zagzebski, linda - divine motivation theory. cambridge university press, new york (2004).pdf

P1: KcSAggragation.xml Zagzebski 0 521 82880 5 May 7, 2004 8:36

ii

This page intentionally left blank


Divine Motivation theory

Widely regarded as one of the foremost figures in contemporaryphilosophy of religion, Linda Trinkaus Zagzebski has written anew book that will be seen as a major contribution to ethical theoryand theological ethics.

At the core of the book lies a new form of virtue theory based onthe emotions. Distinct from deontological, consequentialist, andteleological virtue theories, this one has a particular theological,indeed Christian, foundation. The new theory helps to resolvephilosophical problems and puzzles of various kinds: the disputebetween cognitivism and noncognitivism in moral psychology;the claims and counterclaims of realism and antirealism in themetaphysics of value; and paradoxes of perfect goodness in nat-ural theology, including the problem of evil.

A central feature of Zagzebskis theory is the place given toexemplars of goodness. This allows the theory to assume discretebut overlapping forms in different cultures and religions.

As with Zagzebskis previous Cambridge book, Virtues of theMind, this new book will be sought out eagerly by a broad range ofprofessionals and graduate students in philosophy and religiousstudies.

Linda Trinkaus Zagzebski is Kingfisher College Chair of the Phi-losophy of Religion and Ethics and Professor of Philosophy at theUniversity of Oklahoma.

i


ii


Divine Motivationtheory

linda trinkaus zagzebskiUniversity of Oklahoma

iii

cambridge university pressCambridge, New York, Melbourne, Madrid, Cape Town, Singapore, So Paulo

Cambridge University PressThe Edinburgh Building, Cambridge cb2 2ru, UK

First published in print format

isbn-13 978-0-521-82880-2

isbn-13 978-0-521-53576-2

isbn-13 978-0-511-21170-6

Linda Trinkaus Zagzebski 2004

2004

Information on this title: www.cambridge.org/9780521828802

This publication is in copyright. Subject to statutory exception and to the provision ofrelevant collective licensing agreements, no reproduction of any part may take placewithout the written permission of Cambridge University Press.

isbn-10 0-511-21347-6

isbn-10 0-521-82880-5

isbn-10 0-521-53576-x

Cambridge University Press has no responsibility for the persistence or accuracy of urlsfor external or third-party internet websites referred to in this publication, and does notguarantee that any content on such websites is, or will remain, accurate or appropriate.

Published in the United States of America by Cambridge University Press, New York

www.cambridge.org

hardback

paperback

paperback

eBook (EBL)

eBook (EBL)

hardback


For Ken

v


I think he would need time to get adjusted before he could see thingsin the world above; at first he would see shadows most easily, thenreflections of men and other things in water, then the things them-selves. After this he would see objects in the sky and the sky itselfmore easily at night, the light of the stars and the moon more easilythan the sun and the light of the sun during the day. Of course.

Then at last he would be able to see the sun, not images of it in wateror in some alien place, but the sun itself in its own place, and be ableto contemplate it. That must be so.

Plato, Republic 516a-b

vi


Contents

Preface page xiAcknowledgments xv

Part I. Motivation-based virtue ethics

1. Constructing an ethical theory 3I. Value concepts and the metaphysics of value 3II. Three puzzles to solve 8III. Some confusions I wish to avoid 18IV. A taxonomy of ethical theories 29V. Exemplarism 40

2. Making emotion primary 51I. Starting with exemplars 51II. What an emotion is 59III. Emotion and value judgment 74IV. The intrinsic value of emotion 82V. Conclusion 95

3. Goods and virtues 96I. The good of ends and outcomes 96II. The good of pleasure 107III. The good for human persons 110IV. Virtues 118

vii


Contents

V. Defining the kinds of good 130VI. Conclusion 133

4. Acts and obligation 137I. Acts and the exemplar 137II. Obligation 145III. Defining the concepts of act evaluation 159IV. Moral judgment 166Conclusion to Part I 174

Part II. Divine Motivation theory

5. The virtues of God 187I. A brief history of the imitatio Dei 187II. The personhood of God 191III. The emotions and virtues of God 203IV. The motives of God and the Creation 213V. The metaphysical source of value 223

6. The moral importance of the Incarnation 228I. Must Christianity be an ethic of law? 228II. The Incarnation as an ethical doctrine 231III. The imitation of Christ and narrative ethics 247IV. Divine Motivation theory and Divine

Command theory compared 2587. The paradoxes of perfect goodness 271

I. Three puzzles of perfect goodness 272II. The solution of Divine Command theory 278III. The solution of Divine Motivation theory 282IV. Does God have a will? 290V. Is the ability to sin a power? 295VI. Love and freedom 298VII.Conclusion 301

8. The problem of evil 304I. The intellectual problem of evil 304II. Divine Motivation theory and theodicy 313III. Objections and replies 318IV. The problem of suffering 324Conclusion to Part II 339

viii


Contents

Part III. Ethical pluralism

9. Ideal observers, ideal agents, and moral diversity 347I. The problem of moral disagreement 347II. Ideal observers 349III. Ideal agents 359IV. Rationality in the second person: Revising

the self 372V. Religion and the task of developing a common

morality 382

Bibliography 389Name index 405Subject index 408

ix


x


Preface

There are two very different sensibilities out of which moral dis-course and even entire moral theories arise. One is the idea thatmorality attracts. The other is the idea that morality compels. Theformer focuses on value, the latter on obligation. The former is op-timistic enough to think that human beings are drawn to moralityby nature and by the good and bad features of the world. Thelatter is pessimistic enough to think that only law which is tosay, force can be the source of morality. This is not a negligi-ble difference; it grounds the difference between virtue theoriesand duty theories. I have occasionally heard philosophers won-der whether there is any significant difference between the twokinds of theory and whether the difference matters. For many ofthe purposes of morality, it is useful to ignore the differences orto conceal them; the theory of this book is meant to reveal them.

