words which are eval­ uative ( whether primarily or secoudarily ) are also prescriptive; but there are expres­ sioas which are prescriptive but not eval­ uative ( because they do not carry descrip­ tive meaning as well) , The orCnary singu­ lar i:nperative-or rat:ler, to be strictly ac­ curate, its 'neustic' ( LM 2.1 )-is of this kind. Now the philosophers to whom I re­ ferred just now po:nt out ( rightly) that value-\:vord.s are like ordinary descriptive words in that they both have descriptive meanings, \'Vhich are� moreover, alterable and fle xible in both cases. But the purpose of using the term 'evaluative· is not to deny that value-v.rords have descriptive mean­ ing; that is readily admitted, and argu­ ments vlhich seek to prove that they have descriptive meaning are not arguments against my position1 which a1lo-::vs this. ·Nor are arguments designed to show that we can use the words 'true' and 'false' of value-judgements, or that ;,ve can speak of 'describing' somebody as a good man. We can .say these things of any judgement which has descriptive meaning, prnvided that it is its descriptive meaning that we are adverting to. Nor do I wish to deny that the descriptive meanings of value­ words are alterable and flexible; that this is so fits in very well ,vith :::ny thesis. I am not asserting that value-\vords are in this re­ spect different from descriptive words. VVhat I am asse1ting is that the character of what happens when the descriptive mean­ ing of a value-·word changes is profoundly affected by the fact it h2.s prescriptive meaning as well as descriptive. ·ought', and. the like, which they call �normative judgements'. These two classes certainly need to be distingu:shed for some purposes, as we shall see; and this is a useful way of doing it. But since I have used the word 'evaluative· in its wide sense hitherto, it wouid be confusing to use it :ll a diferent sense from now on; 1 shall therefore con­ tinue to use it to cover 'ought' and •rlghf as well as 'good'.



This can be clar.ified by means of a agreement to use the word 'wrong' to cover simple example. Let us suppose ( to use an certain cases and not others ( LM 3.5) . It example whieh is current) that two people follO\vs that the rules '-Vhich these hvo peo­ differ in where they draw the line between ple are using for determining the applica­ a 1msh' and a 'tree', It is possible to imag­ tion of the word 'wront cannot be merely ine situations ( for example if bushes are to descriptive meaning-rules, although they be cut down bl't trees left sta..-,ding) in do, among other functions, determine the which such a verbal difference might lead descriptive meaning of the term. They are to importr....'1t misunderst2ndings. But these rnles ha,1ng moral substance; in accepting misunderstandings could be cleared up one or the other of them the disputants guite easily by means of an agreement on would be committing themselves, not the· use of the word. In agreeing to draw �erely to a certain use of a \"",1ord1 but to a the line in a ce�ain place they would not matter of moral principle. So vvhen \Ve (flex' be settling anything except a question of our moral words, \Ve have regard, not :::oeanbg-a verbal question. V\!herever the merely to matters of mere convenience in line is drR\vn, the sarne instructions as be­ communication, but to substantial qt:es­ fore ce-'1 be unambiguously giv en: e.g. 'Cut tions 0£ morality, down all bushes below 15 ft. high with the lowest branch less than 3 ft. from the ground', So classifying something as a bush does not by itself entail a prnscription to 3. PRINCIPLES cut it down. I "1sh to contrast such a case of purely 3,1. I sought in the preceding chapter to verbal difference with a case of a moral explain in what sense :noral judgen1ents are difference, thereby showing that typical univcrsalizable. The explanation may he moral dis�;n:tes are not purely verbal, as on summed up as follows: they are universal­ a naturalist account they wou!d be, pro­ izable in just the sa111e ,vay as descriptive vided that the non-moral facts were judgements are universalizable;, namelv the agreed. Let 'US suppose that two people way which follows from the fact that\oth know all about the income-tax lav;rs, and moral expressions and descriptive expres� know, speci:G.cally, that a certain method of sions have descriptive meaning; bt:.t in the tax avoidance is perfectly legal; and let us case of moral judgements the universal suppose that they bow all about the pre­ rLLles v;rhich determine descriptive cise tax si!uation of somebody who is pro­ meaning are not mere meaning-rules, but posing to use this means of avoiding tax. moral principles of substance. In this chap­ One of them may say 'That would be ter� I am going to consider various other ,vrong; it would be going too there are ways in which moral judgements might be ways of avoiding tax that are morally per­ said to be universal or universalizable fect]y legitimate ( for example by claiming -mainly in order to a,;mid futn:re misinM deduction on account of a dependent rela­ terpretation by indieatlng to which of these tive, if you have one ); but this proposal views I su:Jscrlbe and to which I tlo not. got>,,s beyond ,vhat I can condone�. But the It is 1 first of all, most important to dis­ other may say, 'In my view this proposal tinguish the logical thesis which I have cannot be condemned on moral grounds; been putting forward from various moral thez-e are methods of tax avoidance which, theses '\Vith which it is easv to con:!use it. I though legal, I would condemn, but this is sai.d above (2,7) that, bec;use of universal­ not one of them; in mv view there is noth� izability, a person who makes a moral irtg wrong about it'. N�,:v it is obvious tl1at judge:nent commits himself, not mere;y to these two people cannot clear up their a meaning-rule ) but to a substantial moral difference, as in the 'btLsh' case,. by a verbal principle. The thesis of universalizability it-



self, however, is still a logical thesis. It is only the rule 'Live thus: . . . ' followed by a very important not to confuse the t.11.esis 0£ min;.::te descriptio:1, in universal terrns, of universalizability with the substantial mor­ how he has lived. To avoid this trivializatio.:i of the prin­ al principles to which, according to it, a person who 1nakes a moral judgement com­ ciple we are co:isidering, let us stipulate that a man is not to be said to have ad­ rnits himself. By a 'logical' thesis I mean a thesis hered lo a rule, nor to have gocerned his abollt the meanings of words, or dependent conduct in accordance with it, unless he has solely upon them. I have been maintaining in some sense had the rule before his mind tha: the meaniog of the word 'ough: and ( at any rate from time to time) and unless other moral words is such �hat a person his conduct has in sorne sense been moti­ who uses fhem corr:mits himself thereby to vated by the desire to conform to it. c is mediated, and hence qualified, by the generalization principle. That the generalization principle is in­ volved in the generalization argument is no doubt obscured by the fact that in applica­ tions of the argument the qualification "all similar persons in similar circumstances" is left inexplicit. But in valid applications of the argument this restriction is either im­ plicitly understood from the context or is indicated by various linguistic devices. For

532 example, the argument "everyone ought to vote because if no one voted the govern­ ment would collapse" is evidently meant to apply only to those legally permitted to vote. This eondition on the argument I shall call that of restricted universality, and I shall go on to discuss it presently. Before doing so, however, it will be useful to set for+Ji vtith some precision the various steps :involved in the generalization arglL"'Ti ent, in order to illustrate more clearly its logical structure. For the generalization principle is not the only principle involved in it. Let us consider, then 1 the anatomy of the gen­ eralization argument. Section .1. The argument involves, in tile first place, foe principle, "If the conse­ queilces of A's doing x woulti be disastrous, then A ought not to do x." The term "disas­ trous" is a stronger tenn than is actually necessary for the statement of this princi­ ple, as are such roughly synonymous terms as "terrible'� and "catastropl1:.c." It can be replaced by the somewhat weaker and more general term "undesirable."' The consequences of an act can be undesirable without being disastrous. But :f they arc disastrous then they are undesirab:e. Thus this principle, whic'h I shall call the princi­ ple of consequences, can he stated as fol­ lows: ( I ) If tile consequences of A's doing x would be undesirable, then A ought not to do x." This is, obviously, equivalent to "If the consequences of A's not coing x would be undesirable, then A ought to do x." It is not, ho\vever, equivalent to "If the conse­ quences of A's doing x 'INould be desirable, chen A ought to do x." I doubt very much wheaer the latter proposition ls true ( and 'Whether it is or not will be discussed in Chapter VII ). At any rate, it is no part of the generalization argument. The principle of consequences is a necessary ethical or moral principle. It is necessary not only in the sense that its de­ nial involves self-contradiction. It is neces­ sary also in the sense that like the gen­ eralization princi?le, it is a necessary pre­ supposition or precondition of mora] rea-

WHAT ACTS ARE RIGHT? saning. There can be sensible and fruitful disagreement about matters within the field deJL,,-,ited hy it, but there can b e no sensi­ ble or fruitful disagreement about the prin­ ciple itse1£. We might say that, like the ge:i­ eralization principle, it is both necessary and fundamental. I do not wish to imply that a::i.yone ever has seriously tjt:estioned or denied this principle. It may be that no has done so, at least explicitly; though there are probably rilany instances in which it has been denied by implication, just as there are unques­ tionably many cases in which it has been violated or dis:egarded. Yet the principle ean be misunderstood, especially if the term "undesirable" .is :not properly under­ stood. This tenn may be interpreted in eith­ er of two senses, with the consequence that there are two ways of interpreting the principle, Though t.liese two Yvays are con­ sistent wHh each other, they s!lould be kept distinct, One sense of "'undesirable'� is that of "undes:rable o,c, tile whole." On fois in­ terpretation, the principle does not mean that if some of the consequences of A's doing x would be undesirable then A ought not to do x. It is perfectly consistent with it for some of foe consequences of an act to be desirable and others to be unc!esirable, or for then1 to be undesirable in some re­ spects but not in others. And it may well be tl1at while some of the consequences of an act a:-e undesirable} it is not undesirable, on the whole, for the act to be done. For the desirable consequences may outweigh the undesirah:e ones. Or it may be that t.¾e consequeilces of A's not dolng x would be worse ( more undesirable) than the conse­ quences of his doing it. In the second sense of "undesirable" it does not have this proviso of "on the whole." On this interprecation, the fact t.'iat some of the consequences of A's doing x would be undesirable is a reasoa for assert� ing that A ought not to do x, but it is not a conclusive reason. On the basis of this fact one could reasonably presume that it



would be wrong for A to do x. 2 Tbis pre­ It is at the tbird step of the argument sumption can be rebutted by showing that that the generalization principle comes into not all the consequences are undesirable, play: ( 3 ) If not everyone ought to do x, and that the undesirable cons_equences are then no one ought to do x. Tbis can of outweighed by ( are less important than) course be stated in the alternative form: If the desirable ones; in other vvords, by it is wrong for everyone to do x, then it is showing that the consequences of A's doing wrong for anyone to do x. Note that I have x would not be undesirable on the whole. left unstated the necessary qualifications. All of these steps are actually tele­ Thus a more adequate, because less ellipti­ cal, statement of the principle, on this in­ scoped in the generalization argument i t ­ terpretation, would be: If the consequences self, wbich is obviously deducible from ( 2 ) of A's doing x would be undesirable, then and ( 3 ) ; If the consequences of every­ A ought not to do x without a reason or jus­ one's doing x would be undesirable, then tification. Such statements as "A ought to do no one ougbt to do x. x" are usually elliptical in this way. It may be useful to display in one These brief remarks should make it place, in slightly different language, tbis clear that this principle assumes a good deduction of the generalization argument deal less than might at first glance be sup­ from the generalization principle and the posed. It does not by itself determine the · principle of consequences. The principle of meaning of the term "undesirable," or ·what consequences ( C ) states that: If the conse­ is desirable or undesirable, or how the var­ quences of A's doing x would be undesira­ ious consequences of an action are to be ble, then A does not have tbe right to do x. weighed against each other in order to de­ The following principle ( GC ) is what I termine whether they are undesirable on cal] a generalization from C : If the conse­ the whole. Agreement on the principle is quences of everyone's doing x Would be quite consistent ,vith disagreement on undesirable, then not everyone has the these latter questions. Indeed, without right to do x. Now the generalization prin­ agreement on the principle, disagreement ciple ( GP ) may be stated as follows: If not on these other matters would have no everyone has the right to do x) then not point. anyone ( no one) has the right to do x. The Now this first step in the generaliza­ generalization argument ( if the conse­ tion argument is the basis for the second, quences of everyone's doing x would be wbich is a generalization from it: ( 2 ) If undesirable, then no one has the right to the consequences of everyone�s doing x do x ) clearly folJows from GP and GC. would be undesirable, then not everyone Some remarks on this deduction are ought to do x. 3 now in order. In the above generalization from the principle of consequences, ( GC), � I am using the e.\.1)ressions "A ought not to do "everyone" is treated collectively, not dis­ x," "It would be wrong for A to do x," and «A has tributively. The hypothesis "If the conse­ no right to do x," synonymously, and I should say quences of everyone's acting in a certain that this is in general conformity with their ordi­ nary use. Thus I am treating "A ought to do x," as way would be undesirable" differs from "If equivalent to "It \vould be ,vrong for A not to do

x," and "A has no right not to do x," Note that the

contradictory of "A ought to do x" is not ''A ought not to do x," which is rather its contrary ( for neither may hold) , but "A has the right not to do x" ( or "A need not do x" ) . 3 Since e.\. 1)ressions like "not everyone ought" and "no one ought" can be deceptive, perhaps it should be said here that I definitely do not mean by "not everyone ought" the same as "not every­ one is required," but rather "not everyone has the

right" or "it would not be right for everyone." Similarly, by "no one ought" I do not mean "no one is required, or has the duty," but rather "no one has the right," or "it would not be right for anyone." ( If one perfers to translate "not everyone ought to do x" by "it ought not to be the case that everyone does x," I can see no objection to it, except that it is not very idiomatic, and I cannot see that it is helpful. )


the consequences of each and every act of that kind would be undesirable." The latter implies that each and every act of that kind would be wrong. This is the true logical generalization of the principle of conse­ quences, but it is not the one intended, nor is it particularly important. Thus GC has as its consequent "not everyone ought to do x," instead of "every­ one ought not to do x," because supposedly if not everyone does x, the undesirable consequences that would result from every­ one's doing it would be avoided. Hence the generalization argument does not imply that the consequences of each and every act of the kind mentioned would be un­ desirable. By reason of the generalization principle it implies that each and every act of that kind may be presumed to be wrong. Yet from the fact that an act is wrong it does not follow that its conse­ quences would be undesirable. The generalization argument is to be distinguished from what may be called the generalized principle of consequences: If the consequences of doing x would be undesirable ( in general, or usually ), then it is wrong ( in general) to do x. Here "x" re­ fers, not to a specific action, but to a kind of action. The consequences of lying are ·usually undesirable; hence lying is usually wrong. The generalized principle of conse­ quences refers to the indiv-idual conse­ quences of actions of . a certain kind. The generalization argument refers to the col­ lective consequences of everyone's acting in a certain way. These are not always the same. From the fact that the generalization principle is involved in the generalization argument, in the way shown, it follows that all the qualifications required by the for­ mer are required by the latter. They are therefore necessary for any application of the argument to be valid. The first is that �f restricted universality, the restriction to every similar person in similar circumstances." The second is the elliptical nature of the conclusion that no one has the right


to do x. As I mentioned once before, the form of the generalization principle espe­ cially appropriate for the proper under­ standing of the generalization argument is: If not everyone ought to act or be treated in a certain way, then no one ought to act or be treated in that way without a reason or iustification. A more adequate statement of the generalization argument, therefore, is: If the consequences of everyone's acting or being treated in a certain way would be undesirable, then no one ought to act or be treated in that way without a reason. In other words, whoever acts in a way in which it would be undesirable for everyone to act must justify his conduct. The fact that it would be undesirable for everyone to act in that way provides a presumptive reason, and not a conclusive one, for the judgment that his conduct is wrong. One can justify oneself, or show that one is an exception, by showing that one's circum­ stances are relevantly different from those in which the act is wrong. But the discussion of the procedures by which one can justify his ( or someone else's ) acting in a way in which it would be undesirable for everyone to act, or in which it would be generally wrong to act, may be left for later on. \'Vhat I propose to do now is to consider in some\vhat greater detail the condition of restricted universal­ ity. . . . Section 3. When people begin to ad­ monish me that if everyone did as I did, etc., I ans·wer that «humanity would proba­ bly perish from cold if everyone produced food, and would certainly starve if everyone made clothes or built houses."4

1 Morris R. Cohen, The Faith of a Liberal ( New York: Henry Holt and Company, 1946), p. 86, Part of the sentence just prior to the one quoted in the text is: "It would be a poor world if there were no diversity of function to suit the diversity of natural aptitudes." This is true; it does not follow that we have here a valid counterexam­ ple to the generalization argument. Cf, Cohen's Reason and Nature ( New York: Harcourt, Brace and Company, 193 1 ) , p. 433: "Nor is there any force in the argument that lying is morally bad because it cannot be made universal. The familiar argument, 'If everybody did so and so . . .' applies



Thls certainly has the appearance of a Thls condition on the validity of the genuine counterexample to the generaliza­ generalization argument is not something tion argument. Since the consequences of ad hoc, devised just lo meet this kind of everyone's producing food would b e unde­ . case, though even if it were, this would be sirable, on the pattern of the generalization no objection to it. It is another of those argument it would seem to follow that it is conditions implicitly understood but not 1,v:ong for anyone to do so> and this, of explicitly stated, and can readily be focor­ course is absurd. porated into the statement of the argu­ But this actttally does not follow, and ment: �7hen I say "but not everyone", I imply that I feel that everyone ought not to try to straighten the picture together or all at the same time; and this is compatible both with the view that the picture ought to be straightened by someone, and with the view that everyone has a good reason for straightening it and thus ought in a sense to straighten it. As has been noted recently, "everyone" is complex: it has what has been called a collective use in which it is replaceable by something like "everyone together" or "everyone at once", and a distributive use in which it is re­ placeable by "anyone". 16 The "everyone" in "but not everyone, please!" is collec16

A. Phillips Griffiths, "The Generalisation Argument: A Reply to Mr. Braybrooke", Analysis, Vol. 23 ( Apdl, 1963 ) , pp, 113-115.


tive, and the "everyone", in the universal­ ization is distributive. Another (bad) reason for thinking that this "someone ought" is not universal­ izable, would consist in the supposed fact that in this use it entails that there is no one who ought to straighten the painting. Suppose that each of us finds heights dis­ agreeable, and that straightening the painting would require climbing a ladder. Each of us might agree that it would be de­ sirable for the painting to be straightened, even though not one of us felt that he ought to do the job. ''"\Vhy me?", each of us might ask, and our questions would go un­ answered. So it seem, that though some­ one ought to straighten it, since only one person is needed and any person vvill do, there is no one who ought to straighten it. But again an error is involved. The fact is that there is no one in particular who ought to straighten the painting, but this could be so because each of us ought to straighten it. In such a case lots might be dra\\rn, not to select someone to do a job that no one ought to do, but rather to select someone to do a job which no one wants to do and which everyone ought to do. So it seems that this third sense of "someone ought" may be universalizable, but the question remains whether it is uni­ versalizable. In the painting case I believe it is true both that someone ought to straighten the painting and that everyone ought to straighten it. Since it would be desirable for the painting to be straight­ ened, someone ought to straighten it; and since anyone can straighten it by himself ( assumed for this case ) , each and every person has a reason for straightening the painting; and hence there is a sense in which each and every person ought to straighten the painting. So in this case it is true both that someone ought to straighten the painting and that everyone ought to straighten it, but this is not to say that the former en­ tails the latter. In fact, there seems little


reason to think that it does. A crucial feature of this case is that anyone can straighten the painting by himself, and rather than say that "everyone ought to straighten it'' follows from "someone ought to straighten it'', perhaps we should say only that in this case "everyone ought" follows from "someone ought" coupled with the statement that anyone can remedy the situation by himself ( anyone can straighten this picture by himself) . It is still possi­ ble that "everyone ought" does not folJow from "someone ought'. by itself; a slightly different case enforces this possibility. It would take three of us to straighten a painting; if anyone were to try to shift it by himself, he would drop it. And each of us knows the others well enough to know that even if he were to try to straighten it, no one would help him. In this case one could argue that although it is true that someone ought to straighten the painting, there is no sense in ,vhich it is true that each of us ought to straighten it. 17 In fact, it could be argued that there is not one of us who ought to engage in straigh t ­ ening it sihce such an effort on the part of any one of us would be quixotic and under the circumstances disastrous. And this argument v10uld stand unless we were to assume that since the consequences of no one's taking steps to straighten the painting would be a crooked painting, each of us ought to take steps to straighten it ( though the painting fall in the process ) . 11

Perhaps careful speakers would say "some of us" rather than "someone", since more than one person is needed. If so, then the "someone ought to do x" under discussion does not possess feature ( b ) ; it is not entailed by "if no one were to do x the consequences would be undesirable" though it is entailed by this coupled with the stat�ment that these consequences can be avoided by the action of one man. Furthermore, if the "someone ought" under discussion is necessarily singular in the manner suggested, then it does possess feature ( a )-it is universalizable. However, I believe that this "someone ought to do x" is not necessarily "singular", It is, I think, often interchangeable with "x ought to be done": 11/hen we use "some­ one ought to do x" in the sense under discussion we say, I think, that x ought to be done by on; person or by some.



But once again this cannot be assJmed; it is what is to be shown. Even if the jpres­ ent use of "someone ought" is universal­ izable, that it is cannot be claimed i! the course of a defense of the generalization argument; this becomes clear when we con­ sider cases in which, though it would be undesirable in terms of consequences for no one to do x, no man can forestall :f:hese undesirable consequences by doing x by himself. And such cases are of course su­ premely relevant to a discussion of gener­ alization arguments. Undesirable conse­ quences that one man can avoid do not call for generalization arguments. It is the occasional necessity for joint action and co-operation that provides a use for the generalization argument, that makes for its conflicts with the principle of conse­ quences, and that renders its justification difficult. 18 What is needed for the present de­ fense of the generalization argument( is a sense of "someone ought to do x" · that, prior to an establishment of the generali­ zation argument, can be shown to combine features ( a ) and ( b ) . Three senses have been examined, none of which satisfies this condition. The first two clearly pos­ sess feature ( a ) , but before one could say that they possess feature ( b ) it would be necessary to show that the generaliz,ation 1e 1fr. Singer makes a crucial mistake when he says that the consequent of ( 6 ) takes the £�rm it does "because supposedly if not everyone does x the undesirable consequences that would (result from everyone's doing it would be avoided" (p. 67 ) . But the fact is that the undesirable Conse­ quences will only be avoided if enough persons do not do x. And I believe that Mr, Singer's use of the word "supposedly" is an indication that he was aware of this fact. He probably reasoned in this way: Since the undesirable consequences ,..ipl be avoided if enough persons do not do x, x pught not to be done by one or some persons, i.e., some­ one ought not to do x; and since someone ought not to do x, there is at least one person such that he ought not to do x. But there seems to be no sense of "someone ought" in Which both of these inferences are unproblematic; for example, if "someone" is used in my third sense only the first inference is evident, and if it is used in my first sense only the second inference is evident.


argument is valid. The third as I have said seems to possess feature ( b ) , but cannot be shown to possess feature ( a ) prior to the establishment of the generalization argu­ ment. Evidently, there are senses of "some­ one ought to do x" that clearly possess fea­ ture ( a ), and there is at least one sense that seems to possess feature ( b ) ; but it is likely that there is no sense which can be shown to possess both of these features prior to a demonstration that the generalization argument is valid. In other words, there seems to be no way in ,vhich "someone ought» can be under­ stood in ( 6 ) and ( 7 ) such that both ( 6 ) and ( 7) can be shown to be true prior to the establishment of ( 8 ) , though there is a sense of the phrase which seems to make ( 6 ) true ( my third sense ) , and there is a sense of the phrase which makes ( 7) true ( my first sense, as well as my second ) . Much o f the initial plausibility of the de­ fense of the generalization argument under discussion derives, I believe, from the pos­ sibility of equivocating on "Someone ought". 3.4 Mr. Singer has not shown that the principle of the generalization argument can be derived from the principle of con­ sequences coupled with the generaliza­ tion principle. Furthem10re, it seems that no attempt along the lines of his could suc­ ceed: ( 6 ) seems to be incurably unsuited to its task. And it must now be added that any defense of the principle of the general­ ization argument that uses the principle of consequences1 will impose restrictions on their interpretations and will raise special problems. The principle of consequences and the generalization argument can conflict; this ,vas shown in Section II. Thus they cannot both be conclusive principles. If both prin­ ciples are true, at least one of them must yield something less than completely un­ qualified ought-judgments. Otherwise, in cases of conflict, two judgments of the fol­ lowing forms would both be true:



that at least sometimes a person ought to do a thing even though if everyone like him and were to do this sort of thing the conse­ A ought not to do x. quences would be undesirable. Maybe the But this is impossible. \Vhen conclusive generalization argument ( or the principle of consequences ) will always prevail in and unqualified "oughts" are involved, cases of conflict, but pri6r to an indication A ought not to do x of the precise character of their qualifica­ tions one must expect that neither will pre­ entails, vail over the other in all cases of conflicts. It is not the case that A ought to do x. A second possibility is that though the principle of consequences is a conclusive Taken in the conclusive sense, "ought" and principle, the principle of the generaliza­ "ought nof' express logical contraries. tion argument is not. If this is the case, Both may be false (perhaps there is no the principle of consequences prevails in one thing that A ought to do ), but not both all conflict cases, and there ·will be occa­ can be true. And so the principle of conse­ sions ( more occasions than one should ex­ quences and the generalization argument pect under the first possibility) when a cannot both be conclusive ought-principles. person ought to do a thing even though it At least one must be a qualified ought-prin- . \VOuld be undesirable in terms of conse­ ciple, e.g., a presumptive principle of the quences for everyone like him to do this form, thing. A third possibility is that though the If p, then presumably A ought to do x. principle of the generalization argument or a prima facie principle of the form, is a conclusive principle, the principle of If p, then A ought to do x other things consequences is not. In this case, the some­ what paradoxical implications of the first being equal. two possibilities do not obtain. It should If there are reasons for thinking that A be noted, however, .that this alternative is ought to do x, it is still possible that A probably not open to one who wants to use ought not to do x. And if A has a reason the principle of consequences in a defense for doing x and thus ought to do x other of the generalization argument; it is hard things being equal, it is still possible that to see how a non-conclusive principle could A ought not to do x. But if there definitely play a role in the establishment of a con­ are conclusive reasons for doing x -then it clusive one. It should also be noted that, of cannot be the case that A ought not to do x. the three alternatives, the present one as­ So anyone who employs the principle signs the greatest importance to the gen­ of consequences in defense of the principle eralization argument and makes its justi­ of the generalization argument, in fact, fication most urgent. 19 anyone who endorses both principles, must 10 Of the three possible positions, Mr. Singer concede that at least one of them is not a selects the second. He holds that the principle of conclusive principle. There are three possi­ consequences can be interpreted in such a ,Vay bilities. Perhaps neither principle is a con­ that it is a conclusive principle ( p, 64 ) . And he clusive ought-principle. If so, then in the maintains that a completely adequate and explicit of the principle of the generalization conflict cases one must expect that some­ statement argument ends with the phrase "without a reason", times the principle of consequences will Since lvfr. Singer does not think that the two prin­ prevail and that sometimes the generaliza­ ciples can conflict, he is not struck by the rela­ tively minor role that he assigns to the generaliza­ tion argument will prevail. Thus, though it tion argument, and he is not aware of the be somewhat paradoxical, one must expect somewhat paradoxical aspects of his position.

I l


The relations between our two prin­ 1 ciples are apparently complex and many matters remain to be discussed. Bu� this much is clear: these principles cannotJboth be conclusive principles, and if ( though I think it unlikely) the principle of conse­ quences plays a role in the establishment of the principle of the generalization ar­ gument, then at least this latter principle is not a conclusive one, and one should ex­ pect that at least sometimes a person 1ught to do a thing even though it would by un­ desirable for everyone like him to dq this thing.

CONCLUSION The principle of consequences and the prin­ ciple of the generalization argumen� can conflict; this was shown in Section II. j And it is unlikely that the principle of conse­ quences can be used in a defense of the


generalization argument. Mr. Singer has not shown that it can be so used, and there are reasons for thinking that no attempts along the lines of his will prove more suc­ cessful. These points were argued in Sec­ tion III. But of course, even if the principle of consequences plays no role in the validation of the generalization argument, this argu­ ment may still be valid, There are other approaches to it that deserve exploration. Perhaps, for example, it should be under­ stood as an appeal to fairness. Nonetheless, though nothing established in this essay shows that the generalization argument is not valid, it is clear that its validity cannot simply be affirmed. Some demonstra­ tion or defense is rendered imperative by the fact that it can conflict with straight­ forward appeals to consequences. And it is of course possible that no good defense of the generalization argument is avail­ able: it may be just a time-honored fal­ lacy.


Rights, Justice, Punishment, and Responsibility A. Rights B. Justice · C. Punishment and Resp onsibility


A. Rights

On Natural Rights RALPH MASON BLAKE At various times in the history of moral and political philosophy the concept of nat­ ural rights has played an important and prominent role in the thoughts of men. It has frequently, indeed, been the cen­ tral and dominating idea of a whole sys­ tem. At other periods, however- and it is through one of these that we seem at pres­ ent to be passing- it has fallen out of favor. In many quarters it seems just now to be regarded as an outworn and ex­ ploded superstition of the past, and auy appeal to the idea is looked upon as evi- . dence of an antiquated aud unenlightened approach to the problems of the day. We may well ask ourselves, however, whether an idea of such vitality, appealed to at times, indeed, by the most diverse schools of thought as giving ,varrant to their views, and constantly reappearing in men's minds just when it seemed once more finally to have been got rid of, does not really embody some important notion which it would be useful to preserve and danger­ ous to Jose sight of. It is this question which I propose here to examine. It would be admitted, I suppose, that, speaking generally, "rights" are correla­ tive with "duties" and are defined by laws. Thus, if we start vvith the positive law, whether constitutional or statutory, we find that there are many positive laws which define ''rights" vested in certain in­ dividuals or groups of individuals ( one of Reprinted from The International Journal of Ethics ( later renamed Ethics) , 1925-26, pp. 8696, by Ralph Mason Blake by permission of The University of Chicago Press.

which groups may be society as a whole ) , and impose duties correlative with these rights upon other individuals or groups of individuals ( or again upon society as a whole ) . The rights so defined seem usually to be of the nature of claims which their possessors may legitimately ( i.e.i in accord­ ance with the law) make upon others, claims, therefore, which always imply some correlative duty on the part of these others. Thus if in accordance with the law a laborer, upon the performauce of labor, has a right to the wages for which he contracted, this means that he may legitimately claim the sum agreed upon, and that it is the legal duty of the em­ ployer to satisfy this claim by paying the wage. Or again, if a man has a legal right to the exclusive enjoyment of a certain piece of property, this means that he may lawfully claim immunity from any inter­ ference on the part of others with such exclusive enjoyment, and that it is the legal duty of these. others to refrain from such interference. It thus appears that posi­ tive laws may, on occasion, give rise to posi­ tive legal rights and duties; and such rights as thus arise seem to be of the na­ ture of certain liberties or freedoms which their possessors may legitimately exercise. The traditional conception of "nat­ ural" rights, whatever its other content, seems at least always to be that of a sys­ tem of "rights" having a deeper and more fundamental character than any merely legal rights, of a system of rights posses­ sing an ultimate and objective validity





not derived from their relation to any sys­ then, from nah1ral physical laws, but only tem of positive law. Natural riglits, as from natural moral can natural their name indicates, are supposed! to be rights be conceived to arise-from the prin­ based instead upon a law of natµre it­ ciples of natural morality, i£ one may be self, independent of any positive enact­ pen1'.itted to employ so unfashionable a ments of men, and of a higher validity. It term. is therefore considered possible to contrast And now it is necessary to point out merely legal rights with natural ' rights, a very significant difference between the and in case of conflict to assert the supe­ conception of a "law" in the sense of a rior claims of the latter. Just as positi\'e fundamental principle of natural morality, rights are derived from positive la'l"s and and a law in the sense of a positive law or . correlated vdth legal duties, so /lllhm:tl legal precept. The latter is rn essence a rights are conceived to be based µijon the command or imperative issued by some "natural law," prescribed by the ver1 na­ competent authority and backed by the ture of frlings, and to be correlate1d with force of some sanction, some penalty or certain natural duties. And, lik� legal reward to accrue from the authority in rights, natural rights appear usuall}'l in the question, whether this authority be state, form of claims, often to some freedom, church, public opi."1ion, or the customs of immurdty, or privilege 1 made leg'itimate societv. Natural moral laws, on the oL'ier by this '1aw of nature." hand,' are not necessarily to be thought of I But now this conception of a 'i law of as commands or precepts issued by any nature" is certainlv not entirely clear. It anthority whatever. If they chance to be I ments ' . ' Among vvb at b ad::ed by some authority, that is con­ ' further exammatmn. are commonly kno,vn as laws of nature are ceived �o be a wholly accidental circum� certain gene;a1izations of physical1 j cherni· stance. Nor are they necessarily enforced cal, or biological phenomena. These in fact by any sanctions whatever. A positive law are ,vhat we 110\vadays have chie:8,v in mind issued by no authority is a contradiction in when we speak of ;,atural law, �r of the terms. Laws of nature, on the other hand, la\VS of nature. But these laws, it is fairly are conceived to b e independent of auy c1ear1 are precisc1y not what are in �uestio� autho,ity. They remain what they are and when it is a :matter of determining rights. retain their entire validity ...vhether any Laws of nature, in this sense of the t�rm� do authmity commands them and enforces not tell us in the least what is right ir what them with penalties or not A law of na­ ought to be; they merely describe f for us, ture may be enacted into a positive law­ without reference to matters of value, how indeed many have held that in default of natural phenomena as a matter :of fact enactment by any human authority they actually do occur. They thus explicitly ab­ must be conceived to form _ part of the posi­ stract from the vvhole question .J as to tive law of God, and to be enforced by whether the natural occurrences ! which supematural sanctions. But the very con­ they describe embody any sort of right or ception of a law of n ature is that it does justice. not derive its being or its validity from any The natural laws from which jnatural enacbnent or positive command whatever, rights could be supposed to arise must but ,vould remain valid even if every posi­ therefore be of quite another sort. They tive law were to contravene it. It is not must not abstract from matters of value, only independent of positive laws, it is but must instead deal primarily therewith. also deeper and more fundamental than They must be statements of what claims they, and possessed of a higher validity. The conception of a Hnatural law'� various fadividuals or groups of individ­ uals ought to be privileged to make. Not, from which natural rights and natural


duties are derived is, then, the conception of a principle stating tl1at the claim to such and such liherties, freedoms, or privileges ought to be vested in such and such in­ dividuals or groups of individuals, a pr.nci­ ple which is to remain true and valid no matter what the positive law of any authority may decree, and by which any positive la\V which contra\1enes it may be condemned as contrary to natural right and justice-as constituting a violation of natural rights. And by calling these prin­ ciples '1aws of nature" it is implied that they are presc!"ibed and determined by th e very nature of things and are there­ fore independent of any choice or arbi­ trary decree, human or divine. Not only, however> are such natural laws conceived as prescribed by nature h e r ­ self, and a s therefore superior t o any ar­ bitrary decree or enforcing sanction; they are also thought of as CoQnstituting eter­ nal and immutable principles, as posses­ sed of no merely temporary validity, but as remaini::1g permanently valid-so long at least as the present order of nature shall continue in its general features to remaL.'1 substantially what it now is. It might pe,haps be admitted that other principles may be "laws of nature" for be­ ings other t.1-ian human, placed in a di!fer­ a,t general scheme of things; but such laws of nature as affect us human beings are conceived to be permanently valid truths for this our world, lasting un­ changed so long as it shalt last. Moreover, such laws of nature are conceived as being not only permanent in time, but as being also universally applicable in place. Inas­ much as they are determined by the gen­ eral nature of things, it is held that they must be valid for all men everyvvhere, whether as a matter of fact all men have ah:vays recognized or acted upon their validity or not. And finally these natural la\VS are he:d to be rational laws1 tl1at is to say, principles ,vhich right reason, re­ flecting upon the order of nature, must necessarily discover to be generaHy valid.


