Morality, Reason, and Truth: New Essays on the Foundations of Ethics 0847673685, 0847673693

Citation preview

Morality) Reason and Truth

MORALITY, REASON AND TRUTH New Essays on the Foundations of Ethics Edited by DAVID COPP AND DAVID ZIMMERMAN

ROWMAN & ALLANHELD PUBLISHERS

ROWMAN & ALLANHELD Published in the United States of America in 1985 lw Rowman & Allanheld, Publishers (A di\·ision of Littlefield, Adams & Company) 81 Adams Dri,·e, Totowa, New Jersey 07512 Copyright © 1984 by Rowman & Albnhcld, Publishers All rights reserved. No part of this publication may be reproduced, stored in a retric\'al system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior permission of the publisher. Library of Congress Cataloging in Publication Data

Main e11trv under title:

MoL1lity, re;1son .rnd truth. Bibi iography: p. Includes index. 1. Ethics-Addres�es, essays, lectures. I. Copp, D.n·id. II. Zimmerman, Da,·id, 1942170 BJ10 12.M6356 1984 84-13424 ISBN 0-8476-7368-5 ISBN 0-8476-7369-3 (pbk.) 84 8 5 86 / 10 9 8 7 6 5 4 3 2 Printed in the United States of Americ1

Contents Pret:1ce ....................................................... vii Int roduction DaJJid Copp................. : ................... l TRUTH AND THEORY ACCEPTANCE IN ETHICS Is There a Single True Morality? Gilbert Harman ...... 27 Moral Explanations Nicholas L Sturgeon ............... 49 Moral Realism and Explanatory Necessity David Zimmerman .................................... 79 The Explanation of Moral Language Richard R. Brandt ................................... 104 Two Approaches to Theory Acceptance in Ethics Norman Daniels . ..................................... 120 Considered Judgments and Moral Justification: Consen·,uism in Moral Theory David Copp ......... 141 Arguments from Nature Ronald de Sousa ............. 169

PART ONE

1

2 3

4 5 6 7

PART TWO MORALITY AND PRACTICAL REASON 8 Rationality, Reason, and the Good Kurt Baier ........ 193 9 Must the Immoralist Act Contrary to Reason? Kai Nielsen .......................................... 212 10 Reason in Ethics-or Reason \Trsus Ethics? Jan NarJJeson ........................................ 228 11 Justice as Social Choice DaPid Gauthier .............. 251 12 Sociobiology and the Possibility of Ethical Naturalism Richmond Campbell .............. 270 13 The Reconciliation Project Gregory S. Kavka .......... 297 Bibliographv ................................................. 320 Notes on the Contributors .................................... 329 L-

.;

Preface This is a collection of original, recent, and mainly previously unpublished essays on the foundations of morality . Several of the authors are established figures in the field; others are philosophers who are just beginning to make important contributions. Most of them spoke at a conference on reason in ethics organized by the editors a few years ago, and all of them were asked to be responsive in their papers to the work of the other contributors. The result is an anthology that constitutes a coherent multilayered debate among some thirteen philosophers. The introduction aims to identify the major themes and to provide a structure for the discussion. The topics and approaches are varied. It is worth stressing that the familiar issues in the semantics of moral discourse, emphasized so much in the work of ethical theorists for the first two-thirds of this century, are not given pride of place here. They are discussed to some degree. But if any one issue is most emphasized, it is the relation between morality and the theory of moral justification on the one hand and the theory of practical reason on the other. And if one approach is emphasized, it is an approach that seeks to explain the place of morality in a scientific view of the world. We hope that these essays will help to rekindle an interest in meta­ ethical issues. These issues are, after all, among the most ancient of philosophical problems, and they are certainly central to philosophy. Moreover, it is obvious that we need to attach value to our pursuits, both professional and nonprofessional, and so we must come to terms with skepticism about value. We think that if anything has value, then so does the pursuit of the nature of value. That pursuit is meta-ethics. References to the literature are given throughout the book in the form of an author's name and a date . The consolidated list of references at the end of the book contains the needed bibliographical details in each case, following Sen and Williams ( 1982). Among our debts of gratitude, most important is to our contributors, who in many cases revised their essays to our specifications and in all

Jilli

Preface

cases waited patiently while we assembled the book. We are grateful to Merrily Allanson, Dennis Bevington, Ishtiyaque Haji, Eric Pye, and Kalen Wild for their help with many of the tedious details. Simon Fraser University October, 1984

No-one is just willingly but under compulsion, .. . Every man · believes that injustice is much more profitable to himself than justice. Plato, The Republic, 360 c-d, translated by G. M. A. Grube ( 1974). Reason is and ought to be the slave of the passions. It is not contrary to reason to prefer the destruction of the whole world to the scratching of my finger. David Hume, A Treatise of Human Nature, I I, 3, u1. ( 1968), pp. 415-416. The ground of obligation . .. must not be sought in the nature of man or in the circumstances in which he is placed, but sought a priori in the concepts of pure reason. Immanuel Kant, Foundations of the Metaphysics of Morals, Ak 389, translated by Lewis White Beck ( 1959). If the inquirer goes on asking " But why should I live like that?" then there is no further answer to give him, because we have a]ready, ex hypothesi, said everything that could be included in this further answer. We can only ask him to make up his own mind which way he ought to live, for in the end everything rests upon such a decision of principle. R. M. Hare, The Language of Morals ( 1952), p. 69. The theory of justice 1s a part . .. of the theory of rational choice. John Rawls, A Theory of Justice ( 197 1), p. 16.

Introduction David Copp META-ETHICS W hat are we to make of morality? We are thoughtful c 1t 1zens in a modern pluralistic society, and we want to decide how to live and behave. We have deeply felt moral convictions, most of us, and we want to take morality seriously, but only if there is a rational basis to morality. If there is no reason to believe in any moral principle, and if there is no reason to commit ourselves to following any given moral precepts, then we may very well decide not to govern our lives by morality, and in doing so we may be deciding quite rationally. We may come to regard our moral convictions as simply an impediment to the pursuit of what we desire for ourselves and for those we care about. Our problem is that although we have moral beliefs we have some reason to doubt that they are on a secure footing. For one thing, there is no culturally entrenched rationally appealing process, comparable to scientific method, that we can appeal to for an answer to our plea for an articulated and defended morality, or even for answers to specific moral disputes. Moreover, tradition does not single out a privileged morality. In fact, society can be torn by disputes between parties, all of whom see themselves as rational and as possessing truth. Even now our own society faces disputes over questions of varying degrees of significance, and competing traditions are sources of competing answers. In any event, we want a morality with a rational foundation, whether or not it be traditional. The existence of seemingly fundamental disputes between apparently rational and well-intentioned people and the absence of an appealing decision procedure provide us with some reason to doubt the rational basis of the whole enterprise of morality. What then are we to make of it? The issue being raised is not about the specific content of the true, correct, or ideal morality; it is not about which moral principles to take into account in deciding how to live and how to behave. It is about

2

Introduction

the rationality of taking into account any distinctly moral principles at all. Content-related questions do of course arise in moral philosophy. But for present purposes they should be put aside, for the issue being raised here is a logically prior issue about the justifiability of any view as to the content of the ideal morality. It is about the underpinning or basis of moral principles. Actually, a number of questions are being raised: whether there is a true morality; what in the world could determine the truth or correctness of a morality; how we would recognize or have knowledge of the true morality, if there is one; whether "correctness" is even required in order for there to be a rational basis for taking our own moral views seriously; whether the known or rationally believed truth of a moral code is by itself enough to give us sufficient reason for con forming to it. These questions about morality, and other closely related ones, are addressed in the discipline known as meta-ethics. They are the subject of this anthology. Un fortunately, some of the concepts used to define meta-ethics are themselves problematical. The prime examples are the concepts of jus­ tification and of rationality. We ask such questions as whether there is a "rational basis" for morality, whether a "rational person" would commit himself to morality, whether belief in some moral principles is "rationally , justified,, whether there is a "sufficient reason" for con forming to morality, whether any moral code can be "justified." To understand these questions we obviously need to understand the notions of justification and ratio­ nality. But these are contested concepts in philosophy. There are different accounts, and in fact much of the debate in meta-ethics concerns how we arc to understand them. There fore, for the moment we need to rely on our commonsense understanding of matters. Some ideas seem ob­ viously correct. For instance, a rationally warranted or justified belief must, with some exceptions, be based on sound evidence or cogent reasons. The notion of a rational course of action or a rational policy is more controversial, but a common view is that rationality here is simply a matter of the efficient pursuit of what one would aim or desire to achieve overall, given due reflection and full awareness of what one is rational to believe. Another view is that rationality consists in max­ imizing one's "expected utility," a measure of one's preference for outcomes as weighted by their probability. Finally, a rational person is one whose beliefs and actions are by and large rationally warranted. The relationships among these concepts are discussed in several of the essays in this volume. This superficial sketch may help to explain the issue that lies at the bottom of meta-ethics. The fundamental idea is that a justification of morality would be a showing that there is a justification or sufficient reason for taking morality seriously, for governing one's life by it. A

David Copp

3

rational person presumably would not do so unless cogent reasons could be given for accepting some given set of principles as constituting the correct or ideal morality, and unless sufficient reasons could be given for accepting the policy of con formity to that morality. Here, then, is ·the source of the series of questions that we used to define meta-ethics. At root, the issue is whether morality has a basis that would allow it to be rationally justified. I will point out two different schools of thought as to what this should be taken to mean: the epistemic view and the practical view. There has been an understandable tendency to define meta-ethics in terms suggested by the approaches and theories of its most influential figures. For many years meta-ethics was approached from the standpoint of the Language of ethics, and it virtually became a branch of the study known as the semantics of natural language. It studied the meanings of the "ethical terms '" such as "good '" "bad '" "right '" and "ought '" and it studied the logical or semantical structure and status of the kinds of sentence that are distinctive of ethical discourse, such as "You ought not to lie" or " Honesty is the best policy." Early in this century, G. E. Moore wrote that the "most fundamental question in all Ethics" is how the term "good" is to be defined ( 1903, p. 5). Forty years later, Charles Stevenson began his book Ethics and Language with the remark that his first object would be "to clarify the meaning of the ethical terms-such terms as 'good,' 'right,' 'just,' 'ought,' and so on" ( 1944, p. 1). R. M. Hare devoted the first part of The Language of Morals to a study of the imperative mood, arguing that "the language of morals is one sort of prescriptive language," and the ordinary imperative sentence is the simplest form of that language ( 1952, pp. 1-2). And Philippa Foot argued that the expression "a good action" has a "fixed descriptive meaning," at least within "a certain range." 1 It is not surprising that the semantic view of meta-ethics prevailed, for these four writers were among the most influential on meta-ethics in the first two-thirds of the century. And they spoke for the four main types of theory: intuitionism, emotivism, prescriptivism, and naturalism. 2 Of course, there is a deeper reason for the semantic approach. As we saw, the issues in meta-ethics center around the problem of justifying morality to a rational person. Hence, if one's paradigm of a justification is the epistemic justification of a belief or of a theory) one may regard the justification of morality to require the justification of our moral beliefs, or of proposed moral propositions. If so, then the resources, problems, and strategies of general epistemology become germane to ethics. And one type of question will become central: What is the meaning of the sentences we use when we engage in moral discourse? What is the nature and logical status of the propositions, if any, that

4

Introduction

are expressed by us in moral discourse when we give expression to our moral convictions? Unless we can answer these questions, we will be unable to determine the exact content of the beliefs whose justification is of concern. We may not even be certain that our moral "convictions" are beliefs, strange as this may seem. We could put the point in more baroque language by saying that an understanding of the metaphysical or ontological commitments of morals would be prior to a moral epistemology, but then a semantics of moral discourse would be required as a condition of arriving at a reasoned appraisal of the ontology of morals. So, on the epistemic view of the problems of meta-ethics, semantic issues may seem to be central, even if not decisive. However, the epistemic view is not mandatory, for there are non­ epistemological justifications, for instance, of actions.It is fairly standard to distinguish between theoretical and practical reason: theoretical reason or rationality is at issue where the object of a putative justification is a belief, proposition, theory, or statement, and practical reason or rationality is at issue where the object of a putative justification is an action or a desire . Epistemological issues arise within the theory of theoretical reason: What counts as sound evidence or as a cogent reason for a belief? But if we are called upon to justify some action we have per formed, or to justify how we live or what we want, then our problem is not primarily epistemological. Evidence is not at issue, nor is cogency. Instead the issue is perhaps better said to be one of propriety or appropriateness: As we saw, a rational policy or action could be characterized as one that would be decided upon after due reflection and in full cognizance of the facts. At any rate, the issue seems to be of a different kind from that involved in justifying beliefs. And one might regard the problem of justifying a morality to be of this kind, to be a problem for a theory of practical reason: A morality is justified if, roughly, choice of it, or commitment to it, would be rational (for some defined purpose and in some defined context). This distinction between the epistemic view and the practical view was drawn by John Rawls: Moral philosophy includes moral theory, but takes as its main question justification and how it is to be conceived and resolved; for example, whether it is to be conceived as an epistemological problem (as in rational intuitionism) or as a practical problem (as in Kantian constructivism).3

Rawls views his own theory as providing a practical justification for principles of justice, following the lead of Immanuel Kant ( 1785). He says that "principles of justice may be conceived as principles that would be chosen by rational persons," and he adds that the theory of justice is a part of the "theory of rational choice" ( 197 1, p. 16). But the

David Copp

5

pra cti cal con ception is not restri cted to Kantian and neo- Kantian ap­ proa ches. For exa mple, Tho mas Hobb r.s's moral theory is often un derstoo d to exhibit a pra cti cal vie w of moral justifi cation: In vie w of our interest �n sel f-preservation, in for med an d rational persons would contra ct into a so cial syste m that in clude d a set of so cial rules and the po wer to en for ce the m; this is what justifies the syste m an d the rules that are en for ced in it ( 165 1; see Gauthier, 1979). A si milar approa ch was ta ken in the 1950s by Kurt Baier ( 1958). The pra cti cal con ception of moral justifi cation will see m reasonable if a morality is thought to be a set of standards or canons of behavior. On this pi cture, our models of morality would perhaps be ga mbling syste ms, invest ment strategies, and the li ke, although there are obvious differen ces bet ween poli cies of this nature and a morality. A set of rules for parlia mentary debate would perhaps be a better model. The point is that one would justify a poli cy or strategy of one of these kinds on the basis of standards of rational choi ce, and that the pra cti cal con ception vie ws moral justifi cation as another proble m for the theory of rational choi ce. On the other hand, the episte mi c con ception will see m reasonable if a morality is thought to be a theory as to the nature of a putative real m of fa ct. Psy chology and astrology would be our models on this pi cture. If a morality is a theory, then a cceptan ce of it woul d i deally be justified, if at all, on the basis of episte mi c stan dards. The de cision, then, bet ween these t wo con ceptions of moral justifi cation will turn in one 's con ception of morality. In turn, one 's con ception of morality may turn on one 's theory of the se manti cs of moral dis course. The i mportan ce of the distin ction bet ween the episte mi c and the pra cti cal perspe ctives can be questioned. Moreover, the distin ction is so meti mes blurred in people's writings, and it shoul d also be confessed that it is not entirely sharp itsel f. So me theories see m to be bor derline cases . Still, the episte mi c approa ch was co mmon to the meta-ethi cal theories that do minated philosophi cal dis cussion until the 1960s. In­ tu itionis m an d natu ralis m held what e motivis m and pres criptivis m denied -that an episte mi c justifi cation of moral belief cou ld be a chieved . Ho wever, e motivis m and pres criptivis m, and co mmentators on those theories, too k episte mi c justifi cation as the ideal, as if failure to a chieve it entailed a failure to provide any justifi cation that would be of interest to the moral s kepti c. G . E . Moore hel d a vie w usually called "intuitionis m" (Moore, 1903). In general ter ms, he hel d that true moral assertions pi ck out, or i mply the existen ce of, moral fa cts. In Moore's case, these woul d be fa cts about what things are good, fa cts that are sui generis, an d fa cts that can be kno wn to obt ain so meho w indepen dently of s cientifi c or observational methods. In Moore 's view, propositions about what things are intrinsi cally

6

Introduction

good arc synthetic, but, if true, they are know able a priori . Now so­ c alled n atur alists differ with intuitionism on both the ontologic al and epistemolo gic al st atus of mor als. They agree th at mor al assertions e xpress propositions or are c ap able of bein g true or false, and they agree th at there are mor al facts th at determine the truth or falsity of mor al assertions . But t hey hold th at these facts are not sui generis, or speci al ; r ather, they are comple xes of so-c alled n atur al facts, common or g arden-v ariety facts of the sort investig ated in the sciences, prob ably the soci al sciences . How are we to know which facts? How are we to know which mor al st atements are true or false? One str ategy here, the "reductive str ategy," is to hold th at an an alysis of the me aning of our mor al assertions will show them to h ave a "fi xed descriptive me anin g" speci fi able in nonmor al l an gu age and thereby en able a "reduction" of mor al facts to n atur al facts. A kind of nonsem antic n atur alism is another possibility. For inst ance, in his p aper for this volume, Nichol as Stur geon su ggests th at m ateri alists need not suppose th at mor al facts are identi fi able in a nonmor al ter­ minolo gy. If m ateri alism is correct, then any mor al facts are "physic al," but we should not e xpect to con firm this by me ans of sem antic an alysis of mor al assertions or by "reductive de finitions" of mor al terms . Our mor al beliefs constitute our theory about the mor al truth, and this theory is to be ev alu ated by applying st and ards th at do not diffe r in principle from the st and ards we use to ev alu ate other theories about the world. Cle arly there are signi fic ant differences of det ail . For the moment, the point is th at the v arious versions of intuitionism and n atur alism fit the epistemic mold and adopt a "co gnitivist" st ance re garding ethics: Mor al knowled ge is possible. The m ain critics of co gnitivism h ave held th at mor al "convictions" are not belie fs properly so-c alled, and th at mor al " assertions" are non­ proposition al e xpressions th at are not c ap able of truth or falsity . The sem antic issue is joined here. Wh at is the lo gic al form or st atus of mor al " assertions?" Are they proposition al? Stevenson and H are agreed b asic ally th at our mor al convictions h ave an action -guidin g or pr actic al role th at no mere belief could h ave, and th at there fore a mor al conviction is not (simply) a proposition al attitude, is not (merely) an accept ance of a proposition. Th at is, accepting a mor al jud gment is not acceptin g a proposition . But then a mor al jud gment is not a proposition, and mor al sentences do not e xpress such. Emotivists like Stevenson and prescrip­ tivists like H are presented di fferent acc ount s of the sem antics of mor al sentences, and this is not the pl ace to enter into det ail. Nevertheless, emotivism and prescriptivism do not bre ak aw ay from an epistemic conception of wh at justi fic ation of a mor al conviction would consist in if it were possible . H are 's account of mor al re asonin g is modeled on proposition al logic: We deduce specific mor al jud gments from more

David Copp

7

general moral principles co mbined with factual state ments. But be yond this, justi fication of a moral principle, to the extent that this is possible, could consist at best onl y in "a co mplete specification of the way of l ife of which it is [or would be ] a part." A decision to accept or reject a principle would be as justi fied as it could be provided onl y that it was made in light of such a speci fication. A person just has "to make up his o wn mind." 4 The debate bet ween cognitivists and noncognitivists has been thor­ oughl y discussed in the literature, and it is not addressed in detail in these essa ys . But a nu mber of essa ys address episte mic and se mantic issues, mainl y in the fir st section of the book. In , addition, the difference bet ween the episte mic and the practical conceptions of moral justi fication i s illustrate d: Sturgeon 's essa y exe mplifies the for mer, and David Gauthier 's exe mpli fies the latter. But be fore we look at the contributions to this volu me, we need to notice that there is at least one i mportant and traditional meta-ethical proble m that has not yet been mentioned, but it is one that is co m mon to both conceptions of moral justification. Consider the question, " Wh y be moral?" This question can arise no matter which conception one has of moral theory, episte mic or practical. All that is required is space in one's theor y sufficient to allo w an individual to quer y the rationalit y of his being moral ; to quer y the rationality of his co m mitting hi mself to, or of his con for ming al ways to, a moralit y justi fied in light of one 's theor y. Of course, to get the question o ff the ground, we must have a prior account of what it is to "be moral." This account might si mpl y refer to so-called conventional moralit y. But it might also refer to a theor y of moral truth or justi fication, and this could be of either t ype. It will be helpful to consider an exa mple of each t ype. Platonic theor y manifests the episte mic vie w, for it can be understood to justif y moralit y pri maril y on the basis of a moral ontology and episte molog y: Plato postulates the existence of the " for m" The Good, or, as we might say, he postulates the propert y of goodness, and he supposes that its nature full y deter mines what is good and bad, right and wrong. In addition, he proposes a theor y of intellectual apprehension to explain ho w we might co me to kno w the for m. The existence of The Good, and of our abilit y to apprehend it, would explain ho w justi fied moral bel iefs are possible . Ho wever, there is roo m in the theor y for one to ask wh y one should pa y attention to the for m. In The Republic, Glaucon challenges Socrates to sho w that it is better to be just than to be unjust ( 358a-36 0e). The challenge is fro m the standpoint of a happiness -maxi mizing conception of rational choice, according to which the best or most rational course for a person is al wa ys to choose what will most contribute to his happiness . The challenge is not the sel f-

8

Introduction

answe ring que ry : Is it not morally bette r to be mo rally not-good than to be mo rally good? It is the rathe r mo re inte resting que ry: Given a c hoice between being a mo rall y good pe rson and being a pe rson who remains open to acting immo rall y when so acting would se rve one's inte rests, mig ht one not maximize one's expected happiness b y choosing t he latte r and so be enti rel y rational to do so? Glau con 's c hallenge can also a rise within a theo ry that manifests a p ractical conception of mo ral justification. We have al read y sketched the t heo ry of Thomas Hobbes : Ou r inte rest in sel f-p rese rvation is invoked to explain why we would be rational to favo r the existence of mo rality as a social institution in societ y. Howeve r, it is not obvious that ce rtain individuals mig ht not find it in thei r inte rest to c heat on mo rality. Hence, in Leviathan, Hobbes conside rs the position of the " fool," who says in his hea rt t hat he will violate the obligations of justice wheneve r doing so would be in his own sel f-inte rest ( 165 1, 1958, pp. 120- 122). On the face of it, a violation of mo rality could easil y be in one's sel f-inte rest, especially if the re we re a low risk that it would be detected and punished. A violation could appa rently be rational even in t he face of a significant risk of punis hment if the p romised benefit to oneself we re sufficiently g reat . Yet Hobbes a rgues t hat ou r unce rtaint y about the risk of detection, and , mo re gene rally, about the magnitude and p robability of p romised benefits and costs of given acts of violation, gua rantees that violations of duty a re not rational. The a rgument is discussed in detail in G rego ry Kavka 's cont ribution to this volume. The point he re is just that Glaucon's c hallenge is relevant to t heo ries of bot h the epistemic and the p ractical va riety. W hat, then , is the relationship between the issue raised b y Glaucon and t he p roblems t hat a t heo ry of mo ral justification is meant to solve? Ku rt Baie r seeks to fo rge a connection. He a rgues that immo rality is not necessa rily irrational) fo r the minimal standa rds of rationally ac­ ceptable behavio r do not rule it out. Howeve r, immo ralit y does nec­ essa rily fall sho rt of perfect rationalit y. It is cont ra ry to reason, fo r reasonability is a highe r standa rd of compliance with reason than me re rationality. If Baie r is co r rect, then an adequate meta -ethical theo ry must link t he concepts of justification and reason in such a wa y that acting immo rall y comes out as cont ra ry to p ractical reason. The re must always be a reason to avoid doing whateve r mo rality p rohibits. A t heo ry of justi fication must explain t he natu re of t his reason in a way that yields a non -question -begging response to Glaucon. Mo reove r, if Baie r is co rrect, t hen it would seem t hat t he p ractical view of mo ral justification has t he advantage . But Baie r's position is cont rove rsial, and Kai Nielsen's essay a rgues against it .