The theory is a strong form of virtue theory with a theologi-cal foundation, although I will begin with a general frameworkthat can have a naturalistic form. There are many different ways inwhich God can be related to morality, but the one that has receivedthe most attention in the history of ethics is Divine Commandtheory. This is surprising, because quite apart from the famousobjections to it, Divine Command theory has rarely aspired to bea complete moral theory. At best, it gives ethics a theoretical foun-dation, but it is difficult to see how we can move from a foundation

xi


Preface

of that kind to a theory with anything but the most meager nor-mative content. However, my own reason for looking elsewherefor the foundation of ethics has nothing to do with these short-comings. Divine Command theory is an ethic of law, of obligation.It is an ethic based on compulsion, not on the perception of value.I want to investigate a theological virtue ethics in which moralityis driven by the attractiveness of the good.

The theories in which morality attracts are usually forms oftheories that we have inherited from ancient Greece. Theories inwhich morality compels make up most of modern ethics. NaturalLaw theory was a brilliant attempt to have it both ways. That ishow I read the ethics of Aquinas. Aquinas claimed that moral-ity is law, but it is a law based on human nature, a nature thatcontains an innate propensity toward the good. When sufficientlydeveloped, however, it turned out that there was nothing espe-cially natural about natural law. That is not to say that NaturalLaw theory should be dismissed. In fact, I believe that it is oneof the most viable of all the kinds of ethical theory, and one ofits most appealing aspects in its Thomistic form is that, unlikeDivine Command theory, it gives a theological foundation to afull ethical theory. Nonetheless, it is not the kind of theory I willpursue, because it also is fundamentally an ethic of obligation,and my purpose is to see how far we can get with an ethic of thegood.

Another brilliant attempt to have it both ways is the Kantianidea of morality as autonomy: Morality is a law I give to myself.Presumably, if I give a law to myself, that mitigates the sense inwhich morality is force. This is not an ethic of attraction, but atleast it does not subject us to the tyranny of external law. Kantsethics is surely one of the most important ethical theories underdiscussion today, but I have chosen not to pursue a version of thistheory, and again, my reason is that it is essentially an ethic ofobligation. In my view, Kantian ethics does not give a sufficientlyprominent place to the attractiveness of the good.

In Plato, the good attracts, and one of the most potent and en-during Platonic images of the good is the sun. I find it revealingthat the sun not only attracts but also diffuses. The Earth does nothave to move toward the sun in order to reflect its light. Platos

xii


Preface

analogy suggests that the good is not exclusively something exter-nal to us that draws us toward it. The good may also be somethingwe receive. It might even be in ourselves, diffusing itself throughour acts. We too may have the capacity to bestow good upon theworld. Hopefully, the obvious impertinence of this thought is mit-igated by the further thought that we are not the original bestowerof value.

My purpose in this book is to present an ethical theory drivenby the concept of the good. In what follows, I propose an idea forthe consideration of the community of philosophers. The full the-ory as it appears here is proposed to the community of Christianphilosophers, but I have given a lot of attention to its naturalisticversion, which is a form of nonteleological virtue ethics. I hope toengage Kantian, consequentialist, and neo-Aristotelian virtue the-orists in a discussion of the advantages and disadvantages of thiskind of ethics, whether or not they are committed to any religiousbeliefs.

Some ethicists find theory of any kind problematic. It must beadmitted that theory always sacrifices something richness ofdetail, a certain kind of subtlety, and sometimes clarity. But I amconvinced that there is a deep human need to theorize. What iswonderful about theory is that it compensates for the finitude ofthe human mind. It is our human misfortune that we are not ca-pable of conscious awareness of very much at one time, and sowe try to streamline conscious reality so that as much as possiblecan be packed into a single act of understanding. Theory extendsthe scope of our understanding. At its best, it gives us the maxi-mum possible scope consistent with maximum clarity. But theoryinvolves abstraction from particulars, and the act of abstractionnecessarily leaves something behind. What is left behind mightbe important, and if so, that ought also to be the object of inves-tigation. A good theory should be compatible with work on theparticulars of the subject matter, and it should give that work asimple and natural structure. We want it to clarify and resolve themuddles we get into when we focus on one particular at a time,and above all, a theory should strengthen our grasp of the whole.What we should scrupulously strive to avoid is a way of theoriz-ing that leaves behind what is most important. Bernard Williams

xiii


Preface

claims that that is what has happened with the most abstract ethi-cal concepts right, good, and duty, what he calls the thin ethicalconcepts. But even if moral practice could survive the eliminationof these concepts, how could we understand such a morality? Ab-straction gives us scope, and thinness is the price of scope and acertain kind of understanding. Some philosophers would gladlysacrifice scope for something else that they value richness, thick-ness, imaginative power. I suspect that this difference in values canbe largely explained by differences in philosophical temperament.

What is depth? Do we understand moral reality more deeplywhen we concentrate on what theory leaves behind and try toreveal that part of reality that resides in the most subtle detail?Or does theory have its own kind of depth? The theory of thisbook is designed to honor both theory and narrative detail byexplaining the importance of narrative in the structure of the the-ory, but I will not tell many stories. I will attempt to situate theidea of the paradigmatically good person within the metaphysicsof morals. The theory proposes that the most basic moral featureof the universe is the way such a person perceives the world, akind of perception that is affective and that is expressed in thickconcepts. An affective perception of the world is what I believeconstitutes an emotion.

What follows is an ethical theory based on the emotions of aperfect being.