And now, if it be admitted that these are the features which form the core of the ideas of natural laws, and of natural rights and dutif'-5 hased upon such laws, as these ideas have commonly been conceived, we may well ask ourselves ,vheL.½.er there actuallv is or even conceivablv can be anv such thing as a '1aw of nai:ure" in sense of the term. In \:vhat sense, namely, can �'nature" or '"the nature of thing;" determine laws as to what claims ought to be vested in certain individuals or groups of individuals, and as to what duties ought in tum to be imposed upon others? And at first sight we may well doubt whether anything of the sort is at all possi­ ble, for when we turn to history we find the most diverse- and incompatible opinions, not only as to what these "natural rights" and these "laws of nature" as a matter of fact are, but also as to the very way in which "nature" is supposed to supply these principles or standards. The conception of "living according to nature/' according to natiiral right and justice as opposed to and deeper than merely legal, customary, or conventional standards, has played a cen­ tral part in many different philosophies; but the difficulty is that it has supplied so many and such diverse standards-many of them mutually incompatible. Thus Aris­ tot!e found the "natural" in the rational standard of the mean, and in all that dis­ tinguishes man as a rational animal from the brutes; th e Epicureans, on the other hand, found the "natural" precisely in that which men share with the brutes­ the tendency to seek pleasurable and to avo:d painful experiences. Tims Hobbes found the natm1ll man to be actuated onlv by the egoistic \Vill to povver, whilst t� Rousseau he appeared as a model of inno­ cent peaceableness and sympathetic affec­ tion. It seems impossible to bring any real order out of such a e,.'i.aos, and the attempt to find any standard of "what ought to be" fron1 a contemplation of "nature" may , thus well appear to be who"'y futile. And in fact, as this attempt has usu-




ally been conceived, it really is futilJ. I do not believe it to b e possible to deri�'e any principles with regard to what oughito be, a_ny principles of «natural moraliiy'' or "laws of naturet from a contemR1atiOn7 no matter how earnest, disinterested and thoroughgoing, of the facts of nature, as these are reported to us by the ordinary descriptive sciences of nature. There is no road to the derivation of such principles from the consideration, taken merely in themselves1 of the truths of physics . or as­ tronomy, or even of those of biology, psy­ chology, or history. The reason for this im­ possibility seems to be that from "riature" as it is studied in any of these s�iences all matters of VIUUe have from the be­ ginning been carefully excluded. The phys­ icist, the astronomer, the biologist and the psychologist simply abstract from the value aspects of the phenomena which they study. They do not even ask the question­ and a fortiori do not succeed in ansv;;ering it-\vhether this or that phenomeno:;:1 with which they are dealing is good or bad, right or wrong, just or unjust-all such matters are left entirely to one side. And since a consideration of values is no part of the task of these sciences, no prin­ ciples concerning values can possibly be derived from or determined by their re· suits. The "nature" which they study is the natural order deliberately con�idered apart from any of the values which it may embody, S11d natrue in this abstract! sense can determine no standards. Wben ,�e con­ sider nature apart from value we h�ve al­ ready deprived ourselves of any ba�is for rationally preferring one part of natpre or one tendency manifested in nah)re to any other. vVe cannot deduce, solely� from the fact that physical or psychological laws are what they are, any principles r as to what ought to be. But let us now note that if values are left out of fae "nature' studied bv these sciences, it by no means follo,�s that values are not also themselves a part of the natural order, if that order be only


considered in a less partial and abstract fashion, nor that this value aspect of the natural order may not itself legitimately be made the object of a rational consider­ ation. Reflexion upon human experience in the world discerns certain phases of it as "good" and certain others as "bad." ·we find thus as a matter of fact that the world of ''nature"" contains distinctions of value as an actual part of itself; and from an im­ partial consideration of these value aspects of the natural order it is not so entirely chimerical to hope or to expect that some principles or standards may be derived. If we ask what are foe primary or fundamental «values" of human experience, as discerned by such a consideration, I think that we can give a reply which will represent a very considerable measure of agreement on the part of contemporary students of th e subject. It seems, namely, to be vcrv , that in' widelv' held nowadavs trinsic positive value attaches only to those conscious experiences which can be described as e,xperiences of "happiness" or of •tsatisfaction," and that negative value) on the other hand, attaches only to conscious experiences of "unhappiness,·� "misery," or «dissatisfaction." The exact determinatio,i of the mea-'1ing of these phrases is of course still to a considerable extent a matter of dispute. The:re a:re, on the whole, two main schools of interpreta­ tion. According to the one, happiness and unhappiness consist respectively in the psy­ chological states of feeling called pleasure and displeasure, whereas, according to the other they consist rather in what is ca!led ''fulfilment of interest or desire"-a matter which often goes under the name ( implying also certain modifications of the vie'"v) of "'self-realization.', But in spite of the theoretical difference here indicated there appears to be such a very close con­ nection in practical human experience be­ tween experiences of pleasure and experi­ ences of fulfilment of ir1terest, that the two theories seem bound in the end to come out practically at very much the same point.


In any case, if we admit that this ,videly accepted vie"V-''", for instance, in one or the other of its divergent interpreta­ tions is correct, we are supplied at once with pr'.r,ciples of value, And such prin­ ciples, if t.'1ey or anything Hke them be true, are not merely "'ethical standards" imposed on the natural order from without, In fact they are genuinely natural stand­ ards, It seems to me, for example, that the proposition ''misery is in itself an evil/' or the proposition "happiness is in itself a good," is as much a truth aboLtt the natural order of things as is the proposition "fire bums," or "grass is green." Kow from such principles of value we can de­ rive principles of conduct. vVe can say, for insta.-ice, that what any person or group of persons "ought" to do is simply whatever will produce a greater balance of happi­ ness over misery than any other alterna­ tive that is before him; and conversely, that what he "ought not" to do, or what it would be "wrong" for him to do, is just "vhatever would p:::oduce a lesser balance of positive over negative values-having in view, of course, the effects of his ac­ tions upon the experience of all conscious beings '>vhatsoever. Doubtless al! these conceptio:>S lead to great difficulties when we try to apply them to the details of prac­ tical conduct; but I believe that we can and do guide ourselves more or less effectively by reference to such principles, and at any rate the principles themselves seem clear enough. Now supposing p,faciples of conduct to be determined in t.;is way, it seems to me that they would rightly be called "natural principlest for they would be detern1ined by ''natural" sta._"ldards-by standards1 that is, which have a validity of their o-wn, :inde­ pendent of the arbitrary command or pro­ hibition of any authority whatever. They would be ":::1atural laws» or the laws of "natura:! mornl:ty/' and '"natura] rights" would be simply those "rights," those free­ doms or privileges, wr'..ich any individual "ought," i n the sense defined, to be allowed


to claim, i.e., in the sense that to allow these claims is a condition -of the realiza­ tion of the natural standard of, say, the in­ crease of happiness and the decrease of misery or unhappiness. They would be claims which would be valid because they would rest upon the validlty of a "nat­ ural law," because they would be pre­ scribed by the very "nature of things," In accordance \Vith our previous dis­ cussion of the meaning of "natural rights" \Ve must add, however, that "natural laws/' so far as they are to define natural rights, must be principles of permanent, and ::mt of merely transit01) vnlidity, and further­ more that they must be such as are valid for all men everywhere. In ot.'ier words, the truly natural rights must be those claims, liberties, and prh1leges the possession of which by the person or persons in question \vill continue, so long at least as human na­ ture and the laws of the physical universe remain substantially what they now are, to constitute permanent and general condi­ tions of human happiness. But now, finally, are there any such natural rights, and, if so, what in particular are they? I believe tliat it is reasonable to hold that there are in this sense .natural rights. We can scarcely doubt that t'iere are certain general features of the physical order, of the psychological nature of man, and of the association of men in social and political aggregations, which have been, and 1vill so far as we can see for a long time continue to be, permanent forces to be reckoned \Vith in the human pursuit of happiness. And it seems reasonable to hold, these factors being what t.'½ey are, and tl1e values of human experience being also what they are, that so long as all these fac­ tors of the situation remain the same there \Vill be certain permanent and general con· ditions for the attaimnent of values, and faat some of these conditions will take the form of the allohnent to certain individuals or groups of i.:1d-ivitluais of certain claims, pri1.,ileges, and freedoms \vhich all others shall be bound to respect. I shall not, how1:


ever, attempt to say what in particul9;r any such natural rights may be. In fact, i I re­ gard the determination of natural Hghts rather as a problem awaiting solutioJ than as anything as yet at all fully or satisfacto­ rily determined. But I do at least regard it as a legitimate problem, and as one which there is some fair hope of solving. Indeed, I suppose that sociologists and political scientists, however unwilling they may be to state the problem in the terms which we have here employed, are already, at least in some measure, approximating to its; solu­ tion. What has brought the notion of n�tural rights into general, and as it seems to me undeserved, contempt has1 I think, : been the almost universal assumption on th� part of those who have believed in them that the "natural rights of man" and the j "laws of nature" on which they are basecl are readily discoverable either by a simpl� con­ sideration of traditional commandJ and prohibitions, or by some short and: easy




method of insight or immediate intuition. Rather, their determination must be the re­ sult of long and careful investigation and experiment. And the defenders of natural rights have also made other serious errors. They have usually conceived of natural rights too abstractly, as if each stood quite on its own basis, independent of and indif­ ferent to every other. On the contrary, nat­ ural rights must form a system of carefully interrelated and mutually adjusted rights and duties. And finally, natural rights have too often been thought of as vested in cer­ tain individuals apart from all reference to society and the interests of society. The very definition of a "right" should be suf­ ficient to dispose of such a notion. A right is a claim which ought to be allo·vved to an individual in view of the general welfare. Allowed by whom? VVe can only answer, "By society." A claim upon what? Upon the forbearance and support of others. Society is implied at every tum.

Rights SIR DA VID ROSS A general discussion of right or duty would hardly be complete without some discus­ sion, even if only a brief one, of the closely related subject of rights. It is commonly said that rights and duties are correlative, and it is worth while to inquire whether and, if at all, in what sense this is true. The statement may stand for any one, or any combination, of the following logically independent statements: ( 1) A right of A against B implies a duty of B to A. (2) A duty of B to A implies a right of A against B. ( 3) A right of A against B implies a duty of A to B. ( 4) A duty of A to B implies a right of A against B. vVhat is asserted in ( 1) is that A's hav­ ing a right to have a certain individual act done to him by B implies a duty for B to do that act to A; ( 2) asserts the converse implication; what is meant by ( 3 ) is that A's having a right to have a certain act done to him by B implies a duty for A to do another act to B, which act may be ei­ ther a similar act ( as where the right of having the truth told to one implies the duty of telling the truth ) or a different sort of act ( as where the right to obedience im­ plies the duty of governing well ) ; ( 4 ) as­ serts the converse implication. Of these four propositions the first apFrom The Right and the Good by Sir David Ross, pp. 197-203, by permission of the Clarendon Press, Oxford.

pears to he unquestionably true; a right in one being against another is a right to treat or be treated by that other in a certain way, and this plainly implies a duty for the other to behave in a certain way. But there is a certain consideration which throws doubt on the other three propositions. This arises from the fact that we have duties to animals and to infants. The latter case is complicated by the fact that infants, while they are not ( so we commonly believe) ac­ tual moral agents, are potential moral agents, so that the duty of parents, for in­ stance, to support them may be said to be counterbalanced by a duty which is not in­ cumbent on the infants at the time but will be incumbent on them later, to obey and care for their parents. YVe had better there­ fore take the less complicated case of ani­ mals, which we commonly suppose not to be even potential moral agents. It may of course be denied that we have duties to animals. The view held by some writers is that we have duties concern­ ing animals but not to them, the theory be­ ing that we have a duty to behave humane­ ly to our fellow men, and that we should behave humanely to animals simply for fear of creating a disposition in ourselves which will make us tend to be cruel to our fellow men. Professor D. G. Ritchie, for instance,

implies that we have not a duty to animals except in a sense like that in vvhich the owner of an historic house may be said to have a duty to the house.' Now the latter sense is, I suppose, purely metaphorical. 1

Natural Rights, 108.




of a We may in a fanciful mood thi noble house as if it were a conscious being � having feelings which we are bound to re­ spect. But we do not really think tha ·t has them. I suppose that the duty of the owner of an historic house is essentially a duty to his contemporaries and to posterity; and he may also think it is a duty to his ancestors. On the other hand, if we think we ought to behave in a certain way to animals, it is our of consideration primarily for their feelings that we think we ought to behave so; we do not think of them merely as a practising-ground for virtue. It is because we think their pain a bad thing that we think we should not gratuitously cause it. And I suppose that to say we have a duty to so-and-so is the same thing as to say that we have a duty, grounded on facts relating to them, to behave in a certain way towards them. Now if we have a duty to animals, and they have not a duty to us ( which seems clear, since they are not moral agents ), the first and last of our four propositions can­ not both be true, since ( 4 ) implies that a duty of men to animals involves a right of men against animals, and ( 1 ) implies that this involves a duty of animals to men, and therefore ( 4 ) and ( 1 ) together imply that a duty of men to animals involves a duty of animals to men. And since the first prle defines how this presumption may be re­ butted. It might be argued at tbls point that justice requires only an equal liberty. If, however, a greater liberty were possible for all without loss or conflict, then it would be irrational to settle on a lesser liberty. There i:s, of cm..:.rse, \Vi the element of necessity does not render the conception of mun:_a) acknowledgment in­ applicable, although it may make it much more urgent to change unjust than unfair institutions, For one activity in which one can always engage ls that of proposing and acknowledging principles to one another supposing each to be similarly circum­ stanced; and to judge practices by the principles so arrived at ls to apply the standard of fairness to them. Now if the participants in a practice accept its rules as fair, and so have no com­ plaint to lodge against it, there arises a prima facie duty ( and a corresponding prima facie right) of the parties to each other to act in accordance with L'le practice when it falls upon them to comply. When any number of persons engage in a prac­ tice, or conduct a joint undertaking accord­ ing to . rules, and thus restrict their liberty, those who have submitted to these restric­ tions when required have the right to a similar acquiescence on the part of those who have benefited hv their submission. These conditions \Vlll obtain :if a practice is correctly acknowledged to be fair, for in this case a!'. who participate in it will benefit from it. The rights and duties so arising are special rights and duties in that they depenc on previous actions voluntar­ ily undertaken, in this case on the parties having engaged in a COffiL"TIOn practice and J..-nowingly accepted its benefits. 12 It is not, however, an obligation which presupposes a deliberate perforn1ative act in the sense of a promise, or contract1 and the like, 13 a For the deSnition of this prim.a facie duty, and the idea that it is a special duty, 1 am indebted to H. L, A. Hart See his paper ''Are There Any Natural R!ghts?," Philosophical Reciew, LXIV (1955), 185 f. H The se:ise of "performative" here is to be dert\•ed fr.r.n J. L, Austin's paper in the sympo­ sium. "Other ;\Hnds," Proceedings of the Aristotel­ ian Society, Supplementary Volume (194.6), pp. 170-174.

587 An unfortunate mistake of proponents of the idea of the social contract was to sup­ pose that politioal obligation does require some such act, or at least to use language which suggests it. It is sufficient t.1-i at one has kno\\ingly participated in and accepted the benefits of a practice acknowledged to be fair. This prima facie obligation may, of course, be overridden: it may happen, when it comes one's turn to follow a rule, that other considerations will justify not doing so. But one cannot, in general, be re­ leased from this obligation by denying the justice of the practice only when it falls on one to obey. If a person rejects a practice, he should, so far as possible, declare his in­ tention in advance, and avoid participating in it or enjoying its benefits. This duty I have called that of fair play, but it should be admitted that to refec to it in this way is, perhaps, to extend the ordinar,; notion of fairness. Usually acting unfairly is not so much the breaking of any particular rule, even if the frifraction · is difficult to detect ( cheating) , but taking advantage of loop-holes or ambiguities in r,.lies1 availing oneself of unexpected or special circumstances \vhich make it impos­ sible to enforce them, insisting that rules be enforced to one's advantage when they should be suspended, and more generally, acting contrary to the intention of a prac­ tice. It is for this reason that one speaks of the sense of fair play: acting fairly requires more than simply being able to fo11ow rules; what is fair must often be felt, or perceived� one \vants to say. It is not, how­ ever: an unnatl.ll'al extension of the duty of fair play to have it include the obligation v,rhich participants who have knmvingly ac­ cepted the benefits of their common prac" tice owe to each other to act in accordance with it when their perfomiance fans due; for it is usuallv considered unfair if some­ one accepts tJ{e benefits of a practice but refuses !o do his part in maintaining it. Thus one might say of the tax-dodger that be violates the duty of fair play: he accepts tlie benefits of• govenunent but will not do

588 his part in releasing resources to it; and members of labor unions often say that fel­ low workers who refuse to join are being unfair: they refer to them as •free riders," as persons who enjoy what are the sup­ posed benefits of unionism, higher ,vages, shorter hours, job security, and tbe like, but who refuse to share in its burdens in the form of paying dues, and so on. The duty of fair play stands beside other prima facie duties such as fidelity and gratitude as a basic moral notion; yet it is not to be confused with them. 14 These duties are all clearly distinct, as would be obvious from their definition�. As with any moral duty, that of fair play implies a con­ straint on self-interest in particular cases; on occasion it enjoins conduct which a ra­ tional egoist strictly defined would not de­ cide upon. So while justice does not require of anyone that he sac:iflce his interests in tbat general position and procedure whereby the principles of justice are pro­ posed and acknowledged, it may happen that in particular situations ) arising in the context of engaging in a practice, the duty of fair play will often cross his interests in the sense that he will be required to forego particular advantages which tbe peculiari­ ties of his circums:ances might permit him to take. There is, of course, nothing surpris­ ing in this. It is simply the consequence of the firm commitment which the parties ::nay be supposed to have made, or which they would make, :in the general position, u This, however� commonly happens, Hobbes, for example, when invoking the notion of a "tacit covenant,,.,. appeals not to the uatura':. law that promises should be "kept but to his fourth law of nature, Gat of gra.titude. On Hobbes's shift from

fidelity to gratitude, see VVarreJ1der, op. cit., pp. 51-52, 2.33-237. "'0/hile it is not a serious criticism

of Hobbes, it would have improved his argument had he appealed to the duty of fair play, On his premises he is perfectly entitled to do so. Similarly Sidgwick thought that a prbcip1e of justice, such as every man ought to receive adequate requital for bis labor, is like gratitude universalized. See Methods of Ethics, Bk. III, ch, v, Sec. 5. 'There is a gap fa the stock of moral concepts used by phi­ :Oso_phers into which the concept of the duty of fair -play fits quite naturally.

JUSTICE together 'Vlith the fact that they have par­ ticipated in and accepted tbe benefits of a practice which they regard as fair. Now the acknowledgment of this con­ straint in particular cases, which is man!­ fested in acting fairly or wishing to make a::nends, feeling ashan1ed, and the like, when one has evaded it, is one of the fonns of conduct by whicli participants in a com­ mon practice exhibit their recognition of each other as persons with similar interests and capacities. In the same ,vay that, fail­ ing a special explanation, the critelion for the recognition of suffering is helpfog one who sufrers, acknowledging the duty of fair play is a necessary pait of the criterion for recognizing another as a person with simi­ lar interest.s and feelings as oneself. 15 A person vvho never under any circumstances showed a wish to help others in pain would shovv, at the same time� that he did not rec ognize that they were in pain; nor could he have any feelings of affection or friend­ ship for anyone; for having these feelings implies, failing special circumstances1 that he comes to their aid when they are suf­ fering. Recognition that another is a per­ son in pain shnws itself in sympathetic ac­ tion; this primitive natural response of compassion is one of those responses upon t,,;,hich the various forms of moral condt.ct are built. Simllarly, the acceptance of the duty of fair play by participants in a common practice is a reflection in each person of the recognition of the aspirations a."Tld interests H> I am using the concept of criterion here in what I take to be Wittgenstein's sense, See Philo­ sophical In.oostfgatio-ns ( Oxford, 1953 ) ; and Nonna.n Malcolm's review, "\Vittgenstei.··t's Philo­

sophical ln..vestiga:t-ions;' Philosophical llimie--..v, LXIII ( 1954), 543-547. That the response of compassion, under appropriate circumstances, is

part of the criterion for whether or not a person understands what ''pain" means1 is, I think, in the PhUosophicaJ. Investigations. The view in the text is simply �Jl extension of this idea. I cannot, how­ ever, atte:r:r:pt to justify it here. Simi1ar thoughts are to be found, I thh1k, in £\Jax Scheler1 The Nature of Sympathy, tr. by Peter Heath (New Haven, 1954). His ,vay of ½'l'iting is often so obscure that I cannot be certain.



of the others to be realized by their joint upon rational and mutually self-interested activity. Failing a special explanation, their parties who are related and situated in a acceptance of it is a necessary part of the special way. A practice .is just if it is in ac­ criterion for their recognizing one another cordance with the principles which all who as persons with similar interests and capac­ participate in it might reasonably be ex­ ities, as the conception of t..liei.:r relations in pected to propose or to acknowledge be­ the general position supposes t11em to be. fore one another when they are similarly Otherwise they would show no recognition circumstanced and required to make a firm · ithout knovvl­ of one another as persons ,vith similar ca­ cornmitment in advance w pacities and _interests, and indeed, in some edge of what will be their peculiar condi­ cases perhaps hypothetical, they would not tion1 and thus when it meets standards recognize one another as persons at all, but which the parties could accept as fair as complicated objects involved in a com­ should occasion arise for them to debate its plicated activity. To recognize another as a merits. Regarding the participants them­ person one must respond to him and act to­ selves, once persons knm.vingly engage in a ,vards hi.'*11 in certain ways; and these ways practice which they acknowledge to he fair intimately com1ected \",rith the various and accept the benefits of doing so, they prima facie duties. Acknowledging these are bound by the duty of fair play to fol­ duties in some degree, and so having the low the niles when it comes their turn to elements of morality, is not a matter of do so, and this implies a limitation on tbeir choice, or of intuiting n1oral qualities, or a pursuit of self-interest in particular cases, Kow one consequence of this concep­ matter of the expression of feelings or atti­ tudes ( the three interpretations between tion is tliat1 where it applies, there is no which philosophieal oplnion frequently os­ moral value in the satisfaction of a claim cillates ) ; it is simply the possession of one incompatible \\1th it Such a claim violates of t..1.e forms of conduct in which the recog­ the conditions of reciprocity and commu­ nity amongst persons, and he who presses nition of others as persons is manifested. These remarks are unhappily obscure. it, not being willing to acknowledge it Their main purpose here;, howeverl is to when prnssed by another, has no grounds forestall, together with the re::narks in Sec­ for complaint when it is denied; whereas tion 41 the misinterpretation tha½ on the he against '\Vhom it is pressed can com­ vie-w presented1 the acceptance of justice plain. As it cannot be mutoally acknowl­ and the acknowledgment of the duty of edged it is a resort to coercion; granting fair play depends in every day life solely the claim is possible only if the party can on there being a de facto balance of forces compel acceptance of what the other will between the parces. It would indeed be not admit But it m2..l.:es no sense to con­ foolish to underestimate the importance of cede claims the denial of which cannot be such a balance in securing justice; but it is complained of fo preference to claims tbe not the only basis thereof. The recognition denial of which can be objected to. Thus in of one another as persons ·with similar inter­ deciding on the justice of a practice it is ests and capacities engaged in a common not enough to ascertain that it ansvvers to practice must, failing a special explanation, wants and interests in the fullest and most show itself in the acceptance of the princi­ effectiv� manner. For if any of these ples of justice and the acknowledgment of conflict with justice, they should oot be counted, as their satisfaction is no reason at the duty of fair play. The conception at which we have ar­ all for having a practice. It would be irrele­ :ived, then, is that the principles of justice vant to say1 even if true, that it resulted .in may be thought of as arising once the con­ the greatest satisfaction of desire, In tal­ straints of havlng a morality are imposed lying up the merit-; of a practice one must

toss out the satisfaction of interests the claims of which are incompatihle with the principles of justice. 6. The discussion so far has been ex­ cessively abstract. While this is perhaps un­ avoidable, I should now like to bring out some of the features of the conception of justice as fairness by comparing it with the conception of justice in classical utilitarian­ ism as represented by Bentham and Sidg­ wick, and its counterpart in welfare eco­ nomics. T�is conception assimilates justice to benevolence and the latter in turn to the most efficient design of institutions to pro­ mote the general welfare. Justice is a kind of efficiency. 1 6 Nov,r it is said occasionally that this form of utilitarianism puts no resITictions on what might be a just assigmnent of rights and duties in that there might be cir­ cumstances which, on utilitarian grounds, would justify institutions highly offensive to our ordinary sense of justice. But the classical utilitarian conception is not totally 590

YVhile this assimilation is implicit in Ben­ tham's and Sidgwick's moral theory, eA-plicit state­ ments of it as applied to justice are relatively rare. One clear instance in The Principles of Morals and Legislation occurs in ch, x, footnote 2 to section XL: ". . . justice, in the only sense in which it has a meaning, is an imaginary personage, feigned for the convenience of discourse, whose dictates are the dictates of utility, applied to certain particular cases, Justice, then, is nothing more than an imagi­ nary instrument, employed to forward on certain occasions, and by- certain means, the purposes of benevolence, The dictates of justice are nothing more than a part of the dictates of benevolence, which, on certain occasions, are applied to certain subjects . . . ." Likewise in The Limits of Jurispru­ dence Refined, ed. by C, \V. EVerett ( New York, 1945 ) , pp. 117 f., Bentham criticizes Grotius for denying that justice derives from utility; and in The Theory of Legislation, ed. by C. K. Ogden ( London, 1931 ) , p. 3, he says that he uses the words "just" and "unjust'' along with other words "simply as collective terms including the ideas of certain pains or pleasures." That Sidgwick's con­ ception of justice is similar to Bentham's is admi t ­ tedly not evident from his discussion of justice in Book III, ch. v. of ?,,fethods of Ethics. But it fol­ lows, I think, from the moral theory he accepts. Hence C. D. Broad's criticisms of Sidgwick in the matter of distributh'e justice in Five Types of Eth­ ical Theory ( London, 1930 ) , pp. 249-253, do not rest on a misinterpretation. is1


unprepared for this objection. Beginning with the notion that the general happiness can be represented by a social utility func­ tion consisting of a sum of individual util­ ity functions with identical weights ( this being the meaning of the maxim that each counts for one and no more than one ) , 1 7 it is commonly assumed that the utility func­ tions of individuals are similar in all essen­ tial respects. Differences between individu­ als are ascribed to accidents of education and upbringing, and they should not be taken into account. This assumption, cou­ pled with that of diminishing marginal util­ ity, results in a prima facie case for equal­ ity, e.g., of equality in the distribution of income during any given period of time, laying aside indirect effects on the future. But even if utilitarianism is interpreted as having such restrictions built into the util­ ity function, and even it it is supposed that these restrictions have in practice much the same result as the application of the principles of justice ( and appear, perhaps, to be ways of expressing these principles in the language of mathematics and psychol­ ogy), the fundamental idea is very dif­ ferent from the conception of justice as fairness. For one thing, that the principles of justice should be accepted is interpreted as the contingent result of a higher order administrative decision. The form of this decision is regarded as being similar to that of an entrepreneur deciding how much to produce of this or that commodity in view of its marginal revenue, or to that of some­ one distributing goods to needy persons a c 1-1 This maxim is attributed to Bentham by J. S. 1Jill in Utilitarianism, ch, v, paragraph 36. I have not found it in's writings, nor seen such a reference. Similarly James Bonar, Philnso­ phy and Political Economy ( London, 1893 ) , p. 234 n. But it accords perfectly with Bentham's ideas. See the hitherto unpublished manuscript in David Baumgardt, Bentham a,nd the Ethics of Today ( Princeton, 1952) 1 Appendix IV. For exam­ ple, "the total value of the stock of pleasure belonging to the whole conununity is to be obtained by multiplying the number expressing the value of it as respecting any one person, by the number expressing the multitude of such individu­ als" (p. 556 ) .


cording to the relative urgency of their wants. The choice between practices is thought of as being made on the basis of the allocation of benefits and burdens to in­ dividuals ( these being measured by the present capitalized value of their utility over the full period of the practice's exist­ ence ) , which results from the distribution of rights and duties established by a prac­ tice. Moreover, the individuals receiving these benefits are not conceived as being related in any way: they represent so many different directions in which limited re­ sources may be allocated. The value of as­ signing resources to one direction rather than another depends solely on the prefer­ ences and interests of individuals as indi­ viduals. The satisfaction of desire has its value irrespective of the moral relations be­ hveen persons, say as members of a joint undertaking, and of the claims which, in the name of these interests, they are pre­ pared to make on one another; 1 8 and it is this value which is to be taken into account by the (ideal) legislator who is conceived as adjusting the rules of the system from the center so as to maximize the value of the social utility function. It is thought that the principles of jus­ tice will not be violated by a legal system so conceived provided these executive de18 An idea essential to the classical utilitarian conception of justice. Bentham is firm in his state­ ment of ·it: "It is only upon that principle [the principle of asceticism] , and not from the princi­ ple of utility, that the most abominable pleasure which the vilest of malefactors ever reaped from his crime would be reprobated, if it stood alone. The case is, that it never does stand alone; but is necessarily follo\-ved by such a quantity of pain ( or, what comes to the same thing, such a chance for a certain quantity of pain) that the pleasure in comparison of it, is as nothing: and this is true and sole, but perfectly sufficient, reason for making it a ground for punishment" ( The Princi­ ples of Mora� and Legislation, ch. rr, sec. iv. See also ch. x, sec. x, footnote 1 ) . The same point is made in The Limits of Jurisprudence Defined, pp. 115 f. Although much recent welfare economics, as found in such important works as I. M. D. Little, A Critique of tVelfare Economics, 2nd ed. ( Oxford, 1957) and X. J. Arrow, Social Choice and Individual Values (New York, 1951 ) , dis-


c1510ns are correctly made. In this fact the principles of justice are said to have their derivation and explanation; they simply ex­ press the most important general features of social institutions in which the adminis­ trative problem is solved in the best vrny. These principles have, indeed, a special ur­ gency because, given the facts of human nature, so much depends on them; and this explains the peculiar quality of the moral feelings associated ,vith justice. 1 9 This as­ similation of justice to a higher order execu­ tive decision, certainly a striking concep­ tion1 is central to classical utilitarianism; and it also brings out its profound individ­ ualism, in one sense of this ambiguous word, It regards persons as so many sep­ arate directions in ,vhich benefits and bur­ dens may be assigned; and the value of the satisfaction or dissatisfaction of desire is not thought to depend in any way on the moral relations in which individuals stand, or on the kinds of claims which they are willing, in the pursuit of their interests, to press on each other. 7. Many social decisions are, of course, of an administrative nature. Certainly this is so when it is a matter of social utility in what one may call its ordinary sense: that is, when it is a question of the efficient de­ sign of social institutions for the use of common means to achieve common ends. penses with the idea of cardinal utility, and uses instead the theory of ordinal utility as stated by J, R. Hicks, Value and Capital, 2nd ed. ( Oxford, 1946) , Pt, I, it assumes ,vith utilitarianism that individual preferences have value as such, and so accepts the idea being criticized here. I hasten to add, however, that this is no objection to it as a means of analyzing economic policy, and for that purpose it may, indeed, be a necessary simplifying assumption. Nevertheless it is an assumption which cannot be made in so far as one is trying to analyze moral concepts, especially the concept of justice, as economists would, I think, agree. Justice is usually regarded as a separate and distinct part of any comprehensive criterion of economic policy. See, for example, Tibor Scitovsky, TVelfare and Competition ( London, 1952 ) , pp, 59�69, and Little, op. cit., cs, vrr. See J. S. Mill's argument in Utilitarianism, ch. v, pars. 16-25. ie