David Copp

9

So far, this i ntroductio n has att empt ed to lay out a few co mpass li nes to h elp navigatio n through th e t erritory o f meta- ethics. At th e c ent er is th e issu e wh eth er morality, or wh eth er some morality, has a basis that �oul d p er mit its ratio nal justificatio n. How ev er, th er e ar e diff er ent approach es to this issu e. A su bsidiary issu e, t hen, is th e choic e a mo ng co nc eptio ns o f moral justificatio n, a major choic e bei ng betw een th e epist emic a nd practical co nc eptio ns. This is not a fr ee choic e. It will be co ntroll ed by a th eory o f th e natur e o f morality. Mor eov er, o n th e day o f r ecko ni ng, o ne's meta- ethical th eory must a nsw er to a chall eng e o f ov erall plausi bility. It must be co mpati bl e with acc ept ed sociological a nd psychological t heory a nd with plausi bl e philosophical th eori es i n r elat ed fi elds, esp ecially with a plausi bl e epist emology · or th eory o f practical r easo n, as th e cas e may be. It must be co mpati bl e with a plausi bl e accou nt o f th e practical or actio n-guidi ng natur e o f moral co nvictio n, with a plausi bl e accou nt o f moral r easo ni ng, a nd with a vi ew o f th e s ema ntics o f moral discours e that is co mpati bl e with g eneral s ema ntic th eory. Fi nally, o ne's th eory must r espo nd to Glauco n's chall eng e. So a noth er su bsidiary issu e is wh eth er practical ratio nality is o f such a natur e that th er e is always a r easo n to avoid i mmorality. Must a p erso n who is p erfectly ratio nal accord to morality a privil eg ed positio n i n practical d eli beratio n? That is, must h e do so with r esp ect to a morality justifi ed by th e lights o f o ne's th eory? I f not, wh er e is th e sti ng for moral sk epticis m? 5 Th e issu es su mmariz ed i n th e last paragraph ar e thos e that ar e addr ess ed i n this book . Th e book do es not pr es ent a favor ed th eory. Rath er, th e aut hors discuss diff er ent ar eas o f th e meta- ethical t erritory, ent eri ng fro m diff er ent dir ectio ns but meeti ng each oth er a nd d ebati ng th e merits o f th eir r esp ectiv e positio ns. Th e work pu blish ed h er e is larg ely new, although o f cours e it grows out o f oth er work by th e authors a nd by oth ers i n th e fi eld. I f th er e is o ne sort o f approach that is emphasiz ed, it is a n approach that, i n Gil bert Har ma n's words, s eeks to explai n "th e place o f valu e a nd o bligatio n i n th e world o f facts as r ev eal ed by sci enc e," or "th e plac e o f valu es i n th e natural world." Har ma n t er ms this approach " naturalis m." C ertai n Ka ntia n possi biliti es a nd i ntuitio nistic possi biliti es ar e not ca nvass ed her e. Th e i ntuitio nistic strat egi es would i nvolv e i ntroduci ng typ es o f entiti es or facts i nto o ne's metaphysics, or pri ncipl es i nto o ne's epist emology, that ar e not co mpati bl e, or that ar e beli ev ed not to be co mpati bl e, with a naturalist vi ew; a nd th e Ka ntia n strat egi es woul d co nsist i n i ntroduci ng postulat es i nto o ne's accou nt o f practical r easo n that agai n ar e not, or ar e beli ev ed not to be, co mpati bl e with a naturalist vi ew. This is not to say, how ev er, that it would be a si mpl e matt er to stat e exactly what th e naturalist vi ew a mou nts to, nor is it mea nt to

Introduction

rule out cer tain novel stra tegie s in me ta-e thic s tha t migh t be rendered c ompa ti bl e wi th the na turali st view in the l ong run. I t need s to be b orne in mind tha t the que sti on of wha t is c ompa tible wi th science and wi th scien tific me th od can be a ma t ter for re search. A na turali st a ssu me s, for in stance , tha t science i s unified , tha t bi ol ogy, phy sic s, and p sych ol ogy are mu tually c ompa tible and in terrela ted in the order of explana ti on. H owe ,·er, thi s a ssump ti on ha s n ot been vindica ted in every rich de tail. Similarl y, a na turali st in e thic s migh t pr oceed on the a ssu mp ti on tha t e thic s will be vindica ted in a scienti fic w orld view wi th ou t having de vel oped a de tailed scheme . The b ook i s di vided int o tw o par ts. Par t I, " Tru th and The ory Acceptance in Ethic s," addre sse s the require men ts or c on strain ts tha t migh t be imp osed on a the ory of m oral ju sti fica ti on, u sually bu t not e xcl usi vely fr om the epi stemic viewp oint . The se range fr om the na turali st c on straint already menti oned -tha t an adequa te the ory mu st sh ow the place of value in the scien ti fic w orld view - to a differen t kind of "na turali st" c on straint : tha t an argu ment fr om the na ture of mankind to a the ory of m orali ty mu st be in formed by ev olu ti onary bi ol ogy. Par t I I, " M orali ty and Prac tical Rea son," addre sse s the rela ti on ship be tween e thic s and the th eory of prac tical rea son . The e ssay s include an a t te mp t to gr ound ju stice in the the ory of social ch oice, argumen ts tha t illu stra te the prac tical c oncep ti on of m oral ju sti fica ti on, a t te mp ts to deal wi th Glauc on' s challenge, and a ske tch of an appr oach to the the ory of prac tical rea son fr om the standpoin t of soci obi ol ogy .

PART I: TRUTH AND THEORY ACCEPTANCE IN ETHICS Gilber t Harman' s e ssay se ts a maj or topic for the b ook , " I s there a single true m orali ty? '' Of c our se, in acc ord wi th an ecumenical a t ti tude toward the c hoice be tween epi stemic and prac tical c oncepti on s of the ba sic i ssue , we migh t think of the topic a s whe ther there i s a sin gle true or " maximally ju sti fied " m orali ty. Bu t the deba te in any ca se c oncern s the tru th of '�ab solu ti sm," which Harman de fine s a s the view tha t there i s a m oral la w tha t e very one ha s su fficien t rea son to foll ow and tha t there are thing s tha t e very one ha s a su fficien t rea son to h ope for or wi sh for." Rela ti vi sm" i s simply de fined a s the denial of thi s the si s. Notice tha t Harman ha s b uil t i nto hi s de fini ti on of ab soluti sm the n oti on tha t rea son s w ould be "intcrna r' to a ju sti fied or true m orali ty in tha t there w ould alway s and nece ssarily be a sufficient rea son to av oid imm orali ty. In thi s way, suppor t for a bsol u ti sm would be supp or t for the ra ti onali ty of being m or al .

David Copp

11

Harman argues that naturalism naturally leads to relativism and that absolutism will be found most plausible by those who reject naturalism in favor o f "autonomous ethics." Autonomous ethics is said to differ from naturalism in that it is not dominated by a concern to find the place o f ethics in the natural world. Not being dominated by this concern, autonomous ethics can pursue ethics "internally," beginning with our moral beliefs and searching for principles. Nicholas Sturgeon suggests, however, that naturalism is not incompatible with the autonomous methodology. It may be that we need to pursue ethics internally in order to achieve a satis factory naturalist theory. For it may be that an explanation o f the place of ethics in the natural world would r �quire the con frontation o f our best moral theory with our best theory of the (rest of the ) natural world . And it may well be that, in order to arrive at a satis factory ethical theory, we would have to engage in investigating and systematizing our moral beliefs. If this is right, then the so-called internal approach is compatible with naturalism. The fundamental issue, then, would be whether to accept the naturalist constraint -that an adequate moral theory must be compatible with a scientific world view. Naturalism leads away from absolutism, Harman argues, given what he thinks is the philosophically favored theory of practical reason from the naturalist point o f view. This theory, a " Humean" theory (attributed to H ume, 1 739 ), links having a sufficient reason with motivation: There is warranted reasoning that would lead one to decide to do whatever one has a sufficient reason to do. But for any given thing o f proposed value, or any given propose d moral principle, there are people who are not motivated appropriately, and who would fail to decide appropriately even i f they were exposed to what seem to be all relevant pieces o f warranted reasoning. And so there seems to be no absolute value or principle, no principle or value that could be a source o f sufficient reasons for everyone. The key issues, then, assuming that we accept the naturalist constraint, are the theory of practical reason and the sufficient-reason requirement. Perhaps some other theories o f reason are compatible with naturalism, and perhaps a morality could be justified even if it were not a source o f sufficient reasons for everyone. These issues are taken up in many of the essays in the volume. We have already mentioned some o f the issues discussed by N icholas Sturgeon. In fact, Sturgeon intends his essay as a criticism of a skeptical argument presented by Harman in The Nature of Morality ( 19 77), an argument based on the view that we are justified in postulating the existence of entities or facts o f a given kind only if the postulation o f such entities or facts is an essential part of our best explanation of some o f our belie fs or observations. Harman contends that we do not need to assume that there are any moral facts in order to explain observations

12

Introduction

o r belie fs o f any so rt, and he thinks that this tends to suppo rt mo ral skepticism. Stu rgeon thinks on the cont ra ry that mo ral facts are co m­ ponents o f "ou r best ove rall explanato ry pictu re o f the wo rld." But he limits himsel f to a rguin g that we do in fact accept mo ral explanations o f events, that at least some of these a re good explanations, and that they a re "not obviously unde rmined by anythin g else that we know." The bu rden o f Stu rgeon's essay is the explanation, elabo ration, and de fense of this view, to gethe r with a discussion o f the relevance to it of views about the " reduction" of mo ral facts to natu ral facts. The a rgument also is o f relevance to the position taken by Ha rman in his essay fo r this volume. Ha rman claims that natu ralism tends to suppo rt relativism, but Stu rgeon a rgues that a scienti fic wo rld view is compatible with allowin g the existence of mo ral facts that delimit a sin gle t rue mo rality. Still, nothin g said by Stu rgeon shows that such a mo rality would be the mo rality of absolutism, a sou rce of sufficient reasons fo r eve ryone. My co-edito r, David Zi mme rman, disputes Stu rgeon 's pos 1t 1on, as well as ce rtain a rguments by Ma rk Platts and David Wi ggins ; he re ga rds each of these autho rs as a rguin g in favo r of "mo ral realis m." Diffe rent autho rs may use the te rm " realism" in diffe rent ways, but Zimme rman means by it the view that "the natu re of the good and the ri ght can be explicated without essential refe rence" to what any mo ral a gents "app rove o f o r a re able to commit themselves to o r desi re." 6 I will not attempt to su mma rize the debate with Stu rgeon. Suffice it to say that Zimme rman assumes that Stu rgeon 's essay is meant as an a rgument fo r realism, and he claims that the a rgument fails . Platts and Wi ggins seem to be a rguing that "subjectivism" about values would cause the disto rtion o r disappea rance o f featu res of ou r psycholo gy that we would not want to lose. In Platts 's case, the a rgu ment is that subjectivism would lead to the e rosion of ou r "nonappetitive desi res," fo r motivation by a nonappetitive desi re depends on the belief that the object of desi re is desi rable independently o f the ba re fact that it is desi red ( 198 1). David Wi ggins a rgues that subjectivism fo rces on the theo rist a distinction between an inne r pe rspective on values, a pe rspective o f commitment, and an oute r pe rspective, from which the inne r pe rspective is held to be illusory. Subjectivism would lead to the disto rtion of the sense o f life's meaning that we find in the inne r pe rspective on value ( 1976). Zimme rman 's response is to su ggest that the ' tsubjective basis fo r values" consists o f one 's most deeply felt desi res, desi res that a re rooted in one 's psycholo gy, desi res that a re rooted in "the co gnitive, constitutional and histo rical density of the pe rson." Desi res o fthis so rt can su rvive and continue to be a basis fo r commitment ' even given the belief that subjectivism is co r rect. The inne r pe rspective,

David Copp

13

so to speak, is one from which we are motivated by our nonappetitive desire of what is valuable, but we can take up this perspective without having a belief in realism. . Richard Brandt's essay pursues a version of naturalism also argued for in his recent book, A Theory of the Good and the Right ( 1979). Brandt takes up what we have called the practical conception of moral justification, developing an account of rational desire and action, and then arguing that a moral code is justified for a person if and only if the person would support it or choose it for currency in his society, were he to be " fully rational." The notion of "full rationality " is a technical one, and it is explained in detail both in his book and in the essay found here. It is criticized by Kurt Baier i n the second section of this book and defended by R ichmond Campbell. Briefly, the notion is that the full rationality of a person's choice depends on its having two characteristics ( 1979, pp. 1 1- 1 3). First, the choice must be a choice of exactly what the person would have chosen if "every item of relevant available in formation" had been "present to awareness, vividly, at the focus of attention, or with an equal share of attention." Here, in formation counts as "relevant" if, roughly, it would have made a difference and as "available" if it is justified either by the science of the day, by publicly accessible evidence, or by the principles of logic. Second, the desires or aversions that brought about the person's choice must be as they would h ave been had the person undergone "cognitive psychotherapy," had the person's "motivational machinery [been ] fully su ffused by [relevant ] available in formation" (pp. 1 1 1- 1 12). So we could say that one 's choice is " fully rational" just in case it is what it would have been had one's beliefs, desires, and aversions been corrected by one's bringing vividly to mind at the focus of attention, and perhaps repeatedly, every item of relevant available in formation. Brandt's proposal, then, is that justi fying a moral code is a matter of justifying one's choice of it to serve as the pu blic moral code in one's society. Of course, it is not necessarily the case-in fact it is unlikely-that the code it would be "rational" for one person to support is the same as the code it would be "rational" for another to support. So Brandt's naturalist program does not yield an absolutist theory. Brandt un folds this position in his essay here. He begins with the observation that there is a commonsense morality in most, if not all, societies. He holds that such a morality consists in a complex of moral attitudes and concepts, together with the terminology to express those attitudes and the thought that they are justified. Now, as we saw, the epistemic conception of moral justification gives rise to a concern with the semantics of moral language. The notion is that we can determine whether our moral judgments are epistemically justified only if we can

14

Introduction

first determine whether they are propositional, or capable o f truth or falsity, and, second, only i f we can determine their propositional content. It may well be that the pra cti cal conception also requires a con cern with such semantic issues. At any rate, Brandt suggests that there are two traditional approaches to the issues o f clari fying moral language. One aims at describing the ordinary meanings o f moral terms or at least the meanings of moral terms as used by the theorist and the intended audien ce. A second aims at providing recommendations for the construal o f moral language. This is the so-called method o f re forming definitions ( 1979, p. 3). Brandt thinks that this second approa ch is the more promising o f the two, and his essay attempts to work out some o f the constraints on su ch an e xer cise. It is obvious that not just any re formation will do; we are a fter a theory of morality. So Brandt proposes a set o f criteria for "an optimal conceptual framework for moral dis course. " The key term, Brandt says, is "justified, " as it occurs in the phrase "justified moral code," and Brandt sets out seven conditions for a satis factory "e xpli cation " o f this term . He then argues that his own a ccount o f moral justification is the optimal one. He goes on to propose that what we should refer to as the facts o f morality are just the fa cts that constitute the truth conditions of moral assertions given the optimal conceptual framework. Here, then, we see the un folding o f a cognitivist meta-ethical theory via a set o f constraints on meta-ethical theories and an argument for a practical conception o f the justification o f morality. Norman Daniels dis cusses two o f the conditions proposed by Brandt, the "empiricist constraint " and the "disalienation constraint ." The former arises from the requirement that the justification o f a moral code, i f it is to answer practical doubts that people have about morality, must itsel f be value-neutral. Justification must rest on fa cts and logic; not these alone, as Daniels suggests, but these together with our nonmoral pre f­ erences as corrected by facts and logi c through cognitive psychotherapy. The empiricist constraint is that the desires and belie fs constitutive o f our present moral code should not be taken into account in evaluating the justifiability o f candidate moral codes. The second constraint, the disalienation constraint, is that a successful a ccount o f justifi cation for moral codes must close the gap between justifi cation and motivation; it must be su ch that one who believed a code to be justified, in the sense specified by the theory, would find con formity to the code thereby to be recommended. This second constraint is obviously designed to eliminate or redu ce the force of Glaucon 's challenge. It says in effe ct that a theory o f justifi cation for moralities is unsatis fa ctory i f it leaves (too much) conceptual spa ce for the challenge, " Yes, but why should I con form? " Daniels argues that there is a deep tension between these tw o requirements. Our desires are shaped by our culture, in cluding the

David Copp

15

morality that is part of our culture, and so we cannot "step outside our tradition," as required by the empiricist constraint, except at the cost of alienating ourselves from the result. That is, to the extent that a theory of moral justification decreases the effect of our prior moral views on the content of a morality it deems justi fied, it increases our alienation from that morality by increasing the distance between what it requires and what we desire, or what be believe ought to be the case. If Daniels is right, then no theory can meet both requirements, and Daniels argues that in fact Brandt 's own theory does not strictly speaking meet either . Moreover, he argues that "intuitionism," not in G. E. Moore 's sense, but in that de fined by John Ra:wls 's theory of "wide re flective equilibrium," is as successful in meeting the constraints as is Brandt 's theory. The wide reflective equilibrium ( W R E) theory suggests that a moral code or moral principle should be dee med justi fied (for a person) only if it would be accepted (by the person) in a wide reflective equilibrium : a settled and coherent state of belief, where the code or principle in question is coherent with (the person 's) relevant considered moral judgments, and where relevant moral or nonmoral background theories provide support for the code or principle, support that is to some degree independent of the support provided by the match with considered mor al ju dgments. One 's initial moral beliefs are "filtered" to elimin ate obvious sources of error, and they may change quite radically be fore wide reflective equilibrium is reached. Nevertheless, it seems clear that the empiricist constraint is not met, for the content of one's equilibrium point will be affected by the content of one's moral starting point ; and the disalienation constr aint is not met, for one may apparently settle on a particular moral code as most coherent with the body of one 's beliefs without being motivated to con form. The constraints are not met, but the thrust of Daniels 's arguments is that this is no objection to W RE t heory. It is at this point that the present author enters the debate with his essay "Considered Judgments and Moral Justi fication ." W R E theory imposes a constraint of its own on what can be deemed a justi fied moral code or principle, viz., that to be justi fied, a principle must cohere with our considered moral judgments. To be sure, it is not re quired that there be coherence with the considered judg ments with which we begin reflection; it is only required that there be coherence in the equilibrium that may ultimately be reached between a principle and a set of considered judgments. But this is enough, I argue, to call W R E theory into question. The W R E constraint is a cousin of a kind of widely accepted meta­ ethical constraint, a "conservative" constraint to the effect that an ade quate moral theory must conserve our considered moral judgments, those we would have, let us say, were our moral convictions to have been adequately

16

Introduction

con fronte d wi th relevan t non moral considera tions: A theory is to be rejec ted if i t regards as jus tified some thing tha t is and re mains "coun ­ terintui tive." This vie w is a basis on which many philosophers would be incli ne d to evalua te me ta-e thical theories such as Brand t's. For, as is argued by Nicholas S turgeon, Brand t's theory would per mi t qui te un­ in tui ti ve moral principles to be jus tified by i ts ligh ts ( 1982). Ho wever, W R E theory appears to be i mmune to this line o f a ttack. The me ta­ e thical cons train t is buil t into i t, in tha t i t is logically i mpossible for a moral code or principle to be dee med jus tified by W R E theory unless i t coheres wi th our considered judg men ts in W R E. Some argu men t is required to jus tify the privileged s ta tus given in W R E theor y, and in the com mon conserva tive cons train t, to our con­ sidered judgmen ts. Daniels holds tha t no epis te mological priority is given to considered judg men ts in W R E theory ; there is no de ter mina te and fixed se t of judg me nts identified in advance and trea ted as jus tified ab initio. And I mus t agree wi th this. Ho wever, the judg men ts tha t would be our considered judge ments in W R E are given a special s ta tus as s tandards of jus tifica tion. They are able to defea t a moral code, or in the case of the me ta-e thical co ns train t, they are able to defea t an accoun t of moral jus tifica tion. So me argu men t is needed for according the m this s ta tus as defea ters. The final essay in Par t I is by Ronald de Sousa. I t considers the viabili ty of argu men ts fro m our na ture as humans to theories of value. Such arguments would cross the " fac t-value gap," so-called, bu t there is an ancient tradi tion holding tha t so me argu ments of this sor t are valid. For ins tance, Aris to tle argued in the Nicomachean Ethics tha t if there is a function of hu mans, then the hu man good consis ts in fulfilling this func tion in accord wi th the appropria te excellences ( 1097 6 2 31098a 19). We need no t pause to consider Aris to tle's concep tion of our fu nc tion ; de Sousa argues tha t the progra m sugges ted by the Aris to telian func tion argument has been under mined by modern evolu tionary biology. Never theless, biology s till does use the concep t of a func tion, and so i t seems wor th inves tiga ting whe ther a concep tion of our nature can be cons truc ted wi thin the boundaries of accep ted biological theory as a basis for an argument to a theory of value. This is the s tudy carried ou t in de tail in de Sousa 's essay.