Norman, OklahomaMay 28, 2003

xiv


Acknowledgments

The idea for this book originated in a response I gave to a paperby Bill Rowe on John Hicks way of handling the problem of evil,at a conference on Hicks work at Claremont McKenna College,April 78, 1989. At that time, I was writing a book on the dilemmaof freedom and foreknowledge, and I decided to leave for anothertime the task of investigating whether the idea had any merit.Since then, I have worked on parts of the theory in a number ofpapers and have been influenced by the work of many philoso-phers, sometimes without realizing it until I read their work a sec-ond time. I know that this is true of the influence of Bob Adamsand Bill Alston. It is no doubt true of many other people whom Icannot name. I especially want to thank Tom Carson, whom I havenever met. He read and commented on the entire manuscript, andduring the course of e-mail correspondence we discovered manymutual interests and ideas.

I wrote a first draft of the book during the academic year 199899, while I was on leave from Loyola Marymount University as Se-nior Fellow in the Lilly Fellows Program at Valparaiso University.I am grateful to Loyola Marymount and to the Lilly Foundationfor support during that year. In the fall of 1999, I began teach-ing at the University of Oklahoma and taught a graduate seminaron the manuscript. I thank the students in that seminar for theirmany probing questions and objections. On April 1920, 2001, the

xv


Acknowledgments

Philosophy Department and the College of Arts and Sciences atOU hosted a conference on my work to inaugurate my chair. BillAlston gave a paper at that conference critiquing Divine Motiva-tion theory, which was invaluable in getting me to introduce thetheory in a different way.

I spent May 2000 as a guest of the Theology Departmentat the University of Uppsala. Some of the work of this projectwas presented there, and I thank my hosts, Eberhard Hermannand Mikael Stenmark, and the Swedish government for bringingme there for a month of interesting conversation in a beautifulplace.

I also thank John Hare and Calvin Colleges Center for ChristianScholarship for bringing me to Calvin for a week of philosophicalconversation in August of 2000. John Hare, Bob Roberts, and Iread large portions of each others related book manuscripts andenjoyed a lively week of discussion and debate.

In the spring of 2002, I had teaching leave from OU to workon the manuscript, and I thank the Philosophy Department for itssupport. I also thank the National Endowment for the Humanitiesfor a Summer Stipend for the summer of 2002, during which timeI wrote a completely new draft of the first eight chapters.

Baylor University graciously invited me to spend several daysthere in talks and private discussions with members of the Philos-ophy Department in March 2003. I am especially grateful to BobRoberts, Jay Wood, and Steve Evans for their comments on severalchapters of the book.

Finally, I want to thank my research assistants, Kyle Johnsonand Tony Flood, who provided me valuable bibliographic assis-tance and help with preparing the final version of the manuscript.

Ideas from a number of my previously published papers areused in various chapters of the book:

Portions of Emotion and Moral Judgment, Philosophy and Phe-nomenological 66:1 (January 2003), pp. 104124, appear in Chap-ters 2 and 3.

Ideas from The Virtues of God and the Foundations of Ethics,Faith and Philosophy 15:4 (October 1998), pp. 538552, appear inChapter 5.

xvi


Acknowledgments

Portions of The Incarnation and Virtue Ethics, in The Incar-nation, edited by Daniel J. Kendall and Gerald O Collins (Oxford:Oxford University Press, 2000), appear in Chapter 6.

Some of the ideas in Perfect Goodness and Divine MotivationTheory, Midwest Studies in Philosophy 21 (1997) (Philosophy ofReligion), pp. 296310, are used in Chapter 7.

Fragments from An Agent-Based Approach to the Problem ofEvil, International Journal for Philosophy of Religion 39 (June 1996),pp. 127139, appear in Chapter 8.

Some of the principles of rationality proposed in Chapter 9come from Religious Diversity and Social Responsibility, Logos4:1 (Winter 2001), pp. 135155; published in Spanish in Comprenderla religion, edited by Javier Aranguren, Jon Borobia, and MiguelLluch, Eunsa (Pamplona: Ediciones Universidad de Navarra,2001), pp. 6985.

Some discussion of the Divine Command theory of Bob Adamsin Chapter 6 is taken from Obligation, Good Motives, and theGood (symposium paper on Robert M. Adams, Finite and InfiniteGoods), Philosophy and Phenomenological Research 64:2 (March 2002),pp. 453458.

I thank all those from whom I benefited in writing these papersas well. Many people have encouraged the development of myideas; some keep me within the bounds of critical normalcy; mostdo both. I am very grateful for all of these friendships.

xvii


xviii

P1: JzQ0521828805agg.xml Zagzebski 0 521 82880 5 April 18, 2004 11:44

Part IMotivation-based virtue ethics

This book exhibits a way to structure a virtue ethics with a theo-logical foundation. Since the foundation is an extension of virtuediscourse to the moral properties of God, the theory might becalled a divine virtue theory. In Part I, I give the framework fora distinctive kind of virtue ethics I call motivation-based. Thistype of theory makes the moral properties of persons, acts, andthe outcomes of acts derivative from a good motive, the most ba-sic component of a virtue, where what I mean by a motive is anemotion that initiates and directs action. Chapter 1 raises the cen-tral problems involved in providing an adequate metaphysics ofvalue for virtue theory and proposes the methodology of exem-plarism. Chapter 2 gives an account of emotion and its intrinsicvalue. Chapter 3 defines a good end, a good outcome, the good fora human being, and virtue in terms of a good emotion. Chapter 4shows how the moral properties of acts can be defined in termsof a good emotion. In Part II, I will propose a Christian form ofthe theory according to which the motivations of a perfect Deityare the ultimate foundation of all value. I call the enhanced theoryDivine Motivation theory.

1


2


Chapter 1Constructing an ethical theory

The virtuous person is a sort of measure and rule for humanacts.