In this case either the benefits and burdens . may be assumed to be impartially distrib­ uted, or the question of distribution is mis­ placed, as in the instance of maintaining public order and security or national de­ fense. But as an interpretation of the basis of the principles of justice, classical utilitar­ ianism is mistaken. It permits one to argue, for example, that slavery is unjust on the grounds that the advantages to the slaveholder as slaveholder do not counter­ balance the disadvantages to the slave and to society at large burdened by a compara­ tively inefficient system of labor. Nmv the conception of justice as fairness, when ap­ plied to the practice of slavery with its offices of slaveholder and slave, would not allow one to consider the advantages of the slaveholder in the first place. As that office is not in accordance with principles which could be mutually acknowledged, the gains accruing to the slaveholder, assuming them to exist, cannot be counted as in any way mitigating the injustice of the practice. The question whether these gains outweigh the disadvantages to the slave and to society cannot arise, since in considering the jus­ tice of slavery these gains have no ,veight at all which requires that they be overrid­ den. Where the conception of justice as fairness applies, slavery is always unjust. I am not, of course, suggesting the ab­ surdity that the classical utilitarians ap­ proved of slavery. I am only rejecting a type of argument which their view allows them to use in support of their disapproval of it. The conceptioil of justice as derivative from efficiency implies that judging the jus­ tice of a practice is -always, in principle at least, a matter of weighing up advantages and disadvantages, each having an intrinsic value or disvalue as the satisfaction of in­ terests, irrespective of whether or not these interests necessarily involve acquiescence in principles which could not be mutually acknovvledged. Utilitarianism cannot ac­ count for the fact that slavery is always un­ just, nor for the fact that it would be recog­ nized as irrelevant in defeating the accusa-


tion of injustice for one person to say to another, engaged with him in a common practice and debating its merits, that nev­ ertheless it allowed of the greatest satisfac­ tion of desire. The charge of injustice can­ not be rebutted in this way. If justice were derivative from a higher order executive efficiency, this would not be so. But now, even if it is taken as es­ tablished that, so far as the ordinary con­ ception of justice goes, slavery is always unjust ( that is, slavery by definition vio­ lates commonly recognized principles of justice ) , the classical utilitarian would surely reply that these principles, as other moral principles subordinate to that of util­ ity, are only generally correct. It is simply for the most part true that slavery is less efficient than other institutions; and \vliile common sense may define the concept of justice so that slavery is unjust, neverthe­ less, where slavery would lead to the great­ est satisfaction of desire, it is not wrong. Indeed, it is then right, and for the very same reason that justice, as ordinarily un­ derstood, is usually right. If, as ordinarily understood, slavery is always unjust, to this extent the utilitarian conception of justice might be admitted to differ from that of common moral opinion. Still the utilitarian would want to hold that, as a matter of moral principle, his view is correct in giv­ ing no special ,veight to considerations of justice beyond that allowed for by the general presumption of effectiveness. And this, he claims, is as it should be. The every day opinion is morally in error, although, in­ deed, it is a useful error, since it protects rules of generally high utility. The question, then, relates not simply to the analysis of the concept of justice as common sense defines it, but the analysis of it in the wider sense as to how much weight considerations of justice, as defined, are to have when laid against other kinds of moral considerations. Here again I wish to argue that reasons of justice have a spe­ cial weight for which only the conception of justice as fairness can account. 11fore-


over, it belongs to the concept of justice that they do have this special weight. \Vhlle Mill recognized that this was so, he thought tc'iat it cou!d be accounted for by the special urgency of the moral feelings wh:ch naturally support principles of such high utility. But it is a mistake to resort to the urgency of feeling: as >'ith the appeal to intuition, it manifests a failure to pursue the question far enough. The special weight of considerations of justice can be explained from the conception of justice as fairness. It is only necessary to elaborate a bit what has already been said as follows. If one examines the circumstances in which a certaL� tolerance of slavery is jus­ tified1 or p�rhaps better, excused, it turns out that these are of a rather special sort. Perhaps slavery exists as an inheritance from the past and it proves necessary to dismantle it piece by piece; at times slav­ ery may conceivably be an advance on pre­ vious institutions. Nmv while there may he some excuse for slavery in special condi­ tions, it is never an excuse for it that it is sufficiently advantageous to the slaveholder to outweigh the disadvantages to the slave and to society. A person who argues in this way is not perhaps making a wildly irrele­ vant remark; but he is guilty of a moral fal­ lacy. There is disorder in his conception of the ranking of moral principles. For the slaveholder, by his owu admission, has no moral title to the advantages which he re­ ceives as a slaveholder. He is no more pre­ pared th= the slave to acknowledge the principle upon which is founded the re­ spective positions in which they both stand. Since slavery does not accord with prlnci� ples which they could mutually acknowl­ edge, they each may be supposed to agree , that it is unjust: it grants claims which it ought not to grant and in doing so denies claims which it ought not to deny. Amongst persons in a general position who are de­ bating the £om1 of their common practices, it cannot, therefore, be offered as a reason for a practice that, in conceding these very claims that ought to be denied, it neverthe-


less meets existing interests more effective­ ly. By their very nature the satisfaction of these claims is without weight and cannot enter into any tabulation of advantages and disadvantages. Furthemiore, it follows from the con­ cept of morality that, to the extent that the slaveholder recognizes his position vis-8.­ vis the slave to be unjust, he would not choose to press his claims. His no'!: vvanting to receive his special advantages is one of the ways in which he shows that he thinks slaveri is unjust. It would be fallacious for the legislator to suppose, then, that it is a ground for having a practice that it brings advantages greater than disadvantages, if those for whom the practiee is designed, and to whom the advantages .flow, ac­ knowledge that they have no moral title to them and do not wish to receive them. For these reasons the principles of jus­ tice have a special weight: and with re­ spect to the principle of the greatest satis­ faction of desire, as cited in the general po­ sition amongst those discussing the merits of their common practices, the principles of justice have an absolute weight. In this sense they are not contingent; and this is why their force is greater than can be ac­ counted for by the general presumption ( assuming that there is one) of the effec­ tiveness > in the utilitarian sense, of prac­ tices which in fact satisfy them. If one ·wants to continue using the con­ cepts of classical uti1ita,r:ianism i o!le will have to say, to meet this criticism, that at least the individual or social utility func­ tions must be so defined that no value is given to the satisfaction of interests the representative claims of which violate the principles of justice. In this way it is no doubt possible to include these principles within the form of the utilitarian concep­ tion; hut to do so is, of course, to change its inspiration altogether aS a moral concep­ tion, For it is to incorporate within it prin­ ciples which cannot be u!1derstood on the basis of a higher order executive decision aiming at the greatest satisfaction of desire.


It is worth remarking, perhaps, that this criticism of utilitarianism does not de­ pend on whether or not the two assump­ tions, that of individuals having similar utility functions and that of diminishing marginal utility, are interpreted as psycho­ logical propositions to be supported or re­ futed by experience, or as moral and politi­ cal principles expressed in a somewhat technical language. There are, certainly, several advantages in taking them in the latter fashion.2 ° For one thing, one might say that this is what Bentham and others really meant by them, at least as shown by how they were used in arguments for social reform. More importantly, one could hold that the best way to defend the classical utilitarian vie\v is to interpret these as­ sumptions as moral and political principles. It is doubtful whether, taken as psychologi­ cal propositions, they are true of men in general as we know them under normal conditions. On the other hand, utilitarians would not have ,vanted to propose them merely as practical working principles of legislation, or as expedient maxims to guide reform, given the egalitarian sentiments of modem society. 2 1 When pressed they might well have invoked the idea of a more or less equal capacity of men in relevant respects if given an equal chance in a just society. But if the argument above regard­ ing slavery is correct, then granting these assumptions as moral and political princ i ­ ples makes n o difference. To view individu­ als as equally fruitful lines for the alloca­ tion of benefits, even as a matter of moral principle, still leaves the mistaken notion 20

See D. G. Ritchie, Natural Rights (London, 1894 ) , pp. 95 ff., 249 ff. Lionel Robbins has insisted on this point on several occasions. See An

Essay on the Nature and Significance of ,Economic Science, 2nd ed. ( London, 1935 ) , pp. 134-43,

"Interpersonal Comparisons of Utility: A Com­ ment," Economic Journal, XLVIII ( 1938 ) , 635--41, and more recently, "Robertson o n Utility and Scope," Economica, n,s, XX ( 1953 ) , 108 f. 21 As Sir Henry Ivfaine suggested Bentham may have regarded them. See The Early History of Institutions ( London, 1875 ) , pp. 398 ff.


that the satisfaction of desire has value in itself irrespective of the relations between persons as members of a common practice, and irrespective of the claims upon one an­ other ,vhich the satisfaction of interests rep­ resents. To see the error of this idea one must give up the conception of justice as an executive decision altogether and refer to the notion of justice as fairness: that participants in a common practice be re­ garded as having an original and equal lib­ erty and that their common practices be considered unjust unless they accord with principles which persons so circumstanced and related could freely acknowledge be­ fore one another, and so could accept as fair. Once the emphasis is put upon the concept of the mutual recognition of prin­ ciples by participants in a common practice the rules of which are to define their sev­ eral relations and give form to their claims on one another, then it is clear . that the granting of a claim the principle of which could not be acknowledged by each in the general position ( that is, in the position in which the parties propose and acknowl­ edge principles before one another ) is not a reason for adopting a practice. Vie,ved in tl1is way, the background of the claim is seen to exclude it from consideration; that it can represent a value in itself arises from the conception of individuals as separate lines for the assignment of benefits, as iso­ lated persons who stand as claimants on an administrative or benevolent largesse. Oc­ casionally persons do so stand to one an­ other; but this is not the general case, nor, more importantly, is it the case when it is a matter of the justice of practices them­ selves in which participants stand in var­ ious relations to be appraised in accord­ ance ,vith standards which they may be expected to acknowledge before one an­ other. Thus hO\vever mistaken the notion of the social contract may be as history, and however far it may overreach itself as a general theory of social and political obli­ gation, it does express, suitably interpreted,



an essential part of the concept of justice. 22 8. By way of conclusion I should like to make two remarks: first, the original modification of the utilitarian principle ( that it require of practices that the offices and positions defined by them be equal un­ less it is reasonable to suppose that the rep­ resentative man in every office vvould find the inequality to his advantage ) , slight as it may appear at first sight, actually has a different conception of justice standing be­ hind it. I have tried to show how this is so by developing the concept of justice as fairness and by indicating how this notion involves the mutual acceptance, from a general position, of the principles on which a practice is founded, and how this in tum requires the exclusion from consideration of claims violating the principles of justice. Thus the slight alteration of principle re­ veals another family of notions, another way of looking at the concept of justice. Second, I should like to remark also that I have been dealing with the concept of justice. I have tried to set out the kinds of principles upon which judgments con­ cerning the justice of practices may be said to stand. The analysis will be successful to the degree that it expresses the principles involved in these judgments when made by competent persons upon deliberation and reflection. 23 Now every people may he

supposed to have the concept of justice, since in the lffe of every society there must be at least some .relations in which the par­ ties consider themselves to be circum­ stanced and related as the concept of jus­ tice as fairness requires. Societies ,vill differ from one another not in having or in failing to have this notion but in the range of cases to which they apply it and in the em­ phasis v,rhich they give to it as compared with other moral concepts. A firm grasp of the concept of justice itself is necessary if these variations, and the reasons for them, are to be understood. No study of the development of moral ideas and of the differences between them is more sound than the analysis of the fun­ damental moral concepts upon which it must depend. I have tried, therefore, to give an analysis of the concept of justice which should apply generally, however large a part the concept may have in a given morality, and \Vhich can be used in explaining the course of men's thoughts about justice and its relations to other moral concepts. Ho,v it is to be used for this purpose is a large topic which I can­ not, of course, take up here. I mentio� it only to emphasize that I have been dealing with the concept of justice itself and to in­ dicate what use I consider such an analysis to have.

22 Thus Kant was not far wrong when he inter­ preted the original contract merely as an "Idea of Reason"; yet he still thought of it as a general cri­ terion of right and as providing a general theory of political obligation, See the second part of the essay, «on the Saying 'That may be right in theory but has no value in practice' " ( 1793 ) , in Kant's Principles of Politics, tr. by \V. Hastie ( Edinburgh, 1891 ) . I have drawn on the contrac­ tarian tradition not for a general theory of political obligation but to clarify the concept of justice. 23 For a further discussion of the idea expressed here, see my paper, "Outline of a Decision Proce­ dure for Ethics," in the Philosophical Review, LX ( 1951 ) , 177-197, For an analysis, similar in many respects but using the notion of the ideal observer instead of that of the considered judgment of a

competent person, see Roderick Firth, "Ethical Absolutism and the Ideal Observer," Philosophy and Phenomenological Research, XII ( 1952 ) , 317-345. VVhile the similarities between these two discussions are more important than the differ­ ences, an analysis based on the notion of a con­ sidered judgment of a competent person, as it is based on a kind of judgment, may prove more helpful in understanding the features of moral judgment tlmn an analysis based on the notion of an ideal observer, although this remains to be shmvn, A man ,vho rejects the conditions imposed oil a considered judgment of a competent person could no longer profess to iudge at all, This seems more fundamental tban his rejecting the conditions of observation, for these do not seem to apply, in an ordinary sense, to making a moral judgment.

Problems of Distributive Justice NICHOLAS RESCHER THE TASK OF A THEORY OF DISTRIB UTIVE JUSTICE The task of a theory of distributive justice is to provide the machinery in terms of which one can assess the relative merits or demerits of a distribution1 the «assessment" in question being made from the mora1 or ethical point of view. Its objective is to es­ tablish a principle by which the "assess­ ment'' of alternative possible distributions can be carried out. 1 It is clearly not sufficient that a princi­ ple of distributive justice should tell us only about the ideal distribution, the very best of possible alternatives. For to apply any such principle in practice we must know vvhich of several feasible non-ideal 3.ltematives is to be preferred: we must know not only what is the best, but must be able to determine-in the great majority of cases, at any rate-which of several alter­ native possibilities is the "better." A princi­ ple of evaluation is not adequate if it merely depicts a theoretical ideal that we cannot apply in practice to determine which of several putative possibilities comes "closer to the ideal," ( How far has the beginner come toward learning how to evaluate bridge hands when he is told that the ideal holding consists of the four aces,

From Distributive Justice by Nicholas Rescher, copyright © 1966, by The Bobbs-1vferrill Company, Inc., reprinted by permission of the publishers and the author. 1. Henry Sidgwick, The Methods of Ethics ( 7th ed., London: lvlacmillan, 1907) , p. 270.


the four kings, the four queens, and a jack?) What is needed is something to which writers on ethics-and, for that matter, writers on economics also-------have been loath to address themselves, a criterion of merit for suboptimal alternatives. The evaluation criterion of an adequate theory of distribu­ tive justice must be capable not simply of absolute idealization ( i.e., of telling us what the ideal is ) , but also of relative evaluation ( i.e., of telling us which of several possible alternatives is to be regarded as the most satisfactory ) . Distributive justice-exactly like puni­ tive justice-can be brought to realization only in this world, that is, in an imperfect world populated by imperfect men. A per­ fectly just system of punitive justice would apprehend, process, and punish all and only the guilty, and would ignore, leave unprocessed and unpunished, all and only the innnocent. But any realizable system will be such that it cannot fail to depart from the ideal in several ways ( say, by catching some of the innocent and by let­ ting escape some of the guilty ) , And these modes of injustice are interrelated and in­ terlocked: as we modify the system to avoid injustices of the one kind, we ipso facto increase those of another. In evaluat­ ing alternative procedures in criminal law and law enforcement we have to be pre­ pared to make choices among the realiz­ able, and thus less-than-ideal, alternatives; exactly the same is true in evaluating socio-economic arrangements with respect


to their accordance with or violation of the principles of distributive justice.


butions. But at a later stage of tbe inquiry we shall return to the conception of tbe principle of utility as a general arbiter in ethics.

THE PRINCIPLE OF UTILITY We shall base our inquiry into tbe princi­ ple of distributive justice upon an investi­ gation of tbe doctrine ' of utilitarianism. This doctrine is founded upon the principle of utility, which asserts tbat utility ( or, if you wish, simply the good things of this life) should be distributed according to the rule of "tbe greatest good of tbe greatest number." Exact as it sounds, this classic principle is imprecise and indeed inade­ quate. The first objective of our discussion is to exhibit these shortcomings in con­ siderable detail. And when the necessary emendations are made, the resulting posi­ tion will be such tbat tbe label "utilitarian" ( as usually construed) can be pinned to it only \vith serious reservations and qualifi­ cations, if at all. We shall use tbis critique of the principle of utility to illustrate tbe very complex and inherently problematic char­ acter of the concept of distributive justice. Some of tbe classical utilitarians have taken the principle of utility to relate to human actions in general, and not to be con£ned specifically to distributions of goods. Insofar as the majority of human ac­ tions have consequences affecting others fa­ vorably or unfavorably, tbey "distribute" utilities ( or disutilities) to tbe parties con­ cerned. So regarded, all actions affecting others become "distributions of utilities," and tbe principle of utility can be applied, thus -widening its scope from that of a mere principle of distributive justice to a general criterion of right action in human conduct. For the present, however, we must take a limited view of tbe principle of utility, without reference to its oft maintained role as a -criterion of right action in human con­ duct. We shall resist-at any rate for tbe time being-this expansion of its scope, confining its application to genuine distri-

THE PRINCIPLE OF UTILITY IS A TWO-FACTOR CRITERION Suppose that some three particular persons, Messrs A, B, and C, can be given tbe utility shares ( a ) , ( b ) , and ( c ) , respectively, in accordance with either S-cheme I or Scheme II: 2 Share (a) (b) (c)

Scheme I 3 units 3 3

Scheme II 2 units 2 6

vVhich scheme represents the superior mode of distribution? Scheme II yields "the greater good": it distributes ten units as compared witb tbe nine of its rival. Scheme I yields a greater advantage in goods for "tbe greater number" : two persons gain by its adoption and only one loses. The exam­ ple brings out tbe fact tbat the principle of utility is a two-factor criterion ("greater good," "greater number" ) , and tbat tbese two factors can in given -cases work against one another. There is tbus nothing in the principle of utility itself to help us in mak­ ing- let alone in dictating a particular out­ come of-a choice between Scheme I and Scheme II. The principle unqualified is pat­ ently incomplete as an effective means for deciding between alternative distributions of a good. Some utilitarians have, at least seem­ ingly, gone from a two-factor to a one­ factor -criterion, placing their sole reliance upon "tbe greater good," dropping the last

y.,re are to think of the indicated shares here not as representing marginal utility increments added to an otherwise fixed initial amount, but as the total resultant utility distribution after what­ ever distributing mechanism may be supposed operative has done its work. 2


four words from the utilitarian formula. ( Bentham himself inclined to this view in his later days, 3 reasoning, along lines shortly to be described, that the greater good requires greater numbers.) But de­ spite its greater logical tidiness, the view that the eligibility of a proposed action ,vith its consequent distribution of utility turns solely on the total good involved, without any regard whatsoever to the pat­ tern of its distribution, is pretty obviously unacceptable. On the other hand, it would obviously not do to place, in a burst of democratic enthusiasm, an exclusive reli­ ance on "the greater number.'' 4 For con­ sider the distributions: Share (a) (b) ( c) (d) (e)

Scheme I

5 units 5 5 1 1

Scheme II 4 units 4 3 3 3

Doubtless the greater number of recipients would opt for Scheme I and would vote for its adoption as against Scheme II, but it is doubtful ( to say the least) that the first mode of division is to be preferred. In such cases, even the most ardent of democratic theoreticians have ever seen £t to safe­ guard the interests of minorities in ways that preclude an automatic adoption of schemes of type I. 5 3

Compare F. Y. Edgeworth, Mathematical pp. 117-118. Edgeworth strongly endorses the alteration, remarking: "The principle of greatest happiness may have gained its popular­ ity, but lost its meaning, by the addition 'of the


greatest number.' " 4

We wholly ignore the ambiguity that is sin­ gled out by the question "Greater number of what?" i.e., do we have an anthropocentric form of utilitarianism, where only humans (perhaps better, intelligent creatures) count, or a universalistic form, \Vhere other sentient beings are also included? � It is important to qualify by safeguards of this sort the census technique by which D. Braybrooke and C. E. Lindblom seek to replace the classic util­ itarian calculus in their recent book, A Strategy of Decision ( New York: The Free Press of Glencoe, 1963).


One traditional objection to utilitarian­ ism articulated along these lines is pre­ sented in the choice between a less popu­ lous world with a higher per capita aver­ age utility, and a more populous world with a lower per capita average utility, say between: Scheme 1 (b)


( a)

Scheme 2 (f ) ( e) (d ) ( c) (b ) (a)

A contemplation of these alternatives should force an adherent of the principle of utility to decide whether by "the greatest good" he is to mean the greatest total good or the greatest good per capita ( i.e., the greatest average good). That proto­ utilitarian vVilliam Paley wrote:

A larger portion of happiness is enjoyed amongst ten persons, possessing the means of healthy subsistence, than can be produced by five persons, under every advantage of power, affiuence, and luxury . . . ; it follows, that the quantity of happiness in a given district, although it is possible it may be increased the number of inhabitants remaining the same, is chiefly and most naturally affected by altera­ tion of the numbers: that, consequently, the decay of population is the greatest evil that a state can suffer; and the improvement of it the object which ought, in all countries, to be aimed at, in preference to every other political purpose whatsoever. 6 Other- and no\vadays surely more common- sentiments go the other way. Thus C. D. Broad writes :

e YVilliam Paley, The Principles of Moral and Political Philosophy ( 7th edn., London: Baldwin &

Co., 1790 ) , Vol. II, Book VI, ch. 2, pp. 346-347. Bentham, Godwin, and most early utilitarians side with Paley here. See E. Halevy, The Growth of Philosophic Radicalism, tr. Mary Morris, pp. 218--221. Compare also Henry Sidgwick, The Methods of Ethics, pp. 415--416.

PROBLEMS OF DISTRIBUTIVE JUSTICE If Utilitarianism be t:ue it would be o::ie's duty to try to inc:ease the numbers of a community, even though one reduced the average total happiness of the members, so long as the total happbess in the community \Vould be in the least increased. It seems perfectly plain to me that this kind of action, so far from being a duty, would quite certainly be wrong. 7

clause, a "principle preveution" stipulating


of catastrophe­ a minimal utility

fioor for all individuals below which no one

should be pressed. The principle at issue may be regarded as being more or !ess built in to the very conception of a genuinely "minimally acceptable" share of good. For we ,vould not conceive of a given level in just this way unless we were prepared to do A "UTILITY FLOOR" IS NEEDED battle for the rule that an exaltecl priority Let us trv the effect of one facile amend­ should be given to reducing to the lowest ment of ilie principle of utility. One of the feasible number the people who receive less standard textbook objections to the princi­ than this share. We might thus add to the ple is presented by the following variant of initial principle the proviso: provided that nobody receives less of "the gooa' than a the previous example: certain (i.e., some plaU,S/1,le) minimum amount. Clearly one of the most basic ele­ Scheme 2 Scheme 1 ments of our concept of justice is to n1ini­


• (b) i



Here Scheme II not only yields "the greater good," hut works to the advantage o: "a greater number/' since two of the three people involved are obvious bene­

ficiaries of its adoption. But is it reasonable

that we should in all such cases be pre­

pared to sacrifice an �"individual intere..')f' in

"the general benefit," as the principle of utility says we must do? Tbe answer to tbis question cannot be other than no! VVe would surely not want to subject one indi­

vidual to unspeakable suffe::-ing to give

some insignificantly small benefit to many others ( even an innun1erable myriad of them ).' Actual privation offends our sense of iustice i.n a more serious way than do

rnere inequities, These considerations suggest adding to the principle of utility another qualifying 1 C. D. Broad, Five Types of Ethical TheonJ (London: Rout:edge & Kegan Pali:, 1930) , p. 250. 8 A somewhat out-of-the-way example of this line of reasoning is provided by the argument of some modem theologians that creation as a whole is not ,vorthwhile if it has the consequence of eter­ :nal damnation for some creatues.

mize the number of persons in a state of genuine deprivation regarding their share in the available pool of utility. Diminish­ ing the number of those who simply do not have enough is a more fundamental ele­ ment of the concept of justice than dimin­ ishing the gap between the "haves" and the "have-nots."9 And although the idea of an "There is. a bottom level of instrumen!:al good ( money, or in th!S instance, food) below which equality is useless because it is equa�ty in noth­ ingness, or something so near to nothingness that tt would be of no use to a"!ly of the recipients. No good would be achieved by requiring equality under such conditions, On the one !1and, a person who said, "I kno'\v we'll al! starve, but we must share equally an}"vay' would really be running the equality principle into the ground! Precisely the same th!ng has been alleged of a socialistic econ­ omy: though it provides near�equality, t1ie ince:::1tive is so low and, human nature being what it is, the system is inevitably so inefficient that after a while there ,vill not be much left to divide equally: we shall have what has been called a state of 'splendidly equalized dei,titutlon." If it could be shoVi!Il that an economy characterized by equal distribution produced this .result, such an economv \YOdd be almo�t as useless to its mem­ bers as· the situation of ten men on tbe ice floe sharing substarvation rations." John Hospers, Hurnan Conduct ( New York: Harcourt, Brace & World, 1961), p. 428. (This and the following a.:cerpts from John Hospers, Human Conduct, are reprinted with the pennission of the publishe!". ) On the conception of a utility :ioor, see also B. de Jouvenel, The Ethics of Redistribution, pp. 2.3-24 and 8,5-88. 9


acceptable minimum level has traditionally been stressed primarily in survival contexts, the idea has long been applied in such other connections as, for example, educa­ tion. The utilitarians of the eighteenth and early nineteenth centuries recognized and accepted this principle, and it led them to abandon, in the economic (but not the po­ litical) sphere, the egalitarianism to which they were othenvise committed: If the Utilitarians rejected absolute equalitar­ ianism, it was not because they considered society as naturally hierarchical, but because

they thought the quantity of subsistence

actually available was not sufficient to allow all the individuals actually existing to live . . .

[adequatelyJ . 1 0

Apart from some such qua!iflcation, the principle of utility is clearly deficient. It is perfectly conceivable that at some his­ torical juncture an institution of slavery, for example, could conduce to the greater good of a greater number; but we shall not be prepared to let this fact count decisively on its behalf. Again, considerations of a cognate sort led to the economic doctrine that an economy must afford every partici­ pant a '1iving wage."11

AN EQUITY PRINCIPLE IS NEEDED But the amendment proposed in the preced­ ing section vvill not of itself suffice. For even when one inserts such a to t½ink that such a supposedly necessary truth about punishment ( or anything else ) could conslitute a retiibutive or anv other sort of moral justillcatfon oI it ( or a{,ything else ). To attempt to justify from the con­ cept alone is like trying to prove existence from the concept alone-The Ontological Argument. Though there is this at least to be said for it: tbat if "our moral convic­ tions" are to be accepted as the arbiter, then the attempt to justify from the con­ cept alone v/ill amount to an appeal to the popular moral convictions incapsulated in ordinary language. ( c) Onerdetermination.-Sometimes it seems to be assumed that there must be an inconsisumcy in justifying the adoption and e�forcemeiit of a law or a system of lav.ts by both utilitarian and retributive appeals. But there is no necessary inconsistency in this, any 1nore than t:J1ere is in having advo­ cated the nationalization of coal but now 12

"l\-fr, Lyttelton has helped to force through a federation bas:ed 011 the admitted policy of regard­ ing A_frican political opinions as irrelevant'' (Observet, 23/8/53).

631 opposing the nationalization of chemicals or cement: though of course the people making these appeals or combining these policies may say things, or offer sup­ porting reasons, ·which do involve them in inconsistency. Thus., having a law against murder, or against any other sort of behav­ ior which is reckoned to be wrong whether it is made illegal or not, can be defended both on the grounds that this makes for commodious living and against lives nasty, brutish, and short and on the grounds that it makes for wicked men getting their de­ serts. ( I suspect that not only would most people resort to both sorts of arguments� but that Mill too would have tlone so­ thongb he would have said that any "sec­ ondary pr'.nciple" of retribution for ill­ desert had ultimately to be justified by reference to the ''£rst principle", the "Greatest Hap"-J:iness Principle", I cannot hope to mak-;,c good the historical elairn here: but I refer again to Um1Son's most excellent paper; and to the fact that Mill made "the turning point of the distinction between morality and simple expediency" tc'iat for wrong-doing "a person ought to be punished" ( lac. cit., p. 45) . ) Perhaps philosophers have been mis­ led into this assumption by undertaking to find a general comprehensive justification for all justified punishments; deceptively described as "a theory of punisbme"t". This must comrr1it them to producing: either an ethical clru:m which is to be insisted on in every case and to ,vhich no exceptions whatever \yjll be admitted; or a necessary truth which obtains the universa:ity re� quired only at the cost of ceasing to be any sort of ethical justification. IV

·we have tried in this paper to bring out the main features of the logic of the expres­ sion "the justification 0£ punishment"; and to apply the lessons to one outstanding

paper about tbe justification of punish­ ment. We have not attempted, except inso­ . far as was absolutely necessary, incidentally, to do any actual justillcation on our own account. Indeed the entire paper may be considered as prolegomena only: except that tbey are prolegomena of the sort 632


which suggest that the latter enterprise- in

enquiry as opposed to either a series of piecemeal jobs done in particular contexts or tbe attempt to find generally useful principles to which tbere will be occasional exceptions-was misconceived.

the form of a comprehensive and universal

Free- Will and Psychoanalysis JOHN HOSPERS O Thou, who didst with pitfall and with gin Beset the Road I was to wander in, Thou \Vilt not with Predestined Evil round Enmesh, and then impute my Fall to Sin! -Edward FitzGerald,

The Rubaiyat af Omar Khayyam.

It is extremely common for nonprofes­ sional philosophers and iconoclasts to deny that human freedom exists, but at the same time to have no clear idea of what it is that they are denying to exist. The first thing that needs to be said about the free-will issue is that any meaningful term must have a meaningful opposite : if it is mean­ ingful to assert that people are not free, it must be equally meaningful to assert that people are free, whether this latter a�se�­ tion is in fact true or not. \i\1hether 1t 1s true of course, will depend on the meaning that' is given the weasel-word "free." For example, if freedom is made dependent on indeterminism, it may well be that human freedom is nonexistent. But there seem to be no good grounds for asserting such a de­ pendence, especially since lack of �ausat10n _ is the furthest thing from people s mmds when they call an act free. Doubtless there are other senses that can be given to the \Vord "free"-such as "able to do anything we ,vant to do"-in which no human beings are free. But the first essential point about ,vhich the denie,r of freedom must be clear Reprinted from Section 2 and part ?f ,,Section 3 , of the article ":tvfeaning and Free-\V1ll, Philoso­ phy and Phenomenological Research, X, 1950 by permission of the publisher.

is what it is that he is denying. If one knows what it is like for people not to be free, one must know what it would be like for them to be free. Philosophers have advanced numerous senses of «free" in which countless acts per­ formed by human beings can truly be called free acts. The most common concep­ tion of a free act is that according to which an act is free if and only if it is a voluntary act. But the word "voluntary" does not al­ ways carry the same meaning. Sometimes to call an act voluntary means that we can do the act if we choose to do it: in other words, that it is physically and psychologi­ cally possible for us to do it, so that the oc­ currence of the act follows upon the deci­ sion to do it. ( One's decision to raise his arm is in fact followed by the actual raising of his arm, unless he is a paralytic; one's decision to pluck the moon from the sky is not followed by the actual event. ) Some­ times a voluntary act is con9eived ( as by Moore 1 ) as an act which would not have occurred if, just beforehand, the agent had chosen not to perform it. But these senses are different from the sense in which a vol­ untary act is an act resulting from delibera­ tion, or perhaps merely from choice. For example, there are many acts which we could have avoided, if we had chosen to do so but which \Ve nevertheless did not choose to perform, much less deliberate about them. The act of raising one's leg in the process of taking a step while out for a 1

Ethics, pp, 15-16,



walk, is one which a person could have avoided by choosing to, but which, after one has learned to walk, takes place auto­ matically or semi-automatically through habit, and thus is not the result of choice. ( One may have chosen to take the walk, but not to take this or that step while walk­ ing. ) Such acts are free in Moore's sense but are not free in the sense of being de­ liberate. Moreover, there are classes of acts of the same general character which are not even covered by 1vioore's sense: sudden outbursts of feeling, in some cases at least, could not have been avoided by an immediately preceding volition, so that if these are to be included under the heading of voluntary acts, the proviso that the act could have been avoided by an immedi­ ately preceding volition must be amended to read "could have been avoided by a voli­ tion or series of volitions by the agent at some time in the past''-such· as the adop­ tion of a different set of habits in the agent's earlier and more formative years. ( Sometimes \Ve call persons, rather than their acts, free. Stebbing, for example, declares that one should never call acts free, but only the doers of the acts.2 But the two do not seem irreconcilable: can we not speak of a person as free with respect to a certain act ( never just free in general) if that act is free-whatever we may then go on to mean by saying that an act is free? Any statement about a free act can then be translated into a statement about the doer of the act.) Now, no matter in which of the above ways ,ve may come to define "voluntary," there are still acts which are voluntary but which we would be very unlikely to think of as free. Thus, when a person submits to the command of an armed bandit, he may do so voluntarily in every one of the above senses: he may do so as a result of choice, even of deliberation, and he could have avoided doing it by willing not to-he could, instead, have refused and been shot. 2

Philosophy and the Physicists, p. 212.