PART II: MORALITY AND PRACTICAL REASON The essa ys in Par t II emphasize one of the the mes already introduced : the rela tionship be tween morali ty an d prac tical reason . One aspec t of this is the concep tion tha t a morali ty is to be jus tified , if a t all, b v sho wing the ra tionali ty of choos ing i t for some specified purpose i ;1

David Copp

17

some defined context . We have seen Brandt 's arguments in this direction. A second aspect arises out o f Glaucon 's challenge- "What reason do I have to be moral?"-and out o f Brandt 's notion that an adequate meta ­ ethical theory must close the gap between justification and motivation, disalienating people fro m moralit ies seen to be just ified by its light. Recall that Harm an holds also that i f there is one true morality, in any sense worth discussing, then everyone has sufficient reason to con for m to that morality. He defines "absolutism" as the thesis that there is such a morality. A third issue implicit in all o f this is the concept o f a practical reason, and its relationsh ip to the concepts o f mot ivation and obligation. A " Humean" view makes the relationship between reason and mot ivat ion " internal": It is logically necessa ry, i f one has a pract ical reason, that one have an appropriate motivation, or that one would, in specified circumstances. Such a view is held, for instance, by both Brandt and Harman. Is the relationship between obligation and mot ivat ion also internal? Harman holds that it is ( 1975; see Copp, 1982). These are among th e issues exam ined in the essays in Part I I. A major problem addressed in these essays is the conflict that arises in so -called prisoner's-d ilemma situations, where rational sel f-interested behavior on the part o f everyone leads paradoxically to a result that is worse from everyone's sel f-interested po int o f view than some other available res ult. Many situat ions can be o f this general sort, but the example that is typ ically g iven concerns two prisoners. Suppose that two felons are arrested and held on a charge o f ar med robbery. They are in separate cells and cannot co mmunicate with each other. Each is told, and told that t he other is told, that their s ituation is as follows: the prosecution cannot secure a conv iction on the charge o f armed robbery unless one o f them confesses. I f they both refuse to con fess, both will be convicted, but only o f a lesser offense. Each will spend three years in prison. I f one o f them con fesses and the other does not, then the one who cooperates and hel ps to secure the conviction o f the other will be given a very light sentence o f only one year. The other will suffer ten years in prison. I f both confess, both will be convicted o f the serious charge, but they will be g iven some credit for their cooperat ion . They will eac h spend eight years in prison. They must independently dec ide whether or not to confess. What ever one chooses will affect the wel fare o f the other, but ne ither o f them can affect the other 's choice. They care little for each other, and, let us say, there will be no opportunity in the future for retaliation. Each o f them knows that con fess ion assures a shorter sentence for t he person who confesses, regardless o f what the other does, but they also know that i f they both con fess each will rece ive a longer sentence than if both do not confess. In this case, i f each o f them is rat ional to aim for a m ini mu m sentence, then each is rational

18

Introduction

to confess, regard less of what the other docs, for each of them saves two years by confessing. Howe\·er, if they both confess, they both s u ffer for eight years, \vhilc i f neither had confessed, they wou l d only have spent three years in prison, so the optimal result fo r them both considered as a group, the resu l t that each of them would rat ionally prefer to the outcome where both confess, is an outcome that they w i l l not achieve if each of them aims to m i n i m ize h is or her own prison sentence . Situations with this stru ctu re arc not u ncommon . For example, i t may b e that the nuclear confront at ion between the s u perpowers is a prisoner's-d ilemma situat ion . 7 For our p urposes, the prisoner's-d ilemma suggests that there may be a formal problem for the project of j us t i fying morality on the basis of practical reason, ass u m i ng, that is, that rat ionality consists in seeking to maximize one 's self-i nterest . It calls i nto question this maxim izing conception of reason, and s uggests there may be a logical bar rier to showi ng that the sort of cooperat ive beha\·ior d is t i nc t i \·e of moral ity is rat ional from each o f our point s of v iew. Ku rt Baier argues that the conception of pract ical reason deve loped by Richard Brandt is strongly cou nt e rintuit i ve . To be s ur e , B randt regard s his concept ion as a reformi ng sugges t ion fo r a use of ''rat ional," not as a desc r i pt ion of our ord inary use. Neverthe less, B rand t ad m its that a sat isfactory account must preserve the " recommendatory force,, of remarks of the form, "The rational thi ng to do wou l d be . . . '' It must "d isalien.uc , , us from what i t wou ld be rat ional for u s to do ( 1 9 79, pp. 1 60- 1 6 2 ) . Baier argues that B rand t's theory does not solve the problem of a l ienat ion. I n fact, "the closer we come to solving [this] problem, the closer we arc to depriving reason of its gu id i ng funct ion .'' There is a tension between the gu id ing fu nct ion of reason and our d isalienation from reason, a tension of the same nature as the ten s ion noted by Danie l s between the abi l i ty of a meta-eth ical theory to take u s out of our t rad i t ion and gu ide us to a new moral perspect ive and its abil ity to d isalienatc us from the moral ideal . Neverthe less, Baier does not suppose that a theory must i nevitably break under this tension : He argues that "'the concept of rat ionality embedded i n our c u l tur e and o ur language,, .l \'oid s the alienation of reason, while at the same t i me it preserves its gu id ing fu nction . The theory sketched by Baier is ric h , complex, and h igh ly s uggest i \'C of new appro ac hes to deal ing with the theoretical problems we h a\'C bee n d iscuss i ng. Reason is said to be a capac ity to solve problems with the aid of gu ide l i nes pro\·ided by our cul ture. These guideli nes arc not static . The i r purpose is to enable us to ac hieve o ur objectives, both theoret ic1l and pract ica l , at a " " h igher aspi rat ion lc\'C l " t ha n would othe rw ise be possible . The gu id e l i nes can be rev ised and i mprm-ed m·cr t i me i n l ight of the res u l ts of their use. One must su ppose that the

David Copp

19

princi pl es of morality could also b e charact eri zed in this way, according to Bai er, for h e argu es that, although immorality is not n ec essarily irrational, it do es n ec essarily fall short of perfect rationality, b eing contrary to .r eason . But th en th e princi pl es of morality must b e among th e guid elin es of r eason. Oth erwis e th er e would b e no logical guarant ee that actions a pprais ed as falling short of moral r equir em ents would also b e pro perly a ppra is ed as falling short of perfect rationality. It follows that th er e must b e mor e to practical r eason than r easons of s el f-int er est. Th e guid elines of a stabl e soci ety must som etim es r equir e action that runs contrary to s el f-int er est. This is Hobb es's insight ( 165 1), and various of our authors r ein forc e it by r eflecting on prison er:'s-dil emma situations. If w e w er e all "rational egoists," soci ety would b e unstabl e. H enc e, th e r equir em ents of a stabl e social ord er must b e r egard ed as r easons that ov errid e ind epend ent s el f-int er est ed r easons . How ev er, th e probl em of ali enation now r ea ppears. Practical r eason has be en d efin ed in t erms of o ptimal social guid elin es for th e maximal ful fillm ent of th e as pirations of each, com patibl e with a stabl e soci ety . How ev er, Baier says, to show that this is ind eed th e natur e of reason, th e "social r equir em ents must constitut e g enuin e r easons for ev eryon e" in that ev eryon e must hav e "ad equat e" ind epend ent s el f-int er est ed r easons "to want ther e to b e such social r equir em ents and to hav e th em r egard ed as ov erriding." To say this is to introduc e th e probl em of ali enation. Y et Bai er thinks h e can esca pe it: Su ppos e a social ord er that provid es ev eryon e with " equally good r eason" to r egard its r easons as ov err iding s el f-int er est, a social ord er in which no on e 's r easons can b e str ength en ed without som eon e's b eing w eak en ed. Th en " ev eryon e has r easons as good as ev eryon e can d emand," and so ev eryon e has ad equat e r easons. In such a soci ety th er e is r eason for each to con form to social guid elin es, and so th e guid elin es constitut e r equir em ents of r eason. Kai Ni els en argues that Bai er 's ent er pri se fails by its own standards. Cons id er a person with th e obj ectiv e of b eing "a prud ent f ree-rid er w ith stabl e and sound moral institutions in plac e." Bai er has not shown that this obj ectiv e is contrar y to r eason. A person with this obj ectiv e may succ ess fully pursu e his int er ests in full control of his passions , with an accurat e appraisal of th e facts , and making no errors of r eason ing. Such a person tak es into account th e g en eral cultural guid elin es, mani pulating th em to enabl e ma ximal ful fillm ent of his int er ests. H e do es perha ps hav e su ffici ent r eason to want g en eral moral guid elin es to b e entr ench ed in hi s soci ety and to b e socializ ed, but Bai er has not shown that our person has any r eason for wanting th e sociali zation to "stick" in his own cas e . Our person may b e quit e ca pabl e of b eing a fr ee-rid er who maximiz es his own utility. A k ey point rais ed by Ni els en is th e possibility that practical r eason is not uni fi ed in a singl e s et of guid el in es : Perha ps

20

Introduction

there just are both "agent-relative" sel f-interested reasons and " agent­ neutral" re asons such as mor ality entails, with no requirement th at agent­ neutr al reasons alw ays be given priority by a fully ration al person. Baier 's view is that if mor ality is a source of re asons, then everyone must h ave an adequ ate reason to reg ard the requirements of mor ality as overriding sel f-interest; that is, everyone must have an interest , as great as is comp atible with a like interest for all, in its being the c ase th at the requirements of morality are con formed to by everyone. This idea is taken up by Jan Narveson, who sees it as supporting a contractarian progr am in ethics. For if e ach of us is to h ave an interest in everyone 's con forming to morality, an interest equ al to the interest of anyone else, then it would seem that mor ality must consist of rules th at all of us could agree to in a fair b arg ain for mutu al adv ant age. The problem though, given a utility -m aximizing or sel f-interested conception of r ational choice, is th at our interest may lie in everyone else )s con forming to mor ality. If morality is viewed schem atic ally as a set of "rules constr aining individual behavior in the direction of optimality" in prisoner 's-dilemm a situations, as Narveson argues, it m ay be in one 's interest th at others con form, but not in one 's interest that one con form onesel f. This is the free-rider problem. David Gauthier has suggested th at the utility-max­ imizing conception of r ation ality should be repl aced by a conception of r ationality as "constr ained m aximiz ation" ( 1975 a). But Narveson proposes a different strategy for dealing with the problem. One 's ac­ ceptance of mor ality h as two aspects: a m atter of one 's beliefs and a m atter of one 's actions. The latter also h as two aspects: a m atter of con forming to morality and a matter of participating in activities aimed at securing general con formity. Such activities can alter the utilities of those who might otherwise find it in their interest to che at on mor ality. There fore, such acti vities can themselves foster rational con formity . Moreover, the rationality of everyone 's p articipating in activities of this sort is clear, even in light of a utility-maximizing conception of ration al choice, and this is all that is required for a ration al underpinning of morality. A morality is "a body of principles such th at any r ation al person would want to have them generally observed, and would accept th at he should participate, at least weakly, in a social system designed to bring about th at gener al compli ance." It is cle ar that Narveson 's strategy leaves morality open to Gl aucon 's ch allenge, and Gregory Kavk a's ess ay examines the merits of the ch allenge in abund ant det ail. However, be fore we t urn to K avka, we will ex amine t wo other essays : David Gauthier develops a co ntractarian theory of justice th at is an ex ample of the sort of work possible within the practical conception of justi fic ation, and Richmond Campbell examines the n a-

Da Pid Copp

21

t uralistic credentials of the theory of pra ctical reason that has been relied on to s uch a degree by o ur a uthors. David Ga uthier arg ues that the prin ciples of j ustice arc not only the sol ution to a problem of rational choice , as was arg ued by John Rawls in · A Theory of Justice, b ut they are also part of the theory of rational choice. That is , they are prin ciples for making rational choices , prin ciples for ma king rational choices by a socie(v. Now if there is a parallel between the theory of individ ual de cision and the theory of so cial choice , and if an individ ual a cts rationally by maximizing his expe cted utility , then we may s uppose that a so ciety a cts rationally by maximizing e xpe cted , "so cial wel fare _ , Howe ver, Ga uthier contends that so cial rationality does not parallel individ ual rationality in this way. A s ociety does not p urs ue , its "own ' ends , not even a system of ends somehow derived from the ends of individ uals; rather, it p urs ues the several indi\·id ual ends to ens u re mutual advantage. That is , this is what is p urs ued to the degree that so cial de cisions arc rational. B ut prin ciples that wo uld se cure mutual ad vantage arc prin ciples to which rational individ uals wo uld agree even if they were see king to ma ximize their own utility. And so the prin ciples of j ustice, as prin ciples of rational so cial choice , wo uld be the sol ution to a problem of rational agreeme nt for m ut ual ad vantage among sel f. interested individ uals. The Rawlsian theory is that prin ciples of j ustice arc, as it were , chosen behind a Peil of ignorance ( Rawls , 197 1) . B ut if this were right , there wo uld be no g uarantee that the prin ciples of j ustice wo uld be chosen by individ uals choosing rationally in light of what in fact is their expe cted utility. But then there wo uld be no g uara ntee that prin ciples of j ustice wo uld in fact se cure m u tu al advantage. Given the con cl usion that j ustice do cs or wo uld promote m ut ual advantage , the Rawlsian approa ch m ust be reje cted . The alternative that is de veloped by Ga uthier exploits the notion of rational bargaining among actual individ uals in light of their actual interests and capa cities. There is some idealization. The privileged bargain is what would res ult from rational bargaining among ideally competent , in formed , and rational proxies for ea ch of us . The rationales for this , and for certain other idealizations , are stated by Ga uthier. In addition , Ga uthier de velops a theory of rational bargaining at least to the e xtent that he is able to state a set of fo ur conditions on rational bargaining and to derive from them the res ult that rational bargainers will a ct on a prin ciple of minimax concession. The idea is that ea ch bargainer has as his base point the ma xim um that he co uld a chieve witho ut cooperation , and as his ma xim um claim the most he co uld achieve from a s cheme of cooperation , consistent with its remaining advantageo us for the others both to remain in the s cheme themselves and not to e xcl ude him . A rational agreement wo uld a fford ea ch individ ual no more than

22

Introduction

his claim an d no less th an his base point; it would see each individual ac cept some concession, as measu re d by the pro po rtion he gives up o f the maximum that he claime d ove r the basepoint. Now Gauthie r a rgues that rational ba rgaine rs would reach ag reement on a scheme only i f any alternati ve would requi re some in dividual to accept a g reate r concession than the la rgest requi re d un de r that sche me . Rational ba rgaine rs would act on a p rinc iple o f minimax concession . Social choice plays a role in "ensu ring p rotection" fo r individuals in society an d in "inc reasing p roduction." The princi ples o f justice a re the p rinci ples o f rational social choice in this do main, and so they a re princi ples that woul d result from applying the mini max concession rule. Outsi de o f this do main, rational pe rsons seek to maxi mize thei r ex pected utility . Within this domain they a re to follow the p rinciples o f rational social choice, even when doing so would be a depa rtu re from expected ­ utility maximization. He re again we find the proble m o f alienation . Pe rhaps the alienation proble m should be avoided by a mending o r rejecting the view that rationality fo r an individual consists in maxi mizing ex pecte d utility . Davi d Gauthie r has a rgued that this is so. At any rate, it seems that i f the no rms o f mo rality need to be reconciled with a scientific wo rld view, so do the no rms o f rational choice and rational belie f. Richmon d Campbell takes u p pa rt o f this task in his essay . He a rgues fo r the st riking thesis that "sociobiology can have a significant role to play " in suppo rting a conce ption o f rational choice within a scientific conception o f the wo rld. Sup pose we begin with the expected ­ utility -maximi zing mo del o f rational choice, o r with " rational egois m." Campbell 's i dea is that "empi rical studies can illu minate the normative cha racte r o f [this mo del o f ] rational motivation by ma k ing its att ractiveness as a no rm fo r behavio r mo re intelligible." The study o f prisone r's- dilemma situations shows that coo pe rative o r alt ruistic rest raint fo r mutual advantage is not always rational, given the ex pecte d-utility -maximizing model . Cam pbell suggests that a fo rmal simila rity exists between the p roblem posed for the theo ry o f rational egoism by the p risone r's- dilemma and the p roble m pose d fo r the theo ry o f evolution by alt ruistic behavio r. " To as k how recip rocal alt ruism is possible in natu re on the assum ption that natu re has evolved th rough natu ral selection is only su pe rficially diffe rent from asking how recip rocal alt ruis m is rationally possible on the assum ption that rationality is rational egois m ." Robe rt Axel rod an d William Hamilton have a rgued that a st rategy o f limite d coope ration, known as Tit for Tat, has a biological supe rio rity ove r othe r st rategies like pu re coo pe ration an d pu re noncoope ration: It has evolutiona ry stability, robustness, an d viability ( 19 8 1). They a rgue this on the basis o f a model o f the envi ron ment that is realisti c in that it assumes a series o f p risone r's- dile mma inte ractions

David Copp

23

inv olving individuals, where the pr obability of repeat interacti ons between the same individuals is fixed at a p oint between zer o and one. This m odel "diss olves the parad ox of bi ol ogical altruism in P D interacti ons ." It als o "dem onstrates h ow c ontinued mutual c ooperati on in such in­ t �racti ons is c onsistent with rati onal eg oism." Rati onal eg oists can select strategies, and the Tit for Tat strategy pr oduces m ore expected utility for an individual "than any of a wide variety of c ompeting strategies ." Hence, it is a rati onal strategy "given a series of P D's with a sufficient pr obability of the same individuals interacting again." Thus rati onal eg oism is "c onsistent with c ooperative restraint in the pursuit of individual ends -a restraint that is c omm only th ought t o be paradigmatic of m oral acti on." Campbell has argued, foll owing Axelr od and Hamilt on, that certain general formal features of the interactive envir onment force a rec oncil ­ iati on between reas on and altruism, and hence they eliminate the alienati on pr oblem in realistic settings . We have seen other strategies. Harman pr op osed that a m orality sh ould be regarded as true or justified only if it is a s ource of sufficient reas ons for c on formity. Baier argued that the norms of reas on, including th ose of m orality, are cultural guidelines with a given purp ose, and s o that there is a unity of practical reas on, with m oral standards being rati onal standards that override all others . These strategies, with the p ossible excepti on of Ca mpbell's, would serve t o answer Glauc on, for Glauc on demanded an intrinsic c onnecti on between m orality and one's own g ood, or one's own rati onal ch oices. Greg ory Kavka seeks t o expl ore in a systematic way the nature, sc ope, and depth of the extrinsic c onnecti ons between m orality and practical reas on. We assume s ome acc ount of m orality as a backgr ound, calling it "c omm onsense m orality," and we assume that rati onal c onduct is sel f­ interested c onduct or "rati onal prudence." Given this perspective, where we assume m orality is given as a set of justi fied requirements on acti ons, it may n ot see m a seri ous pr oblem that " We cannot expect t o c onvince a clever imm oralist that it pays every one t o act m orally on every specific occasi on in any s ort of s ociety." Kavka's strategy, then, is t o narr ow the pr oject of rec onciling reas on and m orality by identifying s ome of its dimensi ons and selecting for discussi on certain less ambiti ous versi ons of it. S o we argue that adopting the moral way of life is c onsistent with sel f-interest, n ot that all moral actions are; we argue that a kind of "c opper rule" m orality of recipr ocal altruism, as perhaps is modeled by Tit for Tat, is c onsistent with sel f-interest, n ot that the G olden Rule ideal m orality is c onsistent with sel f-interest; we argue that m ost pe ople in actual s ocial circu mstances find m orality t o be in their sel f-interest, n ot that every one always necessarily w ould; we rec ognize that rati onality

Introduction

24

embraces the pursuit of ends that are not strictly "sel f-intereste d." We aim to show that moral people have no reasons for regret, an d that parents have sufficient reasons to raise moral chil dren. An d Kavka concludes, " While it is normally prudent to be moral, it is sometimes rational to be moral even if it is not pru dent ."