Aristotle, Nicomachean Ethics X.5

i value concepts and the metaphysics of value

Let us begin with good and bad. One of the things I will argue inthis book is that the ways of having value are not all forms of goodand bad, but because good and bad are as close to basic as we aregoing to get, I begin with them for simplicity. One of the mostobvious but also most troublesome features of good and bad isthat they apply to things in a variety of metaphysical categories:objects of many kinds, persons and their states and traits, acts,and the outcomes of acts. We also call states of affairs good orbad apart from their status as act outcomes, and we call certainthings designated by abstract names good life, nature, knowl-edge, art, philosophy, and many others. Some of the things in thislast category belong in one of the other categories, but perhapsnot all do.

Do the items in these different categories have anything non-trivial in common? One plausible answer is that they are all relatedto persons. That answer applies to states of persons such as plea-sure or happiness, character traits, motives, intentions, acts and

3


I. Motivation-based virtue ethics

their outcomes, and states of affairs that are valuable to persons insome way, whether or not they are produced by human acts. Buteven if human persons did not exist, some of the items of valuejust mentioned would still exist and would still be valuable forexample, life and nature so the suggestion that everything goodor bad is related to persons is too limiting. But in another way, itmay not be limiting enough, since ultimately everything is proba-bly of some concern to persons. Traditional ethics has been muchmore restrictive. It focuses on the human act and that to whichan act is causally connected, either forward or backward.1 For themost part, I will follow common practice in limiting my subjectmatter in this way, although I am not convinced that there are es-pecially good reasons for doing so. My focus will be mostly on thestates of affairs to which human agents respond when they act, thepsychic states and dispositions that produce acts, acts themselves,and the outcomes of acts. Moral philosophers have generally re-garded these objects of evaluation as particularly important. Theyare also thought to be intimately related. It is hardly controversialthat a good person generally acts from good motives and formsgood intentions to do good acts and, with a bit of luck, producesgood outcomes. What is at issue is not the fact that such relationsobtain, but the order of priority in these relations.

The question of priority arises in more than one way. One isconceptual: Is there a relation of dependency among the conceptsof good person, good motive, good act, and good outcome? If so,what is the shape of that dependency? Is one of these conceptsbasic and the rest derivative from it? Notice that this is a questionnot of conceptual analysis but of theory construction. Theories donot describe so much as they create conceptual relations. The the-orist is concerned with whether a good person should be definedas a person who acts from good motives, or as one who producesgood outcomes, or as one who does good acts. Should a goodact be understood as an act done by a good person, or as an act

1 The new field of environmental ethics may indicate that contemporaryethics is moving away from a focus on human beings, but even that isunclear, because environmental ethics usually emphasizes the ways theenvironment is impacted by human acts.

4


Constructing an ethical theory

done from a good motive, or as an act that produces good states ofaffairs? Is a virtue a quality that leads to the performance of good(alternatively, right) acts, or one that leads to good outcomes, oris a virtue more basic than either acts or outcomes? Of course,these are not the only options for the relationships among theseconcepts, but they are among the simplest.

A related but distinct question is this: Is there a relationshipof metaphysical dependency among the different categories ofthings with value? Are some bearers of value or some moralproperties more basic than others? If so, which is the most ba-sic, and how do the things in other categories derive their valuefrom the more basic ones? According to consequentialism, an actgets its moral value (generally called rightness rather than good-ness) from the goodness of its outcome or the outcome of acts ofthe same type. Consequentialism may be intended as an answerto the first question and hence as a conceptual thesis, but it canalso be intended as a thesis in the metaphysics of value. If it is theformer, it is the proposal that we ought to think of the rightnessof acts as determined by the goodness of their consequences; thisway of thinking is recommended as preferable to alternatives. Ifthe thesis is the latter, it is the claim that the value of an act ac-tually arises from the value of outcomes. Similarly, the thesis ofa certain kind of Kantian ethics can be understood as proposingeither a conceptual or a metaphysical priority between the valueof an act of will and the value of the end the will aims to bringabout. If it is the latter, it is the thesis that the value of the end of anact arises from the value of a property of the will that produces it.Christine Korsgaard expresses this position when she says valueflows into the world from a rational will.2 Here, Korsgaardsthesis is one about the source of value, not about how we ought todefine the concept of a good end. It is a thesis in moral ontology.

Conceptual order may or may not be isomorphic with ontolog-ical order. It would be helpful if it were, but it is also possible thatour concepts do not map ontology. In the first part of this book,I will argue for a certain way of conceptualizing morality. I willpropose a theory in which good motives are conceptually more

2 Kants Formula of Humanity, in Korsgaard (1996a), p. 110.

5



basic than good traits, good acts, and good outcomes of acts andwill outline a metaphysical theory to accompany it. In Part II, Iwill propose a more substantial theory in theistic metaphysics ac-cording to which the motives of God are the ontological basis forthe value of everything outside of God. The two parts of the bookare detachable, but together they outline a moral theory whoseconceptual structure is mirrored in the metaphysics of value.

The realm of value is usually considered to be broader than therealm of moral value, since aesthetic value, epistemic value, thevalues of etiquette, and perhaps the values of health and happinessare nonmoral values. That is possible, but I will have very littleto say about the distinction between moral and nonmoral valuein this book, both because I have never heard of a way of makingthe distinction that I found plausible and because I do not thinkthe distinction is very important. Since the theory of this bookis structured around the traditional units of moral theory acts,motives, ends, and outcomes the values discussed are mainlymoral values, but I will sometimes venture beyond the traditionalcategory of the moral without comment.

It is sometimes said that what makes the territory of the moraldistinctive is a strong notion of obligation. I see no reason to thinkthat is true, but the relationship between value and obligation hasbeen an important issue in modern moral theory. The categories ofthe obligatory or required and the wrong or forbidden are distinctfrom the axiological categories of good and bad. So in addition tosorting out the relationships among the various kinds of thingsthat are good and bad, there is also the problem of specifying therelationship between the good and bad, on the one hand, and therequired and forbidden, on the other. Again, this question can beabout either conceptual or ontological priority. Value is presum-ably broader than the required or forbidden, since it is usuallythought that the latter applies only to the category of acts and in-tentions to perform acts.3 Persons and states of affairs can be goodor bad, but they cannnot be required or forbidden. An act can be

3 A notable exception is that Christians may say that we are obligated tolove. But it is rare in moral philosophy to make an emotion, or any psychicstate other than an act of will, a matter of duty or obligation.