The man who reveals a state secret under torture does the same: he could have re­ fused and endured more torture. Yet such acts, and persons in respect of such acts, are not generally called free. \'Ve say that they were performed under compulsion, and if an act is performed under compul­ sion \Ve do not call it free. VVe say, "He \Vasn't free because he was forced to do as he did," though of course his act was vol­ untary. This much departure from the identi­ fication of free acts with voluntary acts al­ most everyone would admit. Sometimes, however, it would be added that this is all the departure that can be admitted. Ac­ cording to Schlick, for example,

Freedom means the opposite of compulsion; a man is free if he does not act under compul­ sion, and he is compelled or unfree when he is hindered from without in the realization of his natural desires. Hence he is unfree when he is locked up, or chained, or when someone forces him at the point of a gun to do what othen.vise he v/Ould not do. This is quite clear, and every­ one will admit that the everyday or legal notion of the lack of freedom is thus correctly interpreted, and that a man ·will be considered quite free . . . if no such external compulsion is exerted upon him. 3

Schlick adds that the entire vexed free-will controversy in philosophy is so much wasted ink and paper, because com­ pulsion has been confused with causality and necessity with uniformity. If the ques­ tion is asked whether every event is caused, the ans\ver is doubtless yes; but if it is v.rhether every event is compelled, the answer is clearly no. Free acts are uncom­ pelled acts, not uncaused acts. Again, when it is said that some state of affairs ( such as water flowing downhill) is necessary, if "necessary" means "compelled," the answer is no; if it means merely that it always hap­ pens that way, the answer is yes: universal­ ity of application is confused with compula The Problems of Ethics, Rynin translation, p. 150.

FREE-WILL AND PSYCHOANALYSIS sion, And this, according to Schlick, is the

e,id of the matter. Schlick's analysis is indeed clarifying and helpful to those who have fallen vfotim to the confusions he exposes-and this probably includes most persons in their phllosopbfoal growing-pains. But is this the end of the matter? Is it true that all acts, causec:, are free as long as they are not compelled in the sense \vhich he spe­ cifies? May it not be that, while the identification of "free" with "uncompelled" is acceptable, the area of compelled acts is vastly greater than he or ::nost other phllos; ophers have ever suspected? ( Moore is more cautious in Ws respect than �ch Hck.i while for :tvfoore an act is free if it is volun­ tary in the sense specified above, he thinks there may be another sense in which hu­ man beings, and human acts1 are not free at 1111. 4 ) Vie remember statements about human beings being pawns of their early environment1 victims of conditions beyond their control, the result of causal inHuences stemming from their parents, and the like, and we ponder and ask, "Still, are we really free?" To there not something in what gen­ erations of sages have said abm::t man being fettered? Is there not perhaps some­ thing too facile, too sleight-of-hand, in Schlick's cutting of the Gordian knot? For e.x.ample1 when a metropolitan newspaper headlines an article with the \vords "Boy Klller is Doomed Long before He Is Bom,»5 and then goes on to describe how a twe:ve-year-o!d boy has been sentenced to prison for the murder of a girl, and how his p arental background includes records of drunkenness, divorce, social mafa:djust� ment, and paresisJ we stiJl to say t..'lat his act,. though voluntary and as­ suredly not done at the point of a gun, is free? The boy has early displayed a tend­ ency toward sadistic activity to hide an un­ derlying masochism and ""prove that he's a man"; being coddled by his mother only worsens this tendency, until, spun1ed by a • Ethicst Cl1aptcr 6, pp. 217 ff. Ne,v York Post, Tuesday, May 18, 1948, p. 4.


635 girl in his attempt on he kilts her-not simply in a fit of anger, but calculatingly, deliberately. Is he free in respect of hls criminal a.ct1 or £or that matter in most of the acts of his life? Surely to ask this ques­ tion is to ansvter it in the negative. Perhaps I have taken an extreme case; but it is only to show the superficiality of the Schlick analysis the more clearly. Though not everyone has criminotic tendencies, every­ one has heen moulded by influences which in large measure at least determine his present behavior; he is literally the pmduct of t:lese influences, stemming .from periods prior to his "'years of discretion/' giving him a host of character traits that he cannot change now even if he would. So obviously does what a man is depend upon how a man comes to be} that H is small wonder that P"ilosophers and sages have consid­ ered man far indeed from being the master of his fate. It is not as if man's will ,vere standing high and serene above the flux of events that have moulded him; it is itself caught up in this flux, itself carried along on the current. An act is free ,vhen it is de­ termined by the man's character, say mor­ alists; but what if the most decisive aspects of his character were already irrevocably acquired before he could do anything to mould them? What if even lhe degree of will power available to him in shaping his habits and disciplining himself now to over­ co:11e the influence of his ea:ly environ­ ment is a factor over which he has no con­ trol? ' What are we to say of this kind of "freedom"? Is it not rather like the freedom of the machi.'1e to stamp labels on cans when it has been de,1sed for just that pur­ pose? Some machines can do so more efficiently than others, but only because they have been better constn::cted. It is not my pmpose here to establish this thesis in general, but only in one spe­ cific respect which has received compara­ tively little attention, namely, the field re­ ferred to by psychiatrists as that of uncou­ restrict my attention to it because it illus­ scious motivation, In what follows I shall


trates as clearly as anything the points I wish to make. Let me try to summar:ze very briefly the psychoanalytic doctrine on this point. 6 The conscious life of the human being, in­ cluding the conscious decisions and voli­ tions1 is rnerely a mouthpiece £or the un� conscious-not direct1y for the enactment of unconscious drives, but of the compromise between unconscious drives and t:ncon­ scious reproaches. There is a Big Three be­ hind the scenes which the m1tomaton called the conscious personality carries out: the id, ao "eternal gimme," presents its wish and demands its immediate satisfac­ tion; the super-ego says no to the ,vish immediately upon presentation, and the unconscious ego� t1H:: mediator between the two, hies to keep peace by means of compromise.1 To into examples of the functioning of these three ''bosses" would be endless; psychoaualytic case books supply hundreds of them. The important point for us to see in the present context is that it is the im­ conscious that determines what the con­ scious impulse and the conscious action. shall be. Hamlet, for example, had a strong Oedipus wish, \.Vhich \vas violently counter­ acted by super-ego reproaches; these early wishes were vividly revived in an unusual adult situation in which his uncle usurped the coveted position from Hamlet's faber and won his rnother besides. 1'bis situation evoked strong strictures on the part of 0 1 am av.11ro that the theory presented below is not accepted by all practicing psychoanalysts. Many non-Freudians wcu: often i:Jvolving deprivation of attention and affect:on. Generally the child soon learns that this form of rebellion is prof­ itless, and brings him more harm than good. He wan:s to respond to frustrat:on with violent aggression1 and at the same time learns faat he will be punished for such aggression, and that in any case the latter is ineffech,a!, 'What face-saving solu­ tion does he find? Since he must "face facts/''Since he must in any case ''conform)• jf he is to have a:1y peace at all, he tritlS to make it seen1 as if he himself is the source of the commands and Fohibitions: the ex­ terrwl proiiibitive force is intenrnli�ed-a.,'1.d here '\Ve have the origin of conscience. By making the prohibitive agency seem to 1 Edrou.."1d Bergler, The Battle of {he come from V1s con­ ditir>n worsens, and the mother does all she can to keep it alive, ,vithout, however, leaving the train, for she declares that it is absolutely necessarv that she reach her d es� tination. But befor� she gets there the ch!ld is dead. After that, of course, the mother grieves 1 blames berself1 weeps hysterically,


and joins the church to gain surcease from the guilt that constantly overwhelms her when she thinks of how her aggressive be· havior has killed her child. Was she responsible for her deed? In ordinary life, after making a mistake� we say, "Chalk it up to cxperience. Here we should say, "Chalk it up to the neurosis." She could not help it if her neurosis forced her to act this v.rav-·she didn't even knmv what was going o; behind the scenes, her conscious self merely acted out its assigned part. This is far more true chan is generally realized: criminal ac::ions in general are not actions for ,vhich their agents are responsi­ ble; the agents are passive1 not active-they are victims of a neurotic couflict. Their verv hyper-activitv " is unconsciouslv " determ�ed.' To say this is, of course, not to say that we should not punish criminals. Clearly, for our own protection, vve must r0..move them from our midst so that they can no longer n1olest and endanger organized so­ ciety. And, of course, if we use the word "responsible" in such a way that justly to hold someone responsible for a deed is by definition identical with being justified in punishing him, then we can and do hold people respon.sible. But this is like the sense of "'free'' in \Vhich free acts are volun­ tary ones. It does not go deep enough. In a deeper sense ·we cannot hold the person re­ sponsible: we can hold his neu:-osis respon­ sible, but he is not respamible far his neu­ rosis, particularly since the age at which its onset was inevitable was an age before he could even speak. The nec:rnsis is respousible-but isn't the neurosis of pa.rt of him? We have been speaking all the t'.me as if the person and his unconscious were t\vo sep�ate beingsi but isn't he one personality, including con­ scious and unconscious deparbnents to­ gether? I do not wish to deny this. But it hardly helps us here; for what people want when they talk about freedom, and what 0


IJ1ey hold to when they champion it, is the idea th at the conscious v,,ill is the master of their destiny. �-1 am the master of iny fate, I am the captain of my sour'--and they surely :mean their conscious selves, the self that they can recognize and search and in­ trospect. Between an unconscious that ,villy�nilly determines your actions, and an external force which pushes you, there is little if anyfoing to choose. The unconscious is just as if it ,vere an outside fo:ce; and in­ deed, psychiatrists will assert that the inner Hitler ( your super-ego ) can torment you far more than any external Hitler can. Thus the kind of freedom that people waut, the only kind they will settle for, is precisely the kind that psychiatry says that they can· not have. Heretofore it was pretty generally thought that, while we could not rightly blame a person for the coior of his eyes or the morality of his parents� or even for what he did at the age of three, or to a large extent what impulses l::e had and ,vhom he fell in love with, one could do so for oilier of his adult activities, particularly the acts he performed voluntarily and with pre:nedit.ation. Later this attitude ,vas shaken , f.,fany voluntary acts came to he recognized1 at least in some circles; as com­ pelled by the u:iconscious. Some philoso­ phers recognized this too-Ayer 12 taiks about the kleptomaniac being unfree) and about a person being unfree when another person exerts a habitual ascendancy over his personality. But tl1is is as far as he goes. The usual examples, such as :he kleptoma· niac and the schizophrenic, apparently sat­ isfy most philosophers; and with these ex­ ceptions removed,. the rest of mankind is permitted to wander in the vast and allur­ ing fields of freedom and responsibility. So far, the inroads upon freedom left the vast n1ajority of hu..."llanity untoucl1ed; they began to hit home when psychiatrists began to realize, though philosophers did 12 A . J. Ayer1 "Freedo:rn and Necessity," Polemic (September-October 1946), pp, 40-43.



not, that the dmnination of the conscious which are a menace to society, he must be by the unconscious extended, not 1nerely to put into prlson 1 of course) but responsibil­ a few excepti.onal individuals, but to all ity is another matter. The time when the human bein-gs, that the ·��:g three behind events occurred ,vhich rendered his neu­ the scenes" are not respecters of persons, rotic behavior inevitable was a time long and dominate us all, even including that before he was capable of thought and deci­ sanctum saru;torum of freedom, our con­ sion. A s an adult; he is a victim of a world scious w:1.1. To be sure, the domination by h e never made-only this world is inside the unconscious in the case of �'normal"' in� him. dividua�s is somewhat more benevolent What ahout the children who tum out than the tyranny and despotism exercised "all right''? All ,ve can say is that "it's jnst in neurotic cases, and therefore the ·former lucky for them" that what happened to have evoked !ess co1nment; but the princi­ their unfortunate brother didn't happen to ple remains in all cases the same: the un­ them; through no cirtue of their own they conscioes is the master of every fate and are not doomed to the life of unconscious the captain of every soul. guilt, expiation, conscious depression! ter­ We speak of a machine turning out rified ego-gestures for the appeasement of good products most of the time but every a tyrannical super-ego, that he is. The ma­ once In a while it turns out a "lemon." \Ve chine turned !:hem out, with a minimum of do net, of course, hold the product respon­ damage. But if the brother cannot be sible for this,_ but the ::nachine, and via the blamed for his evils, neither can they he machine, its maker. Is it silly to extend to praised for their good; unless, of course, inanimate objects the idea of responsibil­ we should blame people for what is not ity? Of course. But is it any less so to em­ their fault, and praise them for lucky acd. ploy the notion in speaking of huma.>i crea­ dents. We all agree that machines tum out tures? Are not the two kinds of cases analo­ gous in countless importa;::i,t ways? Occa­ "lemons/' we all agree that nature turns out sionally a child turns out badly too, even misfits in the realm of biology-the blind, when his environment a�d training are the the crippled, the diseased; but we hesitate same as that of his brothers and sisters who to include the realm of the personality, for turn out �all right." He is the "bad penny." here, it seems, is the last retreat of our dig­ His acts of rebellion against parental disci­ nity as human heings. Our ego can endure pline in adult life ( such as the case of the anything but this; this island at least must gambler, already cited) are traceable to remain above the encroaching fiood. But early experiences of real or fancied denial may not precisely the same analysis b e of infantile wishes. Sometimes the denial made here also? Nature turns out psycho­ had been real, though many denials are ab­ logical "le:nons" too, in far greater quanti­ solutely necessary if the child is to grow up ties than any other kind; and indeed all of to obsen1e the common decencies of ch'i­ ns are "lemons" in son1e respect or other, lizcd life; sometimes, if the child has an t1:ie difference being one of degree. Some of unusual quantity of narcissism, eveiy event us are 1uc&:y enough not �o have a gam­ that occurs is interpreteC. by him as a de­ bling-neurosis or cr..minotic tendencies or nial of h!S wishes1 and nothing a parent masochistic mother-attachment or overdi­ could do, even granting every humanly mensional repetition-compulsion to make possible ,; would help. In any event, our lives miserable, but most of our actiotL'i 1 the later neurosis can be attributed to this, those usually considered the most impor­ Can the person hlmself be held responsi­ tant> are unconsciously dominated just the ble? Hardly. If he engages in activities same. And1 i1 a neurosis nrny be likened to


a curse of God, let those of us, the elect, who are enabled to enjoy a measure of life's happiness without the hell-fire of neu­ rotic guilt, take this, not as our own achievement, but simply for what it is-a gift of God. Let us, however, quit metaphysics and put the situation schematically in the form of a deductive argument. 1. An occurrence over which v,re had no control is something we cannot be held responsible for. 2. Events E, occurring during our ba­ byhood, \Vere events over which we had no control. 3. Therefore events E \Vere events which we cannot be held responsible for. 4. But if there is something we cannot be held responsible for, neither can we be held responsible for something that inevi­ tably results from it. 5. Events E have as inevitable conse­ quence Neurosis N, ,vhich in turn has an inevitable consequence Behavior B. 6. Since N is the inevitable conse­ quence of E and B is the inevitable conse­ quence of N, B is the inevitable conse­ quence of E. 7. Hence, not being responsible for E, we cannot be responsible for B. In Samuel Butler's Utopian satire Ere­ whon there occurs the following passage, in ,vhich a judge is passing sentence on a prisoner: It is all very well for you to say that you came of unhealthy parents, and had a severe accident in your childhood which permanently undermined your constitution; excuses such as these are the ordinary refuge of the criminal; but they cannot for one moment be listened to by the ear of justice. I am not here to enter upon curious metaphysical questions as to the origin of this or that- questions to which there ,vould be no end were their introduction once tolerated, and ,vhich would result in throwing the only guilt on the tissues of the primordial cell, or on the elementary gases. There is no question of how you came to be wicked, but only this-namely, are you ,vicked or not? This


has been decided in the affirmative, neither can I hesitate for a single moment to say that it has been decided justly. You are a bad and dangerous person, and stand branded in the eyes of your fellow countrymen with one of the most heinous known offenses. rn

As moralists read tbis passage, they may perhaps nod with approval. But the joke is on them. The sting comes when we realize vvhat the crime is for ,vhich the prisoner is being sentenced: namely, consumption. The defendant is reminded that during the previous year he was sentenced for aggra­ vated bronchitis, and is warned that he should profit from experience in the future. Butler is employing here his familiar method of presenting some human tend­ ency ( in this case, holding people responsi­ ble for what isn't their fault ) to a ridicu­ lous extreme and thereby reducing it to ab­ surdity.

Assuming the main conclusion of this paper to be true, is there any room left for freedom? This, of course, all depends on what we mean by "freedom." In the senses sug­ gested at the beginning of this paper,. there are countless free acts, and unfree ones as well. VVhen "free" means "uncompelled," and only external compulsion is admitted, again there are countless free acts. But now we have extended the notion of compulsion to include determination by unconscious forces. VVith this sense in mind, our ques­ tion is, ''"With the concept of compulsion thus extended, and in the light of present psychoanalytic knowledge, is there any freedom left in human behavior?" If practicising psychoanalysts were asked this question, there is little doubt that their answer would be along the fol­ lo,ving lines: they would say that they were not accustomed to using the term "free" at all, but that if they had to suggest -a criterion for distinguishing the free from the unfree, they would say that a person's

Samuel Butler, Erewhon ( :tvlodem Library edition ), p. 107. 18

FREE-vVILL AND PSYCHOANALYSIS freedom is present in inverse proportion to his neuroticisni; in other words, the more his acts are determined by a malevolent unconscious, the less free he is. Thus they would speak of degrees of freedom. They would say that as a person is cured of his neurosis, he becomes more free-free to realize capabilities that were blocked by the neurotic aflliction. The psychologically \Vell-adjusted individual is in this sense comparatively the most free. Indeed, those who are cured 0£ mental disorders are sometimes said to have regained their free­ dom: they are freed from the tyranny of a malevolent unconscious which formerly ex­ erted as much of a domination over them as if they had been the abject slaves of a cruel dictator. But suppose one says that a person is free only to the extent that his acts are not unconsciously determined at all, be the unconscious benevolent or malevolent? If this is the criterion, psychoanalysts vrnuld say, most human behavior cannot be called free at all: our impulses and volitions hav­ ing to do with our basic attitudes toward life, whether we are optimists or pessimists, tough-minded or tender-minded, whether our tempers are quick or slow, whether we are "naturally self-seeking" or "naturally benevolent'' ( and all the acts consequent upon these things ) , what things annoy us,


whether we take to blondes or brunettes, old or young, whether we become philoso­ phers or artists or businessmen- all this has its basis in the unconscious. If people gen­ erally call most acts free, it is not because they believe that compelled acts should be called free, it is rather through not know­ ing how large a proportion of our acts ac­ tually are compelled. Only the compara­ tively "vanilla-flavored" aspects of our lives- such as our behavior toward people who don't really matter to us-are ex­ empted from this rule. These, I think, are the two principal criteria for distinguishing freedom from the lack of it which we might set up on the basis of psychoanalytic lu1owledge. Con­ ceivably we might set up others. In every case, of course, it remains trivially true that "it all depends on how ,ve choose to use the word." The facts are what they are, re­ gardless of what words we choose for la­ beling them. But if we choose to label them in a way which is not in accord with what human beings, however vaguely, have long had in mind in appl)�ng these b.­ bels, as we would be doing if we labeled as "free" many acts which we know as much about as we now do through modern psy­ choanalytic methods, then we shall only be manipulating words to mislead our fellow creatures.

The Hunianitarian Theory ojf Punishm,ent C. S. LEWIS In England we have had a controversy about Capital Punishment. I do not know whether a murderer is more likely to re­ pent and make a good end on the gallows a few \Veeks after his trial or in the prison in­ firmary thirty years later. I do not know whether the fear of death is an indispensa­ ble deterrent. I need not, for the purpose of this article, decide whether it is a morally permissible deterrent. Those are questions which I propose to leave untouched. My subject is not Capital Punishment in parti­ cular, but that theory of punishment in general which the controversy showed to be ahnost universal among my fellow­ countrymen. It may be called the Humani­ tarian theory. Those who hold it think that it is mild and merciful. In this I believe that they are seriously mistaken. I believe that the «Humanity" which it claims is a dangerous illusion and disguises the possi­ bility of cruelty and injustice without end. I urge a return to the traditional or Retri­ butive theory not solely, not even prima­ rily, in the interests of society, but in the interests of the criminal. According to the Humanitarian theory, to punish a man because he deserves it, and as much as he deserves, is mere re­ venge, and therefore, barbarous and im­ moral. It is maintained that the only legiti­ mate motives for punishing are the desire Reprinted from Res Judicatae, VI, 19531 pp. 224-230, by permission of the publisher.


to deter others by example or to mend the criminal. vVhen this theory is combined, as frequently happens, with the belief that all crime is more or less pathological, the idea of mending tails off into that of healing or curing and punishment becomes therapeu­ tic. Thus it appears at first sight that we have passed from the harsh and self­ righteous notion of giving the wicked their deserts to the charitable and enlightened one of tending the psychologically sick. 'Nhat could be more amiable? One little point which is taken for granted in this theory needs, however, to be made explicit. The things done to the criminal, even if they are called cures, will be just as com­ pulsory as they were in the old days when we called them punishments. If a tendency to steal can be cured by psychotherapy, the thief will no doubt be forced to undergo the treatment. Otherwise, society cannot continue. My contention is that this doctrine, merciful though it appears, really means that each one of us, from the moment he breaks the law, is deprived of the rights of a human being. The reason is this. The Humanitarian theory removes from Punishment the con­ cept of Desert. But the concept of Desert is the only connecting link bet\veen punish­ ment and justice. It is only as deserved or undeserved that a sentence can be just or unjust. I do not here contend that the ques-


tion "Is it deserved?" is the only one we can reasonably ask about a punishment. ,Ve may very properly ask whether it is likely to deter others and to reform the criminal. But neither of these two last questions is a question about justiC:e. There is no sense in talking about a "just deterrent" or a "just cure." We demand of a de.terrent not whether it is just but whether it will deter. We demand of a cure not whether it is just but whether it suc­ ceeds. Thus \vhen we cease to consider what the criminal deserves and consider only what will cure him or deter others, we have tacitly removed him from the sphere of justice altogether; instead of a person, a subject of rights, we now have a mere ob­ ject, a patient, a "case." The distinction will become clearer if we ask who will be quali£ed to determine sentences when sentences are no longer held to derive their propriety from the criminafs deservings. On the old view the problem of fixing the right sentence was a moral problem. Accordingly, the judge who did it was a person trained in jurispru­ dence; trained, that is, in a science which deals with rights and duties, and which, in origin at least, was consciously accepting guidance from the Law of Nature, and from Scripture. 1Ve must admit that in the actual penal code of most countries at most times these high originals were so much modified by local custom, class interests, and utilitarian concessions, as to be very imperfectly recognizable. But the code was never in principle, and not always in fact, beyond the control of tbe conscience of the society. And when ( say, in eighteenth­ century England) actual punishments conflicted too violently with the moral sense of the community, juries refused to convict and reform was finally brought about. This was possible because, so long as we are thinking in terms of Desert, the propriety of the penal code, being a moral question, is a question on which every man has the right to an opinion, not because he follows this or that profession, but because


he is simply a man, a rational animal enjoy­ ing the Natural Light. But all this is changed when we drop the concept of De­ sert. The only t\vo questions we may now ask about a punishment are whether it de­ ters and whether it cures. But these are not questions on which anyone is entitled to have an opinion simply because he is a man. He is not entitled to an opinion even if, in addition to being a man, he should happen also to be a jurist, a Christian, and a moral theologian. For they are not ques­ tions about principle but about matter of fact; and for such cuiquam in sua arte cre­ dendum. Only the expert "penologist" ( let barbarous things have barbarous names ) , in the ligbt of previous experiment, can tell us what is likely to deter: only the psy­ chotherapist can tell us what is likely to cure. It vvill be in vain for the rest of us, speaking simply as men, to say, ''but this punishment is hideously unjust, hideously disproportionate to the criminal's deserts." The experts with perfect logic "�11 reply "but nobody was talking about deserts. No one was talking about punishment in your archaic vindictive sense of the word. 1-Iere are the statistics proving that this treatment deters. Here are the statistics proving that this other treatment cures. VVhat is your trouble?" The · Humanitarian theory, then, re­ moves sentences from the hands of jurists whom the public conscience is entitled to criticize and places them in the hands of technical experts whose special sciences do not even employ such categories as rights or justice. It might be argued that since this transference results from an abandon­ ment of the old idea of punishment, and therefore, of all vindictive motives, it vvill be safe to leave our criminals in such hands. I will not pause to comment on the simple-minded view of fallen human na­ ture which such a belief implies. Let us rather remember that the "cure" of crimi­ nals is to be compulsory; and let us then watch how the theory actually works in the mind of the Humanitarian. The immediate

648 starting point of this article wa.s a letter I read in one of our Leftist weeklies. The au­ thor was pleading that a certain sin) nmv treated by our lmvs as a crime, should henceforward be treated as a dLsease. And he complained that under the present sys­ tem Hie offender, after a term in gaol, was simply let out to return to his original envi­ ronment where he would probably relapse. \'!hat he complained of was not the shut­ ting up but the letting out. On his remedial view of punishment the offender shouId, of course, be detained until he was cured. And of course the official straighteners are the only people who can say when that is. The first result of the Humanitarian theory is, therefore, to substitute for a deftnite se�­ tence ( reflecting to some extent the com­ munity's moral judgment on the degree of ill-desert involved) an inde,llnite sentence terminable only by the word of those ex­ perts-and they are not experts in moral theology · nor even in the Law of Na­ ture-who inflict it. · which of us, if he stood in the dock, wou:d not prefer to be tried by the old system? It may be said that by the continued use of the word punishment and the use of the verb "inflict" I am misrepresenting Hu­ manitarians, They are not punishing 1 not inflicting, only healing, But do not let us be deceived bv a name. To be taken without consent fro;n my home and friends; to lose my liberty; to undergo all those assaults on my personality which modern psychother­ apy knows how to deliver; to be re-made after some pattern of "normality" hatched in a Viennese laboratory to \Vhich I never professed allegiance; to know that this process .,,,111 never end until either my cap­ tors have succeeded or I gro\vn \'1tise enough to cheat them with apparent success-who cares ,vhether this is called Punishment or not? That it includes most of the elements f�Jr which any punishm.ent is feared-shame, exile, bondage, and years eaten by the locust-is obvious. Only enor­ mous ill-desert could justify it; but ill­ desert is the very conception which the

PUNISHMENT AND RESPONSIBILITY Humanitarla!.1. theory has thrown over� board. If we tum from the curative co the de­ terrent justi...'ication of punishment we shall find the new theory even more ,:\Th.en you punish a man i-n terrorem, make of him an ""e.xample" to others, you are ad­ mittedly using him as a means to an end; someone else's end. T:1:S, in itself, \Vould be a very wicked thing to do. On the classical theory of Punishment it was of conrse jus­ ti£ed on the ground that the 1nan deserved it. That was assumed to be established be­ fore any question of '"inaking him an exam� ple" arose. You then, as the saying is, killed two birds with one stone; in the process of giving him wh&t he deserved you set an ex� ample to others. But take away desert and the whole moral:ty of the punishment dis­ appears. vVhy, in Heaven's name, am I to be sacrificed ;:o the good of society in this way?···-unless, of course, I deserve it. But that is not the worst. If the justi­ fication of exemplary punishment is not to be hased on desert but solelv on its efficacy as a deterrent 1 it is not absoiutely necessar)' that the man we punish should even have committed the crime. The deterrent effect demands t.\at the public should draw the moral, "If we do such an act we shall suffer like that man." The punishment of a man actually guilty whom the public think inno­ cent will not w:ill not have the desired effect; the punishment of a man actually innocent will, nrovided the nublic think hlm guilty. But every mode,;, State has powers which make it easy to fake a trial. \Vhen a victim is urgently needed for ex­ emplary purposes and a guilty victim can­ not be found, all the purposes of deter­ rence will be equally served by the punish­ ment ( call it "cure" if you prefer) of an in­ nocent victim, provided that the public can be cheated into thinking him guilty. It is no use to ask me why I assmne that our rulers will be so wicked. The punishment of an i....,nocent, that is, an undeserving, man is wicked only if we grant the traditional vievv that righteous punishment means de-


served punishment. Once we have aban­ doned that criterion, all punishments have to be justified, if at all, on other grounds that have nothing to do with desert. Where the punishment of the innocent can be jus­ tilled on those grounds ( and it could in some cases be justilled as a deterrent) it will be no less moral than any other pun­ ishment. Any distaste for it on the part of a Humanitarian will be merely a hang-over from the Retributive theory. It is indeed, important to notice that my argument so far supposes no evil inten­ tions on the part of the Humanitarian and considers only \Vhat is involved in the logic of his position. My contention is that good men ( not bad men) consistently acting upon that position would act as cruelly and unjustly as the greatest tyrants. They might in some respects act even vmrse. Of all tyr­ annies a tyranny sincerely exercised for the good of its victims may be the most oppres­ sive. It may be better to live under robber barons than under omnipotent moral busy­ bodies. The robber baron's cruelty may sometimes sleep, his cupidity may at some point be satiated; but those who torment us for our own good will torment us with­ out end for they do so with the approval of their o,vn conscience. They may be more likely to go Heaven yet at the same time likelier to make a Hell of earth. Their very kindness stings with intolerable insult. To be «cured" against one's ,vill and cured of states which we may not regard as disease is to be put on a level with those who have not yet reached the age of reason or those who never \Vill; to be classed with infants, imbeciles, and domestic animals. But to be punished, however severely, because we have deserved it, because we "ought to have known better," is to be treated as a human person made in God's image. In reality, however, we· must face the possibility of bad rulers armed with a Hu­ manitarian theory of punishment. A great many popular blue prints for a Christian society are merely what the Elizabethans called "eggs in moonshine" because they


assume that the whole society is Christian or that the Christians are in control. This is not so in most contemporary States. Even if it were, our rulers would still be fallen men, and therefore, neither very wise nor very good. As it is, they will usually be un­ believers. And since wisdom and virtue are not the only or the commonest quali­ fications for a place in the government, they will not often be even the best unbe­ lievers. The practical problem of Christian politics is not that of drawing up schemes for a Christian society, but that of living as innocently as we can with unbelieving fel­ low-subjects under unbelieving rulers who will never be perfectly wise and good and who will sometimes be very wicked and very foolish. And when they are wicked the Humanitarian theory of punishment will put in their hands a finer instrument of tyr­ anny than wickedness ever had before. For if crime and disease are to be regarded as the same thing, it follows that any state of mind which our masters choose to call "dis­ ease" can be treated as crime; and compul­ sorily cured. It will be vain to plead that states of mind which displease government need not ahvays involve moral turpitude and do not therefore always deserve forfei­ ture of liberty. For our masters will not be using the concepts of Desert and Punish­ ment but those of disease and cure. vVe know that one school of psychology al­ ready regards religion as a neurosis. ·when this particular neurosis becomes inconven­ ient to government, what is to hinder gov­ ernment from proceeding to "cure" it? Such ''cure" will, of course, be compulsory; but under the Humanitarian theory it \vill not be called by the shocking name of Persecu­ tion, No one will blame us for being Chris­ tian, no one will hate us, no one will revile us. The new Nero will approach us with the silky manners of a doctor, and though all will be in fact as compulsory as the tu­ nica molesta or Smithfield or Tyburn, all will go on within the unemotional thera­ peutic sphere where words like "right'' and ''wrong" or "freedom" and "slavery" are



never heard. And thus when the command very essence involves the recognition of is given, every prominent Christian in the guilt and ill-desert in the recipient. ll crime land may vanish overnight into Institutions is only a disease which needs cure, not sin for the Treatment of the Ideologically Un­ which deserves punishment, it cannot be sound, and it will rest with the expert gaol­ pardoned, How can you pardon a man for ers to say when ( if ever) they are to re­ having a gumboil or a dub foot? But the emerge, But it will not be persecution. Humanitarian theorv wants simplv to abol­ Even. if the treatment is painful, even if it ish Justice and substitute Mercy f�r it. This is life-long, even if it is fatal, that will be means that you start being "kind" to people only a regrettable accident; the intention before you have considered their rights, was purely therapeutic. Even in ordinary and then force upon them supposed kind­ medicine there were painful operations and nesses which they in fact had a right to re­ fatal operations; so in this. But because fuse, and finally kindnesses ,vhich no one they are "'"treatment;' not punis11ment,. they but you vviil recognize as kindnesses and can be criticized only by fellow-experts and which the recipient will feel as abominable on technical grounds, never by men as men cruelties. You have overshot the mark and on grounds of justice. Mercy, detached from Justice, grows un­ This is why I think it essential to op­ mercifu L That is the important paradox. As pose the Humanitarian theory of punish­ there are plants which will flourish only io ment> root and branch, wherever we en­ mountain soil, so it appears that 1\ifercy \Vill counter it. lt carries on its front a sen1- Hower only when Et grows in t.he crannies blance of mercy which is wholly false. Thac of the rock of Justice: tran.spb.nted to the is how it can deceive men of good will marshlands of mere Humanitarianism1 it The error began, perhaps, with Shelley's becomes a man-eating o;,:veed, all the more statement that the distinction between dangerous because it is still called by the mercy and justice was :invented in the same name as the mountain variety. But courts of tyrants, It sounds noble, and was we ought long ago to have learned our les­ indeed the error of a noble mind. But the son. We should be too old now to be .de­ distinction is essential The older view was ceived by those humane pretensions which that mercy "tempered" justice, or { on the have served to usher in every cruelty of the highest level of all) that mercy and ju.stice revolutionary period in ,v1tlch we live. had met and kissed. The essential act of These are the "precious balms" ,vhich ,.,'1Il mercy was to pardon; and pardon in its 'break our heads."