CONCLUSION These essays carry forward the debate on meta -ethical issues, breaking new ground in some cases, cultivating and surveying older territory in others. I hope that the result will be to prompt philosophers to devote more of their attention to foundational questions in ethics, both as thinkers an d as teachers. The questions are ancient in their origin, they are at the center of philosophy, and they are of great cultural importance. They demand all of the sophistication, rigor, and sensitivity that phi­ losophers can bring to bear. 8

NOTES 1 . Foot ( 1 967), " I nt roduct ion," p. 8 . See also Foot ( 1978b ) . 2 . This i s the thesis o f G . J . Warnock ( 1 967) . For d iscussion, see Warnock ( 1967) and Brand t ( 1 9 59), chaps . 7-9. 3. Rawls ( 1980), Lecture 3, "Construction and Objectivity," p. 5 54 . 4 . Hare ( 195 2), p. 6 9 . In more recent work, Hare has attempted t o retain the essentials of this view while arguing that the logic of moral d iscourse u nderwrites ut ilitarianism . See Hare ( 1 9 8 1 ) . 5 . These issues are explored i n Copp, Morality and Society (forthcoming) . 6. The reader should exercise due caut ion throughout to ascertain the precise sense given by an author to technical philosophical terms, such as, in the case of Zimmerman's essay, the terms " realism" and "subjectivism ." One should not assume that t he same meaning is to be understood in every case. For instance, the reader shoul d take care to ascertain whether the three authors d iscussed by Zimmerman are all real ists in the sense he defines. 7. See, for example, Schelling ( 1 960), Gauthier ( 1 967), Lackey ( 1 9 8 2 ) , and H ardin ( 198 3 ) . 8 . I a m indebted t o Richmond Campbell and Spencer Carr for helpful c riticisms o f this essay.

part one Truth and Theory Acceptance in Ethics

cha p ter one

Is There a Single True Morality? Gilbert Harman CONFESSION I have always been a moral relativist . As f ar back as I c an re me mbe r thinking about it, it h as see med to me obvious that the dictates of mor ality arise fro m so me sort of convention or underst anding among people, that di fferent people ar rive at different understandings, and that there are no basic mor al de mands that apply to everyone. For many yea rs , this see med so obvious to me that I assu med it was everyone 's instin ctive view, at le ast eve ryone who g ave the matter any thought "in this day and age." When I first studied philosophic al ethics (in the 1950s), I w as not dis abused of this opinion. The main issue at the ti me see med to be to deter mine exactly what fo rm of "noncognitivis m" was correct. (According to noncognitivis m, mor al judg ments do not function to describe a moral re ality but do so mething else -express fee lings, prescribe a course of action, and so forth.) It is true th at many of the philosophers I studied see med for so me reason to want to avoid c alling the mselves " relativists." This was usual ly acco mplished b y defining mo ral relativis m to be an obvious ly inconsistent position ; for ex amp le, the view both th at the re are no univers al mor al t ruths and also th at everyone ought to follow the dictates of his or he r group, where this l ast clai m is t aken to be a universal moral truth. I w asn't sure what this verba l maneuver was supposed to acco mp lish . Why would anyone want to give such a definition of mora l re lativis m? Moral relati vis m w as obviously cor rect, and the philosophers I w as studying see med all to be moral rel ativists even if they did not want to describe the mselves in th at w ay.

28

Is There a Single True Morality?

In the 1960s I was distressed to hear fro m various people tea chi ng ethi cs that stude nts i n their classes te nded to pro clai m the mselves moral relativists u ntil they had bee n show n how co nfused they were a bou t ethi cs . I suspe cted that what confusio ns there were were not co nfusio ns of the students , but were confusio ns of their tea chers , due per haps to a faulty defi nitio n of moral relativis m. It see me d to me that t he o bvious solutio n was to show that moral relativis m ca n be co nsiste ntly defi ned as a plausi ble view a nd that sta ndar d o bje ctio ns to moral relativis m are mistaken. So , I eve ntually wrote a nd pu blishe d a n essay a bout this (Har ma n, 1975), naively thinki ng it would clear thi ngs up a nd e nd worries a bout moral relativis m. I was surprised to dis cover that this di d not happe n. I was also startled to fi nd that ma ny stude nts i n my ow n et hi cs courses resisted my atte mpt to make clear w hat I thought they i nsti nctively believe d . After so me stu dy I con clu ded that in fact o nly so me of t he students i n my courses were i nsti nctive moral relativists ; a sig nifi ca nt nu mber of the m were insti nctive a bsolutists. I ha d k now n of course that there were philosophers a nd frie nds of mine who were not moral relativists. For a lo ng ti me I attri bute d t his to their perversity a nd love of the bizarre a nd atta ched no sig nifi ca nce to it. But the n I dis covere d that so me of the m thought moral relativis m was the perverse view, a kin d of p hilosophi cal folly like skepticis m a bout other mi nds or the e xter nal worl d (for e xa mple , Nagel , 1980). I was stunned ! How could they thi nk that whe n they k new so ma ny moral relativists (like me) a nd no episte mologi cal skepti cs (at least no ne w ho took su ch skepti cis m seriously i n or di nary life)? It the n o ccurre d to me to wo nder how I coul d thi nk of moral a bsolutis m as su ch a perverse view when I k new so many moral a bsolutists.

THE ISSUE It tur ns out to my surprise that the questio n whet her there is a si ngle true morality is a n u nresolve d issue i n moral philosophy. O n o ne si de are relativists , skepti cs , nihilists , an d no ncog nitivists. O n the ot her si de are those who believe in a bsolute values a nd a moral law t hat applies to everyo ne. Stangely , only a few people see m to be u nde ci de d. Al most everyone see ms to be fir mly o n one si de or the other, a nd al most everyo ne see ms to think his or her side is o bvio usl y rig ht , the other si de represe nti ng a kind of ridi culous folly . This is stra nge si nce everyo ne k nows , or oug ht to know, that many i ntellige nt people are o n ea ch side of t his issue.

Gilbert Harman

29

Two Approaches

I n this essay I wa nt to suggest that part of the expla natio n for this mutual i ncomprehe nsio n is that there are two differe nt ways to do moral philosophy . If o ne approach is take n, moral relativism, no ncog nitivism, or skepticism may seem obviously correct a nd moral absolutism may seem foolish. If the other approach is take n, absolutism may seem clearly right a nd skepticism, relativism, a nd no ncog nitivism may seem foolish. The differe nce i n approaches is, to put it crudely, a differe nce i n attitude toward scie nce. O ne side says we must co nce ntrate o n fi ndi ng the place of value a nd obligatio n i n the world .of facts as revealed by scie nce . The other side says we must ig nore that problem a nd co nce ntrate o n ethics proper. Of course, both sides agree that we must begi n at the begi n ni ng with our i nitial beliefs, both moral a nd no nmoral, a nd co nsider possible modificatio ns that will make these beliefs more cohere nt with each other a nd with plausible ge neralizatio ns a nd other expla natory pri nciples. Eve ntually, we hope to arrive at a "reflective equilibrium" (Rawls, 197 1) whe n no further modificatio ns seem called for, at least for the time bei ng. This process will i nevitably leave ma ny issues u nresolved ; i n particular, we may fi nd ourselves with no accou nt of the place that value a nd obligatio n have i n the world of facts. This will not dismay someo ne who is willi ng to leave that questio n u na nswered, but it will be disturbi ng to someo ne who, o n the way to "reflective equilibrium," has come to thi nk that the basic issue i n moral philosophy is precisely how value a nd obligatio n fit i nto the scie ntific co nceptio n of the world. I will use the term " naturalism" for a n approach to ethics that is i n this way domi nated by a co ncer n with the place of values i n the natural world. I will call a ny approach that is not so domi nated a n i nsta nce of "auto nomous ethics," si nce such a n approach allows us to pursue ethics i nter nally. Of cours e, auto nomous ethics allows that scie nce is releva nt to et hics i n as much as ethical assessme nt depe nds o n the facts of the case. But u nlike naturalism, auto nomous ethics does not take the mai n questio n of ethics to be the naturalistic status of values a nd obligatio ns.

Naturalism

I hope the terms " naturalism" a nd "auto nomous ethics" will not be too misleadi ng. The term " naturalism" is sometimes reserved for the thesis that moral judgme nts ca n be a nalyzed i nto or reduced to factual stateme nts of a sort clearly compatible with the scie ntific world view. I

30

Is There a Single True Morality?

am us ing the term " naturalism" more broad ly in a more t rad it ional and accu rate sense. Naturalism in this sense does not have to lead to naturali s t ic red uction, although that is one possibili ty. Another possibility is that there is no way in which ethics could fit into the scient ific conception of the world . I n that case naturalism lead s to moral nihilism, as in Mackie ( 1 977) . Mackie supposes that ethics requires absolute values which have the property that anyone aware of their existe nce must necessarily be mot ivated to act morally. Since our scientific conception of the world has no place for entities of this sort, and since there is no way in which we cou ld become aware of such entities, M ackie concludes that ethics must be rejected as resting on a false presupposition . That is a version of natu ralism as I am using the term . Naturalism can also lead one to a noncognitive analysis of moral j udgments . I n this view, moral j udgments do not function to describe the world, but to do something else-to express one's attitudes for and against things, as Stevenson ( 1 9 6 3 ) argues-or to recommend one or another cou rse of act ion or general policy, as Hare ( 1 9 5 2, 1 9 8 1 ) propose s . Or a naturalist may decide that moral j udgments do make factual claims that fit in with the claims of science . This can be illustrated by some sort of naturalistic reduction . One example would be an analysis that takes moral claims to be claims about the reactions of a hypothe t ical impartial observer as in H u me ( 1 739) or F irth ( 19 5 2) . More complex positions are possible . M ackie ( 1 9 77) argues in C hapter l that ethics rests on a false presupposi t ion, but then he goes on i n later chapters to discuss part icular moral issue s . I t i s almost a s i f he had first demonst rated that God does not exist and had then gone on to consider whether He is wise and loving. Presumably, M ackie believes that ethics as normally conceived must be or can be replaced with something else . But he does not i nd icate exactly what sort of replacement he has in mind-whether it is an institu t ion of some sort, for example. Nor does he say how moral claims made w ithin this replacement fit in with the claims of science . I suspect he wou l d accept some sort of noncognit ivist account of the j udgments that are to replace the old moral j udgment s . It is possi ble t o be both a naturalist and a n absolut ist, although this is not very com mon . F irth ( 1 9 5 2) defends an absolutist version of the ideal-observer t heory and Hare ( 1 9 8 1 ) defends an absolutist version of noncogn itivis m . But I will argue that the most plausib le versions of naturalism involve a moral relativism that says d ifferent agents are subj ect to d i ffcrent basic moral req u i rements depend ing on the moral conve nt ions in which they part icipate .

Gilbert Harman

31

Autonomous Ethics Naturalism tends toward relativism. W hat I am calling autonomous et.hi es, on the other hand, can have a very different tendency. In this approach, science is relevant, since our moral judgments depend on what we take the facts to be ; but we attach no special importance to saying how obligations and values can be part of the world revealed by science. Rather, we do ethics internally . We begin with our initial moral beliefs and search for general principles. Our initial opinions can be changed to some extent so as to come into agreement with appealing general principles and our beliefs about the facts, but �n important aspect of the appeal of such principles will be the way in which they account for what we already accept. This approach normally (but not always) involves a n initial assumption of moral absolutism, which in this context is of course not the thesis that there are simple moral principles that hold absolutely without exceptions, but rather the thesis that there are basic moral demands that apply to all moral agents. Autonomous ethics tends to retain that absolutist thesis. It may also involve some sort of intuitionism, claiming that each of us has immediate insight into the truths of certain moral principles. It sometimes leads to a fairly conservative morality, not much different from one 's initial starting point . That is not surprising given the privileged position assigned to our initial moral beliefs. But let me stress that conservatism is not inevitable, and autonomous ethics can and often does lead to more radical moralities too. It leads some philosphers to a radical utilit arianism, for example. It leads Rawls ( 197 1) to principles of social justice that appear to be considerably more egalitarian than those most people accept. And Nozick ( 1974), using the same general approach, comes out at a very different place, in which he ends u p denying that any sort of egalitarian redistribution by go vernments is ever morally justified. ( However, the moral theory in Nozick, 198 1, as contrasted with the political theory in Nozick, 1974, insists on the moral re quirement of helping others.) Indeed, there are many different w ays in which ethics can be pursued as an autono mous discipline with its own principles that are not reducible to the principles of any science. I can illustrate this variety by mentioning a few of the many other contemporary philosophers who accept some form of au­ tonomous ethics: Baier ( 1958), Darwall ( 198 3), Donagan (1977), Fran­ kena ( 1976), Fried ( 1978), Gewirth ( 1978), Grice ( 1967), Nagel ( 1970, 1980), and Richards ( 197 1). Each of these philosophers has a somewhat different approach, although all are absolutists who rely on some form of autonomous ethics.

32

Is There a Single True Morality ?

I should say that it is possible to believe i n auto no mous ethics without bei ng a n absolutist. O ne might be i mpressed by the variety o f vie ws he ld by those who accept auto no mous ethics a nd so be led to allo w for relativis m while co nti nui ng to accept the method o f auto no mous ethics , be lievi ng that natura lis m must be rejected. A possible exa mple is Mc Do well ( 1978 , 1979 , 198 1). But the te nde ncy o f auto no mis m i n ethics is toward absol utis m . I n what follo ws I will restrict my discussio n to absolutist versio ns o f auto no mous ethics a nd to relativistic versio ns o f naturalis m .

Teachers of Ethics I might a lso me ntio n that ethics pursued i nter nally , as i n auto no mous ethics , is more i nteresti ng to ma ny peop le tha n ethics as pursued by naturalis m. That is because auto no mous ethics allo ws o ne to spe nd more o f o ne 's ti me thi nki ng about i nteresti ng co mplicated moral pu zzles tha n natura listic ethics does , a nd ma ny people fi nd moral pu zzles more interesti ng tha n "abstract" questio ns about the objectivity o f value a nd its place i n nature. Philosophers attracted by naturalis m te nd not to find ethics as i nteresti ng a subject as do phi losphers attracted by auto no mous ethics. So , relativists te nd t o be less i nterested i n ethics tha n absolutists are. For exa mple , logicia ns , phi losphers o f scie nce , a nd philosophers o f mathe matics ,who te nd to ward naturalis m, are usually not moral absolutists a nd are not very i nterested i n ethics as a philosophical subject. Philosophers who are re latively i nterested i n ethics te nd to be those who favor auto no mous ethics a nd there fore te nd to be abso lutists. This is why teachers o f ethics te nd more tha n their stude nts to be absolutists. It is not merely , as they so meti mes suppose , that ethics teachers have see n through co nfusio ns that affect their stude nts. A more i mporta nt factor 1s that re lativists te nd not to beco me teachers o f ethics.

WHY DO WE BELIEVE WHAT WE BELIEVE? Auto no mous ethics a nd natura lis m represe nt very differe nt attitudes toward the relatio n betwee n scie nce a nd ethics . Co nsider, for exa mple , the questio n o f what exp lai ns our be lievi ng what we i n fact believe . Naturalists see a n i mporta nt differe nce bet wee n our factura l belie fs a nd our moral beli efs . Our ordi nary factua l belie fs provide us with evide nce that there is a n i ndepe ndent world o f objects because our havi ng those belie fs ca nnot be plausib ly explai ned without assu mi ng we i nteract with a n i ndepe nde nt wor ld o f objects exte rna l to ourselves , objects we perceive a nd ma nipu late.But our havi ng the moral belie fs we have ca n be explai ned entirely i n ter ms o f our upbri ngi ng a nd our psychology , without a ny a ppea l to a n i ndepe nde nt real m o f values a nd ob ligatio ns. So our mora l

Gilbert Harman

33

beliefs do not provide us with evidence for such an independent realm of values and obligat ions, and we must choose between skepticism, noncognit ivism, and relativism ( Harman, 1977, chapter 1 ) . Autonomists disagree with this. They claim we often believe that something is good or right or obligatory in part because it is good or right or obligatory. They accuse naturalists of begging the question . When naturalists say that a belief cannot be explained by virtue of something's being right, unless that thing 's being right consists in some psychological or sociological fact, they simply assume that all explanatory factors are part of the world revealed by science . But this is the point at issue. Autonomists argue that it is more o�vious that we sometimes recognize what is right than that naturalism is correct. True, we may be unable to say how a given "moral fact" and someone's recognition of it fit into the world of facts as revealed by science. But there are always unanswered questions. To jump from our current inability to answer this quest ion to skepticism, relativisim, or noncognitivism is to make a more d rastic move than this puzzle warrants, from the point of view of autonomous ethics.

Explanation and Reduction The naturalist seeks to locate the place of value, j ustice, right, and wrong, and so forth in the world in a way that makes clear how they might explain what we take them to explain. A naturalist cannot understand how value, j ustice, right, and wrong might figure in explanations without having some sense of their "location" in the world . We can say that this involves "naturalistic reduction," but it need not involve reductive definitions of a serious sort. I ndeed, reduction rarely ( if ever) involves serious reductive definit ions. We identify tables with clusters of atoms in a way that allows us to understand how tables can hold up the things they hold up without having to suppose the word table is definable using only the concepts of physics ! Similarly, we identify colors with d ispositional properties of objects, namely, their tendencies to look in certain ways to certain sorts of observers in certain conditions, without having to suppose there is a sat isfactory definition in these terms. Similarly for temperatures, genes, and so on . What a naturalist wants is to be able to locate value, j ustice, right, wrong, and so forth in the world in the way that tables, colors, genes, temperatures, and so on can be located in the world . What i s at issue here i s understanding how moral facts might explain something, how the badness of someone's character might explain why that person acts in a certain way, to take an example from Sturgeon's essay. It is not sufficient that one be prepared to accept the counterfactual

34

Is There a Single True Morality ?

ju dg ment that the pe rson woul d not have acte d i n that way if the pe rso n had not had a ba d cha racte r, if o ne does not see how the badness of the pe rso n's cha racte r coul d have such an effect . A natu ralist believes o ne ca n see that o nly by locating ba dness of cha racte r i n aspects of the wo rld which o ne sees can have that effect . Notice that a "natu ralist" as I a m he re usi ng the te rm is not just so meone who supposes that all aspects of the wo rl d have a natu ralistic location in this way, but rathe r so meone who takes it to be of ove r ridi ng i mpo rta nce i n doing mo ral philosophy actually to atte mpt to locate mo ral p rope rties . My clai m is that, when one takes this atte mpt se riously, one will te nd to beco me skeptical o r relativistic . Stu rgeon is not a natu ralist in my se nse, despite his insiste nce that he takes mo ral facts to be natu ral facts.