6



good or bad, but it can also be required or forbidden, obligatoryor wrong. Presumably there is some connection between the twokinds of evaluation. There are moral philosophers who have main-tained that requirement is conceptually more basic than good andhave defined good as that which requires a response of a particularkind for example, the attitude of love.4 Others have maintainedthat good is conceptually more basic than requirement and havedefined wrong and the obligatory in terms of the attitude or be-havior of good (virtuous) persons.5 Both of these positions areconceptual, not metaphysical. Robert Adams (1999, Chapter 10)has recently argued that the good is ontologically more basic thanthe obligatory, but that the latter is not derivative from the former.Of course, there are many other options. I will propose an accountof the way in which obligation derives from value in Chapter 4.

Moral theorists who ask questions about the priority of onemoral concept over another give radically different answers, butthey all share the assumption that it is a good thing to attemptto construct a conceptual framework that simplifies our think-ing about the moral life. I will go through a series of alternativeframeworks in section IV, but as I mentioned in the Preface, somewriters doubt the wisdom of any such project on the grounds thattheory distorts morality.6 I have said that I regard theory as a goodthing. I do not deny that it distorts the subject to some extent, butin compensation, theory helps us understand more with less ef-fort. I mention this now, not to defend the project of developingconceptual frameworks, but to point out that while it can be de-bated whether conceptual moral frameworks are a good thing,the same debate does not arise about the metaphysics of morals.The questions of what value is, of where it comes from, and ofwhether value in one category arises from value in another are all

4 See Chisholm (1986), pp. 52ff.5 Rosalind Hursthouse does this in several places, most recently in her book

On Virtue Ethics. I present a similar way of defining a right act in Virtuesof the Mind, at the end of Part II. I will pursue a version of this approachin Chapter 4 of this book.

6 There is a substantial literature on anti-theory since Williams (1985), whichhas been very influential in leading some ethicists to eschew theory. Seealso the collection by Clarke and Simpson (1989).

7



important philosophical questions. Of course, we may doubt thatwe will ever get plausible answers to these questions, but that isnot the worry that the anti-theorists have about theory construc-tion. In what follows, I will present both a conceptual theory anda metaphysical theory of value. Objections to the two projects willdiffer, but my intention is to enhance the plausibility of each byits relation to the other.

ii three puzzles to solve

There are three sets of puzzles that drive the project I am describingin this book. One of my purposes is to propose a theory that solves,or at least makes it easier to solve, these three sets of puzzles. Thefirst set of puzzles is in moral psychology. The second is in themetaphysics of value. The third is in natural theology. Each ofthese puzzles has a large literature, and my purpose in this sectionis not to discuss them in any detail but rather to call attention tothem and to the way the need to resolve them constrains what isdesirable in an ethical theory.

1 A puzzle in moral psychology: cognitivism versus noncognitivism

One of the most enduring legacies of David Hume is his claim inthe Treatise of Human Nature that cognitive and affective states aredistinct and independent states. The former is representational,the latter is not (Book II, section 3, p. 415). The latter motivates, theformer does not (p. 414). The terminology for describing psychicstates has changed since Hume, but the moral commonly drawnfrom Humes arguments is essentially this: No representationalstate (perceptual or cognitive) has the most significant propertyof affective states, the capacity to motivate. An affective state mustbe added to any cognitive state in order to motivate action, andthe motivating state and the cognitive state are always separable;they are related, at best, causally.

This position immediately conflicts with the intuition thatmoral judgments are both cognitive and motivating. Moral judg-ments seem to be cognitive because they are often propositionalin form, have a truth value (and are not always false), and when

8



a person makes a moral judgment, he asserts that propositionand others may deny it. On the other hand, we typically expectmoral judgments to be motivating. A simple way to see that is toconsider our practices of moral persuasion. If we want to con-vince someone to act in a certain way for moral reasons, wedirect our efforts toward convincing her to make the relevantmoral judgment herself. If we can get her to do that, we nor-mally think that she will thereby be motivated to act on it. Ofcourse, we know that she may not be sufficiently motivated toact on it, because she may also have contrary motives, but thepoint is that we think that we have succeeded in getting her tofeel a motive to act on a moral judgment as soon as we get herto make the judgment. If the Humean view is correct, however,a moral judgment can motivate only if it is affective that is,noncognitive. The Humean view therefore compels us to choosebetween the position that a moral judgment is cognitive and theposition that it is motivating. The problem is that we expect it tobe both.

The phenomena of moral strength and weakness highlightsome of the problems with the Humean psychology. It often hap-pens that a moral agent struggles before acting when he makes amoral judgment. Sometimes he acts in accordance with his judg-ment and sometimes he does not, but the fact that he strugglesindicates that a motive to act on the judgment accompanies thejudgment. When he is morally strong, a motive sufficient for ac-tion accompanies his judgment; when he is morally weak, a mo-tive insufficient for action accompanies his judgment. Either way,we think that a motive in some degree accompanies the judg-ment. But if the making of a moral judgment is a purely cognitivestate, and if cognitive and motivating states are essentially distinct,the motive must come from something other than the judgment,something that is not an intrinsic component of it. Moral strengthand weakness therefore pose a problem for cognitivism.

It may also happen that the agent acts on a moral judgmentwithout struggle, but that case does not help the cognitivist, be-cause we tend to think that when struggle is unnecessary, the rea-son is that the moral judgment carries with it a motive sufficientlystrong to cause the agent to act without struggle. So whether or

9



not there is struggle, and whether or not the agent acts in accor-dance with her judgment, there is a strong inclination to expectmoral judgments to be motivating.