The Humanitarian Theory of Punishnient

J. J. C. SitlART

I wish to discuss one or two logical points whfoh arise out of C. S. Lewis's article on "The Humanitarian Theory of Punish­ menf'1 . . . Lmvis has got at cross purposes with him.self in a way very similar to that in which Intuitionists and Utilitarians in moral philosophy have often entangled themselves when arguing about the nature of obligation. Consider these two sorts of questions: ( 1 ) "Ought Smith to leave his wife?" . "Ought I to return this book?" "Ought I to drive on this side of the road?" and (2) "Are our present marriage custon1s for the best?" "Is the institution of promise making a good one?" "Ought we to have a rule of the road and if so what?': I shall call the first sort 0£ questions ''first-order questions" and the seco:id sort of questions "second­ order questions". \Ve can now say, roughly, that the Intuitionists were right when they dealt with the first-order questions, but hopelessly at sea when they dealt with the second-order questions, while the Utilita­ rians were able to talk a great deal of sense when they discussed the second-order questions but were most strained and un­ plausible when they dealt with the first­ order questions. The dispute in moral phi­ losophy was so £rnitless because each had ( roughly) the right answer to one sort of ReprinteC: from Res Jud{citae bv " permission 9f the author a..od Res Judicitae. 1 6 Res Judicatae, 224-30,

question but not to the other. No one in his senses would weigh up the social conse­ quences of returning a book he has bor­ rowed: there is a moral rule that plainly covers the case and so he knows immedi­ ately what he should do. It is about mies and social institutions that we ask the Utili­ tarian type of question, not about individ­ ual actions. Of course in exceptional cases we have to think as Utilitarians about individual ac­ tions. This is eithe: when rules conflict or when there is no rule that covers the c-ase in question. But by and large we just "see" what to do in the individual cases ( we have been brought up so to do ) : it is only when we co!:isider the effect of certain rules or institutions on society as a whole) \vhen we consider modifying or augmenting these mies and institutions, that the Utilita­ rian pattern of thought becomes appropr i ­ ate. Philosophers like Butler and Kant are at their happiest when discussing how we deal with the first- order type of questioo, those like Bentham and Mill when dealing with the second-order type of question. ( Though note that Butler in one place2 seems to say that God made our con, sciences as He did because He is a Utilita­ rian> even if we must not be. And I !llVself believe, rather heretically, that only a slight rephrasing of Kant is needed in . order to turn him into a Utilitarian abollt rules.) = Dissertation upon the Nature of Virtue, § 8.



Bentham's Utilitarian methods of argument work smoothly when he is considering rules ( legislation) but he is quite silly where he tries to talk in the same way about indiv.:d� ual actions. The point, then, is this1 that it is not a question of inh.:.itionism or utilita­ rianism but of both ( in different p:aces ) , The actions ,v! we ''see» to be right are those which con1e under rules we have been trained to obey: the justification 0£ the rules, but not, in general, of the indi­ vidual actions, is utilitarian. I have sketchet! out the above theory of morals ( which you can find more fully worked out in Touhnin's book The Place of Reason in Ethics) becat:Se i t appears to me tbat theories of punishment have got at cross purposes in a precisely si�ilar ,vay to t':iat in which the intuitionists and utilita­ rians get into cross purposes about right and wrong, From the point of view of the legisla­ tor, we ask: ''Is this the best punishment to assign for this type of offence?" It seems to me that the only way in which this ques� tion can be rationally discussed is the utili­ tarian way: that is by considerhig d1e consequences for society of adopting or not adopting the penal law in question. vV:oat otber type of argument is relevant? Admit­ tedly one could appeal to Scripture, but the Nei,,v Testament was not intended as a text-book of penology, and some of the penal ideas of the Old Testa:nent are bar" barous. Certainly if we knew tbat God had said that such-ant!"such was the law we should adopt we should be foolish not to adopt it. But how does God how that it is the best law? God is rational and must have argued ratio,ially to His decision. How else, then, than by arguing in the way we should, if we vvere rational, that is1 in the U tilita:rian way? ( Cf. Butler again. ) There is sometbing else that Lewis might put in the place of Utilitarian argument: an appear to the Law of Natme. I do not know what this is. But I think I know what the use of the e1rpression ('Law of Nature"' is. I t Ls this: "this is the Law of Nature" =


"this is the rule that ought to be adopted", said by someone who wishes to disguise his own dog1natism and to conceal the fact that he is either unable or too lazy to search for a rational ( i.e. a Utilirarlan ) justification of the proposed meass.:re. From a U!:ilitarian point of vfovv, then1 we discuss a measure by asking "Will this n.1:easure or will some alternative one tend most to promote the welJ-being of society?" If the proposed measure is a penal law there seem to be only three ways in which it can be of value: (1) To deter people; ( 2 ) To protect society by eliminat­ ing o: removing criminals; ( 3 ) To reform t'.1e criminal. The first two of these are bv far the most h:nportant I t is :not a1ways p�ssible to refom1 the criminal. And I should say that ( l) is of greater importance than ( 2). Lewis discusses ( 1) and ( 3 ) but ignores ( 2 ) . There may be other \vays in \Vhich the institution of punishment may benefit so" ciety and which could be cited to justi!y it. I do not know of any. It might be argued that punishment satisfies the desire of cer­ tain members of the society for revenge. Howeve,r} the desire fo: revenge is some­ thing which is perhaps better left unsa­ tisfied. It is d:fficu !t to believe that society wouid not be happier if it thought less abo-;.;t revenge. 1vforeover I do not see how the principle of revenge itself codd possi­ bly be justiilec!. "If we adopt the principle 'An eye foe an eye and a tooth for a tooth' we \Vill make society happier."' How? We see then that Utilitarian considera­ tions are relevant in discussing what !?enal legislation we should adopt. But just as in the case of rightness, analysed earlier in thLs paper, we fb:1d a totally different situa" tion when \Ve come to the individual ac­ tion1 the action of the judge or magistrate. The judge or magistrate must not argt:e as a Utilitarian; save per acc-idens 'W:lere the law leaves some margin for choice1 when decit!ing what punishment to impose. The just punishment for murder is death. That


is, death is the punishment laid down by law. It is totally beside the point for the judge to argue about ,vhat action, in this particular instance, would promote the greatest general happiness. I now make the following suggestion. A lot of what Lewis says is perfectly true. As judges or magistrates we must not think as Utilitarians. But this has not the slightest bearing on the question of whether legisla­ tion should or should not be governed by Utilitarian criteria. Lewis lays stress on the

concept of Desert, and it is here, in the

thinking of the judge or magistrate, that this concept comes in. The concept of De­ sert is quite inapplicable so far as the thinking of the legislator is concerned. O r ­

dinarily we know what is meant by "the deserved punishment". It is that laid down by la\v, But how can "desert" have a mean­ ing when ,ve discuss what punishment the law ought to lay down? If we try to apply the idea of desert here we are either driven

back on to Lewis's personal preferences ("I


should like to see murderers hanged", say) or we have to fall back on some crude equation of punishment with amount of damage done: an eye for an eye. Why the damage-retribution equation should be thought a sound principle of legislation I do not know. I do not see how it could pos­ sibly be justified. Why should society be happier if we adopt this principle? Indeed it is quite easy to see that society will be happier if we do not adopt this principle. To sum up: Le,vis shows quite clearly that we do not always think about punish­ ment in the Utilitarian way. My reply is that it is when we think of ourselves in the situation of magistrates that we are quite right not to think as Utilitarians. In this sit­ uation \'Ve are concerned with the first­ order questions. But it is in considering the penal laws themselves, in considering the second-order questions1 that we must think as Utilitarians. Levi!is's argument derives a great deal of its plausibility from confusing the first-order and second-order questions.

Determinism and Jlforal Perspectives ELIZ.ABETH L. BEA.RDSLEJT Can determinists £nd a satisfactory ration­ and blame? On this ale for moral question, determlnists themselves have long been divided. Although the affirma­ tive answer has enjoyed the status of a majority opinion, the negative anS\ver has at times found very eJfective support. The force of the negative answer emerges clearly in certain recent writings� in which writers sympathetic to determinism vigor­ ously defend th e thesis that determinism removes from the concepts of moral praise­ worthiness and blameworthiness all legit­ imate application whatsoever.i The negative answer to the question posed here is unsatisfactory, I think; but in some ways it is preferable to the affirma­ tive answer as the latter is usually given Reprinted from Philosophy and Plumomerwlogi­ cal Research, XXI. 1960� pp. 1-20, by permission of the author and Philosophy mul Phenomenologi­ cal, i See Paul Ed'n->ards, "Hard and Soft De:ennin­ is:n," and John Hospe:s1 "" !>foans: This Fre e ­ dor:11'", bofa in Sidney Hool,; Determinism and Freedom {New York University Press, 1958) ; 2.lso

VV. I. Matson, "On the frrelevance of Free-will to Moral Responsibility/' Mind) Vol. LXV ( .1956 ) , pp. 489-497. Although these \.vrlters frame their

argu:inent more explicitly in terms of the concept of moral Tesponsibility than in tenns of moral praise� worthiness and blamewor.lliness, the application to the !atter concepts is clear, .Matson's chief thesis,

moral responsibility no more successfolly de­ termi:tlsm can, \\'l:l not be dealt y:Hh in the present paper; the part of his ;uticle ,vhich bears most di­ rectly on what I shal1 have t-0 say is found in sec­ tions 2 and 3, 654 that libertarianism can validate the concept of 1

and suppo::ted. In this paper, I shall argue that judgments of moral praise and blame, affirmative as well as negative, can be made within the framework of detemrln­ ism, provided that we accept a more com­ plex account of these judgments and their foundations than is supplied or assumed. I shall maintain that judgments concerning the presence or absence of moral praiseworthiness and blameworthi­ ness are made from several different stand­ points, · which I shall call "moral perspec­ tives." :My pri1nary purpose is to show hm,v an understanding or these perspectives and their relations can contribute substan­ tially toward relieving the tension widely felt ( eveu by some who are reluctant to admit it) to exist between determinism and certain of our basic ethical concepts. The terms "praise" and Olame'> will be used here ,vith the meaning of "moral praise� and ""moral bla..'T..e." Praise and blame will be treated as correlative con­ cepts such that, for everything that is said about one, a corresponding statement about the other could be made, though it will usually be unnecessary to make it. The term "afHl'mative judgment of praise" will be used to refer to any explicit attribution of praiseworthiness to a person. A "nega­ tive judgment of praise" is an explicit de­ nial t.1-iat a person is praiseworthy. T}1e general term "judgment of praise" will re­ fer b'.ldifferently to either an affirmative or

DETERMINISM AND MORAL PERSPECTIVES a negative judgment of praise, and simi­ larly for "judgment of blame." Judgments of praise and blame ,;,,111 be treated h.ere as assertions which are true or false, and not as acts which may be useful or useless to perform.



sents no particular problem for the deter­ miaist It is trc.e that certain libertarians have apparently seen something pro­ foundly incongruous L'1 the application of any normative predicates at all to the con­ stituent parts of a determined universe; hut tlris line of t110ught has persuaded so few that it may be disregarded. Much less harmony p,evails among philosophers who have reflected on the relation between determinism and the con­ cepts of moral praise and blame. Liber­ tarians: of sC-ourse1 maintain that because foe truth of determinism would invalidate affirmative judgments of praise and blame, determinism is false, and criteria for praise­ \Vorthiness and blameworthiness must in­ clude the requirement that an agent should have performed his act "freely." Though determinists are united in rejecting the conclusion. of the libertarian argument as false, they cH.:fer sh arply concerning the acceptability of the conditional premise. Among leading determinists who be­ lieve that valid affirmative judgments of praise and blame can be made, a fairly clear account of the criteria for praise­ worthiness and blameworthiness seems to have emerged. I shall call those who sub­ scribe to this account ''Group I determin­ ists." Details of the account vary, but a substantial area of agreement remains. It is commonly held that if an agent has acted \Vrongly, without external constraint ("voluntarily" ), "1thout ignorance of rele­ vant facts, and from a motive or because of a trait that is undesirable, then, and only then, the agent deserves blame for his act.2 Sb1ilar conditions are held to govern praisev;rorthiness, 3 Group I determinists deny that there is anything here to conflict with the truth

Before discussbg judgme11ts of praise and blame, it \\111 be helpful to consider briefly cc::tain moral judgments of a different kind. The standpoint from which we affirm or deny that acts are objectively right or wrong I shall call the "perspective of ob­ jective rightness or wrongness." A judg­ ment of objective rightness or wnmgness is a judgment made about an act, not an agent; and it does not carry with it any irnplication about the praiseworthiness or blameworthiness of an agent. Statements like "Smith's act was objectively right, but he deserve.;;; no praise for it" not only are self-consistent, but are often true; objec­ tively right acts can be committed inad­ vertently, or from reprehensible motives. The judgment that an act is objec­ tively right furnishes insufficient evidence for a. judgment that its agent is praise� worthy, because certain key facts concern­ ing foe causal antecedents of the right act remain to be supplied. The objective rightness or v.rrongness of an act does not depend in any v,ray on its causal anteced� ents, but on other coILeiderations, such as its consequences (for teleologists), or its harmony with the v.1ll of God, moral rules, or the like ( for formalists ) . It is therefore appropriate to call this perspective a "non­ causar' one, for it takes no account of whether an act had ,;ausal antecedents of one kind rather than anoL1ier1 or indeed I have founG the presentation of this general had completely determining causal ante­ position by P, H. NowellwSmith in Ethics ( London, 1954 ) particularly helpful he:re. ced=ts at all. ;; Note that the terms ''blameworthy" and Most philosophers, I think, would ";,raisewortl'iy" have not been defined here and agree that the use of the moral perspective that no definitions of these terms wit be offered of objective rightness or ,vrongness pre- in ::his paper. · 2




of determinism. Yney point out that those for a voluntary act and not an involuntary who make judgments of praise and blame one, when we know full well that even must indeed attend to several kev factors the voluntary act can be traced back to among the causal conditions that produced causes-environmental or hereditary-be­ the acts whose agents are judged, But any longing to a world the agent never made? I believe that there are eiements of other causal conditions that may have been present, and, in particular, antecedents of truth in each of these brands of detem1in­ antecedents, are to be completely disre­ ism, and I sbal! try to show that this is the garded. Moral praisers and blamers, on case. this view, are s!mply not concerned v;rith the nature, or even the existencc1 of suc:J. additional facto:-s. Determinism is thus THE PERSPECTIVE OF fully compatib1e with attdbutions of praise­ MORAL WORTH worthiness and blameworthiness. 4 To deterlO'inists of a second group­ Surely there is no doubt that the conditions ''Group II deterrninists';-this acco1:..11.t for praisevmrthiness and. bla.rnewort:liness seems seriously ove,simpli'led. 5 They con­ set forth in the Group I detem1inist's tend that the same reasoning which leads account do in fact constitute one important us to withhold praise and blame from and familiar standard accorcbg to which agents whose acts were commit:ed invo l ­ we make judgments of praise and blame. untarily will, when combined v,ith the the­ It is highly convenient to introduce a spe­ sis of determinism, lead on inexorablv to ciaJ term for the cha-:acteristic of moral t.lie conclusion that no one ever des�rves value t.l,at may be said to belong to an praise or blame for anything. They are agent who has performed an act that meets haunted by the k.,owledge that many of the conditions specified. I shall say t.'iat an the causal antecedents of acts have not agent has "positive moral worth'' :if and been investigated by those who mete oc:t only if he bas acted rightly, voluncarily, praise and blame on the grounds specified with luiowledge of releva11t facts, and above; and most particularly they are frolT! a desire that is bad in its situation.6 haunted by the la1owledge that not all of The term "moral worth" vvrn be µsed to the causal antecedents of voluntarv acts refer to either positive or negative moral are voluntary acts. Thus they con1e "to be­ worth ind:fferently, and the standard by lieve that no distinction between "volun­ which agents are judged to have moral tary" and "involuntary" acts th-at a deter� worth ( positive or negative) will be called rninist can consistently make can sustain the "standard of moral worth." Elsewhere 7 the moral weight that it must bear if we I have discusseG certain features of the arc to judge men praiseworthy or blamew concept of moral worth in some detail1 and worthy. How, they ask, could we ever be have indicated hov;r conditions for the justifled in blaming or praising someone I prefer to fommlate this last condition in • 1fany would argue, of course1 that determb­ :erms of a "desire" than of a "trait/' ism is much more than merely "compatib!e" with be�'l.use we sometfrnes make judgments of this kind these judgments, since we cannot speak of an without having sufficient evidence to ascertain the agent's act as "his" act or as arising "from" a :;:resence of a trait; but this point is not of central 1

ir.otive or trait, unless determinism is assumed to be true. This argument wil: be deliberately set .aside here. :a My distinction between "Group I" rm2. "Group ll" determinism is plainly very .i:;im!lar to Edwards' d!stinction behveen "soft" nnd "hard" determin� ism. See Edwards, op. cit., and also Edwards and Pap, A li-fodem fotroduction- to Philosophy (Free Press, 1957), p, 380.

importance here. 1 "Moral \Vorth and Mora� Credit," The Philo• sophical Review, Vol. LXVI ( 1957), pp. 30'i-328. In the present t:'eatment I have left the term "moral worth" undefined, and 1 have also intro­ duced it here to refer to any attribute of agents who t:re praiseworthy or blar.1'3worthy for acts rather than to refer to a.1 attribute 0£ acts them� selves.



presence of degrees of moral worth may we should avoid the confusion of believing that persons who happen to form part of be set up. A "judgment of moral worth," which fae pleasant or unpleasant gestalts just may be affirmative or negative2 is a judg­ mentioned deserve praise or blame for ment in ,vhich moral worL11 is asserted to what they do. Because the crucial distinc­ be present or absent \:Ve must of course tion between voluntary and involuntary· distinguish between a negative judgment acts is bound to collapse in the end, no of positive n1oral worth (" Agent A is not one ever deserves praise or blame. Per­ morally wo:c+Jiy for act A") and an affirma­ haps judgments of praise and blame are tive judgment of negative :noral worth niade frorr. the perspective of moral worth, ("Agent Y is morally unworthy for his act but they should not be.8 B" ) . To this the Group I detenninist will I shall call the standpoint from which :reply that, since the conditions for "moral we make judgments of moral worth the wortlr were originally taken directly from "perspective of moral worth.'' This is an analysis: of conditions for praiseworthi­ plainly not a wholly no11-causal perspec­ ness and blameworthiness, it is highly tive, as is the perspective of objective arbitrary\ to say the very least, to attempt rightness o:r wrongness. Because some ( a to purge judgments of moral worth of all strictly limited set) of the circumstances association wHh judgments of praise and causally relevant to the performance of an blame, �,foreove:-1 he will continue, the act are taken into account when the moral assertion that human beings have "'feel­ worfo of its agent is being judged, this ings" which are merely "positive" or "neg­ perspective may accurately be called a ative/' when they encounter persons ex­ "causally limited" perspective. The factors hibiting positive or negative moral worth, taken into account in makbg judgments is decidedly misleading. The "feelings" from this moral perspective will b e termed referred to consist of deflnite reactions of a specific sort, to which are added, for most the '\vorth-determ!ning"' factors. Group II determinists are likely to feel moral judges, quite explicit rellective con­ that the introd;:ction of the term "moral victions, Human beings feel-and reserve \VOrt...1-i" is unobject:onable) and perhaps -a very special kind of approval and dis­ even usefol, provided that judgments of approval for those members of their species moral wo:i:h are not held to imply judg­ who perform acts tha� have certain salient ments of praise or blame. TI1inkers of this features. Furtherrnore, the majority of group may be disposed to admit that hu­ those who have ::-e::3.ected on the matter' man beings do indeed have a strong psy­ seem to have been convinced that ap� chological tendency to experience positive proval and disapproval of this specia.l kind feelings when co!lfronted by the gestalt are reactions to '-Vhich the persons in ques­ agent-perfonning -act-under-conditions-for­ tion have a morally justified claim. It is positive-moral-worth, and to experience this claim which is put for+J1 in affirmative negative feelings when confronted by the judgments of praise and blame. In view of corresponding negative gestalt. They may these considerations; a heavy burden of contend that, since these feelings cannot � It is conve:1ient to speak of the point a.t :ssue he rationally justified, human beings had between the Group 1 and Group II detenninisto; as better trv to elt-nina!e them from their concerned primarily with the validity of affirma­ tive judgments of praise and blame; an'1thout limits of any kind. There­ fore we may call this a "causally unlimited" perspective. As a moral perspective it is, of course, strikingly different in some re­ spects from the others that we have ex­ amined. Judgments made from the perspective of moral worth and the per­ spective of moral credit are judgments of discrimination. This is obvious in the case of comparative judgments, but it is also true of noncomparative ones. Our interest in knowing that X possesses positive moral worth for his honest act, and Y negative moral credit for his cowardly one,. stems in large part from the fact that there are honest acts whose agents do not possess positive moral worth, and cowardly acts whose agents earn no negative moral credit. Judgments of praise and blame based on moral worth and moral credit are answers to questions which can in principle be answered either affirmatively or nega­ tively. This is not true of judgments made from the perspective of ultimate moral equality. Here all are on the same moral footing: none has any ultimate claim to praise or blame, and the judgments made from this perspective are all negative. No matter what acts a person has pedormed, all that we can say of him from this final moral perspective is that he deserves no praise for what he has done, or that he deserves no blame. The statement "X is not ultimately praiseworthy for A'' is a negative judgment of praise made from the perspective of ultimate moral equalit)'1 whereas "Y is not ultimately blameworthy for B" is a negative judgment of blame made from the same perspective. Judgments of praise and blame made from this perspective will



here be limited in scope to persons whose than the differences, and they are never acts have earned for them moral worth or more signillcant than they become ,vhen moral credit That :s to sayi the statement we are in danger of assuming that the "X is not ultimately praiseworthy for A" differences tell the whole factual story. So will be nermissible if and only if A is an it is with moral appraisals of human beings. act for �vhich X possesses eifher positive For the whole moral story, judgments of moral worth or positive moral credit. And praise and blame based on ,noral worth the truth�condition for this statement can and moral credit need to b e supplemented be stated very briefly : the statement is by judgments made from the perspective hue if and onlv if A has ultimate causes. of ultimate moral equality. Similarly, "Y .is �ot ultimately blameworthy Because this is true, we are justified for B" is per,nissible if and only if B is an in regarding the perspective of ultimate act for which Y possesses negative moral moral equality as a genuinely "mocal" per­ \\'Orth or neg�tive moral credit, and true spective, even though it eradicates ::noral ·discriminations. The · knowledge that when if and only if B has ultimate causes. But, if determinism is true, we know persons are viewed in relation to the of any event that it has ultimate causes, ultimate causal factors of their behavior and ""·ve know this without any specific moral discriminations no longer apply to investigation. The behavior of all men is them is a piece of moral knowledge, at causally determined: and the nature of least in being knowledge about moral mat­ what we have called the "ultimate" causes ters, It is curious that as we go from a is equally unknown in each case. This causally lin1ited perspective to a causally eradication of all 2.istinctions in the causal extended one \YB increase our pov.rer to status of acts erases all distinctions in the make moral discriminations-, whereas when moral status of their agents. Therefore in we come to a causally unlimited perspec· one way it can never be news that Brov;'ll tive these moral discriminations stop alto­ does not ulti.rnately deserve praise for his gether , But the knowledge that this is so kind deed, or that Robinson does not is moral knowledge, and it has important ulti.'Ilately deserve blame for his unkind bearings on the rest of our mora1 one. knowledge. The relation that holds between the In another wa•:r, however, these asser­ tions are news, an·d !.mnortant news. The perspective of ultimate moral equality and fact that Erm� and Robinson are uiti· the other motal perspectives from which mately moral equals is a vital part of the judgments of praise ·and blame are made whole moral truth about them, Compare is analogous jn certain ways to the relation the _ situation for a factual aecount.11 In bet\veen the perspective of moral credit factual descriptions of human beings we and that of moral worth. Judgments based · are interested in the qualities in which on moral credit, as we have seen, set they diffe!\ to be sure; but we are also Hmits to judgments based on moral worth. interested in the qualities in -..vhich they Similarly;. the knowledge that are a1:Gce. For some purposesj and in some beings can be view"ed fron1 a perspective conte."\".ts, the similarities may be 1egiti­ which will show them to be morally equal mately disregarded; but this does not mean will remind t.1:iose who make judgments that they can always be left out of ac­ based on moral worth and moral credit count Someti...""11es they a:re more signilicant that these judgments of moral inequality do not tell the whole story about the in­ n By this :a.tanner of speaking I do not mean to dividuals being judged. This knowledge, rule out the possibility that a naturalistic account in tum, will affect tl:ie attitudes of those of the meaning of basic ethical terms can be who have it: they ,,111 regard themselves given.


and each other with more tolerance than before. Feelings of admiration, contempt, guilt, and pride, will all be experienced more moderately by those wbo know that no man is ever the first cause of evil or good deeds, or finally responsible for win­ ning or losing when confronted by moral odds. But this is not to say that such feel­ ings will not be experienced at all, or that they should not be. For the perspective of ultimate moral equality cannot give us the whole truth about the praiseworthiness and blamewor­ thiness of human beings either. The fact that X has negative moral worth for his · act, or that Y has positive moral credit for his, is not cancelled by saying that X does not ultimately deserve blame, or that Y does not ultimately deserve praise. vVe value in a special way those whose acts meet the standards of moral worth and moral credit, and this is something that we cannot change. As Spinoza saw, it is true­ even in a determined universe-that "we desire to form for ourselves an idea of man upon which we may look as a model of human nature."18 The idea of a man who performs a right act voluntarily, know­ ingly, and from a good desire, and the idea of a man who, when confronted by odds, can still do these things-these are the models we have formed. Conformity to these patterns is what we regard as worthy of praise, and deviation from them in certain ways is what we regard as worthy of blame. We cannot feel about persons who thus conform or deviate as we do about animals or inanimate objects which measure up or fail to measure up to certain other standards. AU this being so, judg­ ments of praise and blame based on moral \VOrth and moral credit are not only legiti­ mate but vitally necessary parts of moral discourse. They are answers to questions that we cannot help asking. The full moral truth about a man and his act,. then, might· run as follmvs : that he deserves a low degree of praise for it by � B. Spinoza, Ethics ( Oxford Press, 4th ed., 1930 ) , p, 179. 1


the standard of moral worth, a high degree of praise for it by the standard of moral credit, and ultimately no praise for it when he is judged from the perspective of ultimate moral equality. There is no reason why the three statements cannot 9e true simultaneously. Also, these perspectives seem to be genuinely coordinate, and com­ plementary: we need them all. And, if we distinguish bet\veen moral perspectives, we shall be able to avoid the doubling of metaphysical perspectives which Kant found necessary in order to reconcile cau­ sality and morality. It is easier to regard a man as blameworthy from one point of view but not from another than to say that his act is both caused and uncaused. The acquiring of moral wisdom7 at least as far as moral appraisals are con­ cerned, does not consist only in learning how to make sound judgments from each moral perspective. It consists also in learn­ ing under what circumstances each of the moral perspectives should be used-a large and fundamental problem that cannot be dealt with bere. 19 Here let us note only that most of the questions about the praise­ vrnrthiness and blameworthiness of human beings that are actually asked are questions to which the appropriate answer is a judg­ ment based on moral \vorth or one based on moral credit. \Vriters on ethics 20 have pointed out that we feel something pecul­ iarly objectionable in an attempt by a wrongdoer to exculpate himself on the ground that all }:tis acts were caused and therefore he deserves no blame. Here an inquiry into his blameworthiness is launched from one moral perspective and a reply is made from another. But moral 19

One important question to consider in this connection is whether it is ever justifiable to employ the perspective of ultimate moral equality when thinking of one's own future acts and their moral status. Some reflections which bear on this question ( though they are not e�'Pressed in the language of the present paper) are offered by H. Fingarette in his interesting article "Psychoanalytic R e -Evaluation," Philosophy and Phenomenological Research, Vol. }."VI ( 1955-1956 ) , pp. 18-29. 20 See, for example, Nowell- Smith, op. cit., pp. 297-300.