MORAL ABSOLUTISM DEFINED I now wa nt to be mo re specific about what is to count as mo ral absolutis m . Va rious things might be meant by the clai m that the re a re absolute values a nd one t rue mo rality . Mo ral absolutists i n one se nse might not be mo ral absolutists i n 0the r senses . We must be ca reful not to mix up real issues with pu rely ve rbal issues. So let me stipulate that I will take mo ral absolutis m to be a view about the mo ral reaso ns people have to do things a nd to want o r hope fo r things. I will unde rsta nd a belief about absolute values to be a belief that the re a re things that eve ryone has a reason to hope o r wish fo r. To say that the re is a mo ral law that "applies to eve ryone" is, I he reby stipulate, to say that eve ryo ne has su fficient reasons to follow that law . It is t rue that many philosophe rs pu rsue so mething that rese mbles auto no mous ethics whe n they ask what p ri nciples a n "ideal" mo ral code of one o r anothe r so rt woul d have, quite apa rt fro m the question whethe r people now have a ny reaso n to follow that code . Depe ndi ng o n what so rt of idealization is being consi de red, the re may o r may not be a u nique "i deal " co de of that so rt. But I a m not goi ng to cou nt as a fo rm of mo ral absolutis m the clai m that the re is a u ni que i deal mo ral co de of such a nd such a type. Relativists a nd absolutists in my sense might ve ry well ag ree about this clai m without that havi ng any effect at all on what I take to be the basic iss ue that sepa rates the m, si nce this clai m has no i mme diate relevance to questions about what reaso ns people actually have to hope fo r ce rtain things o r do ce rtain things. Si mila rly, I a m not going to count as a fo rm of mo ral absolutis m the clai m that the re is one t rue mo rality that applies to eve ryone in that e ve ryone ought to follow it, if this is not taken to i mply that eve ryo ne h as a su fficient reaso n to follow it . I a m not su re what ought is supposed

Gilbert Harm an

35

to mean if it is disconnected in this way from reasons to do things. If what is meant is that it ought to be the case that everyone followed the one true morality-in other words that it would be a good thing if . they did-then this is a version of the view that there is a uni que "ideal" moral code. I am not sure what else might be meant, although a great deal more could be said here ( Harman 1978a). Rather than try to say it, however, I simply stipulate that this sort of claim is not a version of what I am counting as moral absolutism. I should note that, of the contemporary philosophers I have identified as absolutists, Baier, Darwall, Donagan, Frankena, Gewirth, Grice, Nagel, and Richards, clearly advocate moral absoluti sm in this sense . They all think that there are basic moral demands that in some sense every competent adult has reasons to adhere to. I believe the others I men ­ tioned-namely Rawls, Nozick, and Fried-also a gree with this, although they do not explicitly say so in the works I have cited.

DOES A SINGLE MORAL LAW APPLY TO EVERYONE? Consider the issue between absolutism and relativism concerning reasons people have for doing things. According to moral absolu tism about this, there is a single moral law that applies to everyone; in other words, there are moral demands that everyone has sufficient reasons to follow, and these demands are the source of all moral reasons. Moral relativism denies that there are universal basic moral demands and says different people are subject to different basic moral demands depending on the social customs, practices, conventions, values, and principles that they accept. For example, a moral absolutist might suppose there is a basic moral prohibition on causing harm or injury to other people. This prohibition is in one sense not absolute, since it can be overridden by more compelling considerations and since it allows exceptions in order to punish criminals, for instance. But the prohibition is supposed to be universal in the sense that it applies to absolutely all agents and not just to those who happen to participate in certain conventions. The abso lutist claims that a bsolutely everyone has sufficient reasons to observe this prohibition and to act as it and other basic moral requirements dictate. A moral relativist denies this and claims that many people have no reasons to observe this prohibition. Many people participate in moralities that sharply distinguish insiders and outsiders and do not prohibit harm or injury to outsiders, except perhaps as this is likely to lead to retaliation against insiders. A person participating in such a morality has no reason to avoid harm or injury to outsiders, according to the relativist, and so the general prohibition does not apply to that person. Such a person

36

Is There a Single True Morality ?

may b e a m emb er of som e p rim 1t 1v e t ribal g roup, but h e o r sh e n eed not b e. H e o r sh e might also b e pa rt of cont empo ra ry soci ety, a succ ess ful p rof essional c riminal, say, who recogniz es va rious obligations to oth e r m emb ers of a c riminal o rganization but not to thos e on th e outsid e . Acco rding to th e mo ral relativist, the succ ess ful c riminal may w ell have no reason at all not to ha rm his o r h er victims.

A n Argument for Relativism L et us conc ent rat e on this ca se. Th e mo ral absolutist says th e d emands of th e on e t ru e mo rality apply as much to this succ essful c riminal as to anyone els e, so this c riminal do es hav e a reason no t to ha rm a gi ven victim. Th e relativist d eni es th e c riminal has any such reason and so d eni es th e rel evant mo ral d emand is a univ e rsal d emand that appli es to ev eryon e. H ere natu ralism t ends to suppo rt rela tivism in th e following way. Consid er wha t it is fo r so meon e to have a su ffici en t reason to do som ething . Natu ralism requi res tha t this should b e explained in t erms cong enial to sci enc e. W e cannot simply t rea t this as i rreducibly no rmativ e, saying, fo r exampl e, that som eon e h as a su ffici ent reason to do som ething if and only if h e o r sh e ought to do it. Now, p resumably, som eone has a suffici en t reason to do som ething if and only if th e re is wa r ran ted reasoning that p erson could do which would l ead him o r h er to d ecide to do that thing . A natu ralist will suppos e that a p erson wi th a su ffici ent reason to do som ething migh t fail to reason in this way to such a d ecision only b ecaus e of som e so rt of empi rically discove rabl e failu re, du e to inatt ention, o r lack of tim e, o r failu re to consider o r app recia te c ertain a rgum ents, o r igno ranc e of certain availabl e evidenc e, o r an e r ro r in reasoning, o r som e so rt of i rra tionality o r un reasonabl en ess, o r w ea kn ess of will . If th e p erson do es not in tend to do som ething and that is not b ecaus e h e o r sh e has fail ed in som e such empi rically discov e rabl e way to reason to a d ecision to do that thing, th en, acco rding to th e natu ralis t, that p erson cannot hav e a su ffici ent reason to do that thing. This is the fi rst p remis e in a natu ralistic a rgum ent in suppo rt of th e relativist. Th e oth er p remis e is that th ere a re p eopl e, such as c ertain p rofessional c riminals, who do not act in acco rdanc e with th e all eg ed requi rem ent not to ha rm o r inju re oth ers, wh ere this is no t du e to inatt ention o r failu re to consider o r app reciat e c ertain a rgum ents, o r igno ranc e of c ert ai n evid enc e, o r any erro rs in reasoning, o r any so rt of i rra tionality o r un reasonabl en ess, o r w ea kn ess of will. Th e a rgum ent fo r this is simply that th ere cl ea rly a re p eopl e who do not adh e re to th e requi rem ent in qu estion and who do not seem to hav e fail ed in any of th es e ways. So,

Gilbert Harman

37

in the absence of special theoretical reasons, deriving, say, from psychology, to think these people must have faiL�d in one of the specified ways, we can conclude they have not done so. F rom these two premises it follows that there are people who do not have sufficient reasons, and therefore do not have sufficient moral reasons, to ad here to the general prohibition against harming or inj uring others . I n particul ar, a successfu l criminal may not have a sufficient reason not to harm his or her victims. The moral prohibition against harming others may simply fail to apply to such a person. It may fail to apply in the relevant sense, which is of course not to say that the principle makes an explicit exception for crim inals, allowing them but not others to inj ure and harm people without rest raint . Rather, the principle may fail to apply in the sense that the criminal in question may fail to have sufficient reason to act in accordance with the principle .

A n Absolutist Reply Moral absolutism must rej ect this argument . It can do so by invoking autonomous ethics at the place at which moral relativism invokes nat­ uralism . Autonomous ethics does not suppose that we must give s.o me sort of naturalistic account of having a sufficient reason to do something, nor does it suppose that only a science like psychology can d iscover the conditions under which someone has failed to reason in a certain way because of i nat tent ion, irrationality, unreasonableness, or any of the other causes of failure ment ioned in the relativistic argument. Autonomous ethics approaches this issue in the following way. We begin with certain bel iefs . Presumably these imply that everyone has a sufficient reason to observe the prohibition against harm to others, includ ing, in particu lar, the successful criminal who does not part icipate in or accept any practice of observing this general prohibition . At the start we therefore believe that the criminal does have sufficient reason not to harm his or her vict ims. Fol lowing autonomous ethics, then, we shoul d continue to believe this unless such continued belief conflicts with generalizat ions or other theoretical principles internal to ethics that we find at tractive because they do a better job at making sense of most of the things we originally believe . Taking this approach, the absolutist must claim that the relativistic argu ment does not provide sufficient reason to abandon our original absolutism. It is more plausible, accord ing to the absolutist, that at least one of the premises of the rel ativistic argument is false than that its conclusion is t rue.

38

Is There a Single True Morality ?

Assessing the First Premise The firs t premise of the rela tivis tic argumen t is tha t for someone to have a sufficien t reason to do some thing there mus t be warran ted reasoning available to tha t person tha t leads to a decision to do tha t thing, so tha t if the person fails to in tend to do tha t thing i t mus t be because of ina t tention, lack of time, failure to consider or apprecia te cer tain ar­ gumen ts, ignorance of relevan t evidence, an error in reasoning, irra­ tionali ty, unreasonableness, or weakness of will. The absolu tis t migh t objec t tha t this is oversimpli fied . If a person wi th sufficie nt reason to do some thing does no t do i t, then some thing has gone wrong, and i t migh t be one of the things the rela tivis t men tio ns, bu t i t migh t be some thing else as well. There migh t be some thing wrong wi th the person in ques tion. Tha t person migh t be bad, immoral. The failure migh t simply be a failure no t to care enough abou t o ther people. A person ough t to care abou t o thers and there is some thing wrong wi th a person who does no t care, even if tha t person is no t ina t ten tive, ignoran t, rushed, or defec tive in any o ther of the par ticular ways the rela tivis t men tions. So, even if some peo ple fail to observe the prohibi tion agains t harming o thers no t because of ina t ten tion, lack of time, a nd so for th, bu t simply because of lack of concern and respec t for o thers, such people s till do have sufficien t reason no t to harm o thers. ( This response on behalf of absolu tism was sugges ted to me by Thomas M . Scanlon.) This response to the rela tivis tic argumen t is a response wi thin au­ tonomous e thics. I t does no t e xplain having a sufficie nt reason to do some thing in terms tha t are accep tably fac tual from a na turalis tic per­ spec tive. I t appeals also to the no tion of some thing 's being wro ng wi th someone, where wha t migh t be wrong is simply tha t the person is bad or immoral . I t is like saying one has a sufficien t reason to do some thing if and only if one ough t to do i t, or if and only if i t would be wrong no t to do i t. The rela tivis t claims tha t the only plausible accoun ts of these nor ma tive no tions are rela tivi stic ones.There is no prohibi tion on harm to ou tsiders in the criminals ' morali ty. There is such a prohibi tion only in some o ther morali ty. In tha t o ther morali ty some thing is wrong wi th a person who has no compunc tion abou t injuring someone else ; bu t no thing is wrong wi th such a person wi th respec t to the criminal morali ty, as long as those injured are ou tsiders . Bu t how can i t be a sufficien t reaso n for the cr iminal no t to harm his or her vic tims tha t this is prohibi ted by somebody else 's morali ty? How can i ts being bad, immoral , or wrong in this o ther morali ty no t to care abou t and respec t o thers give the crimina l, who docs no t accep t tha t morali ty, a sufficien t reason to do a ny thing?

Gilbert Harm an

39

The a bsolutist 's answer is that failure to respect others is not just wrong according to so me morality the cri minal does not accept , it is also wrong , period. So mething is really wrong with lack of respect and _concern for others. It is not just wrong in relation to one or another morality. Of course , the relativist will not be satisfied with this answer and , appealing to naturalis m, will ask what it is for so mething to be wrong in this way. The a bsolutist supposes that the failure to care a bout and respect others does involve so mething the a bsolutist points to by saying this failure is wrong. But what is this thing that is true of such a failure to care and that can give the cri minal a sufficient reason not to har m and injure others? The relativist can see no aspect of such a failure that could provide such a reason . This · of course is because the relativist , as a naturalist , considers only aspects of the failure that are clearly co mpati ble with a scientific world view. The relativist disregards putative aspects that can be specified only in nor mative ter ms . But the a bsolutist , as an autono mist , can specify the relevant aspect of such a failure to care a bout others: It is bad , i mmoral , wrong not to care ; the cri minal ought to have this concern and respect and so ought not to har m and injure others , and there fore has a sufficient reason not to har m and injure the m.

Assessing the Second Premise We have been discussing an argu ment for relativis m concerning moral reasons. We have seen that naturalis m supports the first pre mise of this argu ment and that autono mous ethics allows the rejection of this pre mise . The sa me thing is true of the second pre mise , which says that there are people , such as the successful cri minal , who do not o bserve the alleged require ment not to har m or injure others and this is not due to inattention , failure to consider or appreciate certain argu ments , ignorance of relevant evidence , errors i n reasoning , irratio nality , unreasona bleness , or weak ness of will. Naturalis m supports this because ther e do see m to be such people , and no scientifically accepta ble grounds exist for thinking this is an illusion. On the other hand , autono mous ethics allows other grounds , not reduci ble to scientific grounds , for thinking this is an illusion. In autono mous ethics we begin by supposing that we recognize the wrongness of har ming others , where this is to recognize a sufficient reason not to har m others . If that is so mething we recognize , then it must be there to be recognized , so the successful cri minal in question must be failing to recognize and appreciate so mething that is there. The a bsolutist might argue that the cri minal must be irrational or at least unreasona ble. Seeing that a proposed course of action will pro ba bly cause serious injury to so me outsider, the cri minal does not treat this

40

Is There a Single True Morality ?

as a reason not to undertake that course o f action. This must be irrational or unreasonable, because such a consideration si mply is such a reason and indeed is an obvious reason, a basic reason, not one that has to be derived in so me co mple x way through arcane reasoning. But then it must be irrational or at least unreasonable for the cri minal not to care sufficiently about others, since the cri minal 's lack o f concern for others is wh at is responsible for the cri minal 's not taking the likelihood o f harm to an outsider to be a reason against a proposed course o f action. This is one way an absolutist might argue. The relativist 's reply to such an argu ment is that, on a ny plausible characterization o f reasonableness and unreasonableness (or rationality and irrationality) as notions that can be part o f the scienti fic conception o fthe world, the absolutist 's clai m is just false. So meone can be co mpletely rational without feeling concern and re spect for outsiders. But o f course this reply appeals to naturalis m. The absolutist who rejects naturalis m in favor o f autono mous ethics relies on an unreduced nor mative char ­ acterization o f rationality an d irrationality (or reasonableness and un ­ reasonableness). Now the argu ment continues as be fore. The relativist argues that, i f rationality and irrationality (or reasonable ness a nd unreaso nableness) are conceived nor matively, they beco me relative notions. W hat one morality counts as irrational or unreasonable, another does not. The cri mi nal is not irrational or unreasonable in relation to cri minal morality, but only in relation to a morality the cri minal rejects. But the fact that it is irrational or unreasonable in relation to this other morality not to have concern and respect for others does not give the cri minal who rejects that morality any reason to avoid har ming or injuring others. The absolutist replies that relative irrationality or unreasonableness is not what is in question. The cri minal is irrational or at least unreasonable, period. Not just irrational or unreasonable in relation to a morality he or she does not accept. Since it is irrational or unreasonable for anyone not to care sufficiently about others, everyone has a sufficient reason not to injure others, whether he or she recognizes this reason or, through irrationality or unreasonableness, does not recognize it. The naturalist is unconvinced by this because the naturalist can fin d no aspect o f th e- cri minal the absolutist might be referring to in saying the cri minal is "irrational" or "u nreasonable," i f this aspect is to give the criminal any reason to care about others. This o f course is because the naturalist is considering only naturalistic aspects o f the cri minal, whereas the absolutist, as an autono mist, is thinking about an u nreduced normati ve aspect , so mething the naturalist cannot appeal to.

Gilbert Harman

41

So, a s was true o f the fir st pre mi se o f the rel ativi stic argu me nt abou t re aso ns, the seco nd pre mi se depe nd s o n an assu mp tio n o f natur ali sm. By appe ali ng to au to no mou s e thic s, an ab solu ti st c an rejec t thi s pre mi se. A n ab solu ti st may i n fac t ac tu ally accep t o ne or the o ther o f the pre mi se s o f the rel ativi stic argu me nt (al though o f cour se no t bo th). A give n ab solu ti st migh t rejec t ei ther the fir st pre mi se or the seco nd or bo th pre mi se s. A n ab solu ti st migh t eve n be u ndecided, holdi ng merely th at o ne or the o ther pre mi se mu st be rejec ted, wi thou t sayi ng which. There i s no thi ng wro ng wi th bei ng u ndecided abou t thi s. Re fle ctive e quilibriu m le ave s many i ssue s u nre solved .

ARE THERE ABSOLUTE MORAL VALUES? The si tu atio n i s si mil ar i n the theory o f v alue. N atur ali sm te nd s to suppor t the co nclu sio n th at all v alue i s rel ative and th at so me thi ng i s al way s good for o ne or ano ther per so n or group o f people or i n rel atio n to a specified se t o f purpo se s or i ntere sts or ai ms. Au to no mou s e thic s allo ws al so for ab solu te v alue s, thi ng s th at are good, period, and no t ju st good for so meo ne or so me group or for so me purpo se. The i ssue here co ncer ns the good ne ss or v alue o f a po ssible state o f affair s, no t the good ne ss or v alue o f so me thi ng as a thi ng o f a give n sor t. The i ssue i s no t wh at i t i s for so me thi ng to be a good thi ng o f a ki nd, a good k ni fe, a good watch, a good b ack swi ng, a good apple, a good far mer, a good poe m. The i ssue i s r ather wh at i t i s for an eve nt or si tu atio n to be a good thi ng ; wh at i s i t, for ex ample, to be a good thi ng th at i t i s r ai ni ng or th at Egyp t and I sr ael sig ned a pe ace tre aty. I t i s u nco ntrover si al th at thi s sor t o f good ne ss i s so me ti me s rel atio nal. A si tu atio n i s good for so meo ne or so me group o f people, good fro m a cer tai n poi nt o f vie w, i n rel atio n to cer tai n purpo se s or i ntere sts. Th at i t i s r ai ni ng i s a good thi ng for the far mer, bu t no t for the v ac atio ner. Th at Egyp t and I sr ael sig ned a pe ace tre aty migh t be good fro m their poi nt o f vie w, bu t no t fro m the poi nt o f vie w o f the P L O. Give n a fixed poi nt o f re fere nce, we can ev alu ate state s o f affair s as be tter or wor se. The v alue o f a state o f affair s i n rel atio n to th at re fere nce poi nt repre se nts the degree to which so meo ne wi th the relev ant purpo se s and i ntere sts h as a re aso n to try to bri ng abou t, or want, or at le ast hope for th at state o f affair s. No w i t c an be argued th at there i s al so a ki nd o f ab solu te v alue. The cl ai m i s th at state s o f affair s c an be good or b ad, period, and no t merely good or b ad for so meo ne or i n rel atio n to give n purpo se s or i ntere sts. O n he ari ng o f poi ntle ss p ai nful experi me nts o n l abor atory ani mal s, for ex ample, o ne i mme di ately re ac ts wi th the though t th at thi s i s b ad and

42

Is There a Single True Morality ?

it wo uld be good to eliminate s uch practices. Clearly, one does not simply mean that these tort ures are bad for the animals involved and that these animals wo uld bene fit if s uch experiments were ended. A heartless experimenter might agree that what he does is bad for the animals witho ut having to agree that it wo uld be a good thing to eliminate this sort of experimentation. Similarly, it seems intelligible to s uppose that it wo uld be better if there were no ineq ualities of wealth and income in the world even tho ugh this wo uld not be better for everyone, not for those who are now relatively wealthy, for instance . And this seems to say more , for example, than that the average person wo uld be better o ff if there were no s uch ineq ualities, since an elitist might agree with that b ut not agree that the envisioned state of a ffairs wo uld be better, period, than o ur present sit uation. Again, we can consider which of vario us pop ulation policies wo uld lead to the best res ulting state of a ffairs even tho ugh these policies wo uld res ult in di fferent pop ulations, so that we cannot be simply considering the interests and p urposes of some fixed gro up . It may seem, then, that we can consider the absol ute val ue of a possible state of a ffairs.