Among those who accept a Humean psychology, the noncog-nitivists are better placed than the cognitivists to explain moralstrength and weakness, since the former see moral judgment asintrinsically motivating. But noncognitivists face a related prob-lem, the problem of moral apathy.7 The morally apathetic personmakes a moral judgment while completely lacking any motiveto act on it. Given what has already been said, we would expectthis phenomenon to be rare, but it probably does exist, and it isa problem for both cognitivism and noncognitivism. Given thatthe cognitivist maintains that a moral judgment is a purely cog-nitive state, he has the problem of explaining why we find moralapathy surprising. But the noncognitivist cannot explain why itexists at all. There should be no such thing as apathy, according tononcognitivism, insofar as noncognitivism takes the motivationalforce of a moral judgment to be an essential feature of each suchjudgment.

The Humean view on the essential distinctness of cognitive andmotivating states forces us to give up something in our ordinaryways of thinking about moral judgment, yet I believe that thatview is less plausible than what it forces us to give up. Nonethe-less, the phenomena of moral strength, weakness, and apathy sug-gest that what we intuitively expect is complicated. It should turnout that a moral judgment is both cognitive and intrinsically mo-tivating in enough central cases that we can see why we find thephenomena of strength, weakness, and apathy surprising. Thesephenomena indicate that the strength of the motivational force of ajudgment varies, and that it is possible for the motive to disappearentirely. In what follows, I will aim for an account of moral judg-ment according to which there is a primary class of moral judg-ments that express states that are both cognitive and intrinsicallymotivating. I will later give an account of the thinning of moral

7 Alfred Mele (1996) calls this problem moral listlessness. See also MichaelStocker (1979).

10



judgment that permits the motivational force of a moral judgmentto be detachable from it in such a way that moral strength, weak-ness, and apathy may occur.

2 Some puzzles in the metaphysics of value

Philosophers often find evaluative properties more problematicthan descriptive properties. The reason for the worry is unclear,but perhaps we do not need reasons to find something peculiar.Peculiarity is only one of the problems, however. Even if thereis nothing especially odd about value, valuable objects, or eval-uative properties, there is something in need of explanation ifsome things (properties) are evaluative and some are not. At aminimum, we want to figure out where value comes from andhow it relates to the natural or descriptive or to whatever valueis contrasted with. If the evaluative differs from the nonevalu-ative in some significant way, that may mean that we come toknow it in a different way. The issue of the way we come to makevalue judgments is therefore related to the issue of the nature ofthe objects of such judgments. Difficulties in finding a plausibleaccount of moral judgment are closely connected with difficul-ties in finding a plausible account of what those judgments areabout. The problem in moral psychology of choosing betweencognitivism and noncognitivism therefore leads us into the prob-lem in metaphysics of choosing between value realism and valueantirealism.8

Value realism is the position that value properties exist in aworld independent of the human mind. I assume that value real-ism is the default position for the same reason that realism aboutsensory properties is the default position: Objects outside the mindplainly appear to have (some) evaluative properties just as muchas they appear to have (some) nonevaluative properties. If I seesomeone taking advantage of a weaker subordinate, it may be justas apparent to me that there is badness in the act as that the act

8 Realism about value is commonly called moral realism, but the issue ismore general than the nature of moral value.

11



causes distress. The judgment That act is bad is on a par with thejudgment That act causes distress in the clarity of its meaningand the conviction of its truth. There is allegedly a puzzle aboutthe grounds for the truth value of the first judgment that does notarise about the second, but I am not much taken by this worry. I donot see anything more mysterious about the reality of value thanabout the reality of causes. What does sound mysterious is the in-tuitionism that usually accompanies value realism, for we plainlydont have an account of how we come to make value judgmentsto anything like the extent to which we have an account of howwe come to make descriptive judgments, particularly the subsetof descriptive judgments that are perceptual. For this reason, Ithink that a theory that explains our ability to detect value with-out referring to unanalyzed intuition has an advantage, and inwhat follows I will attempt to begin identifying the capacities andprocesses through which we form moral judgments.

The more serious problem for value realism is that evaluativeand nonevaluative properties appear to differ in a way that needsexplanation. Nonetheless, the distinction is not clear-cut. Considerthe following list of properties: square, salty, yellow, smooth, reliable,brutal, honorable, contemptible, pitiful, offensive, funny, exciting, nau-seating. Which properties on this list are evaluative and which aredescriptive? Most of them appear to be both, which raises the fur-ther question of how the two aspects come to be combined in somany properties if they differ in some metaphysically fundamen-tal way. But they do seem to differ, and it is commonly thoughtthat they differ in that some exist in a world independent of themind, but most do not. Furthermore, it is also commonly thoughtthat their degree of independence of the mind is related to theirdegree of perceiver variability. Allegedly, the less variation thereis among observers in the perception of a property, the more inde-pendent of human minds the property is, and hence the more realit is in some pre-theoretic sense of the real. Usually, this view isthought to have the consequence that square is more real than anyof the other properties on the list, that yellow is less so, that pitifulis even less so, and that nauseating is least of all. It is surprisingthat this conclusion is so common, since it depends upon at leasttwo disputable theses: (1) that perceptual variability is inversely

12



proportional to degree of existence in a mind-independent world(apparently there are degrees of reality), and (2) that square is lessperceptually relative than the other properties on the list.

The simplest form of value realism is the position that valueproperties are like square. This is highly implausible, but not be-cause it is obvious that value properties are not as much a part of anindependent world as is square, nor because the thesis that valueproperties are like square must reject at least one of the assump-tions just mentioned. Simple value realism is implausible becauseit attempts to maintain the reality of value properties by ignoringthe differences between evaluative and descriptive properties. Theproperties on the list differ from square and from each other in anumber of ways. Some are more observer-variable than square.More importantly, many of them are not detectable through sen-sory powers alone. To say that contemptible is like square does notexplain what value is, and more importantly, it does not explainthe fact that whatever value is, it is contrasted with something thatis not value. It seems to me, then, that the project of defending theplace of value properties in a mind-independent world shouldnot depend upon the view that there are no significant differencesbetween value properties and nonvalue properties.