DETERMINISM AND MORAL PERSPECTIVES •perspectives, hmvever coordinate, are cer­ tainly not interchangeable. Questions -about praiseworthiness or blameworthiness should be answered from the perspective from which they are asked, whenever it is possible to tell what this is. Sometimes it will be appropriate, and even very desir­ able, to add to this answer a judgment from another moral perspective; but often it will not be. Particular caution ·must be exer­ cised in advancing judgments made from the perspective of ultimate moral equality. These are illuminating. and even inspiring, when made in the right context, and by those who knmv how to make accurate discriminations by the standards of moral wortb and moral credit. Otherwise they are apt to seem shallow, and somehow sen­ timental, or cheap. 2 1

SOME OBJECTIONS AND REPLIES The account which has been given of the perspective of ultimate moral equality and its relation to our other perspectives of praise and blame seems likely to arouse objections from all sides. Three of these appear to me to be particularly striking, and in this concluding section I shall try to reply briefly to each one in tum. ( 1 ) The first objection that I shall consider is one that will be raised by Group II determinists. Some of what has been said in the preceding will presum­ ably be acceptable to members of this group; but they will want to know. bow it can be maintained that the perspective of ultimate moral equality is merely one of several perspectives from which agents are appraised. It is rather the moral perspec­ tive, ,vhich, because of its special natu�e, zi Judgments made from the perspective of ulti­ mate moral equality are most effective when directed to individuals whose moral credit has been ascertained, as well as their moral worth, If we go directly from the perspective of moral worth to the perspective of ultimate moral equality ,vithout passing through the perspective of moral credit, it appears that something important has been left out.


invalidates all others. It is superior to the others because it is broader in its scope. This is the only perspective tbat is causally unlimited, the only one from which we view acts in the context of all their causes unknown as well as known. And since w; are seeing more broadly, it follows tbat we are seeing more accurately. The answer here must be that it is not clear that this does follow. Do we see "better" from a height tbat takes in a large part of tbe surrounding territory? The only reply can be "Yes and no." Details which are not seen from a height spring back into 'View when we climb back down and look at objects in the context of a smaller part of their surroundings; and things which looked alike from above manifest striking differences ·when seen from a position farther down. In this case, one standpoint does not reveal the "real nature" of tbe objects better than another one does. But we must not be lured into pres­ sing our optical metaphor of "perspectives" too far. In any case, the Group II deter­ minist may vvish to support his basic con­ tention- that the perspective of ultimate moral equality has a privileged status- in another way. He may claim that it will be impossible, psychologically, for those who have viewed persons from the perspective of ultimate moral equality to go on making the same old judgments of . praise and blame from other perspectives. How, he will ask, can we throw ourselves into the task of sorting the worthy from the un­ worthy, the creditable from the discredit­ able, when we l.'1low that from another standpoint these distinctions will disappear altogether? ·wm not the view from every perspective but that of ultimate moral equality take on the aspect of something unreal� a mirage without power to deceive for more than a moment? Part of the answer here is that ,ve cannot, indeed, make moral discrimina­ tions in exactly the "same old" way, and that we cannot "throw ourselves" into the sorting and grading processes with quite

666 the zeal of those who have never seen hu­

man beings as ultimate moral equals. The

difference, however, will be in the quality

and intensity of the emotions accompany­ ing judgments of praise and blame based on moral worth and moral credit. The judgments themselves will go on being made; human acts will go on being looked at in causal contexts of varying scope. Our models of human nature, in short,. will go on being used. To see whether it is psy­ chologically possible to return from the perspective of ultimate moral perspectives, no determinist needs to look farther than his own experience. It is in truth not pos­ sible to do anything else. ( 2 ) A second objection is likely to be raised by certain libertarians. It is the charge that our so-called "perspective of ultimate moral equality" is ignoble and degrading. Such a perspective, it will be said, affords a particularly deplorable ex­ ample of the levelling tendency which seeks to destroy standards of merit in all areas. Plato's charge against political de­ mocracy- that it makes equals of unequals -applies a thousandfold to the "moral democracy" that is claimed to be visible from this distorting perspective. The moral world is hierarchical to the core; to remove the sheep and the goats from the moral landscape is to destroy it. So runs this second charge. The ansyver to it divides into t\vo parts. The first point to be made is that moral discriminations have not been per­ manently eradicated. Viewed from the other moral perspectives ( which, as ,ve have seen, are not eliminated by the per­ spective of ultimate moral equality ) , the moral landscape S\Vims back into our ken with sheep and goats intact. Nforeover, it is sometimes more appropriate to look at the moral world in this way. But some "�11 feel that this part of the ans,ver is not enough: that it is wrong ever to see all persons as moral equals. Now even though one may be deeply con­ vinced that the perspective of ultimate


moral equality, far from being ignoble, is -when appropriately used- exalted and inspiring, it is not altogether easy to know how to argue for this conviction. One may point to the increase in compassion, toler­ ance, equanimity,. that come to those who know how to look at themselves and each other on occasion from the perspective of ultimate moral equality. But it may con­ ceivably be said in reply that these are regarded as benefits only by persons who are antecedently convinced that determin­ ism is, as a matter of fact, true. This reply is not without force. That portion of Spinoza�s defense of determinism which shows «what service to our o,vn lives a knowledge of this doctrine is" has not lost its power to move deterrninists; but what power do his ,vords have over others? If all our acts go back to ultimate causes, then we should indeed look at all human beings \ compassion and tolerance; but what if they do not? Libertarians may con­ tend that compassion and tolerance are not spiritual goods if these attitudes are directed toward humans who1 because they have misused their freedorni simply do not deserve to be pitied or tolerated. And equanimity in the face of moral iniquity is nothing but extreme moral callousness, particularly unforgivable, it will be said, when the wrongdoer is oneself. It may prove impossible to disabuse some extreme libertarians of their convic­ tion that human beings should never be regarded as being all morally equal. But it is hard to believe that most people, what­ ever their metaphysical beliefs, will not find something to which they can respond positively in the attitudes engendered by the perspective of ultimate moral equality. The making of all our judgments of praise and blame with less finality, Jess assurance that they represent the whole truth, must seem to many an end to be welcomed.22 Religious teachings which have kept be2� This end is particularly desirable for those more sweeping judgments in which persons are praised or blamed, not for specified acts, but for


fore our eyes the ·view from something like the perspective of ultimate moral equality ( "There but for the grace of God go I") have performed a great service for our moral outlook.23 ( 3 ) Finally, I want to take note of the contention-in which those of all meta­ physical persuasions will doubtless heartily concur-that this account of praise and blame is simply too complicated to be ac­ ceptable. How could the average unspec­ ulative mortal ever find his way among such a bewildering variety of moral per­ spectives? How could he ever make a judgment of praise or blame? In reply to these questions, two points must be made. First, a single moral judg­ ment arrived at from a single moral per­ spective is not necessarily made more complicated by the present account than by other accounts. 24 But, secondly, it must be admitted that difficulties do arise when a moral judge is asked to remember that other moral perspectives exist and set limits to the one that he is using at any given time, and \Vhen he is asked, as he sometimes must be, to decide on the moral perspective that should be used in a par­ ticular situation. vVe have seen that moral wisdom, on the present view, consists not merely in the ability to make correct moral appraisals from a single perspective, but also in the ability to correlate the per-

their whole characters. Space limitations have pre­ cluded the consideration of such judgments here; but their treatment forms an important part of a more detailed examination of moral perspectives. :rn Note that, even though the grace of God may be regarded as being "freely" given, in so far as it is held to constitute a causal determinant of human action, it is treated as essentially similar to what have here been called "ultimate causes." Of considerable interest for further study would be a comparison of the perspective of ultimate moral equality-here described in purely naturalistic terms-,vith such religious concepts as "equality in the sight of God." 2 • Indeed, the distinction between moral worth and moral credit makes it possible to give a sim­ pler account of the judgments made by each standard than the account which proponents of the single standard view would have to give if their position were adequately worked out.


spectives, and, on occasion, to choose among them. It may well be that few attain this kind of \visdom, yet it is by no means clear that it cannot be attained by unspeculative persons. Perhaps such persons can and do make concrete judgments of moral praise and blame in a balanced and large-minded way, keeping the various relevant consid­ erations in due proportion, and governing their own attitudes accordm'gly, despite a lack of any grasp of a theoretical basis for what they are doing. But, if it should turn out that we cannot really evade the con­ clusion that the present account makes moral wisdom harder for an unspeculative person to attain, this conclusion would not necessarily vitiate the account. Why should it not be the case that moral wisdom de­ mands considerable resources of intellect as well as of character?

In this paper, I have been arguing that the question with which we began, "Can determinists find a rationale for moral praise and blame?", can be answered affirmatively. I have tried to show how­ ever, 'that the unrecognized assu�ption behind the typical and influential affirma­ tive answers that have been given- the assumption that judgments of praise and blame are made from a single moral per­ spective-is mistaken. I have maintained that those determinists who give a nega­ tive answer to our original question have caught sight of some important truths that the others have missed. In the end, how­ ever, with their attempts to set up the perspective of ultimate moral equality as the sole valid perspective for judgments of praise and blame, they have fallen into the same fundamental error as the others. One group eternally confronts the other with the question "How can you deny that hu­ man beings can be said to be praiseworthy and blameworthy, in view of the fact that they commit acts that are right or wrong and at the same time done voluntarily, knowingly, and from good or bad desires?" To which the second group incessantly 1


hurls back a question of its own: «HO\:v can you assert that human beings can be said to be praiseworthy or blameworthy, in view of the fact that their acts, like all other events, are wholly subject to causal laws, and must be traced back, in the end, to factors wholly beyond the agents' con­ trol?" The account given here, which may be called the "theory of multiple moral perspectives," is designed to help put an end to this durable impasse. I have tried to show that the first group is speaking from the perspective of moral worth, ,vhile the second replies from the perspective of ultimate moral equality. Both perspectives are valid; but each perspective is incom­ plete. Three moral perspectives are neces­ sary, I have contended, if we are to tell the whole about the praiseworthiness and blameworthiness of human beings. One of


these, the perspective of ultimate moral equality, takes form as a consequence of assuming determinism to be true; but its adoption is not without moral and spiritual benefits. The other perspectives can be exhibited in an examination of judgments of praise and blame conducted quite inde­ pendently of any determinist assumptions; and we can then see that determinism is­ at the very least-fully compatible ,vith the use of these moral perspectives. It seems to me that considerable work remains to be done in clarifying and refining these concepts and principles, and in exploring their implications in many directions. But if the claims made here are in essentials justilled, it follows that determinists need not feel that old familiar uneasiness when confronted by the concepts of moral praise and blame. On the contrary, it may be that we stand here on solid ground.

V Ethics and Psychology A . Reasons and Causes B. Obligation and �lotivation C. Why Be Moral?

A. Reasons and Causes

Reasons and Causes S. I. BENN RICIIARD PETERS Problems connected with freedom and determinism ace best clarified by examin­ ing what we wish to convey when we say that an action is 'detem1ined'. For there are two clliferent things, which are often confused, which 'determined' can mean. Firstly, there is what we might call 'causal explicability' and, secondly, 'unavoidabil­ ity, Many people have failed to distin­ guish tbese hvo very different strands in the meaning 0£ the tenn "determined\ and they have often thought that 'determined' involves both of these thlngs. 'When w e say that our behaviour is determined, therefore, it is often assumed both that our behaviour has causes and that it is un­ avoidable. Let us, therefore, consider these t\vo strands in the meaning of 'determin­ fam� in turn, &."1d we can then later show that simply because our behaviour has causes it does not necessarily follow that it is unavoidable.


Determinism to a scientist conveys the general proposition that every eveot has a cause. vVhether this general proposition is true is a very clifficult question to decide, but it is certainly assumed to be true by Reprinted frou1 Social P1incip'les and the D6mc� cratic St-ate bv S, L Benn and Richard Peters by permission of the authors and George Allen & Unwin Ltd, The Ame--"fotm edition is entitled Prin-­ ciples of FoUtictil Thought.

most scientists. To say that au event has

a cause is to say that there are universal laws together with statements about initial conditions prevailing at particular times, and that from these hvo together we can predict an event which we eall an 'effect'. F'or example, given that under the con­ ditions xs,z, fron expands when it is heated, and given that the conditions x,y,z prevail and that this is a case of iron being heated, we can make the prediction that iron will expand. Here we have a typical causal relation. The so-called 'cause' is then the eve11t referred to in the statement of initial conditions. And these conditions are regarded as being sufficient to e,q,lain the effect, if it is a full-blooded causal explanation. Have we such relations in human affairs? The initial difficulty about saying that we have is that it is difficult to main­ tain that there are any psychological or sociological laws which ,vould enable us to make such definite predictions. TI1ere are also difficulties connected v,ith our knowl­ edge of particular situations which con­ stitute ilie initial conditions; for when we are dealing with stones and bodies falling, their past history is scarcely part of the present situation, But when we are deal� ing with human beings, their past histmy is very much part of the present situa­ tion, and it is very difficult to know whet,l:,er a given case is really of the type to which the particular law we have in mind applies. Nevertheless, there are some generalizations i n psychology and the 671



social sciences which are reasonably well viction. To perceive something is to be established. They do not enable us to make right in our claims about what is before detailed predictions; they merely enable our eyes; to learn something is to improve us to state the sort of thing that will tend at something or to get somed:ilng right. All to happen under certain typical conditions. such concepts have norms written into In this respect psychology is in no worse them. In a similar way1 as we have pre­ plight than other sciences like meteorology. -viously argued, 1 a human. action is typi­ The difficulties arise fro!TI the complexity cally something done in order to bring of the subject-matter, and, it might be about a ::·esult or in accordance \.Vith a standard. Such actions can be sa;d to be argued, can be remedied in time. If, however, we look more closely at done more or less inteliigently and n1ore these so-called laws in psychology we llnd, or less correctly only because of the norms in the main> that they do not give sufficient defining what are ends and v,hat are explanations of human actions, of what efficient a.."lart of of emotion'. Men do not d::eam or forget a his brain being stimulated, or that l�ar:1ing name 'on purpose' any more than they are is a functio�,. in part, of antecedent 'ten� deliberately sub;ect to impulses or gusts siont, But the very meaning of Kremember­ of emotion. One class of laws in psychol­ ing' and 'learning' prech:tles a sufficient ogy, then, gives causal explanations which explanation in these sorts of naturalistic seem sufficient to account for \vhat hap­ terms. pens to a man, bt:t not for what he does. Furthermore the proble,n of the free­ There is another class of laws, how­ dom of the will arose mainly in connection ever, which concern not what happens to with a type of action that is palpab:y men) but \vhat they do-their actions, per­ different froc1 a mere movement or pro­ formances and achievements, But such cess- an action t.i'iat is preceded by laws state necessarv rather than sufficient deliberation and choice. For, roughly conditions. vVe hav� ln mind here the con­ speakiag, a \villed action� \Vas usually tributions made by physio�ogical psychol­ taken to mean i one in "'"vhich we thbk ogists and those who have studied before we act) \vhen we make up our cognitive skills like learning, remembering, n1inds in terms of considerations whkh are and perceiving. Part of ,vhat we mean by relevant to the matter in hand before '-Ve such terms is that human beings attain a act. There are difficulties about developing norm or standard. Remembering is not ;ust 1 See Gha:;:>ter 1. a psychological process; for to remember t ·whether the concept of 'a willed action' is a is to be correct about what happened in useful or clear one is another matter. Reference is the past. Kno,ving is not just a mental here made to traditional oontrove:rs!es about free­ dom of the -...vill such as that between Hob'�es and state; it is to be sure that we are correct Bish OJ) Bramhall, ( See R. S. Peters, Hobbes, 1956, and to have good grounds for our con- pp. 178--59. )

REASONS AND CAUSES causal laws for actions of this type which are additional to those already stated about actions in general. Such difficulties ar?, simila: to those which the sodal scientist, as well as the psychologist, has in pre­ dicting what human beings will do. This is connected with the fact that into the hu..'Ilan being's deliberations about what he is going to do will be introduced considera­ tions about what he is likely to do, which the social scientist may hav·e published, A scientist 1nay discover a causal la,v con­ nect! the properties of clover with a cer­ tain effect upon the digestive organs of sheep. But, when he publishes his findings, the sheep cannot take account of them and modify their behaviour accordingly. But with 1nen it is different. :tvfany causal con­ nections discovered by psychologists may only hold good provided that the people whose actions are predicted in accordance \Vith the law remain ignorant of "vhat it asserts. And it is practically impossible to ensure tbat this is the case. So, if people know the causes on which a prediction o! a certain type of behaviour is based, and if they deliberate before acting, tl1ey may do something dlffere,it from what is pre­ dicted) just because they recognize these causes. A prediction may thus be valid only on the assumption that the people concerneG remain unconscious of the causes on which it is based. Otherwise it may be no 1nore than a warni::lg. But why cannot causal explanations afao be given of such informed deli::Jera­ C.ons which precede actions? \'Ve are here confronted with the difficulty of accounting for logical thought in causal terms, of giving a causal explaaation for rational actions done after deliberation which involves logically relevant considerations. This is an extreme case of fae difficulty already cited of giving suf:icient explanations in causal terms for actions and performances which involve :!Orms and standards. Yett as has already been pointed out, such premedi­ tated actions are particularly important in the free-will controversy, as the exercise

673 of 'will' has usually been associated with rational deliberation before acting. When a man is solving a geometrical prob:em and his thoaghts are proceeding in accord­ ance \Vith certzin logical canons ) it is logicaHy absurd to suggest that any causal explanation in terms of movem.ents in his brain, his temperament, his bodily state1 and so on, is sufficient by itself to explain the movement of his thought. For logical canons are normative and cannot be sufficiently explained in terms of sta:es and processes which are not. Of course there are any number of necessary conditions which must be taken account of. A man cannot think withcnd a brain, for instance. But any sufficient explanation wodd have to take account of the reasons for his ac­ tions. \Ve would have lo know the rules ol chess, for instance ) w:llc:1 gave some point to a chess-player's move. Indeed we would only ask for the cause of a chess­ players b·ehaviour if he did something which could no� be explained in tez>ms of the rules of chess and the objective at ,vhich he was aiming. If, for instance 1 he refrained from taking his opponent's queen, when this was the obvious and the best move, we might ask "What made him do that?' and we would be asking for a causal explanation, like "he was tired'. But this vmuld novv be an explanation of what happened to him, not of what he did deliberately. '\Ve would not ask for such an explanation :if there was an obvious reason for his move. 3 This example can be generalized and the point m2de that behaviour is explicable not because we b1ow its causes, but because people act in accordance with :; Of course the category of ".;.ction' is much ,;vider than that of premeditated action, though it may be co-exter.sive vvith that of 'rationality'. For this covers the sort of things for which a man· could have a reason-Le. which fall under what we call !he purposive ru:e-followfag model, Pre­ meditated action is a particu�ar case of action where action is preceded by rehearsals and deli­ beration; but often reasons can be given by peo;;ile for v:hgt they do even though they do not del:ber­ ate before they act.


certain known rules and conventions and adopt appropriate means to objectives which are accepted as legitimate goals. We know why a parson is mounting the pulpit not because we know much about the causes of his behaviour but because we know the conventions governing church services. V\Te would only ask what were the causes of his behaviour if he fainted when he peered out over the congregation or if something similar happened to him. Most of our explanations of human behaviour are couched in terms of a purposive, rule­ following model, not in causal terms. Moral behaviour, above all other sorts, falls into this purposive, rule-following category. For, as Aristotle put it in his Ethics, it is not a man's passions which are the object of moral appraisal nor his capacity to be sub­ ject to such passions; rather we praise or blame a man for what he does about bis passions, for the extent to which he con­ trols or fails to control them ·in various situations. Deliberation and choice may not precede every action, but habits are set up as a result of such deliberation and choice. It is for the exercise of such habits that men are praised and blamed- for the ends which they seek and for the means which they adopt to bring about their ends. Punishment, too, as we have pointed out, presupposes that men can foresee the con- · sequences of their actions and that they can learn to avoid those to which penalties are attached. Praise and blame, reward and punishment, act as rudders to steer human actions precisely because men de­ liberate and choose and can be influenced by considerations of consequences. There is a radical difference between actions of this sort and cases where things happen to a man-where he acts 'on impulse', has a dream, a vision,_ or lapse of memory, or where he is afflicted by a feeling of nausea or hysterical paralysis. Questions of the 'freedom of the will' do not arise where things happen to a man; only where a man acts and can be praised or blamed, pun­ ished or rewarded for what he does. Yet


it is precisely in these cases of human ac­ tions, as distinct from passions, that causal explanations seem inappropriate as suffi­ cient explanations. Two sorts of objection might be mounted against this attempt to limit the role of causal explanations of human be­ haviour. In the first place it might be said that by substituting concepts like rule­ following and the pursuit of objectives we were in fact introducing other sorts of causes. Now the word 'cause' can be used in this very wide sense. But the termino­ logical question is largely irrelevant; for two sorts of explanations which are logi­ cally quite different would then be in­ cluded under the enlarged concept of 'cause'. To follow rules, to take steps which are seen to be necessary to reach some sort of objective, to see the point of some­ thing, these may be 'causes'; but they are causes in quite a different sense of 'cause' from things like stomach contractions, brain lesions, acute pains, and so on. The types of explanation must be distinguished whether we use the term 'cause' to cover both or not. And certainly seeing the point of something is quite different- even if it is called a 'cause'-from the causes preva­ lent in the physical world. In the early days of the determinist controversy phi­ losophers like Spinoza and Kant used the term 'self-determined' to distinguish ra­ tional actions from those which could be explained in terms of mechanical causes like movements of the. brain and body. Indeed Kant's suggestion that man lives in two worlds, and is subject to two different sorts of causation, is a metaphysical \vay of bringing out the logical distinction between these t\vo sorts of explanation. The second objection is the suggestion that all reasons might be rationalizations­ a smoke screen for what we are going to do anyway. "\Ve are, as it were, pushed by causes in the mechanical, physical sense, whatever we do; but sometimes we throw up an elaborate smoke screen of excuses which make no difference to what we in


fact do. If, hmvever, \Ve say that all reasons are rationalizations, we make no difference between the behaviour of an obsessive or a compulsive and that of a rational man. If a compulsive believes that his hands are covered in blood and spends his time con­ tinually ,vashing them; no relevant consid­ erations will make any difference to his behaviour. All the known tests fail to show blood; yet he still goes on washing his hands. But a civi! servant making a com­ plex decision about policy does not pro­ ceed like this. He will change his mind and alter policy in the light of relevant considerations. Indeed it is only because people sometimes alter their behaviour be· cause of relevant considerations that it makes any s&nse to talk of rationalizations as well as of reasons. A term like "rationali­ zation'. which casts aspersions on the reasons given for action, is a verbal parasite. It flourishes because there are cases of genuine reasons \vith which l'&­ tionalizations can be contrasted. Thus even if all behaviour has causes, m the sense of necessary conditions, there are objections to saying that all behaviour- especially rationa!. behaviour-can be s1.1,ffi clently ex­ plained by causes of the sort suggested by physical scientists, and by mechanistic philosophers like Hobbes. 11/hether this means that there is also a case for freedom depends on whether 'free can be equated, in any of its various senses; with 'not sufficiently explained in causal tenns'. Kant and Spinoza} who spoke of self-determinism in the case of rational action, claimed t!:tat this is the proper sense in which men could be said to be free. A man is free or self-determined, they said, in so far as his behaviour is explained in terms of his rational decisions rather than in terms of purely mechanical causes. Such causes ,vere viewed as rnovements which somehow pushed or impelled a man to act. In other words there was an implied contrast behveen rational actions and those which occurred as the result of some sort of i:iternal push which acted as a quasi-


constraint on a man. And this notion, that a man was somehow the victim of internal forces, fitted in very well with the other meaning of 'determined' which suggested that there vtas some sort of unavoidahility about a ni:an· s actions.

(B) UNAVOIDABILITY If a man is told that his actions have causes, he will probably agree. But if he is told that they are cacsally determined, he will demur. He will not compla:.n about the verbal redundancy of the assertion but will picture himself somehow as a "?risoner or a victim. This illustrates our ... contention that the concept of 'determinism' conveys more than the suggestion that behaviour has causes, or that it can be sufficiently explained in terms of its causal anteced­ ents, To most people to say that actions are determined conveys the suggestion that they are unavoidable, that the individual in question cannot 1"'lp doing what he ' does. Furthermore, it is often assumed that wherever causes C',_an be found for a ::nan,s actions, then also his actions are unavoid­ able. And it has been argued� as a corol­ l aiy, that 011:ce the causes of an action are known, blame and punishme,it are ahsurd. This assumption that any action for which causes can be produced is there­ fore unavoidable is surely a mistake o c ­ casioned b y the peculiar circumsta:1ces of the rise 0£ science. It so happened that scientific advance, ,vhlch consi5ted in the discovery of far-reaching causal laws, coin­ cided with the widespcead theological doc­ trine of predestination and with the meta­ physical picture of the universe as a vast piece of clocl.-work in which human beings, like cog-wheels, were pushed on in a set pattern of movernent. God, as it were, constructed the clock and set it going. If the clock could be seen as a whole 1 men could see ,;,vhat tbe future had in store for them and what movements


determined that their fate should be this and no other. Causal discoveries revealed the springs and levers which pushed men towards their appointed destiny. The tacit assumption therefore developed that wher­ ever causes could be found for actions, they were also unavoidable. Causes, being pictured always as internal pushes and pulls, were thought somehow to compel a

REASONS AND CAUSES man. And this picture suggests compul­ sion whether such causes are properly to be regarded as necessary or as sufficient conditions for human action. Nfen were therefore regarded as being not free be­ cause they ,vere the victims of a peculiar internal sort of compulsion exercised by the causes of their behaviour. They were thus not able to avoid doing what they did.

Reasons and Causes A. J. AYER It is novv al:nost a commonplace among actions of that type are perfonned they are philosophers that motives are not causes. done from the same sort of motive. Finally, a point is made of the fact that motivated But this is not to say tl1at it is true. Why is it tho'ught to be true? There action often consi.:;ts in follo\'\.ing or at� are various reasons) not all of :hem of equal tempting to follow a rule; that is to say, weight. The most simple of them is that the action may b e one to which normative motives operate a fronte whereas causes criteria are applicable; the question arises operate a tergo; to put it crudely, that vvhether it has been performed correctly; causes push while motives pull, A more but this means, so it is argued, that we sophisticated argument is that cause and somehow impoverish the motive if ,ve re­ effect are distinct events: so:1 if the motive gard it merely as a cause. Let us now examine these arguments for an action caused it, it \VOuld have to be a separate occurrence which preceded the in turn. The first of them need not detain action or at any rate accompanied it; but us long. If the eontention is that purposive in many, perhaps i.."'1 most) cases o! moti­ beha vioux is to be accounted for, not as vated actions, such separate occurrences the response to any past or present stimu­ are sirnply not discoverab!ei the specifica­ h.:s, but rather in terms of the future state tion of the motive is part of the descriptio:::i of affairs towards the realization of which of the action} not a reference to anything t..1ie behaviour is directed, the argument outside it, and certainly not a reference to fails for the simple reason that there may any distinct event. Thirdly, it is argued not in fact be any such future state of that in the scientific sense of v, hich are aU that we have seen to be required for the ap1


plicability of a purposive explanation, do not flt the standard model of the relation of cause and effect. Once more this is partly a question of terminology. If ,ve construe the causal re­ lation in a strictly Humean fashion so that its terms can only be distinct events, then the objection holds. For we have seen that in many quite typical cases the motive may be present in the form of a disposition which, though distinct from the behaviour which i t n1otivates, is still not exactly a distinct event. It see:ns to me� however, that even from the point of view of doing justice to the ordinary, let alone the scientific1 use of causal language, this conception of the causal relation may be too restrictive. For one thingt YVe often ,;vant to be able to re­ gard the absence of some circumstance as a causal factor, that is, to admit negative as well as positive conditions; and even this does not flt tidily into the Humean scheme. I have, of course} no quarrel with Hume's fundamental idea that causation ::nust in the end be a matter of regular concomitance} but I suggest that causal relations should be regarded as holding between facts rather than events, where 'fact' is understood in the wide sense in which true propositions of any form can be taken as expressing facts. This involves no sacrific0 1 since in any cases where it is appropriate these causal statements about facts can be translated into statements about events; i t merely ex­ tends the field of causal relations a little more v,idelv. Then the sense in which an agent's motive may be said to be the cause of the action which it rnotivates is that given certain conditions the fact that the agent performs the action is inferable from the fact that he has tbe motive in virtue of a causal law. But just as the first" of the arguments which we listed dissolved into the secontl1 so1 if I am right in what I have just been saying, the second dissolves into the third; for now the question arises whether there reaIIy are such causal lavls. On the face of it1 it see.rns a t leas: very doubtful. As I

REASONS AND CAliSES said when I first referred to this argument, it is surely possible for someone to act from a given n1otive on a particular occasion \Vithout its being tbe case either that when­ ever anyone has a motive of this kind he acts in this ,vay, or that whenever aayone acts in this way he does so from this ldnd of motive. No doubt there must be some degree of regularity in the way in which motives lead to actions, for us to find the connection inte!Hgible. If people hold very queer beliefs, they may indeed take means to a given end that others would not take; in this sense the connection between mo­ tive and action may- even be quite idiosyn­ cratic. It remains true, however, that people in general do what they believe ,vill enable them to achieve their ends) a'!ld that these beliefs, though they may be false, are usually backed by a fail' amount of evi­ dence. The result is that there is at least a tendency for similar motives to be cone­ lated with similar actions. This tendency is especially marked in the case of standing motives. The range of beha,1our which we are prepared to ascribe to jealonsy, or greed t or ambition is fairly narrow. Even so, it will be objected, such tendencies1 at their very strongest, fall a long way short of bei:1g causal laws. All this is true, but possibly not de­ cisive; :for it may be that \Ve are looking for our causal laws in the wrong place. The point from which I think that we should start is that when a man is said to have acted in consequence of having such and such a motive, it is irnp1ied at least that j£ he had not had this motive he would not in the particular circumstances have acted as he did. But this is to say that the existence of the motive is taken to be a necessary condition of the action; not !ndeed a neceS� sary condition of anyone's performing 'the action at any time, or even of the agent's pe:r:forming it at any time, but a necessary condition of his performing it a t just this juncture. The quesaon then arises vthether it is a!so taken to be part of a sufficient con� dition, and this is not e2sy to answer. The


ground for arguing that it must be is that othenvise the ascription of the motive would not properly account for the action; we should have to allow that even granting the agent's motive and the rest of the attendant circumstances, including all the other aspects of the agent's mental and physical condition at that time, we could not entirely rely on the action's taking place; and to this extent its occurrence ,vill still be unexplained and indeed inexplica­ ble. There may, however, be those who are prepared to accept this consequence, so long as they can hold that there is a high probability in this situation of the action's taking place: that is, they may be satisfied with the hypothesis that if the situation were repeated a great number of times the action would take place very much more often than not. But since this leaves an ele­ ment of arbitrariness, in that we have n9 answer to the question why it should ever not take place, it seems preferable to make the stronger claim, unless it can be shO\vn to be untenable. The suggestion then would be that -vvhenever an agent can properly be said to have acted exclusively from a given motive, the circumstances must be such that in any situation of this kind, indispensably including the presence of such a motive, an action of this kind invariably follows. It is clear that if so strong a claim as this is to be made to appear even plausible, a great deal will have to be included in the situation, both in the way of positive and negative conditions. VVe must, however) avoid including so much detail in the de­ scription either of the situation or the action that our claim becomes trivial; our ground for saying that there is an invariable con­ nection between situations and actions of the sorts in question must be not that either the situation or the action is unique, In other \Vords, the types of fact which our laws connect must be envisaged as repeat­ able. But where are these connections to be found? Surely it is idle to maintain that these laws exist if \Ve are unable to produce


any examples. Well, perhaps we can pro­ duce examples of a rather humble kind. Let us begin with the hypothesis that when­ ever a person has a desire for the existence of a state of affairs S and believes that it is immediately in his power to bring about S by performing the action A, but not by any other means,. and there is no state of affairs S1 such that he both prefers the existence of S 1 to that of S and believes it to be immediately in his povver to bring about S 1 , but not conjointly with S, then unless he is prevented he will perform the action A. This is not quite a tautology, since the exist­ ence of the agent's desire is supposed to be established independently of his taking any steps to satisfy it. On this account, indeed, it may even not be unconditionally true: there may be cases of inhibition which would have to be specially provided for. But even if this difficulty can be overcome, the hypothesis still falls short of what we want because its consequent is subject to a general proviso: the person who satisfies the antecedent will perform the action A unless he is prevented. The question is whether this proviso can be dispensed \vith. The way to dispense with it would be to list all the things that might prevent the action from being done, and insert them in the antecedent in the form of negative con­ ditions. This would seem indeed to be an impossible undertaking at this level of generality: we could hardly hope to draw up an exhaustive list of negative conditions which would at this point apply to any ac­ tion whatsoever. But in its application to particular instances, our general hypothesis will in any case dissolve into a number of more specific ones, according to the nature of the case: and once the relevant type of action and perhaps also certain features in the situation of the agent have been speci­ fied, it does not seem to me obvious that the list of negative conditions cannot be com­ pleted. Thus if the action is one which involves making a certain sort of hand movement, there may be a finite number of types of bodily disorder which would


prevent it from being carried out; if it in­ volves handling certain physical ob;ects, there mav be a finite number of wavs in which th�y could become intractable; if the condition of the is specified, the ty?es of psychological impediment to which he is then subjcdmay again be fin ite in nwnber: and if the number of these various factors is so limited, there seems to be no com­ pelling reason to hold that they cannot be discovered and listed. Of course we shall have no guarantee that the list is complete; but then we do not have such a guarantee in the case of laws of any other type. However carefully a generalization is formulated it must at least remain conceivable that it ho]ds only under cer+..ain further conditions which we have failed to specify. Technically, if there are found to be such conditions, the gener­ alization is falsified, though sometimes we prefer to regard it as having been in­ co:npletely stated. I do not, however1 agree with those who would read into everv generalization of law a ceterfa parib,;s clause which tacitly protects it from being falsified through the operation of factors which the proponent of the generalization did not foresee. No doubt it is too much to require, as John Stuart Mill did, that causal laws should hold unconditionally, if this ls ,mderstood to imply that they would still be true no matter what else were true; for one law often depends upon others, so that its truth would not be preserved if these other laws were false, It iS, however, not too much to reoufre that the law should hold under any c'°ircumstances whatever that ac­ tually an.:.e. If the field which it is designed to cover is restricted, this limitation can and should be made explicit If ceteris paribus clauses were allow­ able, the task of finding laws in the sphere of human action 1.voultl be very much easier, but as our example has shmvn, these pro­ visos would only increase the laws· security at the expense of their scienti£c interest. Even as it is1 our hypothesis contained a stipulation which would often not be satis-

REASONS AND CAUSES fied. It is by no means invariably true that when someone is airniJ:1g at a given end he believes that there is only one means of at­,,_g it which it is i.:Umecl.iately in his power to realize. Ver; often he will be pre­ sented with a choice of such means, so that it needs to be explained why he selects one of them rather than the others. I have no doubt that a number of examples could be found i n which I should not know how this was to be done, but I suggest that quite a lot cf cases would be covered by the follow­ ing hypot1:esis: whenever the antecedent of our first hypothesis is satisfied, with the difference that the agent believes that it is immediately in his power to bring about S not only by performing the action A but also by performing the action A1 , but that he cannot perform both, then if he believes that A and A 1 are equally efficacious in br:nging about S, but prefers what he ex­ pects to be the other consequences of A to those which he expects of A 1 , he will, un­ less he is prevented, perform the action A. Again, if this is not to be tautologous, there must he evidence for his preferring the other expected consequences of A to those of A', independently of the fact tlmt he does perfonn A, but I think it fair to assume that this will usually be available. Our hyp o:hesis does not, o�f course, commit us to holding that whenever an agent has a choice between actions of the kind A and A' as means to a given type of end S, he will choose A. A n1an may well decide to walk to work on one occasion and take a taxi on another. But then the assumption is that there is a change in the circU!l1Stancesi in the state of his health or his finances or the weather or son1e other combination of factors, ,vhich '\Vill sufficiently account for the variations in his preference. I admit, however, that until a set of such bypof:heses has been formnlated and tested, the degree of strength. that we can attrlbue to gene:·­ alizations about human conduct remains an open question. The most that I would claim at this stage is that the difference between these



generalizations and those that can be found The same applies, in my view, to the to govern other natural phenomena is no­ argument that actions most often need to where more than a difference of degree. be understood in terms of their social con­ \\1hat I have tried to show, in arguing that texts. A great deal is made by some philoso­ motives may be causes, is that there is no phers of the fact that an action is not a mere warrant for regarding explanation in terms physical movement. It has a significance of motives as something of a different order which depends not only on the agent's in­ from the explanations that occur in the tention and motive, but very frequently also physical sciences. There is nothing about on a complex of social factors. Think of the human conduct that would entitle us to social norms and institutions that are in­ conclude a priori that it was in any way less volved in such commonplace actions as lawlike than any other sort of natural signing a cheque, signalling that one is g�ing to turn when one is driving a car, process. But what of the argument that human saluting a superior officer, playing ,a card, actions conform to rules, that we are often shaking hands with someone to whom one more interested in judging whether and has just been introduced, waving good-bye how far they are up to standard than in dis­ to a friend. To represent these merely as covering how they came about? I cannot different sorts of hand-movements, which see that it is relevant. From the fact that of course they also are, is to miss their we can estimate an action in terms of its significance; it is indeed to fail to represent conforming to a rule, it no more follows that them in their character as actions. All this is true, except that it seems to the performance of the action is not caus­ ally explicable than it follows that the ap­ be an arbitrary question what we are to pearance of a rainbow is not causally ex­ regard as constituting an action: whether, plicable from the fact that it can be made for example, we choose to say that the the subject of an aesthetic judgement. To motorist's action is one of putting out his explain something causally does not pre­ hand or one of signalling that he is going to clude assessing it in other ways. .But per­ tum. Earlier on, I referred to the action of haps the suggestion is merely that to relate drinking a glass of wine as an instance of an action to a rule is one way of accounting the way in which the same action can have for it, and, in the present state of our a different significance in different social knowledge, a better way of accounting for contexts : it would have been no less, but it than trying to subsume it under dubious also no more correct if I had described causal laws. I cannot even agree with this this as an instance of the way in which the because I think that it presents us with a same physical process can in such different false antithesis. The only reason why it is contexts become a different action. VVhere­ possible to account for the performance of ever we decide to draw the line behveen an action by relating it to a rule is that the the characterization of an action and the recognition of the requirements of the rule assessment' of its significance, the point re­ is a factor in the agent's motivation. He mains that the physical movement has to may attach a value in itself to pedorming a be interpreted, and that in order to inter­ certain sort of action correctly; he may see pret it correctly it will often be necessary its correct performance as a means towards to understand its social as well as its per­ some further end; or it may be a combina­ sonal implications. But if we grant this point, what tion of the two. In any event this is as much follows? Certainly not, as the philosophers a causal explanation as any other explana­ tion in terms of motive. The invocation of who lay stress upon it seem to think, that rules adds nothing to the general argu­ these actions cannot be explained in causal terms. For when it comes to accounting for ment.