Skepticism about Absolute Values The relative val ue of a possible state of a ffairs in relation to given p urposes and interests is a meas ure of the extent to which someone with those p urposes and interests has a reason to try to bring abo ut, or want, or hope for that state of a ffairs. The absol ute val ue of a possible state of a ffairs is a meas ure of the extent to which anyone, apart f rom having a personal stake in the matter, has a reason to try to bring abo ut, or want, or hope for that state o f a ffairs. Nat uralism leads to skepticism at this point. How co uld we ever be aware of absol ute val ues? How co uld we ever kno w that everyone has a reason to want a certain possible state of a ffairs? F urther reflection along nat uralistic lines s uggests that apparent ab ­ sol ute val ues are often ill usory projections of one 's personal val ues onto the world. Sometimes this sort of projection yields pla usible res ults, b ut us ually it does not. To begin with the most pla usible sort of case, in hearing abo ut the pain involved in animal experimentation, o ur sympathies arc immediate l y and vividly engaged ; we immediately side with the animals against the experimenters. In saying " That is awful !" we are not j ust saying "That is a wful for the animals," since o ur remark expresses o ur sympathetic identi fication with the point of vie w of the animals. We do not merely state a fact, we express o ur feelings and we expect an awareness of this state of a ffairs to call forth the same feelings of dismay in e veryone . This expect ation seems reasonable eno ugh in this

Gilbert Harm an

43

c ase, since it m ay well be, as Br andt argues, th at everyone h as a sy mp athetic re action to suffering ( 1976, p . 450) . But pl ausibility v anishes as soon as the c ase becomes even a little . complex . Suppose the anim al experiments are not pointless but are an essenti al p art o f a kind o f medic al rese arch th at promises to allevi ate a cert ain amount o f hum an suffering. Or suppose th at, although the experiments promise no pr actic al benefit o f this sort, they are relev ant to a theoretic al issue in psychology. A given person m ay st ill feel th at it is b ad th at the experiments should occur and th at it would be good i f they were not done, the g ain not being worth the cost. Ag ain, the person is not just s aying th at the experiments. are b ad for the anim als, something to which everyone would agree . He or she is also expressing over all dis approv al o fthe experiments, e xpecting others also to dis approve i f they consider the issue in an imp arti al w ay. The trouble is th at people re act differently to these c ases. Consider the question whether it is good or b ad to experiment p ainfully on anim als in order to resolve cert ain theoretic al issues in psychology . The extent to which this is (absolutely) good is the extent to which everyone (ap art from any person al st ake in the m atter) h as a re ason to try to bring it about th at such experiments are done, or to w ant them to be done, or hope t hat they are done . The extent to which this is (absolutely) bad is the extent to which everyone (ap art from any person al st ake) h as a re ason to try to end the experiments, or w ant them to end, or hope they end . But n atur alism suggests th at there is no unique answer here and th at wh at a person h as a re ason to w ant will depend on the rel ative v alu e he or she att aches to anim al suffering, to using anim als as me ans, and to theoretic al pro gress in psychology. Different people att ach different v alues to these things without h aving overlooked something, without being irr ation al or unre ason able, and so on . So it seems th at some people will h ave re ason to be in favor o f the experiments and others will h ave re ason to be opposed to the expe riments, where this is determined by the person al v alues o f those people . I f we suppose th at our answer is the right answer, we are merely projecting our own v al ues onto the world.

The Issue Joined O f course, autonomous ethics sees nothing wrong with projecting our own v alues onto the world, holding in fact th at th at is ex actly the right method ! We should begin with our initi al v alu ations and modi fy them only in the interests o f theoretic al simplicity . I f we st art out believing in absolute v alues, we should continue believing this until forced to believe otherwise.

Is There a Single True Morality ?

44

Cl early th e controv ersy ov er absolut e valu es parall els th e co ntrov ersy about r easons to do things . Th e argu ment against absolut e valu es has th e sa me structur e as th e r elativistic argu ment about r easo ns to do things. Its first pr emis e is that a perso n has a r eason to want or ho pe for or try to bring about a particular stat e o f a ffairs only to th e ext ent that h e or sh e would b e irrational or u nr easo nabl e not to wa nt that stat e o fa ffairs unl ess h e or sh e was unawar e o fso me r el evant consid eratio n, was con fus ed, or had so me oth er s pecified d e fect. Its s eco nd pr emis e is that, exc ept for th e si mpl est cas es, a perso n can fail to want a giv en stat e o f a ffairs without b eing irratio nal or u nr easo nabl e or ig nora nt or what ev er . Th e conclusion is that, exc ept possibly for si mpl e cas es, wh er e, for exa mpl e, th e only thing r el evant is that a cr eatur e suff ers, th er e ar e no r easons ev eryo ne has to want or ho pe for or try to bri ng about a giv en stat e o f a ffairs. So th er e ar e no nontrivial absolut e valu es . As b efor e, th e two pr emis es ar e d efended in each cas e by a n a pp eal to naturalis m: W e must giv e a naturalistic accou nt o f r easo ns an d w e must giv e empirical grounds for sup posing so meon e to b e irratio nal or unr easonabl e. Th e absolutist r ej ects th e argu ment as b e for e by i nvoking autono mous ethics, perha ps by r ej ecting th e naturalistic account o f r easons, perhaps by r ej ecting th e r equir ement that sci entific grounds must b e giv en for a judg ment o firrationality or u nr easonabl eness, possibly r emaining undecid ed b etw een th es e alt ernativ es.

NATURALISM VERSUS AUTONOMOUS ETHICS So th e issu e b etw een r elativis m and absolutis m co mes dow n to th e disput e b etw een naturalis m and autono mous ethics . W hich is th e b est approach in moral philoso phy? Should w e conc entrat e o n th e plac e o f valu es and r easons in th e world o f sci entific fact, as naturalis m r ec ­ o mmends, or should w e start with our initial moral b eli efs an d look for g en eral princi pl es and moral th eori es that will ev entually yi eld a r e fl ectiv e equilibriu m, not putting too much w eight on th e qu estion o f th e plac e o f valu e in th e world o f facts.

Religious Beliefs In thinking o f t he issu e b etw een naturalis m and auto no mous ethics, it is us efu l to consid er analogous issu es that aris e in oth er ar eas. Co nsi d er r el igious b eli efs . Our sci entific conc eption o f th e world has no plac e for gods, ang els, d e mons, or d evils. Naturalists hold that th er e is no e mpirical ev id enc e for th e exist enc e o f such b eings nor for any sort o f divin e int erv ention in hu man history. Naturalists say that peo pl e's r eligious b eli efs can b e ex plain ed in t er ms o f th eir upbringing and psychology

Gilbert Harman

45

without any supernatural assumpt ions, so these beliefs provide no evidence whatsoever for the truth of religious claims. Natu ralists therefore i ncline toward skepticism and atheism, although naturalism might also lead to . a kind of rel igious noncognitivism which supposes that religious language makes no factual claims about a supernatural realm but has a different function, for example, in religious ritual . Another approach to rel igion is for a believer to start with his or her initial religious beliefs, includ ing beliefs in the authority of certain writi ngs, and then to develop general principles and theories that would accommod ate these beliefs, allowing modificat ions in the interest of more plausible general principles . This will continue until no further modi­ fic ations seem useful in improving the organization and coherence of that person's views . I nevitably, many questions will remain unanswered, and these will include issues concern ing the relation between that person's rel igious views and his or her scientific views, for example, as regards creatio n . But this is not a serious worry for autonomous religion, which will say this shows merely that science is not everything, or at least that there are th ings we do not now and perhaps never w il l understand . Naturalists say there is no reason to accept religious claims, because the fact that people have the religious bel iefs they have can be explained w ithout any supernatural assumptions . Rel igious autonomists say there is reason to accept religious claims, at least for someone who begins w ith rel igious beliefs, since the process of generalization, systematization, and theory construction internal to religion will give that person no reason to abandon more than a few, if any, of those religious beliefs . F urthermore, certain supernatural events might b e part of the correct explanat ion of the appearance of sacred texts, the occurrence of m iracles, and particu lar rel igious experiences. There is at present no way to say how these religious explanations mesh with ordinary scientific concep­ t ions, but that by itself is no more an obj ect ion to religion that it is an obj ection to science. Naturalists in ethics might urge this rel igious analogy as an ad hominem argument against those defenders of autonomous ethics who are not w illing to take the same l ine with respect to rel igion .

Beliefs about the Mind There is another sort of issue in which an autonomous position comes off looking rather good, even in an irreligious age, namely, the so-called m ind-body problem . Here the naturalistic position corresponds to the thesis of physicalism, accord ing to which all real aspects of mind must be features of the physical brain and central nervous system, its atomic or neural structu re, or some more complex structure that the brain and

46

Is There a Single True Morality ?

ne rvous system instantiate. This may involv e b ehavio rism o r som e so rt o f functionalism that t reats th e b rain as an in formation-p roc essing syst em like a compute r in a robot. A fe w defende rs of this app roach , lik e Skinn er ( 1974) , conclude that th ere a re no mental ev ents , no mind , no con­ sciousness, no sensation. ( Ro rty , 1965, sympathetical ly desc rib es a simila r vie w, "eliminative mate rialism.") But most physicalists suppose that menta i events and othe r aspects o f mind do exist and can be identi fi ed with ce rtain physical o r st ructu ral o r functional asp ects o f th e b rain and cent ral ne rvous system. On the othe r sid e is autonomous mentalism , which holds that th e physicalist hypoth esis clea rly l eav es som ething out. In this vi ew we clea rly kno w we a re conscious , ca n initiate action , and have exp eri ences o f a distinctive ph enomenological cha racte r and feeling. Th e physicalist hy­ poth esis does not account for this. A compute r o r robot is not conscious. Although a ro bot can move , it does not act in the way p eople can act. And a robot has no sensuous exp erience. Indeed , something could have exactly the functional st ructu re o f the human b rain and n e rvous syste m without b eing conscious. Blo ck ( 1978) d esc rib es a case in which one billion people in radio communication with each othe r model a pa rticula r b rain fo r an hou r, each p erson co r responding to a pa rticula r n eu ron in the b rain. Block takes it to be absu rd to suppose that this vast collection o f p eople would have a g roup consc iousness that was phenomenologically the same as the consciousn ess o f th e p erson whose b rain and cent ral ne rvous system was b eing modeled. Nag el ( 1979) obs erv es that we might kno w eve ry thing th ere was to kno w about th e n eu ro physiological st ructu re and functioning o f the b rain and cent ral n e rvous syst em o f a bat without kno wing what th e exp eri enc e o fth e bat was like. D e fend ers o fautonomous mentalism ag ree that this leaves a mind-body p roblem , sinc e they a re unable to say ho w consciousness , fr ee will , and senso ry expe rience can be pa rt o f the wo rld desc rib ed by physics. But they d eny that this means we must stop believing in consciousness o r must id enti fy it with some aspect o fphysical o r functional st ructu re. Fo r th ey clai m, with consid erable plausibility, that it is much mo re reasonable to believe in consciousness , free will , and senso ry exp erienc e, and to b elieve that th ese a re not aspects o f n eu rophysiological functional st ructu re , than it is to b eliev e in physicalism. I am not say ing that autonomous mentalism is mo re plausibl e than physicalism . A fte r all is said an d done , I find a physicalistic functionalis m mo re plausible than autonomous ment alism. My point is that autonomous mentalism is a pe rfectly respectable philosophical position. A d efende r o f autonomous ethics might even a rgu e that natu ralism in ethics loses much o f its plausibility once autonomous mentalism is recognized as plausible . Fo r that c asts doubt on the unive rsal applicability

Gilbert Harm an

47

of the natu ralistic app roach a nd the re fo re casts dou bt on the natu ralist 's a rgu me nt that a belief that so methi ng is right ca nnot be explai ned by that thi ng 's bei ng actually right u nless that thi ng 's bei ng right co nsists 1.n so me psychological o r sociological fact . The natu ralist 's o nly a rgu me nt fo r this, it might be said, depe nds o n accepti ng the ge ne ral applica bility of natu ralis m. But it is not o bvious that this app roach is ge ne rally applica ble, si nce it is not o bviously co r rect as co mpa red with auto no mous me ntalis m . The re is at least so me plausi bility to the clai m that o ne's awa re ness of what red looks like is to be explai ned by appeal to a n expe rie nce of red ness that does not co nsist e nt irely i n so me neu rophys ­ iological eve nt . I t might be said that the nat uralist has no a rgu me nt agai nst auto no mous ethics, si nce the natu ralist ca nnot take fo r g ra nted the ge ne ral applica bility of natu ralis m.

Ethics Defe nde rs of auto no mous ethics a rgue that thei r app roach rep rese nts the o nly u ndog matic way to p roceed. They say that natu ralis m begs the questio n i n supposi ng that eve rythi ng t rue must fit i nto a scie ntific accou nt of the wo rld a nd by supposi ng that the ce nt ral questio n a bout mo ral ity is ho w, if at all, mo rality fits i nto such a scie ntific accou nt. Defe nde rs of natu ralis m reply that natu ralis m itself is the result of followi ng the method of re flective equili briu m, a nd that auto no mous ethics begs the questio n by assig ni ng a specially p rotected status to initial mo ral beliefs as co mpa red, say, with i nitial beliefs a bout the flat ness of the ea rth o r the i nflue nce of the sta rs o n hu ma n histo ry. Natu ralists say that, sta rti ng with ou r i nitial beliefs, we a re led to develop a sc ie nt ific co nceptio n of the wo rld as a n accou nt of eve rythi ng the re is. I n do ing so, we also acqui re beliefs a bout how we lea rn a bout the wo rld a nd a bout ho w e r ro rs ca n a rise i n ou r thi nki ng. We co me to see ho w supe rstitio n a rises. We begi n to wo r ry a bout ou r mo ral vie ws: A re they me re supe rstitio ns? We note ce rtai n so rts of disag ree me nt i n mo rality a nd ext re me diffe re nces i n mo ral custo ms. We o bse rve that so me people a re not much influe nced by what we co nside r i mpo rta nt mo ral co nsid ­ e rat io ns. All this leads us to raise as a ce nt ral questio n a bout mo rality ho w mo rality fits i n with ou r scie ntific co nceptio n of the wo rld. Natu ralis m is no me re p re jud ice i n favo r o r sc ie nce ; it is a n i nevita ble co nseque nce of i ntellige nt thought. This, at least, is what a defe nde r of natu ralis m will say . A defe nde r of auto no mous ethics will reply that mo ral disag ree me nts, diffe re nces in custo m, a nd the behavio r of c ri mi nals p rove nothi ng. All these thi ngs a re co mpati ble with mo ral a bsolutis m .

48

Is There a Single True Morality ?

The naturalist retorts that any view can be ma de compatible with the evi dence ; astrology, for exa mple, is per fectly co mpatible with the evi dence . Th e issue is not what is co mpatible with the evi dence, but what best accounts for it. The naturalist argues that relativis m accounts for the evi denc e bett er than absolutis m does, since relativis m is able to say how reasons an d values are part o f the worl d science describes, whereas absolutis m is not able to do that . The defen der o f autono mous ethics replies that such an argu ment is no better than the correspon ding argu ment for behav ioris m. Behavioris m is able to say how m ental states (as it conc eives the m) are part o f the worl d physics describes an d autono mous mentalis m is not able to say how mental states (as it conceives the m) are part o f the wor ld physics describes ; but one shoul d not for th is reason alone aban don one 's initial view that one is consc ious, makes decisions, has feelings, an d so on, wh er e this is not just b eing dispose d to act in various ways (since so mething coul d have the dispositions without being conscious an d coul d be conscious without having the dispositions) . Si milarly, one shoul d not accept the naturalistic argu ment an d give up one 's belie f in absolute values an d universal moral reasons . I see no knock down argu ment for either side. A question o f ju dg ment is involve d, " Which view is more plausible, all things consi dere d?" To me, the relativistic naturalist position see ms more plausible . Others fin d the absolutist position o f autono mous ethics more plaus ible. I have not trie d to show that one si d e is correct. I have trie d to bring out the central issue.

cha p ter two

Moral Explanations Nicholas L. Sturgeon There is o ne arg u me nt for moral skept ic is m that I respect eve n tho ugh I re ma in u nco nv inced. It has so met imes bee n called the arg u me nt fro m moral d ivers ity or relat iv ity, b ut that is so mewhat mislead ing, for the proble m ar ises not fro m the d ivers ity of moral v iews, b ut fro m the appare nt d iffic ulty of settling moral d isagree ments, or eve n of k now ing what wo uld be re quired to settle the m, a d iffic ulty tho ught to be not iceably greater tha n a ny fo und in settl ing d isagree me nts that ar ise in, for exa mple, the sc ie nces. Th is prov ides a n arg ume nt for moral skept ic is m beca use o ne obv io usly poss ible expla nat io n for o ur d iffic ulty in settl ing moral d isagree me nts is that they are really unsettleable, that there is no way of j ust ify ing o ne rather tha n a nother co mpet ing v iew o n these iss ues ; a nd a poss ible further expla nat io n for the unsettleab il ity of moral d isagree me nts, in t ur n, is moral nih il is m, the v iew that o n these iss ues there j ust is no fact of the matter, that the imposs ib il ity of d iscover ing a nd establ ish ing moral tr uths is d ue to there not be ing a ny. I a m, as I say, u nco nv inced: partly beca use I th ink th is arg u me nt exaggerates the d iffic ulty we act ually fi nd in settl ing moral d isagree me nts, partly beca use there are alter nat ive expla nat io ns to be co ns idered for the d iffic ulty we do fi nd . For exa mple, it certa inly matters to what exte nt moral d isagree me nts depe nd o n d isagree me nts abo ut other quest io ns wh ich, however d isp uted they may be, are nevertheless regarded as hav ing object ive a nswers: quest io ns s uch as wh ich, if a ny, rel ig io n is tr ue, wh ich acco unt of h u ma n psychology, wh ich theory of h uma n soc iety. A nd it also matters to what exte nt co ns iderat io n of moral quest io ns is in pract ice skewed by d istort ing factors s uch as perso nal interest a nd soc ial ideology. These are large iss ues. Altho ugh it is poss ible to say so me useful th ings to p ut the m in perspect ive, 1 it appears imposs ible to settle the m quickly or in a ny a pr ior i way. Co ns iderat io n of the m is l ikely to have to be p iece meal a nd, in the short r u n at least, fr ustrat ingly indec is ive.

50

Moral Explanations

These la rge issues are not my topic here . But I mention them, and the difficulty of settling them, to show why it is natural that moral skeptics have hoped to fin d some quicker way of establishi ng their thesis. I doubt that a ny exist, but some have of course bee n proposed. Veri­ ficationist attacks on ethics should no doubt be seen in this light, a nd J. L. Mackie 's recent "argument from queerness" is a clear i nsta nce (Mac ki e, 1977, pp . 3 8-42). The quicker argume nt on which I shall concentrate, however, is neither of these, but i nstead a n argument by Gilbert Harman designed to bring out the "basic problem" about morality, which in his view is "its apparent immu nity f rom observational testi ng" and "the seeming irrelevance o f observational evidence" ( Harma n, 1977, pp.vii, viii. Parenthetical pa ge references are to this work). The argume nt is that reference to moral facts appears unnecessary for the explanation of our moral observations and belief s. Harman 's vi ew, I should say at o nce, is not i n the end a skeptical one, and he does not view the argument I shall discuss as a decisive defense of moral skepticism or moral nihilism. Someo ne else might easily so regard it, however. For Harma n himself regards it as creating a strong prima facie case for skepticism a nd nihilism, stro ng enough to justify calling it "the problem with ethics." 2 A nd he believes it shows that the only recourse for someo ne who wishes to avoid moral skepticism is to fi nd defensible reductive de finitions for ethical terms; so skepticism would be the obvious conclusion for a nyone to draw who doubted the possibility of such de finitions. I believe, however, that Harma n is mistake n on both c ounts. I shall show that his argume nt for skepticism either rests on claims that most people would find quite implausible (and so cannot be what co nstitutes, for them, the problem with ethics); or else it becomes just the application to ethics of a familiar general skeptical strategy, one which, if it works for ethics, will work equally well for u nobser vab le theoretical entities, or for other minds, or for an external world (and so, again, can har dly be what constitutes the disti nctive problem with ethics). I n the course of my argument, moreover, I shall suggest that one (an in any case be a moral realist , and indeed an ethical naturalist, without believing that we are now or ever will be in possession of reducti ve naturalistic de finitions for ethical terms.

I. THE PROBLEM WITH ETHICS Moral theories are often tested in thought experiments, against imagined examples ; and , as Harman notes, trained researchers often test scienti fic theori es in the same way. Th e problem, though, is that scienti fic theories can also be test ed agai 11st the world , by observations or real experiments;

Nicholas L. Sturgeon

51

and, Harman asks, "can moral principles be tested in the same way, out in the world?" (p. 4) This would not be a very interesting or impressive challenge, o f course, i f it were merely a resurrection o f standard verificationist worries about whether moral assertions and theories have any testable empirical implications, implications suita ble in some relatively austere "o bserva­ tional" voca bulary . One pro blem with that form o f the challenge, as Harman points out, is t hat there are no "pure" o bservations, and in consequence no purely o bservational voca bularly either. But there is also a deeper pro blem t hat Harman does not mention, one that remains even i f we shelve worries a bout "pure" o bservations and, at least for the sake o f argument, grant the verificationist his o bser vational language, pretty much as it was usually conceived : that is, as lacking at the very least any o bviously theoretical terminolog y from any recognized science, and o f course as lacking any moral terminology. For then the difficulty is that moral principles fare just as well (or just as badly) against the verificationist challenge as do typical scientific principles. For it is by now a familiar point a bout scientific principles-principles such as Newton 's law o f universal gravitation or Darwin 's theory o f evolution ­ that they are entirely devoid o f empirical implications when considered in isolation. 3 We do o f course base o bservational predictions on such theories and so test them against experience, but that is because we do not consider them in isolation. For we can derive t hese predictions only by relying at the same time on a large background o f additional assumptions, many o f which are equally theoretical and equally incapa ble o f being tested in isolation. A less familiar point, because less o ften spelled out, is that t he relation o f moral principle� to o bservation is similar in both t hese respects. Candidate moral principles- for example, that an action is wrong just in case there is something else the agent could have done that would have produced a greater net balance o f pleas ure over pain-lack empirical implications when considered in isolation. B ut it is easy to derive empirical consequences from them, and thus to test them against experience, i f we allow ourselves, as we do in the scientific case, to rely on a background o f other assumptions o f compara ble status. Thus, i f we conjoin the act-utilitari an principle I just cited with the further view, also untesta ble in isolation, that it is always wrong deli berately to kill a human being, we can deduce from these two premises together the consequence that deli berately killing a human being always produces a lesser balance o f pleasure over pain than some availa ble alternative act ; and this claim is one any positivist would have conceded we know, in principle at least, how to test. If we found it to be false, moreover, then we would be forced by this empirical test to a bandon at least one o f the moral claims from which we derived it.