For this reason, the situation is no better if we go the other wayand claim that value properties are like nauseating or exciting inthat they are not part of an independent world. This is often as-sociated with the further position that value properties express orproject properties of the observer. Again, this is highly implau-sible, not because it is obvious that these properties are part ofan independent world, but because this position does not explainthe fact that the properties on the list are not all the same. Theyare not all detectable through the same faculties, and we need anexplanation for this difference. This seems to me to be a more se-rious problem than the objection commonly given to antirealismabout value, namely, that it makes value trivial. If value propertiesare nothing but properties expressing a response in or an attitudeof the observer, then, the objection goes, there is no more reasonto be interested in them than there is to be interested in what isnauseating or exciting. My interest does not extend beyond whatis nauseating, exciting, good, or bad to me. I find this objection

13



unconvincing. It seems to me that the position that value prop-erties are projections or expressions of observer responses in factguarantees that these properties are of interest to us, the observers.Of course, it follows on the antirealist position that value prop-erties are not of interest from Sidgwicks point of view of theuniverse, since they are not in a mind-independent world. But itis not obvious that we should care about that more than we careabout what is nauseating, exciting, boring (etc.) to ourselves. I amnot denying that an argument can be given that we should careabout that, but it does take an argument. And whether or not onecan be satisfactorily given, the fact that the properties on the listare not all on a par remains a puzzle in need of explanation.

Many moral theorists aim at a position somewhere betweenrealism and antirealism. This seems sensible if we take the con-servative approach of accepting both the thesis that degree of re-ality in an independent world is inversely proportional to degreeof perceiver variation, and the view that the evaluative proper-ties on the list are somewhere between square and nauseating intheir degree of perceiver variation. I have already said that I findboth assumptions questionable, but what makes this task partic-ularly daunting is that it is very hard to see how there can beany such position. The reason is that realism is usually associatedwith cognitivism, and antirealism with noncognitivism. Granted,there is no necessary connection between the metaphysical thesisand the thesis in moral psychology, but suppose that we accept theHumean position that cognitive and affective states are necessarilydistinct, and suppose also that we assume that the objects of cog-nitive states are necessarily distinct from the objects of affectivestates, if the latter have objects at all. Suppose also that accord-ing to value realism, moral properties are the objects of cognitivestates, and that according to value antirealism, moral propertiesare the objects of, or are constructed out of, affective states. It fol-lows that we have to choose between realism and antirealism forthe same reason that we have to choose between cognitivism andnoncognitivism.

This argument also has disputable assumptions, but one way toavoid the conclusion is to begin with the desideratum I identifiedfrom the first puzzle. If I am right that there are moral judgments

14



expressing states that combine the cognitive and the affective ina way that is not simply causal, the objects of such states mayalso differ from the objects of purely cognitive or purely affectivestates. If so, this would give us hope of getting a theory that isgenuinely distinct from both realism and antirealism. One of mypurposes in this book is to propose a way to think about valueand the detection of value that leaves intact the pre-theoretic intu-ition that evaluative properties are properties in an independentworld, but that also explains the difference between descriptiveand evaluative properties.

3 Some puzzles in natural theology

The first two sets of puzzles are problems with the property ofgoodness and related properties. The third set of puzzles are prob-lems with the property of perfect goodness. The idea of perfectgoodness has a long history in Christian philosophy, one withstrong Platonic roots. Usually, but controversially, perfect good-ness is thought to entail the maximal degree of goodness. In ad-dition, perfect goodness has traditionally been thought to entailimpeccability, the property of being unable to do anything bad.But impeccability appears to conflict with the attributes of om-nipotence and freedom. If God is impeccable, there are things hecannot do, namely, acts that are evil or that express evil traits.But for the same reason that perfect goodness is thought to en-tail the maximal degree of goodness, omnipotence is thoughtto entail the maximal degree of power. There are many differ-ent accounts of what maximal power consists in, but it has oftenbeen understood as something close to the power to do anythingpossible.9 But since doing evil is a possible thing to do, if a per-fectly good being lacks the power to do evil, such a being lacksthe power to do something possible, and hence is not omnipo-tent. This puzzle was brought into the contemporary literature

9 This assumption has been challenged by many writers on omnipotence,but it is important to see that it is an assumption that is given up onlybecause of logical puzzles. The starting point is the assumption that om-nipotence entails the ability to do anything possible.

15



by Nelson Pike (1969), but it was discussed in the Middle Ages,and Aquinass way out is well known (ST I, a. 3, q. 25, obj. 2 andreply).

The reasoning behind the alleged incompatibility of perfectgoodness and omnipotence leads to a second puzzle. Under theassumption that a perfectly good being is incapable of doing evilor of willing anything but good, the will of such a being does notappear to be free in any morally significant sense. On a standardinterpretation of the conditions for moral praise and blame in thehuman case, persons are morally praised because they choose thegood when they could have chosen evil, and they are morallyblamed because they choose evil when they could have chosengood. Of course, the understanding of moral praise and blame asconditioned upon the ability to do otherwise is a modern idea,and the idea that the ability to do otherwise is morally meaning-less unless it includes the ability to choose something with thecontrary value is open to dispute, but both of these assumptionsare ones that many philosophers accept. But if perfect goodnessinvolves the inability to choose evil, a perfectly good being is notfree in the morally significant sense. Further, it follows that a per-fectly good being cannot be praised in the moral sense of praiseand hence cannot be good in the moral sense of good. This leadsto a third problem. If the concept of perfect goodness is meant toinclude moral goodness, and yet the concept of perfect goodnessis inconsistent with the concept of moral goodness, as allegedlydemonstrated by the foregoing argument, it apparently followsthat the concept of perfect goodness is self-inconsistent.