684 an action, the ooly way in which the social context enters the reckoning is through its influence upon the agent. The significance of the action is the signillcance that it has for him. That is to say, his idea that this is the correct, or expedlent, or desirable thi.'lg to do in these circumstances is part of his motivation; his awareness of the socfal con­ text and the effects which this has on him are therefore to be included in the list of initial conditions from which we seek to derive his performance of the action by means of a causal law. 'Whether such laws a:!'e discoverable or not may be an open question; but the fact that these items figure among the data has no important bearing on it. That human behaviour has a point or meaning, in tWs senset is not even an argu­ ment against the materialist thesis that it is all physiologically determined. This thesis is indeed high:y specul�tive; ,:ve are very far from hayjng a physiological theoq which would account for people's actions in specific detail, let alone from being in a position to apply one. But if the motives which impel men to act are, let us say, projections of the state of their brains, there is no reason why t,'iis should not apply to their social responses as much as to any­ thing else. But surely no purely physiologi­ cal account could be an adequate descrip­ tion of an action. Obviously it could not; even if the study of the agent's brain could give us all the information that we needed beyond the observation of his physical movements, we should still have to decode it. But this is not an objection to holding that actions can be exnlained in these terms, any mo:e than the"' fact that to talk about wave-lengths is not to describe col­ ours is an objection to the science of optics. This also shows that even if I am vvrong in assimilating motives to causes) it v-.rill not follow from this alone that human beha­ vfo� is not entirely subject to causal explanation. For the fact that we can explain an action in one manner by refer� ring to the agent's motive leaves it a fully

REASO:-!S k"l'D CAl:SES open question whether it cannot also be explained more scientiflcally in terms, say, of a physiological tbeoq. None of this setties the issue of deter­ miniSTil. I do not inde'ed think that it can be settled at this level, since I agree w�th those who hold that it shm.:ld not be fater­ preted as an a priori question. It is of course true that not CVE;ry event in human history could in fact be predicted, Not only would the making of each prediction itself then have to he predicted, and so ad infinitum, but as Professor Popper and others have rightly pointed out} :t would :follow that no one could ever have a new idea. But how­ ever comforting this may be to those who dislike conceiving of themselves as subjects for science) it does not go any way to prove that not all events in human history are susceptible of lawlike e:qJlanations.' The strength of the determinists lies in the fact that there seems to be no reason why the reign of law should break dow:1 at this point, though this is an argument which seemed more convincing in the age of classical. physics than it does today. The strength of the indeterminists lies in the fact that the speciflc theories which alone could vindicate or indeed give any sub� stance to their opponents' case have not yet been more than sketched, though this is not to sav that thev never will be. Until such theo;ies are 'properly elaborated and tested, I think that there is little more about this topic that can be usefully said. A philosophical question which I have not here discusse� partly because I do not think that I have anything new to say about it, is whether the denial of determinism is implied in our usual ascriptions of moral and · legal responsibility. In common '-Yith many other philosophers I used to hold that it was not� that in this respect the antithesis between the clafrns of free wiU and deter­ minism was illusory1 but in so far as this is a question of wha: people actually believe, I now think i� more likely that I \Vas wrong. This is indeed a matter for a social survey which, as I said before, would probably not


yield a very clear result. I should, however, eAl?ect it to indicate that if it were shown to them that a man's action could be explained in causal terms, most people would take the view that he was not responsible for it. Since it is not at all clear \Vhy one's respon­ sibility for an action should depend on its being causally ine;1cl)licable, this may only


prove that most people are irrational, but there it is. I am indeed strongly inclined to think that our ordinary ideas of freedom and responsibility are very muddle-headed: but for what they are worth, they are also very firmly held. It would not be at all easy to estimate the social consequences of dis­ carding them.

B. Obligation and Motivation

Remarks on jPsy,cliological Iledonism Vle can now deal with the question whether Psychological Hedonism be itself true. Let us begin with certain undoubted facts which must be ad.,nitted. The belief that a future experience will be pleasant is pro tanto a motive for trying to get it, and the belief that it will be painful is pro tanto a motive for trying to avoid it. Again, the felt pleasantness of a present pleasant experi­ ence is pro a motive for trying to ma�e it last, whilst the felt painfulness of a pres­ ent e>..-perience is pro tanto a motive for trying to make it stop. The question is whether the expected pleasantness of a future experience is the ouly feature in it which can make us want to get it, whet.\ the felt pleasantness of a present experi­ ence is the onlv feature in it which can make us \Vant lo prolong it, -..vhether the expected painfulness of a future experience is the only feature in it wbich can make us want to avoid it, and whether the felt pain­ fulness of a present experience is the only feature in it which can make us want to get rid of it. I must begin with one explaoatory re­ mark which is necessary if the above proposition is to be taken as a pedectly accurate statement of Psychological Hedo­ nism. No sane Psychological Hedonist woukl deny that a pleasure which is be­ lieved to be longer and less intense may be preferred for its greater duration to one From Five Types of Ethf,eal Theory by C, D. Broad, 1930. Reprinted by penru.ssion of Routledge & Kegan Paul Ltd. New York: Hurrumi­ ties Press Inc.



which is believed to be shorter and more intense. Nor would he deny that a _:iearer and less intense pleasure rnay be preferred for its greater nearness to a more intense but remoter pleasure. And this implies that duration and remoteness in some sense factors which affect our desires as well as pleasantness and painfulness. This compli­ cation may be dealt with as follows. There are G-ertain determinable characteristics which every event, as sach, must have. Date of beginning and duration are exam­ ples. There are others which an event may or may not have. Pleasanhiess, colour, and so on, are examples. Let us for the present call them respectively "categorial" and "non-categorial" determinable characteris­ tics of events, Then the accurate statement of Psychological Hedonism would be as fol­ lows. No non-categorial characteristic of a present or prospective experience can move our desires for or against it except its hedonie quality; but, granted that it has hedonic quality, the effect on our desires is detennined jointly by the determinate foim of this and hy the determinate forms of :ts categorial characteristics. Notv, so far as I am a"l.vare, no argu.-.uent has ever been given for Psychological He­ donism except an obviously fallacious one v,rhich j\1ill produces in his Utilitarianism. He says there that "to desire" anything and "to find" that thing "pleasant"' are just two different ways of stating the same fact. Yet he also appeals to car6''u l introspection in support of Psychologica! Hedonism. Sidg­ "�ck points out that, if Mill's statement were


true, lhere would be no more need of intro­ soection to decide :in favour of the doctrine than there is need for :introspection to decide that "to be rich" and "to be wealthy" are tw'o different expressions for the same Eact. But, as he also points out, Mill is de­ ceived by a verbal ambiguity. There is a sense of "please" :in English in which the two phrases "X pleases me" and "I desire X" scand for the same fact. But the verb "to please" and the phrase "to be pleasant" are not equivalent in English. In the sense in which "X pleases me" is equivalent to "I desire X" i t is not equivalent to "I find X pleasant." I f I decide to be martyred rather than to live in comfort at the expense of concealing my opinions, there is a sense in which mart;Tdom must "please me� more than living in comfort under these condi­ tions. But lt certainly does not follow ex vi termini faat I belie;e that martyrdom wlll be "more pleasant" than a comfortable life of esternal conformity. I do not think that "pleasantness" can be defined, or even de­ scribed unambiguously by referen ce to its relations to desire. But I think we can give a fairly satisfactory ostensive definition of it as that characteristic which is common to the expeI:ence of smelling roses, of tasting chocolate, of requited affection1 and so on> and which ls opposed to the characteristic which is common to the experiences of smelling sulphuretted hydrogen, of hearing a squeaky slate-pencil, of being hurt, of un­ requited affection, and so on, And it is certainly not self-evident that I can desire only e,qieriences which have the character­ istic thus ostensively defined. I trunk that there is no doubt that Psy­ chological Hedonism has been rendered plausible by another confusion. The experi­ ence of having a desire fulfilled is always pro tanto and for the moment pleasant. So, whe.aever I desire anythingj I foresee that if I get it I shall have the pleasure of fulfllled desire, It is easy to slip from this into tbe view that my motive for desiring X is the pleasure of ful61:ed desire which I foresee that I shall enjoy i f I get X. It ls clear that


this \\ill not do. I have no reason to antici� pate the pleasure of fulfilled desire on get­ ting X unless I already desire X itself. It is evident then that there must be so11Ul desires which are not for the pleasures of fu!Bl!ed desire, Let u s call them "pr:imary desires;� and the others "'secondarv, Butler has abundantly shown that therd' must be some primary desires. But, as Sidgwick rightly points out, he has gone to exh·emes in the matter which are not logically justi­ fied. The fact that there must be primary desires is quite compatihle with Psycholog­ ical Hedonism, since i t is quite compatible with the view that all primary desres are for primary pleasures, i.e., for pleasures of taste, touch, smell, etc., as distinct from the pleasures of fuHllled desire. Still, introspect­ tion shows that this is not in fact so. The ordinary man at most times plainly desires quite directly to eat when he is hungry. In so doing he incidentally gets primary plea­ sures of taste and the .secondary pleasure of fuHllled desire. Eventuallv he mav become a gourmand. He will then eat because he desires the pleasures of taste and he may even make himself hungry in order to enjoy the pleasures of fuIBlled desire. There is a special form of Psychological Hedonism of which Locke is the main e"'P onent. This holds that all desire can he reduced to the desire to remove pain or uneasiness. The one conative experience -is aversion to present pain, not desire for future pleasure. The position is as follows. When I am said to desire some future state X this means that the contemplation by me of my non-possession of X is painful. I feel an aversion to this pain and try to remove it by trying to get X, Since in the case of some things the contemplation of my non­ possession of them is painful, whilst in the case of others it is neutral or pleasant, the question would still have to he raised as to why there are these differences. Perhaps the theory under discussion should not be counte d as a form of Psychological-. Hedo­ nism unless lt holds that mv av;rareness of the absence of X is painful if and only if I 1 ;


or) as perhaps we ought to say: Now a particular proof of this kind, as an element in vlhat \vill render him such as Plato's, naturally provokes hvo happy. In the Republic this view comes to



light in the sixth book. He there speaks of Tf'i O.:ya00v as that which every soul pur­ sues and for the sake of which lt does all it does� divining that it is something but being perplexed and unable to grasp ade­ quately ,vhat it i later if not immediately, give you peace of mind, the implication turn out to be to our interest to do so." still is that if being helpful did not give you This motive is appealed to so constantly , peace of mind, there would be no obliga­ that we are hardly aware of it. '.'Ve are told tion to be so. If those who utter these to be honest, but not because honesty is a precept, do not mean them egoistically, good thing: we are told that "Honesty they are misleading their listeners, for the pays" and "Honesty is the best po!icy"-the general impression that !.s often left !tpon best policy, of course, being the one that children and others who hear them is that most benefits us in the long run . "Be help- one should do good deeds simply in order to get something out of it for himself,

From Hunum Conduct by Jolm Hospers, 1961, by Harcourt, Brace, & VYorld, Inc. antl reprinted with their permission.


i Life, VcL 43, No. 8 (Augmt 19, 1957), _p. 108.


whether crude rewards like money and services, or intangible rewards like peace of mind. Morality, in short, is presented as an instntment of "enlightened self-interest.'� Of course it is possible that justice does pay; it may be that when you do acts that are right, you yourself will always be the gainer thereby, even though at the time the act is one of heroic se!f.sacri£.ce, At least this possibility is worth examining. \Vhen you do something good, are you al­ ways rewarded for it sooner or later? Is right action always to the advantage of the doer? Offhand, it would certainly seem ob­ vfous that the answer is '"'No." ·vve hear that the good die young and that the big-time crooks are the ones who get away with it. Yet a great many people, moral philoso­ phers and ot..1-iers, have ansvvered our ques· tion wi::h a "Yes." They have held that in the long run, when all the factors are con­ sidered, virtue does pay, crime never pays, and right actio:1 always brings benefit back to the door of the doer. Plato was the first to have held this view, and he devoted his 1ongest dialogue, the Republic, primru·­ ily to an attempt to prove it. Let us see briefly how be defended his position. First of all, according to Plato, it pays to live a just Ee because only in that way can you have the respect of your neighbors and friends- in fact, only io that way can you have any friends. If you are not a tru.s"hvorthy sort of person who pays his debts and keeps his promises, other people will not trust you and wm cease to have dealings with you. If you are trustworthy, you vvill earn the 1·espect and esteem of those around you. Therefore, if you expect others to do decently hy you, you would be wise to behave decently toward them. Only if you give \vill you also receive. Here, surely, is a pe:fectly sensible reason for being moral, a reason that \vi.11 appeal strongly to most peop'.e. In fact, it is prob­ ably the main reason in practice why peo­ ple are moral. Nevertheless, Plato does not set much


store by this argument for morality. After all, public opinion and public esteem are unreliable and quixotic. Your neighbors may respect you for things they shouldn't respect you for, such as raking in a million dollars in an illegal gambling operation; or they may feel contemptuous of you for not making lots of money, by whatever means. Or they may love you for virtues that they mistakenly think you possess; or th ey may hate you for misdeeds which you have never done. Any;,vay, if you are popular vvith them today, you may be unpopular to­ morro>' York:

creates mutual good feeling which is far more of a contribution to happiness than one might think until he has personally experienced suph feeling. The ability to share the joys of other human beings will be likely, then, to heighten the level of happiness possible in one's own life. If only one person can win the European trip and one cares only about himself, then all the other competitors for this prize will be disappointed and frus­ trated. But if the losers are able sincerely to congratulate the winner and even to share in their imagination the enjoyment of the holiday, their own satisfactions will be greatly increased-perhaps not as much as if they had gone themselves but yet far more than if they had been able to e x ­ perience only envy and resentment. If you lose the trip, you will at any rate have some satisfaction in the knowledge that someone will enjoy it; and if you win it, you will be able to enjoy it more in the knowledge that others are not hating you for having the opportunity that is denied to them. It is, then, to your own interest to develop within yourself attitudes that make it pos­ sible for you to share the joys of others. Your life will be richer, less mean and grasping, and no longer characterized by the corrosive feelings of hatred and envy. The success of one person will no longer be built upon the resentment and hostility of others. vVe have, then, another self­ interested reason for "being moral" : one which asks of us that we not merely behave morally \Vith regard to other human beings but that we develop attitudes to­ ward them from vvhich the actions will flow without effort and as a matter of course.

B. DIVINE COMMAND Let us tum, then, to a second ansvver to the question, ''Why should we be moral?"


-namely, "Because God will reward me if I am and punish me if I'm not." 1. "Let's grant," one may say, "that in this world the just are often unhappy and that the unjust are happy. But this truth doesn't apply to the next world, where all these scores will be set straight. Indeed, the very fact that there are such gross inequities in this world is one of the reasons why many people have argued that there must be a next world in which the inequities are overcome." 0£ course, the fact that this world is unjust doesn't prove that there is another world that is just, any more than the fact that people are hungry proves that there will always be food. As an argument for an afterlife, this one is as much of a non sequitur as any argument could be. But it would take us too far afield here to con­ sider the various arguments for immor­ tality. However we may have arrived at the conclusion that there is another life in which all the discrepancies between vir­ tue and happiness will be set right, doesn't our conclusion solve our problem? If our conclusion is true, we need only smile at Plato's hopeless attempt to prove the im­ possible thesis about justice in this world; and when we hear more and more exam­ ples of men being good but unhappy or bad but happy we need only say, "But you see, it doesn't always pay to be moral, when you consider only this life; but when you consider the life to come, you will realize that it does pay to be moral. It is only in the life to come that the prophecies will come true, that the bread that you cast on the ·waters will come back to you, and vvith compound interest." The first thing to observe about this second answer is that it will appeal only to those who already accept the doctrine of an afterlife in which a just God dis­ tributes rewards and punishments accord­ ing to merit. Those who deny the doctrine or aren't sure will not be moved by it as a reason for morality. Since the argument depends on the truth of this doctrine, one


had best be quite sure that it is true and that all competing claimants to the true doctrine are false. If one staked his entire moral life upon the truth of a doctrine announcing eternal rewards, and this doc­ trine turned out to be false ( or some com­ peting doctrine true instead ) , this result would surely seem to b e something of a dirty trick. Moreover, if the only reason you have for being moral is that God will punish you if you aren't or reward you if you are, then the moment you doubt or no longer believe there is a God who will do these things, your only reason for being moral will vanish. Many people,. it seems, are in precisely this situation. Even for those who are already con­ vinced of the afterlife promised by their religion, there is another point they must squarely face: 'When a man says, '1 should be moral because if I'm not, I'll be pun­ ished," he is appealing as much to seIBsh motives, to self-interest, as our first an­ swer did. It is just self-interest pushed into the next world instead of being confined to this one. The person who acts from this motive is just playing the game for higher stakes. He is declaring his willingness to postpone his reward a bit longer in order to collect at a higher rate of interest in the neA-t world- which is as selfish a's it could be. It is like working longer hours in order to collect time and a half at the end of the day. Nor is the moral life conceived as something desirable in itself; it's not that the man enjoys or takes pride in the work, it's only that he wants the money. Indeed, his motive for being moral is not much different from his motive in following the commands of a dictator before whom he cringes and trembles. One may or may not approve or like what the dictator com­ mands, but one follows the commands be­ cause if one doesn't, he will be beaten or tortured. Once remove the threatened punishments and the promised rewards, and the person will no longer do as he was commanded.

WHY BE MORAL? There are many sincere and consci­ entious people who have believed that there was no good reason for being moral apart from divine punislm,enls and re­ wards and that once these '\.Vere removed, the moral . fabric of humanity would cr,1mble into dust. Of course, whether the moral fabric of humanity would crumble into dust if these threats and promises were removed, is an empirical question,. which could be argued pro and con indefinitely. It is possible that most people do have to be prodded constantly ,;ith spears to make them go fonvard. There are: however, reasons for belie,iug ( as we shall see in the next chapter) that in general people's morality would not crumble if religious sa...1ctions were - removed, so .long as the people had not become so accustomed to these religious sanc�ions that they could no longer operate without them; people who have not been brot:ght up to depend upon religious sanctions seem to be none the less moral. Whatever tho answer to the empirical question may be, hope of reward or dread of pt:nishment may offer an in­ centive to those who are too ,veak or selfish to behave moraLly ,vithout them, but it cannot offer a good reason ( Justi£cation ) for being moral If this statement begs the question of what we mean by a good reason, we ca..'"1 say that it offers a good reason only if completely selfish reasons are good reasons; it is simply an appeal to sel:-interest. "Be good or Tll punish you" -nothing could be a clearer appeal to naked and unbridled power than this. In fact the moral goodness of anyone who said it might be questioned simply because of the appeal to power not supported by reasons. Tbe moral irrelevance of the appeal to power can be brought out best, 2erhaps, in this way. If there is a God who is the $Onrce of moral cormnands, then either God had a reason for commanding what He did or He did not a. One alternative is to believe that God had no reason for commanding yvhat

739 He did but the commands are simply the result of an arbitrary whim or fiat. If soJ then what possible evidence do we have for believing that they are good or that we should obey them? We are told to obey them because we'll be punished if we don't. But that is exactly the same kind of reason-a prudential reason-for obeying the commands of a cruel dictator. \Ve obey him in fear and trembling because we fe;r the consequences of not doing so, not be­ cause we revere him or his commandment What sort of peop'.e should we strive to become so that vve as individ­ uals can be happy?'. The individual egoist, is ( or at least as well as the moral can be) vitally concerned ;v:ith this ques­ tion. Such a:u egoist, if he is ,vise, has considered the clain1s of moraJ.ity and has decided thac he will not attain genuine and lasting happiness by striving to be a morally good rnan, though it may be good tactics usually to be a man of good morals. It is his belief that the way of morality is not usually the way of happiness. And if and when it is, it ought to be pursued only because it will bring the pursuer happiness. To be sure, an intelligent :ndividual egoist ,'lill not go around proclaiming that every" Medlin1 op. cit,, p. 113.



one should only look after himself. He may, perverse moral view. But a consistent in­ if he is so inclined, pass on his insight to dividual egoism, intelligently pursued, is his family and some close friends, but he not a doctrine; it is not something that will not try to become an ethical egoist or would be articulated by an intelligent try to base conventional morality on ego­ egoist. Yet privately a person might adopt ism. This would be the very epitome of it as his policy of life. vVhy shouldn't he? ( Recall the foolishness. In certain contexts, he may even find it expedient to mouth "the high­ 'shouldn't' here is not a moral 'shouldn't'. ) minded pomposities of this morning's edi­ Why is he ( or is he?) irrational or mistaken torial". Such behavior, so to say, gives him if he follows this egoistic policy? Surely it a good press. But he has decided to act on is not in our interest for him to act the personal principle: Always look after immorally, but why shouldn't he or I or yourself and no one else, unless looking even you? Why should the "existing in­ dividual" who is trying to decide how to after someone else will benefit you. live happily, or significantly, opt for the True, there cannot be an egoistic way of life or vVeltanschauungphilosophie but point of view of morality rather than an there could be a deliberate, rationally intelligent and carefully controlled individ­ thought out and consistently adhered to ual egoism? Imagine yourself studying all the meta­ personal policy of individual egoism. Brun­ ethical treatises, the systems of normative ton correctly notes, «There can be intel­ ligent, self-controlled people, with a plan ethics, the sage advice of the vvise men, in of life, who care only for themselves". 7 short, all the claims of morality. Then Egoism cannot be an ethical doctrine but imagine yourself in the quiet of your the man committed to individual egoism study weighing up-not for others but for still has a use for 'I ought to consider only yourself alone- these considerations against my own good' as distinguished from 'I only the considerations in favor of individual care about myself'. The former nonnative egoism. Why should you choose to act ( though not moral) sentence indicates a morally rather than non-morally? This question, which is at least as old settled policy of action. The latter, by con­ as Plato, has been traditionally imbedded trast, indicates what may be only a mo­ mentary or very impermanent reaction. in the thick muck of metaphysics. 8 Often The token 'ought' when used in such a it has been confused ,vitl1 a lot of other context has more than just the common 8 It received a new coat with Donald \1/alhout's mark or noise in common with the token essay, "\Vhy Should I Be :tvforal? A Reconsidera­ 'ought' used in a moral context. In both tion," The Review of Metaphysics, vol. XII, No. 4 (June, 1959 ) , pp. 57Q.......88. Consider only ". . the instances they are only properly used if final theoretical answer" to our question is that in some way they indicate a settled policy ". . , one should be moral because this fits into a as distinct from a momentary whim, emo­ pattern of universal harmony of all things . . ." and the "universal harmony of all things can be tion or impulse. regarded as the ultimate culmination of all exist­ Thus, I do not see anything logically ence, not indeed as a description at any particular inconsistent about individual egoism so moment of time, but as an all-pervasive ideal.'' But such an ideal is not left to the whims of long as we don't try to extend it into a new mortal will for we are told "it may be regarded as rival morality or into an iconoclastic world rooted in the ultimate power of being that pro­ view. A view that exhibits a contempt for duces what is." Apparently it is too much to e>..'Pect that the days are over when this kind of all moral considerations whatsoever, could philosophy could be written. l'Valhout sees there is not possibly be a moral view, not even a a problem about justifying the moral point of view 7

J. A. Brunton, "Egoism and Morality," Philo­ sophical Quarterly, vol. VI ( 1956), pp. 298-9.

that was not adequately met by Bradley and · Pri­ chard but in answering what he calls "the ultimate question" he gives us this nonsense.

WHY SHOULD I BE MORAL? questions and recently it has been too lightly dismissed as nonsensical or absurd. The feeling emerges that finally there is no real argument here, one \vay or another; one must just opt for one policy rather than another. Here Sartrean or Kierkegaardian talk about decisions and anxiety seems correct Subjectivism again raises its ugly head. \Ve are tempted to say that here decision or commitment is king. Emotional energy may go i:1to our comrnitment to morality but in a "cool hour" we cannot discov;r decisive reasons for acting in either way. There seem to be no decisive reasons for our choice here; nor can we conceive of a non-question-begging general procedure that would enable us to · decide between these conilicting policies.• Reflective people, uncorrupted by philosophical theories, can be brought by ordinary reflection over morality to recog­ nize the point I have just made. The non­ philosophical idioms 'It's a value judgment' or 'It's finally a matter of what sort of a person you want to be', reflect just this point, The· very anxiety that any slight reference to subjeclivisn1 arouses in some people's breasts counts ( I believe) for rat.'1er than against my claim. \Ve do not come to a conclusion of tlii.s sort unarn­ bivalently. b reflecting about morality and human conduct, we are tempted finally to say that you must just decide what sort of person you want to be. No intellectual con­ siderations will settle the matter for you here. It is just this belief that seems to me to be the comr.:wn sense core of subjectiv� ism. But is it a be!ief that we can and should accept as clear-minded, rational human beings? Can we rationally defend taking a moral point of view? Are there decisive reasons for accepting the claims of morality such that any rational "un-moved spectator of the actual" would have to as0 See )V. H, 'vValsh, "Scepticism About Jvforals and Scepticism About KnO\vledge," Philosophy, vol. XJG{V (July 1960 ) , pp, 218-34.


sent to them? In the next section I shall tum to this question. JI

·why then be moral? We need initially to note tliat this question actually ought to be broken dov;'Il into two questions, namely; 1 ) '\.Vhy should people be moral?• or 'Why should .there be a morality at all? ' a:id 2 ) 'Why should I be moral? '. As will become evident, these questions ought not in the name of clarity, to be confused. But they have been run together; in .asking for a justification for tbe institution of morality both questions are relevant and easily con­ fused. 'W11y be Moral? ' nicely straddles these questions. In this section I shall first examine some traditional, and I believe unhelpful, answers to the above general questions. There the general question is not broken down as it should be and in examining these views I shall not break it down either. A£ter noting the difficulties connected with these approaches, I shall state what I believe to be a satisfactory answer to the question, 'Why should there be a morality at all? ' and indicate why it leaves untouched the harder question, 'Why should I be moral? • There is a prior consjderation that \Ve must first dispose of. In consideri:::ig both 0£ these questions we must be careful to distinguish the causes of a man's being moral from the reasons he gives for being moral. If one is a little careful about the implications of the word 1ikes, Bradley seems perfectly right in saying: "A man is moral because he likes being moral; and he likes it, partly becauBe h e has been brought up to the habit of liking it, and. partly because he finds it gives him what he ,wants; ,vhile its opposite does not do so". 10 In other words people are moral 1

" F. H. Bradley, Ethical Studies {The Liberal A:is Press, ::951 ); p, 7.


primarily because they have been condi­ tioned to be moral. The human an:imal is a social animal and ( as Butler and Hume observed ) people normally tend to con­ sider the welfare of others as well as their m,v n iv. elfare. People indeed act selnshly but they also take out life insurance, feel anxiety over the troubles of others, and even have moments of mild discomfort at the thought that life on this planet may some day be impossible. People react in this way because they have been taught or conditioned to so react But: the 'because� here is explanatory a;:id not justificatory. It explains m a very general way what makes or causes people to b e moral. But the ques­ tion I an1 concerned ,vith here is a quite different one. In asking, 'Why shocld peo­ ple be moral?', I am asking the question, 'What good reasons do people have for be­ mg moral?'. In asking about the justification for acting morally, I am only incidentally concerned with_. an explanation of the causes of moral behavior. VVhat good reasons are there for being moral? And if there are rrood reasons for being moral are they suffi�ient or decisive reasons? Tiiere is a short, snappy answer to 111y question. The plain man might well say: 'People ought to be moral because it is vv:icke I strive to be one sort of a person rather than another without any sufficient rational guides to tell me what J am to do? Does it come to just tha"�finally? Subjec­ tivists say ( at such a juncture) that there are no such guides. And this tirne there se6!:Ils to be a strong strand of common sense or hardheaded street '-visdom to back up the subjectivists' position. I do not believe that we are that badly 11:, 1vfedlin, op. cit,, p. 1I3; italics mine. 21

Ibid,, p, 114.

off. There are weighty considerations of a mundane sort in favor of the individual's taking the moral point of view. But I think the subjectivists are right in claiming that it is a mistake to argue that a nrnn is simply irrational if he does not at all times act morally. It is indeed true that if a man deliberately refuses to do what he acl.-nowl­ edges as morally required of him, we say he is irrational-or better, unreasonable. But here 'irrational' and 'unreasonable' haye a distinctively nwral use. There are other quite standarc! employments of the '-VOrd in which ,ve would not say that such a 1nan is irrational. 22 In all contexts the \:vorG irrational' has the evaluative force of strongly condemning something or other. In different contexts the criteria for \.vhat is to be called 'irrational' differ. In Toul­ min's terms the c:iteria are field-de:,endimt and the force of the word i� field­ independent In saying a man acts irration­ ally in not assenting to any moral consider­ ations whatsoever i,ye need not be claiming t.>iat he makes any mistakes in observation or deduction. Rather we are condemning him for not accepting the moral point of vie'-\'". But he is asking ,vhy he, as an indi­ vidual in an ongoing community, should al-vvays act as a 1noral agent He is not asking for motivation but for a reason- for being a morally good man. He ,wants to know what intellectual mistake the man who acts non-morally must make. To be told such a man is immoral and in that sense is unreasonable or irrational is not to the point. The subjec+..ivist I am interested in contenc!s that in the nature of the case there can be no reasons here for being moral rather than non-moral One must just decide to act one way or another withoi;.t reasons. There is much to be said for the subjectivist's claim here but even here I think there are rational considerations in favor of an individual's opting for morality. 1

I have discussed this issue in my "Appealing to Reason/' Inquiry, vol. 5 ( Spring, 1962}, pp. 65-84. 22