52

Moral Explanations

It might be thought a wo r risome featu re o f this e xample , howeve r, an d a fu rthe r opening fo r skepticism , that the re coul d be cont rove rs y about which mo ral p remise to abandon , an d that we have not e xplaine d how ou r empi rical test can p rovi de an answe r to this question. And this may be a p roblem . It shoul d be a familia r p ro blem , howeve r, because the Duhemian commenta ry inclu des a p recisel y co r responding point about the scientific case: that i f we a re at all cautious in cha racte ri zing what we obse rve, then the requi rement that ou r theo ries me rel y be consistent with obse rvation is an astoun dingl y weak one. The re a re always many, pe rhaps in de finitel y many , diffe rent mutuall y inconsistent ways to a djust ou r views to meet this const raint. Of cou rse , in p ractice we a re o ften con fi dent o f how to do it: If you a re a freshman chemist ry stu dent , you do not conclu de from you r failu re to o btain the p re dicte d value in an e xpe riment that it is all ove r fo r the atomic theo ry o f gases. And the decision can be equall y easy , one shoul d note , in a mo ral case. Consi de r two e xamples. F rom the su rp rising mo ral thesis that A dol f Hitle r was a mo rall y a dmi rable pe rson , togethe r with a mo dest piece o f mo ral theo ry to the effect that no mo rall y a dmi rable pe rson woul d , fo r e xample, instigate an d ove rsee the deg ra dation an d death o f millions o f pe rsons , one can de rive the testable consequence that Hitle r di d not do this. But he di d, so we must give up one o f ou r p remises; an d the choice o f which to aban don is neithe r di fficult no r cont rove rsial. Or, to take a less monumental e xample , cont rive d a roun d one o f Ha rman 's own , suppose you have been thinking you rsel f luck y enough to live in a neighbo rhoo d in which no one woul d do an ything w rong , at least not in public; an d that the mo dest piece o f theo ry you accept , this time, is that malicious c ruelty , just fo r the hell o f it , is w rong. Then, as in Ha rman 's e xample , "you roun d a co rne r an d see a g roup o f young hoo dlums pou r gasoline on a cat an d ignite it ." At this point , eithe r you r confi dence in the neighbo rhoo d o r you r p rinciple about c ruelty has got to give way. But the choice is easy , i f dispi riting , so easy as ha rdl y to requi re thought. As Ha rman says , "You do not nee d to conclude that what the y a re doing is w rong ; you do not nee d to figu re an ything out ; you can see that it is w rong" (p. 4) . But a skeptic can still won de r whethe r this p ractica l confi dence , o r this "seeing," rests in eithe r so rt o f case on an ything mo re than deepl y ing raine d conventions o f thought- respect fo r scienti fic e xpe rts , say , an d fo r ce rtain mo ral t ra ditions -as oppose d to an ything answe rable to the facts o f the matte r, an y reliable st rateg y fo r getting it right about the wo rl d. Now, Ha rman 's challenge is inte resting pa rtl y because it does not rest on these ve ri fication ist doubts about whethe r mo ral belie fs have o bse r­ vational implications , but even mo re because what it does rest on is a pa rtial answe r to the kin d o f gene ral skepticism to which, as we have

Nicholas L. Sturgeon

53

seen, reflection on the verificationist picture can lead . Many of our beliefs are justified, i n Harman's view, by their providing or hel pi ng to provide a reasonable explanation of our observing what we do. It wou ld be consistent w ith your failure, as a beginning student, to obtain the experimental result pred icted by the gas laws, that the laws are m istake n . That would even be o n e explanat ion of your failure . But a better explanat ion, in light of your inexperience and the general success experts have had in confirming and applying these laws, is that you made some m istake in running the experiment. So our scient ific beliefs can be j ustified by their explanatory role; and so too, in Harman's view, can mathematical beliefs and many commonsense beliefs about the worl d . Not so, however, moral beliefs : They appear t o have n o such explanatory role. That is "the problem with ethics." Harman spells out his version of this cont rast : You need to make assumptions about certain physical facts to explain the occurrence of the observations that support a scientific theory, but you do not seem to need to make assu mptions about any moral facts to explain the occurrence of the so-called moral observat ions I have been talki ng about . In t he moral case, it would seem that you need only make assumptions about the psychology or moral sensibility of the person making the moral observat ion . ( p. 6)

More precisely, and applied to his own example, it might be reasonable, in order to explain your j udging that the hood lums are wrong to set the cat on fire, to assume "that the children really are pouring gasoline on a cat and you are seeing them do it." But there is no obvious reason to assume anything about "moral facts," such as that it is really wrong to set the cat on fire . . . . I ndeed, an assumption about moral facts would seem to be totally irrelevant to the explanation of your making the j udgment you make . It would seem that all we need assume is that you have certain more or less well articulated moral principles that are reflected in the j udgments you make, based on your moral sensibility. (p. 7)

And Harman thinks that if we accept this conclusion, suitably generalized, then, subj ect to a possible qualificat ion I shall come to shortly, we must conclude that moral theories cannot be tested against the world as scientific theories can, and that we have no reason to believe that moral facts are part of the order of nature or that there is any moral knowledge (pp. 2 3 , 3 5 ) . M y own view i s that Harman i s quite wrong, not i n thinking that the explanatory role of our beliefs is i mportant to their j ustificat ion, but in thinking that moral beliefs play no such role. 4 I shall have to say something about the init ial plausibility of Harman's thesis as applied to

54

Moral Explanations

his o wn exampl e, but part of my reason for dissenting should be apparent from th e oth er example I just gave . We fin d it easy (and so does Harman [p. 108 ]) to conclude from the evi dence not just that Hitler was not morally admirable, but that he was morally depraved. But isn 't it plausible that Hitler's moral depravity -the fact of his really having been morally deprave d-forms part of a reasonable explanation of why we believe he was depraved? I think so, an d I shall argue concerning t his and other exampl es that moral b eliefs commonly play the explanatory role Harman denies them . B efor e I can press my case, ho wever, I need to clear up several pr eliminary points about just what Harman is claiming and just ho w his argum ent is intended to work.

II. OBSERVATION, EXPLANATION, AND REDUCTION ( 1) For ther e are several ways in which Harman 's argument invites misunderstanding. One results from his focusing at the start on the q uestion of whether there can be moral observations. 5 But this question turns out to be a side issue, in no way central to his argument that moral principles cannot be tested against the world. There are a couple of reasons for this, of whi ch the more important 6 by far is that Harman does not really require of moral facts, if belief in them is to be justified, that th ey figur e in the explanation of moral observations. It would be enough, on the one hand, if they were needed for the explanation of moral beliefs that are not in any interesting sense observations. For example, Harman thinks belief in moral facts would be vindicated if they wer e n ee de d to explain our drawing the moral conclusions we do when we reflect on hypothetical cases, but I think there is no illumination in calling thes e conclusions obse rvations . 7 It would also be enough, on th e other hand, if moral facts wer e needed for the explanation of what were clearly observations, but not moral observations . Harman thinks math ematical beliefs are justified, but he does not suggest that there are math ematical observations ; it is rather that appeal to mathematical truths helps to explain why we make the physical observations we do (p. 10). Moral beliefs would surely be justified, too, if they play ed such a role , wheth er or not there are any moral observations. So the claim is that moral facts are not nee ded to explain our having any of th e moral beliefs we do, whether or not those beliefs are observations, and ar e equally unnee ded to explain any of the observations we make, wh eth er or not those observations ar e moral. In fact, Harman's vie w appears to be that moral facts aren't n eed ed to explain anything at all: although it woul d p erhaps be qu estion -begging for him to begin with this strong a claim , since he grants that if there were any moral facts , th en ap peal to oth er moral facts, more general ones, for example,

Nicholas L. Sturgeon

55

mig ht be nee de d to explai n them (p. 8). But he is ce rtai nl y clai mi ng , at t he ve ry least , t hat mo ral facts a re n't nee de d to explai n a ny no n mo ral facts we have a ny reaso n to believe i n. . T his clai m has see me d plausi ble eve n to so me p hilosop he rs w ho wis h to defe nd t he existe nce o f mo ral facts a nd t he possi bilit y o f mo ral k nowle dge. T hus , T ho mas Nagel has re ce ntl y ret reate d to t he repl y t hat it begs the q uestion to assume that explanatory necessity is the test of reality i n this area. . . . To assume that only what has to be included in the best explanatory pictu re of the world is real, is to assume that there are no irreducibly normat ive truths . 8

But t his ret reat will ce rtai nl y make it mo re diffi cult to fit mo ral k nowle dge i nto a nyt hi ng like a causal t heo ry o f k nowle dge , w hi ch see ms plausi ble fo r ma ny ot he r cases , o r to follow Hila ry Put ma n's suggestio n t hat we "appl y a ge ne rall y causal a ccou nt o f refe re nce . . . to mo ral te rms" ( Put na m, 1975 , p. 290) . I n a dditio n, t he co ncessio n is p re matu re i n a ny case , fo r I s hall a rgue t hat mo ral facts do fit i nto ou r expla nato ry view o f t he wo rl d, a nd i n pa rti cula r i nto expla natio ns o f ma ny mo ral o bse rvatio ns a nd belie fs. (2) Ot he r possi ble misu nde rsta ndi ngs co nce rn w hat is mea nt i n aski ng w het he r re fere nce to mo ral facts is needed to explai n mo ral belie fs. O ne wa rni ng a bout t his questio n I save fo r my co mme nts o n re du ctio n below ; but a not he r, a bout w hat Ha rma n is clea rl y not aski ng , a nd a bout w hat so rt o f a nswe r I ca n atte mpt to defe nd to t he questio n he is aski ng , ca n be spelle d out fi rst. Fo r, to begi n wit h, Ha rma n's questio n is clea rl y not just w het he r t he re is an expl anatio n o f ou r mo ral belie fs t hat does not me ntio n mo ral facts . Al most su rel y t he re is. Equall y su rely , howeve r, t he re is an expla natio n o f ou r co m mo nse nse no n mo ral belie fs t hat does not me ntio n a n exte rnal wo rl d: o ne w hi ch cites o nl y ou r se nso ry expe rie nce , fo r exa mple , toget her wit h w hateve r nee ds to be sai d a bout ou r ps yc holog y to explai n w hy wit h t hat histo ry o f expe rie nce we woul d fo rm just t he beli efs we do. Ha rma n mea ns to be asking a questio n t hat will lea d to skepticis m a bout mo ral facts , but not to skepticis m a bout t he existe nce o f mate rial bo dies o r a bout well-esta blis he d s cie ntifi c t heo ries o f t he wo rl d. Ha rma n illust rates t he ki nd o f questio n he is aski ng , a nd t he ki nd o f a nswe r he is seeki ng , wit h a n exa mple fro m p hysi cs w hi ch it will be useful to keep i n mi nd . A p hysi cist sees a vapo r t rail i n a clou d cha mbe r a nd t hi nks , "T he re goes a p roto n." W hat explai ns his t hi nki ng t his? Pa rtly , o f cou rse , his ps yc hological set , w hic h la rgel y depe nds o n his belie fs a bout t he appa ratus a nd all t he t heo ry he has lea rne d ; but pa rtl y also, pe rhaps , t he hypot hesis t hat "t he re reall y was a p roto n goi ng t hroug h t he clou d c ha mbe r, causi ng t he vapo r t rail , w hic h he saw as a p roto n."

56

Moral Explanations

We will not need this latter assumption, however, "if his having made that observation could have been equally well explained by his psycho­ logical set alone, withou t the need for any assumption abou t a pro ton" (p. 6).9 So for reference to moral facts to be needed in the explanation of our beliefs and observa tions, is for this reference to be required for an explanation that is somehow better than competing explana tions. Corr espondingly, reference to moral facts will be unnecessary to an explana tion, in Harman's view, not jus t because we can find some explanation that does not appeal to them, bu t because no explanation that appeals to them is any better than some compe ting explana tion tha t does not. Now, fine discriminations among competing explanations of almost anything are likely to be difficult, controversial, and provisional. For­ tunately, however, my discussion of Harman's argument will not require any fine discrimina tions. This is because Harman's thesis, as we have seen, is not that moral explanations lose out by a small margin ; nor is i t tha t moral explanations, although sometimes initially promising, always turn out on further examination to be inferior to nonmoral ones. It is, rather, that reference to moral facts always looks, right from the start, to be "completely irrelevant" to the explanation of any of our observations and beliefs. And my argument will be that this is mistaken: that many moral explanations appear to be good explana tions, or componen ts in good explanations, that are not obviously undermined by anything else that we know. My suspicion, in fac t, is that moral fac ts are needed in the sense explained, that they will turn out to belong in our bes t overall explanatory picture of the world, even in the long run, but I shall not at tempt to establish that here. Indeed, it should be clear why I could not pretend to do so. For I have explici tly pu t to one side the issue (which I regard as in capable in any case of quick resolution) of whe ther and to what extent actual moral disagreements can be set tled sa tis factorily. Bu t I assume it would coun t as a defect in any sort of explanation to rely on claims about which ra tional agreement proved unattainable. So I concede that it could turn out, for anything I say here, that moral explanations are all defective and should be discarded. W hat I shall try to show is merely that many moral explanations look reasonable enough to be in the running ; and, more specifically, tha t no thing Harman says provides any reason for thinking they are not. This claim is surely strong enough (and controversial enough) to be worth defending. (3) It is implicit in this stateme nt of my projec t, but worth noting separately, that I take Harman to be proposing an independent skeptical argument-independent no t merely of the argument from the difficulty of settling disputed moral questions, but also of other standard arguments for moral skep ticism. Otherwise his exposition is e ntirely misleading

Nicholas L. Sturgeon

57

(and his argument, I think, not worth independent d iscussion) . For any of these more familiar skeptical arguments will of course imply that moral explanations are defect ive, on the reasonable assumption that it wou ld be a defect in any explanat ion to rely on claims as doubtful as these arguments at tempt to show all moral claims to be . But if that is why there is a problem with moral explanations, one should surely j ust cite the relevant skept ical argument, rather than this derivative d ifficulty about moral explanations, as the basic "problem with ethics," and it is that argument we should d iscuss. So I take Harman's interesting sug­ gestion to be that there is a different d ifficulty that remains even if we put other arguments for moral skepticism aside and assume, for the sake of argu ment, that there are moral facts (for' example, that what the child ren in his example are doing is really wrong) : namely, that these assumed facts still seem to play no explanatory role. This understand ing of Harman's thesis crucially affects my argumen­ tat ive strategy in a way to which I should alert the reader in advance. For it should be clear that assessment of this thesis not merely permits, but requires, that we provisionally assume the existence of moral facts . I can see no way of evaluat ing the claim that even if we assumed the existence of moral facts they would still appear explanatorily irrelevant, without assuming the existence of some, to see how they would look. So I do freely assu me this in each of the examples I d iscuss in the next section . ( I have tried to choose plausible examples, moreover, moral facts most of us would be inclined to believe in if we d id believe in moral facts, since those are the easiest to think about; but the precise examples don't mat ter, and anyone who would prefer others should feel free to subst itute his own . ) I grant, furthermore, that if Harman were right about the outcome of this thought experiment-that even after we assu med these facts they still looked i rrelevant to the explanation of our moral beliefs and of other non moral facts-then we m ight conclude with him that there were, after all, no such facts. But I claim he is wrong: O nce we have provisionally assumed the existence of moral facts, they do appear relevant, by perfectly ord inary standards, to the explanation of moral beliefs and of a good deal else besides. Does this prove that there are such facts ? Wel l of course it helps support that view, but here I carefully make no claim to have shown so much . What I show is that any remaining reservat ions about the existence of moral facts must be based on those other skeptical arguments, of which Harman's argument is independent. In short, there may still be a "problem with ethics," but it has nothing special to do with moral explanat ions . ( 4) A final preliminary point concerns a qualificat ion Harman adds himself. As I have explained his argument so far, it assumes that we could have reason to believe in moral facts only if this helped us "explain

58

Moral Explanations

why we observe what we observe" (p. 1 3) ; but, he says, this assu mption is too strong, for we can have evi dence for the truth o f so me belie fs that play no such e xplanatory role. We might, for e xa mple, co me to be a ble to explain color perception without saying that o bjects have colors, by citing certain physical an d psychological facts. But this woul d not show that there are no colors; it woul d show only that facts a bout color are " so mehow re duci ble" to these physical an d psychological facts. An d this leaves the possibility that moral facts, too, even i f they ulti mately play no e xplanatory role the mselves, might be "re duci ble to certain other facts that can help e xplain our observations" (p. 14) . So a crucial question is: What woul d justi fy a belief in re duci bility? What makes us think color facts might be re duci ble to physi cal (or physical an d psychological) facts, an d what woul d justi fy us in thinking moral facts re duci ble to e xplanatory natural facts o f so me kin d? Har man 's answer is that it is still the apparent e xplanatory role o f color facts, or o fmoral facts, that matters; an d hence that this quali fication to his argument is not so great as it mig ht see m . We know o f no precise re duct ion for facts o f either sort. We believe even so that re duction is possi ble for color facts because even when we are able to e xplain color perception without saying that objects are colore d, we will still sometimes refer to the actual colors of obj ects in explaining color perception, if only for the sake of simplicity. . . . We will conti nue to believe that objects have colors because we will continue to refer to t he actual colors of objects in the explanations that we will in practice give .

But Har man thinks that no co mparable point hol ds for moral facts. " There does not ever see m to be, even in practice, any point to e xplaining so meone's moral observations by appeal to what is actually right or wrong, just or unjust, goo d or ba d" (p . 22). Now I shall argue shortly that this is just wrong : that sober people frequently offer such e xplanations o f moral observations an d belie fs, an d that many o f these e xplanations look plausible enough on the evi dence to be worth taking seriously. 1 0 So a quick reply to Har man, strictly a dequate for my purpose, woul d be si mply to accept his concession that this by itsel f shoul d lea d us to regar d moral facts as (at worst) re duci ble to e xplanatory facts. 1 1 Concern about the nee d for, an d the role of, re ductive defi ni tions has been so central to meta-ethical discussion in this century, however, and ha s also prove d enough o f a sticking point in discussions I have ha d o f the topic o f this essay , that I shoul d say a bit more . As a philosophical naturalist, I take natural facts to be the only facts there are.1 2 I f I a m prepare d to recognize moral facts, there fore, I must take them, too, to be natural facts : But which natural facts? It is wi dely

Nicholas L. Sturgeon

59

thought that a n ethical natu ralist must a nswe r th is quest io n by p rovid i ng re ducti ve natu ral istic defi nitio ns 1 3 fo r mo ral te rms a nd ' i ndee d ' that u ntil o ne has supplie d such defi nitio ns o ne's c re de ntials as a naturalist a bout . any supposed mo ral facts must be i n dou bt. O nce such defi nitio ns a re i n ha n d , howe ve r, it see ms that mo ral e xpla natio ns shoul d be dispe nsa ble , si nce a ny such e xpla natio ns ca n the n be pa raph rased i n no nmo ral te rms ; so it is ha rd to see why a n ethical natu ralist should attach a ny i mpo rta nce to the m. Now, the re a re se ve ral p ro ble ms w ith this re aso ni ng , but the mai n o ne is that the wi dely held view o n wh ich it is based is mistake n: mistake n a bout whe re a sche me of re ducti ve natu ralistic defi nitio ns woul d be fou nd , if the re we re to be o ne , but a lso a bout_ whethe r, o n a natu ralistic view of ethics , o ne shoul d e xpect the re to be such a thi ng at all. I shall take up these po i nts i n re ve rse o rde r, a rgu i ng fi rst (a) that it is a mistake to re qui re of ethical natu ralis m that it e ve n p ro mise reducti ve defi nitio ns fo r mo ral te rms , a n d the n (b) that e ve n if such defi nitio ns a re to be fo rthco mi ng it is , at the ve ry least , no special p ro ble m fo r eth ical natu ralis m that we a re not now i n co nfi de nt possessio n of the m. (a) Natu ralis m is i n o ne clea r se nse a " reductio nist" doct ri ne of cou rse , fo r it hol ds that mo ral facts a re nothi ng but natu ral facts. What I de ny , howe ve r, is that fro m this metaphysical doct ri ne a bout what so rt of facts mo ral facts a re , a nythi ng follows a bout the possibility of reductio n i n a nothe r se nse (to which I shall he nce fo rth co nfi ne the te rm) mo re familia r fro m the philosophical lite ratu re: that is , a bout whethe r mo ral e xpla natio ns ca n be gi ve n re ducti ve defi nit io ns i n so me disti ncti ve no nmo ral voca bul a ry , i n wh ich a ny plaus ible mo ral e xpla nat io ns could the n be recast. The d ifficulty with supposi ng natu ralis m to re qui re this ca n be see n by p ressi ng the questio n of just what this disti nct ive voca bula rly is suppose d to be. It is co mmo n to say me rely that this re duci ng te rmi nology must be " factual" o r " desc ript ive" o r must desig nate natu ral p rope rties; but u nless ethical natu ralis m h as al ready bee n ruled out , this is no help , fo r what natu ral ists of cou rse co nte nd is that mo ral discou rse is itself factual a n d desc ripti ve (although it may be othe r th i ngs as well) , a n d that mo ral te rms the msel ves sta nd fo r na tu ral p rope rties. The i dea , clea rly , is suppose d to be that the test of whethe r these natu ralistic clai ms a bout mo ral discou rse a re co rrect is whethe r this discou rse is reduci ble to so me othe r; but what othe r? I co nsi de r two possi bilities. (i) Ma ny woul d ag ree that it is too rest ricti ve to u nde rsta nd ethic al natu ral is m as re qu iring that mo ral te rms be defi na ble i n the te rmi nology of fu nda me ntal physics . O ne reaso n it is too rest ricti ve is that philosophical natu ralis m might be t rue e ve n if physic alis m, the view that e ve rythi ng is physical , is not . So me fo rm of e me rge nt dual is m might be co rrect , fo r e xa mple. A diffe re nt reaso n , which I fi n d mo re i nte resti ng (because

60

Moral Explanations

I think physicalism is true), is that physicalism entails nothi ng i n a ny case about whether e ven biology or psychology, let alo ne ethics , is reducible to physics. There are a number o f reaso ns for this , but a cardinality proble m noted by Richard Boyd is su fficie nt to secure the point (Boyd , forthcomi ng a). I f there are (as there appear to be) a ny continuous physical parameters, then there are co nti nuu m ma ny physical states of the world , but there are at most cou ntably ma ny predicates i n any language, including that o f e ven ideal physics ; so there are more physical properties than there are physical expressio ns to represe nt the m . Thus , although physicalism certai nly e ntails that biological a nd psycho ­ logical properties (and ethical properties , too , i f there are a ny) are physical, nothing follows about wheth er we have a ny but biological or psychological or ethical terminology for represe nti ng these particular physical properties . (ii) O f course, not many discussio ns o f ethical naturalis m have focused on the possibility o f reducing ethics to physics ; social theory , psychology , and occasionally biology have appeared more promisi ng possibilities . But that facts might be physical whether or not all the disciplines that deal with them are reducible to physics, helps gi ve poi nt to my questio n o f why we should think that i f all ethical facts are natural (or, for that matter, social or psychological or biological) , it follows that they ca n e qually well be expressed in some other, no nmoral idio m ; a nd it also retur ns us to the question o f just what this alternati ve idio m is supposed to be. The answer to this latter questio n simply assu med i n most discussio ns o f ethical naturalism, I think, is that there are a nu mber o f discipli nes that we pretty well know to deal with a single natural world , for exa mple , physics, biology, psychology , and social theory ; that it is a matter o f no great concern whether any o f these disciplines is reducible to so me o ne o f the others or to anything else ; but that the test o f whether ethical naturalism is true is whether ethics is reducible to so me (non moral) combination o f them.1 4 But what rationale is there for holding ethics alone to this reducti ve test? Perhaps there would be one i f ethics appeared i n so me salie nt respect strikingly dissimilar to these other discipli nes: if, for exa mple , Harman were right what whereas physics , biology, a nd the rest o ffer plausible explanations o f many ob viously natural facts, includi ng facts about our beli efs and obser vations, ethics ne ver does. Perhaps ethics could then plausibly be re quired to ear n its place by so me alter nati ve route. But I shall o f course argue that Har man is wrong about this alleged dissimilarity , and I take my argume nt to pro vide part o f the defense re quired for a n aturalistic but nonreducti ve view o f ethics. (b) A naturalist, h owe ver, will certainly want (and a critic o f naturalis m will likely demand) a fuller account than this o f just where moral facts

Nicholas L. Sturgeon

61

are supposed to fit in the natural world . For all I h ave shown, moreover, this account might even provide a scheme of reduction for moral d iscourse : My argument has been not that ethical naturalism could not take this . form, but only that it need not . So where should one look for such a fu ller account or ( if it is to be had ) such a reduction? The answer is that the account will have to be derived from our best moral theory, together with our best theory of the rest of the natural world-exactly as, for example, any red uctive account of colors will have to be based on all we know about colors, includ ing our best optical theory together with other parts of physics and perhaps psychology. If hedonistic act­ u tilitarianism (and enough of its associated psychology) turns out to be t rue, for example, then we can define the good as pleasure and the absence of pain, and a r ight act ion as one that produces at least as much good as any other, and that will be where the moral facts fit. If, more plausibly, some other moral theory turns out to be correct, we will get a d ifferent account and ( if the theory takes the right form ) different reductive definitions. It would of course be a serious obj ect ion to ethical naturalism if we discovered that the only plausible moral theories had to i nvoke supernatural facts of some kind, by making right and wrong depend on the will of a deity, for example, or by implying that only persons with immortal souls could have moral obligations . We wou ld then have to choose between a naturalist ic world view and a belief in moral facts. But an ethical naturalist can point out that there are familiar moral theories that l ack implicat ions of this sort and that appear defensible i n the l ight of all we know about the natural world ; and any of them, if correct, could provide a naturalistic account of moral facts and even ( if one is to be had ) a naturalistic reduction of moral d iscourse . Many philosophers will balk at this confident talk of our d iscovering some moral theory to be correct. But their objection is j ust the familiar one whose importance I acknowledged at the outset, before putt ing it to one side : For I grant that the d ifficulty we experience in settling moral issues, including issues in moral theory, is a problem (although perhaps not an insuperable one) for any version of moral realism . All I contend here is that there is not, in addition to this acknowledged d ifficulty, any special fu rther ( or prior) problem of find ing reductive definitions for moral terms or of figuring out where moral facts fit i n the natural worl d . O u r moral theory, if once w e get i t , w i l l provide whatever reduct ion is to be had and will tell us where the moral facts fit. 1 5 The suspicion that there must be more than this to the search for reductive definitions almost always rests, I believe, on the view that these definitions must be suited to a special epistemic role: for example, that they will have to be analytic or conceptual t ruths and so provide a privileged basis for the rest of our theory. But I am confident that

62

Moral Explanations

mor al re asoning, like re asoning in the sciences , is inevit abl y di alectic al and l acks a priori found ations of this sort. I am also sure th at no ethic al n atur alist need think otherwise.1 6 The relev ance of these points is this: It is true th at if we once o bt aine d correct re ductive definitions for mor al terms , mor al e xpl an ations woul d be in principle dispens able; so if ethic al n atur alism h ad to promise such definitions , it woul d also h ave to promise the elimin ability in principle of expl an ations couche d in mor al terms. But note three points. First , it should be no surprise , and should be reg ar de d as no speci al di fficulty for n atur alism even on a re ductionist conception of it , th at we are not now in possession of such definitions , an d so not now in a position to dispense with any mor al expl an ations th at seem pl ausi ble.To be confi dent of such definitions we woul d nee d to know just which mor al theor y is correct; but ethics is an are a of gre at controversy , and I am sure we do not yet know this.Second , if some mor al e xpl an ations do seem pl ausi ble , as I sh all argue , then one import ant step tow ard improving this situ ation in ethics will be to see wh at sort of theory emerges if we attempt to refine these expl an ations in the light both of empiric al evi dence an d theoretic al criticism. So it is e as y to see , ag ain even on a re ductionist underst an ding of n atur alism th at promises the elimin abilit y of mor al e xpl an ations in the long run , wh y an y n atur alist will think th at for the foresee able short run such e xpl an ations shoul d be t aken seriousl y on their own terms. The third and most import ant point , fin ally , is th at the elimin ability of mor al e xpl an ations for this re ason , if actu ally demonstr ate d, woul d of course not represent a triumph of ethic al skepticism but woul d r ather derive from its defe at. So we must ad d one further c aution , as I promise d, concerning H arm an 's thesis th at no reference to mor al facts is needed in the expl an ation of mor al beliefs. For there are , as we c an now see , two ver y different re asons one might h ave for thin king this. One ­ H arm an 's re ason , and my t arget in the rem ain der of this ess ay-is th at no mor al e xpl an ations even seem pl ausi ble , th at reference to mor al facts alw ays strikes us as "completely irrelev ant" to the e xpl an ation of mor al beliefs. This cl aim, if true , would tend to support mor al s kepticism . The other re ason -which I h ave just been considering , an d with which I also dis agree -is th at an y mor al e xpl an ations th at do seem pl ausi ble c an be p ar aphr as ed without expl an atory loss in entirel y nonmor al terms. I h ave argued th at it is a mist ake to underst an d ethic al n atur alism as promising this kind of re duction even in principle; an d I think it in any c ase absurd overconfidence to suppose th at an yone c an spell out an adequ ate re duction now. But any re ader unconvince d by m y arguments should note also th at this second re ason is no version of mor al s kepticism: For wh at anyone convinced by it must think , is th at we either are or

Nicholas L. Sturgeon

63

will

be abl e t o s ay , in enti rely n onm o ral t erm s , ex actly whic h n atu ral pro perti e s m o ral t erms refe r t o . 1 7 S o H arm an i s rig ht t o pre s ent re­ ducti onism a s an alt ern ativ e t o sk eptici sm; part o f wh at I h av e t ri ed t o . s how i s ju st t h at it i s n eit her t he only n o r t he m ost pl au si bl e suc h alt ern ativ e, and t h at n o ethic al n atu rali st ne ed be c ommitt ed t o it.

III. MORAL EXPLANATIONS Wit h th e s e prelimin ary point s asid e, I tu rn t o my argum ent s ag ain st H arm an ' s t he si s. I s h all fir st add t o my ex am ple o f Hitle r' s m o ral c h aract er s ev eral m o re in whic h it s eem s pl au s ibl e t o cit e m o ral fact s a s part o f an ex pl an ati on o f n onm o ral fact s , and in particul ar o f pe o pl e' s fo rming t he m oral opini on s t hey d o. I s h all th en argu e t h at H arm an giv es us n o pl ausi ble rea s on t o rej ect o r ign o re th e s e ex pl an ati on s; I s h all cl aim , in fact , t h at th e s am e i s t ru e fo r his own ex am pl e of t h e c hild ren igniting t he c at. I s h all c onclud e, fin ally , by att em pting t o di agn os e the sou rce o f t he di s ag reem ent bet ween H arm an and m e on t he s e i s su e s. My Hitl er ex am pl e sugg ests a w hol e rang e o f ext rem ely c omm on c as e s t h at appear n ot t o h av e occu r red t o H arm an , c a s e s in whic h we cite s om e on e' s m o ral c h aracter a s part o f an ex pl an ati on of hi s o r her d eed s , and in which t h at whol e st o ry i s t hen av ail abl e a s a pl au si bl e fu rt her ex pl an ati on o f som e on e's arriving at a c o r rect a s s e s sm ent o f t h at m o ral c h aract er. Tak e ju st on e oth er ex am ple. Be rn ard D e Vot o, in The Year of Decision: 1846, d e sc ri bes t he eff o rt s o f Am eric an emig rant s al ready in C ali fo rni a t o rescu e an ot her party o f emig rant s , th e D onn er P arty , t rapped by sn ows in t he Hig h Si erras , onc e t hei r plig ht bec am e kn own . At a m eeting in Y erba Bu en a (n ow S an F ranci sc o) , t he reli ef eff o rt s were put under t he di recti on o f a rec ent arriv al , P a ss ed Mids hi pm an Selim W ood wo rt h , d e sc ri bed by a previ ou s acqu aint anc e a s " a g reat bu sy b ody and am biti ous of t aking a c omm and am ong t he emig rants." 1 8 But W ood wo rt h n ot only fail ed t o l ead re scu e parti es int o th e m ount ain s him s el f, wh ere oth er rescu er s w ere c ounting on him (l eaving c hild ren t o be picked u p by him , fo r ex am pl e) , but h ad t o be " s h am ed , t h reat en ed , and bullied" ev en int o o rg anizing t he eff o rts o f ot her s willing t o t ak e t he risk ; he s pent tim e arranging c om fo rts fo r him s el f in c am p, preening hims el f on t h e im p o rt anc e o f hi s p o siti on; and a s a predict abl e re sult o f hi s c ow ardic e and his ex erci ses in v aingl o ry , m any died who mig ht h av e been s aved , including fou r kn own still t o be aliv e when h e tu rn ed back fo r t he l a st tim e in mid- M arc h . D e Vot o c onclude s : " P a ss ed Mid ­ shi pm an W ood wo rt h was ju st n o d amned g ood" ( 1942 , p. 442). I cite t his c as e partly bec au s e it h a s so cl early t he st ructu re o f an in ferenc e t o a rea s on abl e ex pl an ati on . On e c an t hink of c om peting ex pl an ati on s ,

64

Moral Explanations

but the evidence points against the m.It isn 't, for e xa mple, that Woo d worth was a basically dece nt person who si mply prove d too weak whe n thrust into a situation that place d heroic de man ds o n hi m. He volu nteere d, he put no serious e ffort even into tasks that re quire d no herois m, a nd it see ms clear that concern for his o wn position a nd reputatio n playe d a much larger role in his motivation tha n di d any co ncern for the people he was e xpecte d to save . If De Voto is right about this evi de nce, moreover, it see ms reaso nable that part of the e xpla nation of his believi ng that Wood worth was no da mne d good is just that Woo d worth was no da m ne d good. De Voto writes of course with more moral inte nsity (a nd with more of a flourish) than acade mic historians usually per mit the mselves, but it would be difficult to fin d a serious work of biograph y, for e xa mple, in which actio ns are not e xplaine d by appeal to moral character : so meti mes by appeal to speci fic virtues an d vices, but often enough also by appea l to a more general assess me nt. A differe nt question, a nd perhaps a more difficult one, concer ns the sort of e xa mple o n which Har ma n co nce ntrates, the e xplanation of ju dg me nts of right a nd wro ng. Here agai n Har ma n appears just to have overlooke d e xplanatio ns in ter ms of moral character : A judge 's thinking that it woul d be wro ng to se nte nce a parti cular o ffen der to the ma xi mu m prison ter m the law allo ws, for e xa mple, may be due in part to her dece ncy a nd fair min ded ness, which I take to be moral facts if any are . But do moral features of the actio n or i nstitutio n being ju dge d ever play an e xpla natory role? Here is an e xa mple i n which they appear to. An i nteresti ng historical questio n is why vigorous a nd reasonably wi desprea d moral oppositio n to slavery arose for the first ti me in the eighteenth an d nineteenth centuries, eve n though slavery was a very old institutio n; and why this oppositio n arose pri marily i n Britai n, France, an d in French- a nd E nglish-speaki ng North A merica, even though slavery e xisted throughout the Ne w Worl d . 1 9 There is a standard ans wer to this question. It is that chattel slavery i n British a nd French A merica, a nd then i n the U nited States, was much worse tha n previous for ms of slaver y, an d much worse than slavery i n Latin A merica . This is, I shoul d a dd, a controversial e xpla natio n. But as is ofte n the case with historical e xplanations, its propo ne nts do not clai m it is the whole stor y, a nd many of its oppone nts gra nt that there may be so me truth i n these co mparisons, a nd that they may after all for m a s mall part of a larger e xplanatio n.20 This latter concession is all I re quire for my e xa mple. E qually good for my purpose woul d be the more li mite d thesis that e xplai ns the gro wth of antislavery se nti me nt i n the U nite d States , betwee n the Revolution a nd the Civil War, i n part by sayi ng that slavery i n the Unite d States beca me a more oppressive i nstitutio n duri ng

Nicholas L. Sturgeon

65

that t i me . The appeal in these standard explanations is st raightforwardly to moral facts . What is supposed to be wrong with all these explanat ions? Harman . says that assumptions about moral facts seem " completely irrelevant" in explaining moral observations and moral beliefs (p. 7), but on its more natural read ing that claim seems pretty obviously m istaken about t hese examples. For it is natural to think that if a particular assumpt ion is completely irrelevant to the explanat ion of a certain fact, then the fact would have obtained, and we could have explained it j ust as well, even if the assumption had been false. 2 1 But I do not believe that Hitler would have done all he did if he had not be�n morally depraved, nor, on the assumption that he was not depraved, can I think of any plausible alternat ive explanat ion for his doing those things . Nor is it plausible that we would all have believed he was morally depraved even if he hadn't bee n . Granted , there is a tendency for writers who do not at tach much weight to fascism as a social movement to want to blame its evils on a single maniacal leader, so perhaps some of them would have painted H itler as a moral monster even if he had not been one. But this is only a tendency, and one for which many people know how to d iscount, so I doubt that our moral belief really is overdetermined in this way. Nor, similarly, do I believe that Wood wor th's actions were overdetermined , so that he would have done j ust as he did even if he had been a more ad mirable perso n . I suppose one could have doubts about De Voto's obj ect ivity and reliabil ity; it is obvious he dislikes Woodworth, so perhaps he would have thought him a moral loss and convinced his readers of this no matter what the man was really like . But it is more plausible that the d islike is mostly based on the same evidence that supports De Voto's moral view of him, and that very different evidence, at any rate, would have produced a d ifferent verd ict. I f so, then Woodworth's moral character is part of the explanation of De Voto's bel ief about his moral character. I t is more plausible of course that serious moral oppos ition to slavery would have emerged in Britain, F rance, and the United States even if slavery hadn't been worse i n the modern period than before, and worse in the Un ited States than in Lati n America, and that the American antislavery movement would have grown even if slavery had not become more oppressive as the n ineteenth century progressed . But that is because these moral facts are offered as at best a partial explanat ion of these developments in moral opinion . And if they really are part of the explanat ion, as seems plausible, then it is also plausible that whatever e ffect they produced was not entirely overdetermined ; that, for example, the growth of the antislavery movement in the United States would at least have been somewhat s lower if slavery had been and remained less

66

Moral Explanations

bad an institution . Here again it hardly see ms "co mpletely irreleva nt" to the explanation whether or not these moral facts obtai ned . It is more puzzling, I grant, to co nsider Har ma n's o wn exa mple i n which you see the children ign it ing a cat a nd react i m me d iately with the thought that this is wro ng. Is it true, as Har ma n cla i ms, that the assu mption that the children are really doing so methi ng wro ng is "totally irrelevant " to any reasonable expla nat io n o f your maki ng that judg me nt? Would you, for exa mple, have reacted i n just the sa me way, with the thought that the action is wrong, even if what they were do ing hadn )t been wrong, and could we ex plain your react io n e qually well o n this assu mption? Now, there is more tha n one way to u ndersta nd this counter factual question, a nd I shall retur n belo w to a rea ding o f it that might appear favorable to Har ma n's vie w. W hat I wish to poi nt out for no w is merely that there is a natural way o f tak i ng it, parallel to the way in which I have been understa nding s imilar cou nter factual quest io ns about my o wn exa mples, on wh ich the ans wer to it has to be s imply : It depe nds . For to ans wer the quest io n, I take it, we must co nsider a situation in which what the children are doi ng is not wro ng, but wh ich is other wise as much like the actual situat io n as possible, a nd the n deci de what your reaction would be i n t hat situatio n . But s ince what makes the ir action wro ng, what its wrong ness consists in, is presu mably so meth i ng like its being an act o f gratu itous cruelty (or, perhaps we shoul d a d d, o f i ntense cruelty, and to a helpless vict im), to i magi ne the m not doi ng so methi ng wrong we are going to have to i mag ine their actio n di ffere nt i n this respect . More caut iously an d more generally, if what they are actually doing is wro ng, and i f moral properties are, as ma ny writers have held, supervenient o n natural ones, 22 then i n order to i magi ne the m not doing so mething wro ng we are go ing to have to suppose the ir act io n di fferent fro m the actual one i n so me o f its natural features as well . So our quest ion beco mes: Even i f the chil dren had bee n do ing so meth i ng else, so mething just di fferent enough not to be wrong, would you have taken them even so to be do ing so methi ng wro ng? Surely there is no one ans wer to this question: It depe nds o n a lot about you, includi ng your moral vie ws a nd ho w good you are at see i ng at a glance what so me ch ildren are do ing. It probably depe nds also o n a debatable moral issue ; na mely, just how di ffere nt the ch il dre n's actio n would have to be in order not to be wro ng . ( Is u nk ind ness to a ni mals, for exa mple, also wrong?) I believe we can see ho w, i n a case in which the ans wer was clearly affir mat ive, we m ight be te mpted to agree with Har man that the wrong ness o f the act ion was no part o f the expla nat io n o f your reactio n. For suppose you are l ike this. You hate chil dren. What yo u especially hate, moreover, is the sight o fchildren e njoyi ng the mselves ; so much so that whenever you see ch ildren hav i ng fu n, you imme d iately

Nicholas L. Sturgeon

67

assu me they are up to no good. The more they see m to be enjoying the mselves, further more, the readier you are to fasten on any pretext for thinking the m engaged in real wickedness. Then it is true that even · if the children had been engaged in so me robust but innocent fun, you would have thought they were doing so mething wrong ; and Har man is perhaps right 2 3 about you that the actual wrongness of the action you see is irr elevant to your thinking it wrong. This is because your reaction is due to a feature of the action that coincides only very accidentally with the ones that make it wrong.24 But, of course, and fortunately, many people aren't like this (nor does Har man argue that they are). It isn 't t rue of the m that, in general, if the