What is the best thing to do' into the question vVhat is the best thing for me to do". In moraJity ·we are concerned ,vith Before I state and examine tb.ose consider­ what is right, �vhat i s good and what is ations I \VOuld like to show ho\v t"vo recent supported by the best reasons, period; but tantalizinglv straightfonvard answers \Vill recall that the individual egoist ls challeng­ not do. Baier has "iiffered one and Hospers lllg the sufficiency of moral reasons v/hich the other. we, as social beings, nonnally grant to the Baier says that "\Vhen we ask ··vvhy moral enterprise. ( VVe need to reflect on should I be moral? , we are asking '\,Vhich the sense of 'sufficiency' here, The egoist is the course of action supported by the is not challenging the point of having moral best reasons? �. Since we can show a1ong codes, He is challenging the sufficiency of Hobbesian lines that men generally have the moral life as a device to enha.>1ce his better reasons for being moral than for happiness. But is this "a goal of morality'' ? being non-moral the individual has "been It is not. ) He is asking for reasons for his given a reason fo:r being moral, for follow� acting morally and unfortunately Baier's ing moral reasons rather than any other short answer does not meet the question . . .". The reason is simply that "they are Baier sets out to ans\ver, though as I have better reasons than any other". Bat in lhe already indicated it does answer the ques­ above type situation) when I am asking, tion, 'Why should people be moral? ' ".Vhy should I be moral? ', I am not con­ Hospers has a different argument cerned ,1rith ,vhich course of action is sup­ ,vhich, v:hile wrong, carries a crucial in� ported by the best reasons sans phrase or sight that takes us to the very heart of our with what is the best thing to co for all argument. Like Baier, Hospers does not concerned. I am onlv conce�ned ,vith v;rhat keep apart the question 'Why should I be is a good reason fo"'r me. I want to know moral? , from vv11y should people be what is the best tMng for me to do; that is, rnoral? '. After giving a psychological ex­ l want to know what ,vill make for my planation of what motivates people to b e moral, Hospers considers what reasons greatest good. Baier might point out that an indi­ there are for being moral. vidual has the best reasons for acting Virtue is its ovvn re\vard and if a!!. act mora1Iy because by each man's acting mor­ is in deed right this is a sufficient reason for ally the greatest possible good "ill be performing the act. \Ve have been oper­ realized. Yet, if the reference is to men ating on the wrong assumption-an severally and not to them as a group, it assumption that we inherited from Plato might well be the case that an individual's -namely, that if it isn't in our interest to acting immorally might in effect further the behave 111orallv we have no reason to do it total good, for his bad example might spur But i t does no't follow that if a right action ot..½.ers on to greater acts of moral virtue. i� not in our interest we have no reason for But be that as it may, the individual egoist doing it. If we ask 'Why should we do l:bis could still Iegifanately reply to Baier: 'All act rather than other ai:,"ts vie m.ight have of what vou sav is irrelevant unless real­ done instead?' the ansvver 'Because it is the ization of the g;eatest total good serves my right act is> says Hospers, �the best anS\ver best interests. vVhen and only when the and ultimately the only answer". 23 reasons for all involved are also the best It is indeed true that if we are reason­ reasons for me am [ personally justified in ing from the moral point of view and if art adopting the moral point of view'. Hospers, Human Conduct: An lntroduc­ We cant of course> criticize a so-called tiott John to the Problems of Etht'.cs ( New York, 196l), ethical egoist :or translating the question p. 194.





act is genuinely the right act to do in a all but at hest a non-rational expression of given situation, then it is the act we a personal predicament Onr problem has should do. Once a moral agent knows that been dissolved- the ""common sense core such and such an action is the right one to of subjectivism" has turned out to be the do in these circumstances he has eo ipso core of the onion. But has it reaI:y? Is any further ques­ been supplied with the reason for doing it tion here but a confused request for But in asking Why should I be moral? ' an individual is asking why he should ( non­ moti'.vatfon to do what we know we have moral sense of 'should' ) reason as a moral the best reasons for doing? Let us take agent. He is asking, and not as a moral stock. Hospers has in effect shown us: ( 1) agent1 what reason there is for his doing That x's being right entails both x should be done ( where 'should' has a moral use) what is right It is at this point that Hospers reply and there is ( from the moral point of ,iew) -··and his implicit defense of his simple a sufficient" reason for doing x ('I ought answer--exhlhits insight. It will, Hospers to do what is right' is a tautology where points out, be natural for an individual to 'ought' is used morally ); ( 2) That from the ask this question only when "the perfor­ point of view of self-interest the only rea­ mance of the act is not to hi,; U'Wn inter­ sons tbat can be sufficient reasons for acting est".24 I t is also true that any reason we are self-interested reasons. This again is an give other than a reason which ,viii show obvious tautology. The man asking 'Why that what is right is in his rational self­ should I do what is right ,vhen it is not in interest will be rejected by him. Hospers my self-interest? ' has made a self­ remarks '11Vhat he wants, and he will ac­ contradictory request when he is asking cept no other answer, is a self-interested thi, question as a self-interested question. reason" for acting as a moral agent.25 But These two points must be accepted, this is like asking for the taste of pink for but what if an individual says: As [ see it, "the situation is ex hypothesi one ln which there are tvrn alternatives: either I act the act required of him is contrary to his from the moral point of view, where log­ interest. Of course it is impossible to give ically speaking I must try to do what is him a reason in accordance with his. inter­ right, or I act from the point of view of est for acting contrary to his interest"." 1 rational self-fr1terest, where again I must have a reason for acting in accordance seek to act accordi.':lg to my rational self. vtith my interest which is contrary to my interest. But is there any nw.son for me interests� is a contradiction. The man always to act from one point of view rather who requests an answer to ,.Why should I than another when I am a member in good do \vha t is dght when it is not in my standing in a moral cornmunity? True interest? ' is making a "'self�contradictory enough, Hospers has shown me that from request'\ \'Ve come back once more to the moral point of uiew I have no alterna­ Prichard and Bradley and see that after tive hut to try to do what i s right and from all our "question" is a logically absurd one a self-interested point of view I have no -no real question at all. The person asking rational alternative but to act according to "'the question�· cannot '\vithout self­ what I judge to be in my rational self. conb:adiction, accept a reason of self­ interest. But what I want to mow is what interest for doing what is contrary to his I am to do: \.Vhy adopt one point of view interest and yet he \\1ll accept no reason rather t.1ian another? Is there a good reason for me, placed as 1 am, to adopt the moral except one of self-interest."27 » His �·question is no real question at point of view or do I just arbitrarily .u Ibid., p, 194. choose, as the subjectivist would argue? =� Ibid, I do not see t..."'-lat Hosoers� maneuver Ibid. has shown this question to �be senseless or Ibid., p. 195. 2'l 7 �

WHY SHOULD I BE MORAL? an expression of a selfcontradictory request. Rather his answer in effect brings the question strikingly to the fore by showing how from the moral point of view 'Because ifs right' must be a sufficient answer, and how it cannot possibly be a sufficient an­ swer from the point of view of self-interest or from the point of view of an individual challenging the sufficiency of the whole moral point of view, as a personal guide for his actions. It seems that- we have two strands of discourse here -with distinct canons of justification. We just have to make up our minds which point of vie-..y we wish to take. The actual effect of Hos­ pers' argument is to display in fine rational order the common sense core of subjectiv­ ism: at this paint we just choose and there can be :r;io reasons for our choice. It will not do for Hospers to argue that an individual could not rationally choose a non-moral way of life or ethos, for in choosing to act from a self-interested vantage point an individual is not choosing a way of life; he is, instead, adopting a personal policy of action in a very limited area for himself alone. Such an individual might well agree with Hospers that a r a ­ tional way o f life is one, the choice of which, is ( 1 ) free, ( 2 ) enlightened, and ( 3 ) impartial.28 This remark, he could contend, is definitive of what we mean by 'a rational way of life'. An intelligent egoist would even urge that such a way of life be adopted but he could still ask himself ( it wouldn't be prudent to ask others ) what reason there would be for him, or any single individual living in a community committed to such a way of life, to act in accordance \Vith it. ( This need not be a question which logically speaking requires a self-interested answer. An existing indi­ vidual is trying to make up his mind what he is to do. ) To reply, 'If it's rational then it should be done\ is to neglect the context­ dependent criteria of both 'rational' and Lshould'. There are both moral and non2!I

Ibid., p. 585.

759 moral uses of 'should' and 'rational'. In the above example Hospers is using 'rational' in a moralistic sense; as Hospers puts it, "Let me first define 'rationality' with re­ gard to a way of life" and while a way of life is not exhausted by moral considera­ tions it essentially includes them.29 Only if 'rational' and 'should' belong to the same strand of discourse is 'If it is rational then it should be done' analytic. Something could be rational from the mDral point of view ( morally reasonable) and yet impru­ dent ( irrational from the point of vie\V of self-interest) . If we were asking what we should do in terms of self-interest, it would not follow in this case that we should do what is rational in the sense of Lmorally reasonable'. Conversely, where vVhat is rational' means 'What is prudent' it would not follow that what is rational is what, morally speaking, we ought to do. 3 0 Thus, it seems to me that neither Baier's nor Hospers' answers ,vill do. \Ve are left with our original question, now made somewhat more precise, 'Is there a good reason for me as an individual in a moral community to always act morally no matter how I am placed? '. There is no room in moraiity for this question but this question can arise ,vhen we think about how to act and when, as individuals, we reflect on what ends of action to adopt. But as a result of Hospers' analysis, must we now say that here we must 1 ) simply make a choice concerning how to act or 2 ) vvhere there is no live question concern­ ing how to act it is still the case that there can be no non-question begging justification for an individual, were he faced with such a choice, to act one way rather than another? ( Of course there is the very best moraZ -justification for his acting as a moral agent. But that is not our concern here, for we are asking: why reason morally? ) Here the pull of subjectivism is strong z,i 30


See here VVilliam Dennes, "An Appeal to Reason," in Reason-, University of California Publi­ cations in Philosophy, vol. 7 ( Berkeley, Califor­ nia, 1939 ) , pp. 3-42 and Some Dilemmas of Natu­ ralism ( New York, 1960 ) , Chapter 5.

760 - and at this point it has an enlightened common sense on its side. But I think there is something more to be said that will take the bite out of such subjectivism. In trying to bring this out, I am in one sense going back to Plato. It is, of course, true that we 7 can t ask for a self-interested reason for doing what is right where ex hypothesi the action is not in our self-interest. But in actual moral situations it is not so clear what is in our self-interest and what is not, and often what is apparently in our self­ interest is really not. Part of my counter to the subjectivist, and here I am with Plato, is that if a man decides repeatedly to act non-morally where he thinks he can get away with it, he will not, as a very general rule, be happy. This isn't the whole of my case by any means, but I shall start ,vith this considera­ tion.

V Suppose that I, in a fully rational frame of mind, am trying to decide whether or not to adopt individual egoism as my personal policy of action. I ask myself: 'Should I pursue a selfish policy or should I consider others as well even when in my best judg­ ment it doesn't profit me?'. In my de­ liberation I might well ask myself: 'Will I really be happy if I act without regard for others? '. And here it is natural to consider the answer of the ancients. Plato and Aristotle believe that only the man who performs just actions has a well-ordered soul. And only the man with a well-ordered soul will be "truly happy". If I am thrown off course by impulse and blind action I will not have a well-ordered soul; I ,viII not be genuinely happy. But the alternative I am considering is not betvveen impulsive blind action and rational, controlled action, but between two forms of deliberate, ra­ tionally controlled activity. vVhy is my soul any less well-ordered or why do I realize

WHY BE MORAL? myself ( to shift to Bradley's idiom) any the less if I act selfishly than if I act morally? If it is replied, 'You will "realize yourself more" because most people have found that they are happiest when they are moral', I can again ask: ·But what has that to do ,vith me? Though I am one man among men, I may not in this respect be like other men. Most people have neurotic compulsions about duties and are prey to customary taboos and tribal loyalties. If I can free myself from such compulsions and superstitions will I be any the less happy if I am selfish? I should think that I would be happier by being intelligently selfish. I can forget about others and single mindedly go after what I want'. To this last statement Plato and Aris­ totle would reply that by always acting selfishly a man will not fully realize his distinctively human arete. By so acting, he simply will not be responding in a fully human way. Vve say of a man that he is a 'good man, a truly happy man' when he performs his function well, just as we say a tranquilizer is a 'good tranquilizer' when it performs its function well; that is to say, when the tranquilizer relaxes the tense, harrassed individual. But can we properly talk about human beings this way? We do speak of a surgeon as 'a good surgeon' when he cures people by deftly performing operations when and only when people need operations. Similarly, a teacher is ·a good teacher' if he stimulates his students to thought and to assimilate eagerly "the best that has been thought and said in the world". We can indeep speak of the arete or '\rirtue" of the teacher, £reman, preacher, thief or even ( as Maclver re­ minds us ) of the wife or unmarried girl.3 1 People have certain social roles and they can perform them ill or well. ..In this sense we can speak of 'a good husband', 'a good father', 'a good Chancellor of the Ex­ chequer . . .", but- Maciver rightly cona1. A. M. Maclver, "Good and Evil and Mr. Geach," Analysis, vol. 18, No. 1 ( October, 1957 ) , pp. 7 -13.


eludes- hardly of "a good man". 32 People, qua human beingsi do not seem to have a function, purpose, or role. A child can sensibly ask: 'What are hammers for? ', 'What are aspirins for? ', 'What are dentists for? ', but if a child asks 'What are people for? ', we must point out to him that this question is not really like the others. 'Daddy, what are people for?' is foolish or at the very least- even for the Theist -an extremely amorphous question. At best ,ve must quickly strike some religious attitude and some disputed cosmology must be quickly brought in, but no such exigency arises for the cosmologically neutral ques­ tion, 'Daddy, what are napkins for?' or 'Daddy, what are policemen for?'. After all, what is the function of man as such? In spite of all his hullabaloo about it, is not Sartre correct in claiming that man has no "essence"- no a priori nature-but that hu­ man beings are what human beings make of themselves? If a human being acts in an eccentric or non-moral way are we really entitled to say he is any less of a human being? If we counter that we are indeed en­ titled to say this, and we then go on to say 'By not acknov.rledging that we are so entitled, we are in effect overriding or ignoring man's "distinctively human qual­ ities" ' are we not now using 'distinctively human qualities' primarily as a grading label? In such contexts, isn't its actual linguistic function primarily moral? We are disapproving of a way of acting and at­ tempting to guide people away from pat­ terns of behavior that are like this. If we say the consistently sel£sh man is less human than the moral man, are we not here using 'less human' as a moral grading label and not just as a phrase to describe men? 'fl/lore human', on such a use, would not be used to signify those qualities ( if there are any) which are common to and distinctive of the human animal; but would be used as an honorific moral label. And if it is used only to describe how people Ibid., p. 8. 1



have behaved then it is perfectly possible for me to ask, '\Vhy should I be more human rather than less?'. Niost moderns would not try to meet the question 'Why should I be moral?' in this Greek way, though they still would be concerned with that ancient problem, 'How should I live in order to be truly happy? '. A rational man might make this elementary prudential reflection: 'If I am thoroughly and consistently sel£sh and get caught people will treat me badly. I will be an outcast, I will be unloved, all hands will be on guard against me. I may even be retaliated against or punished as an "irredeemable moral beast". All of this will obviously make me suifer. Thus, I better not take up such a seIBsh policy or I will surely be unhappy'. At this point it is natural to take a step which, if pushed too far, cannot but lead to a "desert-island example". It is natural to reply: 'Clearly it would be irrational to appear sel£sh. But I don't at all propose to do that. I only propose to look out for '"number one" and only "number one", I will do a good tum for others when it is likely, directly or indirectly, to profit me. I will strive to appear to be a man of good morals and I will do a good deed when and only when it is reasonable to believe there will be some personal profit in it. Surely, a policy of unabashed, outright sel£shness would be disastrous to me. Obviously, this is something I will strive to avoid. But I shall keep as the maxim of my actions: Al­ ways consider yourself first. Only do things for others, when by so acting, it will profit you, and do not be frankly seIBsh or openly no harm is likely to befall you for so acting. Take great pains to see that your selfishness is undetected by those who might harm you'. But, at this point our hypothetical rational egoist would need to consider the reply: 'You will regret acting this way. The pangs of conscience will be severe, your superego will punish you. Like Plato's aggressive except in those situations where



tyrant you 'Nill be a miserable, disordered power of Gyges and that power included man. Your very mental health will be the power to still the nagging voice of my superego, would it not then be reasonable endangered.� Imperceptibly dra,vit1g nearer to a for me to ahvays act in my O'\Yll self-interest desert.island example, the egoist might re­ no matter what the effect on others? If ply, 'But the phrase "mental health" is used there were some non-hannful pill-some to describe those well adjusted people who moral tranquilizer--that I could take that keep straight on the tracks no matter what. would "kill" my conscience but allow me I don't intend to be "·nealthy" in that sense. to retain my prudence and intelligence And, I do not recognize the authority of why then, under those circumstances, conscience. lvly conscience is just the in� should I act morally rather than selfishly? temalized demands of Father and Tribe. VVhat good reason is t..�ere for me in that But why should I assent to those demands, situation to act morally if I don't want to?' It is not sufficient to be to!d that if when it doesn't serve my interests? They are irrational, compulsive moralistic de­ most people had Gyges' ring ( or its mod­ mands, and I shall strive to free myself em, more streamlined, equivalent) they would go on acting as they do now, The from them'. To this it might be countered, question is not '\'Vhat would most people 'Granted that conscience has no moral or do i: they had Gyges' ring?' or even 'What even rational authority over you, you un­ would I do if I had Gyges' ring?' The fortunate man, but practically speaking, question is rather, 'What should I do?'. At you caunot break these bonds so easily. this point can reasons be found \Vhkh Consciously you may recognize their lack would convince an intelligent perso::: that of authority but unconsciously they have even in this kind of situation, he ought to and always will continue to have-in spite act morally'? That is J "\Vould it serve his of all your ratiocination-a dominating grip �'true interests" ( as Plato believes) for him on you. If you flaunt them, go against them, to be 1noral1 even in the event these con­ ignore them, it will cost you your peace of ditions obtained? It is just here, I believe, that subjec­ miud, you will pay in psychic suffering, happiness will be denied you. But as a tivisrn quite legitimately raises its ugly rational egoist happiness is supposedly your head. If the above deserUsland situation goal. And it is ,ruhful thinking tc, think did in fact obtain, I think we would have some psychiatrist will or can take you to say that whefoer it would or would not around this comer. Neither psychoanalysis be in your "true interests" to be moral or nor any other 1.ind of therapy can obliterate non-moral would depend on the sort of the "voice of the superego". It can at best person you are. With the possible exception diminish its demands when they are of a fe\V St Anthonys, we are, as a matter excessive. Your conditioning was too early of fact, partly egoistic and partly other­ and too pervasive to turn your back on it regarding in our behaviour, There can be now. If you are rational you vv:ill not strug­ no complete non-personal, objective justi­ gle in such a wholesale fashion against fication for acting morally rather than nonthese ancient, internalized demands. Thus, 1norally. In certain circumstances a person you should not act without regard to the of one temperament would :find it in his dictates of morality if you really want to interests to act one way and a person of another temperament to act in another. be happy', It is at this stage that the rational We have two policies of action to choose egoist is likely to use his visa to Desert fr01n1 with distinct criteria of appropriate� Island. He might say: 'But if I had the ness, and which policy of action will m ake


us happy will depend on the sort of person we happen to be. It is here that many of us feel the '�existential bite" of our question. Students, who are reasonably bright and not a little versed in the ways of the world; are often ( and rightly ) troubled by the successive destruction of first psychological egoism and then ethical egoism. They come to see that individual egoism can't be a moral view, but they feel somehow cheated; somehow, in some way, they sense that something has been put over on them. And I think there is a point to this rather com­ mon and persistent feeling and I have tried, in effect, to show what this is. I would not, of course, claim that it is al­ ways the "'l'Vhy-should- I -be-moral?" ques­ tion that troubles as a reflective student at this juncture but frequently, like Glaucon and Adeimantus, the student wans to know · why, as a solitary, flesh and blood individ­ ual, he should be moral. He feels that he should be moral, but is he somehow being duped? He wants a reason that will be a good and sufficient reason for his being moral, quite apart from his feelings or at­ titudes about the matter. He does not want to be in the position of finally having to decide, albeit after reflection, what sort of person to strive to be. It seems to me that the subjectivists are right in suggesting that this is just what he finally can't avoid doing, that he doesn't have and can't have the kind of objectivity he demands here. vVe need not have existentialist dramatics here, but we do need to recognize the logical and practical force of this point. Most rationalistic and theological ethical theories seem to me mythmaking devices to disguise this prima facie uncomfortable fact.


correctly placed this irreducible element of choice in reasoning about human con­ duct? Perhaps some will despair but since it is not the job of a philosopher to be a kind of universal Nannie I don't think he need concern himself to relieve this de­ spair. But, I think, if he will remind people of the exact point on the logical map where this subjectivism correctly enters and make them once more aware of the map as a whole they will- now able to see the forest as well as the trees- be less in­ clined to despair about the rationality of their acting morally. If one is willing to reason morally, nothing we have said here need upset the objectivity and rationality . of moral grading criteria. More importantly here, to admit subjectivism at this point does not at all throw into doubt the Hob­ besian defense of the value of morality as a social practice. It only indicates that in the situation in which an indioidual is 1 ) very unlikely to be caught, 2 ) so rationally in control that he will be very unlikely to develop habits which would lead to his punishment, and 3 ) is free from the power of his conscience, it might, just might ( if he were a certain kind of person) make him happier to be non-moral than moral. But this is not the usual bad fellow we meet on the streets and the situation is anything but typical. A recognition of the irrelevance of desert-island examples will provide further relief from moral anxiety, over such sub­ jectivism. Critics of utilitarianism invent situations in which a social practice is, as we use moral language, regarded as ob­ ligatory even though there is no advantage in acting in accordance with it in this par­ ticular kind of circumstance. They con­ struct desert-island examples and then crucify the utilitarian with them. They point out, for example, that promises made VI on desert islands to a dying man to dispose of his effects in a certain way are But need we despair of the rationality of considered obligatory even if it is clear the moral life once we have dug out and that 1 ) some other disposal of his effects


would be more beneficial and 2 ) that there is no reasonable chance that the breach in trust would be detected. The usual utilitar' ian answer is that disregarding promises of this sort would \Veaken our moral charac­ ter; and, in addition, we cannot be quite sure that such · a breach in trust would not be detected or that it would really do more good than harm. Further, to ignore a prom­ ise of this sort is bad, for it would tend to weaken the utility of the social practice of promise-keeping. No\vell-Smith, however, is quite cor­ rect in saying: "The relentless desert­ islander can ahvays break such utilitarian moves by adding stipulations to the terms of the original problem". 33 That is, he will say to the utilitarian, 'But what would you say if breaking a trust in situations of this type would not weaken the utility of the practice of promise-keeping? Surely it is intelligible to suppose that such acts would not weaken people's moral fiber, would not be detected, and would not do more total good than harm. To this the utilitarian can only say that this statement of the desert­ islander is a very "iffy proposition", indeed. Nowell-Smith rightly remarks: "The force of these desert-island arguments . . . depends expressly on the improbability of the case supposed".34 "It is difficult to assess their force precisely because the case is improbable and therefore not catered for in our ordinary Ianguage". 35 The language of human conduct has the structure it has because the world is as it is and not otherwise. If people and things were very different, the structure of moral codes and the uses of evaluative language presumably would be different. The very form of our talk about human conduct "re.B.ects empirical truths that are so general and obvious that we can afford to ignore exceptions". 3 6 If through desert-island examples we withdraw that pervasive con"" P. H. Nowell-Smith, Ethics, p. 240. -' Ibid. a:i Ibid. M Ibid., p. 132. 3


te>.tual background it is difficult to know what is the logically proper thing to say. The logic of the language of human con­ duct did not develop with such wildly improbable situations in view. It, after all, has a wide range of distinct, practical uses, and it only has application in a certain type of setting. If one of these desert-island situations were to obtain, we would have a good reason, as Wittgenstein clearly saw, to make a linguistic stipulation, that is, we would have to decide what is to be said here and our linguistic decision would in­ deed be an intervention in the world, it would indeed have normative import. But it is neither possible nor necessary that we make all such stipulations in advance and we can hardly reasonably accuse the lan­ guage of conduct of inadequacy because it does not cater to desert-island cases. It would be like saying that "the language of voting". is inadequate because it does not tell us what to do in a situation in which a senior class, consisting of a thousand, tries to elect a president from four candi­ dates and each time a vote is taken each candidate gets exactly 250 votes. This in­ deed is a logical possibility, but that this logical possibility is not considered in setting out the procedures for voting does not at all indicate an inadequacy in our voting procedures. Our "Gyges' ring situations" are just such desert-island cases. In fact, Nowell­ Smith is quite correct in remarking that the Gyges' ring example in the Republic is a paradigm of all such desert-island argu­ ments. 'Would I be happier if I were intel­ ligently selfish in a situation in which I could free myself from guilt feelings, avoid punishment, loss of love, contempt of family and friends, social ostracism, etc.?'. To ask this is to ask a desert-island ques­ tion. Surely we can and do get away with occasional selfish acts- though again note the usual burden of guilt-but given the world as it is, a deliberate, persistent though cunning policy of selfishness is very



likely to bring on guilt feelings, punish­ decision to act one way rather than ment, estrangement1 contempt� ostracism another is a matter of arblb:ary choice? Ar.e and the lfae. A clever man might avoid one there paradigm cases whicli establish the or another of these consequences but it subjectivist's case-establish that it is alto- · \Vould be very unlikely that he could avoid gether likely that some clear-headed peo­ them all or �ven moit of them. And it is ple "�II be happier if, in some non-desert. trnistic to remark that we all want com­ island circumstances, they deliberately do panionship, love, approval, comfort; secur� what they acknowledge is wrong. . . Let us examine three prima facie ity and recognition. It is very unlikely that the consistently selfish man can get cases. Suppose a man, believing it to be those things he wants, At this point, :it may ,be objected: 'But suppose someone doesn't wrong, decides to be unfaithful to his wife want Lliose things, then what are we to when it is convenient, non-explosive and say? '. But this is only to burgeon forth unlikely to be discovered. Usually it is with anotl1er desert-island example. The not, on the part of the knight-erra.'1t hus­ proper thing to reply is that people almost band, a deliberate and systematic policy, universally are not that way and that in but it might be and sometimes is. Bored reasoning about whether I should or husbands sometimes dav-dream that this is should not be se:fuh, I quite naturally ap­ a retun1 to paradise; thit is to say, it might peal to certain very pervasive facts ( in· earn, at least in anticipation, a good score eluding facts about attitudes) and do not, ln a felicific calculus. In order to make the and need not, normally, try to find an example sufficiently relevant to the argu­ answer that would apply to all conceivable ment, we must exclude those cases in worlds and all possible human natures. which the husband believes there is nothing To think that one must do so is but to wrong in this behavior and/ or gives reasons exhibit another facet of the genuinely ir· or rationalizations to excuse his behavior. I must also exclude the guilty weak-willed rational core of rationalism. man with the Pauline syndrome. The case demands a man who deliberately- though with sufficiently prudent moderation-com­ VII mits adultery. It is important for our case that he believes adultery to he immoral. It seems to me that the above considera­ Nonetheless, while believing people ought tions cm.:at heavily against adopting a not to be adulterers, be asks himself, thoroughly consistent policy of individual ·should I continue to live this wav anywav? egoism. But do such considerations at all Will I really be happier if I go the way �f touch the individual who simply, on oc­ St. Paul? '. He dof'.s not try to universalize casion, when his need is great, acts in a his decision. He believes that to choose to way t.½.at is inconsistent \ the dictates of remain an adulterer is ,immoral, but the im­ morality? 'Will such a person always be moral choice remains for him a live option. happier-in the long run-if he acts con­ Though people may not put a!J. this to scientiously or is this a myth foisted on us, themselves so explicitly, such a case is not perhaps for good social reasons, by our an impossibility. People may indeed be­ religions and moralities? Are all the situa­ have in this way. My example is not a tions desert-:.Sland situations in which we desert-island one. I admit there is some­ can reasonablv claim that there could be thing odd about my adulterer that might rational men ,;ho would be happier if they make him seem like a philosophical papier acted non-n1oraUv rather than morally or mtlch-8 figure. There is a!so something con­ in which we wo{i]d have to say that· any ceptually odd about sayfag that a man


believes x to be wrong and yet, without guilt or ambivalence and without excusing conditions, rationally decides to do x. W'ith good reason we say, 'If he knows it to be ,vrong or really believes it to be wrong, he will ( everything else being equal ) try to avoid it'. Still there is a sense in which he could say he believes x to be wrong even though he seeks x. The sense is this: he would not wish that people generally choose or seek x. ·vvhen this is the case he says �x is vvrong' even though he makes a frank exception of himself without attempt­ ing to morally justify this exception. It is important to note that this is a special though perfectly intelligible use on my part of 'He believes it to be wrong'. While it withdraws one essential feature, namely that non-universalizable exceptions are in­ admissible, it retains something of the gen­ eral sense of what we mean by calling something morally wrong. Yet, for the sake of the argument at least, let us assume that we do not have a desert-island case. Assuming then that there are such men, is their doing what is wrong here also for them the per­ sonally disadvantageous thing? Can any individual who acts in such a way ever be reasonably sure he won't be caught- that one of the girls won't tum up and make trouble, that he won't run into an acquaint­ ance at the wrong time? Even if these seem to be remote possibilities, can he ever be free enough from them in his dream life? And if his dreams are bothersome, if he develops a rather pervasive sense of un­ easiness, is it really worth it? He must again consider the power of his conscience ( superego ) even though he rationally de­ cided to reject its authority. Will it give him peace? Will the fun be worth the nag­ ging of his conscience? It is difficult to generalize here. Knowledge of onselsef, of people, of human psychology and of im­ aginative literature is all extremely relevant here. I think the individual egoist can correctly argue that it is not always clear that he would be unhappier in such a situ­ ation if he did \Vhat was v,rrong. A great


deal depends on the individual and the exact particular circumstance, but the moralist who says it is never, or hardly ever, the case that a person will be happier by pursuing a selfish policy certainly over­ states his case. Let me now take a different paradigm for which much the same thing must be said. It is important to consider this new case because most people would label this man a "veritable moral beast" yet he stands to gain very much from acting immorally. The case I have in mind is that of a very intelligent, criminally experienced, well­ equipped, non-masochistic but ruthless kidnapper. He is a familiar type in the movies and thrillers. Now, Hollywood to the contrary, why should it not sometimes be the case that such a kidnapper will be happier if he is successful? Indeed, he may have a murder on his hands but the stakes are very high and when he is successful he can live in luxury for the rest of his life. vVith good reason our folklore teaches he would not be happier. It is of the ut­ most value to society that such behavior be strenuously disapproved. And given the long years of conditioning \Ve are all subject to, it remains the case that most people ( placed in the position of the kidnapper) would not be happier with the successful completion of such a kidnapping if it in­ volved murdering the kidnapped child. But then most people are not kidnappers. They have very different personalities. Such brutalities together ,vith fear of de­ tection would haunt them and it is proba­ bly the case that they also haunt many kid­ nappers. But if the kidnapper were utterly non-moral, very, very clever, etc., why wouldn't he really be happier? He could live in comfort; he could marry, have chil­ dren and attain companionship, love, ap­ proval, etc. 'VVell', we would say, 'his con­ science would ahvays bother him'. But, particularly with modern medical help, v,rhich he could now well afford, would it bother him enough? 'Well, there would al­ ways be the awful possibility of detection and the punishment that might follow'.


But, if the stakes were high enough and if he were clever enough might it not be better than a life of dull routine, poverty or near poverty? And think of the ''kicks" he would get in outwitting the police? We all have a little adventure in our souls. 'But'- the dialogue might go on- 'if he were intelligent enough to pull off this job suc­ cessfully, he would certainly be intelligent enough to avoid poverty and to avoid making his living in a routine, boring way'. The dialogue could go on interminably but I think it is clear enough again that even here there is no one decisive, clearcut an­ swer to be given. The case for morality here is stronger than in the previous para­ digm, but it is still not decisive. Yet there are paradigms in which doing what is clearly wrong ( and understood by the in­ dividual in question to be wrong) is in the rational self-interest of some individuals. Our first more typical paradigm is not com­ pletely clear, but the following third and less typical paradigm given by Hospers is a clearer example of a case in which it is in a man's self-interest not to do what is right. There is a young bank clerk who decides, quite correctly, that he can embezzle $50,000 without his identity ever being known. He

fears that he will be underpaid all his life if he doesn't embezzle, that life is slipping by with­ out his ever enjoying the good things of this world; his fiancee will not marry him unless he

can support her in the style to which she is accustomed, he wants to settle dmvn Vilith her in a suburban house, surround himself with books, stereo hi-fi set, and various objects a:art, and spend a pleasant life, combining cul­ ture "\vith sociability; he never wants to commit a similar act again. He does just what he wanted to do: he buys a house, invests the remainder of the money Vi!isely so as to enjoy a continued income from it, marries the girl, and lives happily ever after; he doesn't worry about detection because he has arranged things so that no blame could fall on him; anyway he doesn't have a worrisome disposi­ tion and is not one to dwell on past misdeeds; he is blessed with a happy temperament, once


his daily comforts are taken care of. The degree of happiness he now possesses would not have been possible had he nOt committed the immoral act. 3 7

Clearly it was in his rational self-interest to do what is wrong. Someone might claim that it is too much to expect that he could arrange things so that no blame would fall on him. This could happen only in de