An even harder problem for the attribute of perfect goodness isthe apparent incompatibility between perfect goodness and om-nipotence, on the one hand, and the existence of evil, on the other.Not only is the problem of evil the single most difficult prob-lem in natural theology, it also poses a serious challenge to thereligious belief of ordinary people. The logical form of the prob-lem is the putative conceptual inconsistency among the followingpropositions:

(1) A perfectly good being would be motivated to eliminate allevil.

16



(2) An omnipotent being would be able to eliminate all evil.(3) There is a being who is both perfectly good and omnipotent.(4) Evil exists.

It is now widely held that these propositions are not logically in-consistent, and in recent years greater attention has been focusedon the problem that these propositions seem to be jointly improb-able. They therefore pose a problem for the rationality or justifica-tion of religious belief even if they are not logically inconsistent.10

The apparent inconsistency or joint improbability of (1)(4)needs to be resolved. There have been many attempts to showthat (1)(4) are not jointly improbable or that it can be rationalto believe them. I find some of these arguments plausible, and Iwould not find it surprising if there is more than one way to showthe rationality of a given set of beliefs, even a set as apparentlythreatening as (1)(4). But as I see it, the problem of evil is seriousenough that the more central to the theory of value a given solu-tion is, the better. It is important that the rationality of believing(1)(4) not be an ad hoc solution invented to fix the problem, butrather that it follow naturally from the theory itself. I will aim foran approach of that kind.

The same point applies to the problem that perfect goodnessappears to be incompatible with omnipotence and divine freedom,and that the concept of perfect moral goodness appears to be self-inconsistent. If the metaphysics of value in conjunction with anaccount of the divine attributes generates a puzzle that can besolved by amending something either in value theory or in naturaltheology, that may be acceptable; but it would be preferable if,given the metaphysics of value and natural theology, the problemdid not arise. I will aim for a theory on the nature and originof value from which the puzzles do not arise, or do not arise intheir most threatening form. This will be the task for Chapters 7and 8.

10 This has been recognized from the beginning of the contemporary dis-cussion stemming from J. L. Mackies famous paper (1955). More recentexamples appear in Howard-Snyder (1996); see especially the paper byRichard Gale in that volume.

17



iii some confusions i wish to avoid

1 States of affairs versus outcomes

I want to make it clear at the outset that I do not think of an outcomeas the same as a state of affairs. The outcome of an act is a state ofaffairs, but there are states of affairs that are not the outcomes ofacts for example, a meteor crashing into the Earth. Of course, no-body is likely to confuse an astronomical event with the outcomeof an act, but the confusion is just as mistaken when a state of af-fairs is also the outcome of an act. The same state of affairs can beevaluated either as an act outcome or independently of its statusas an outcome, that is, as a state of affairs simpliciter. Consider thestate of affairs of some human or animal feeling pain. I would agreewith most others that that is a bad state of affairs. It is bad no mat-ter where it comes from, and we know that because we say it is badwithout first inquiring into its origin. But suppose that the pain isthe outcome of an act. Suppose, in fact, that it is intentionally in-flicted by a human agent. We can evaluate that state of affairs as theoutcome of an act, and as an outcome it is bad in a different way.It has another sort of badness than the badness of the mere stateof affairs of someones feeling pain. It does not actually matter formy point that it is a different sort of badness, although I think thatit is. What matters is that it is an additional badness. The evalua-tion comes out differently when the state of affairs is evaluated asan act outcome than when it is not. In both cases, it comes out bad,but as an outcome it has an additional badness that it would nothave if the pain were accidental. The analogous point applies topleasure. Pleasure intentionally produced is better than pleasureaccidentally produced, even when the state of affairs is otherwisethe same. There is one state of affairs of Eves feeling pleasureat a particular time, but that state of affairs can be evaluated asthe outcome of Adams act or just as a state of affairs simpliciter.11

11 The analogous point applies to true belief. I have argued elsewhere thatin order to understand what makes knowledge better than true belief, weneed to look at the source of true belief in human agency. See Zagzebski(2001a, 2003a)

18



This point will be important when we get to my account ofhuman acts in the rest of Part 1. I will argue that typically an act isa response to the affective perception of some preexisting state ofaffairs, and that the act produces another state of affairs that is itsoutcome. The evaluation of the outcome of the act is different inkind from the evaluation of the state of affairs to which the agentis responding. The former is evaluated in relation to the prior actthat produced it; the latter is not.

2 Intrinsic good versus the good of ends

Christine Korsgaard (1983) has reminded us of the distinction be-tween intrinsic good and the good of ends in a very interestingpaper, Two Distinctions In Goodness. Later, I will argue that theconcept of intrinsic good is ambiguous, and the sense Korsgaarduses is not the one I will use. But the point I want to make initiallyis just that an intrinsic good should not be confused with a goodend, and that this is the case whether we mean end in the senseof a natural telos, or end in the sense of a conscious aim. Thedifference between intrinsic and extrinsic goods is a difference inthe location of value. An intrinsic good gets its goodness fromitself; an extrinsic good derives its goodness from something else.By contrast, the difference between goods as ends and goods asmeans is a difference in the way we value things. It is a distinctionin goods considered as objects of desire or choice. Something canbe valued as an end in itself that is, not as a means even thoughit is not an intrinsic good. For example, Korsgaard says that hap-piness is such a good according to Kant. That is because happinessis sought as an end, not as a means to something else, even thoughthe value of happiness is conditional, and Korsgaard infers fromthat that it is extrinsic. I will argue later that Korsgaard is mistakenin thinking that something is an extrinsic good because its good-ness is conditional, but she is right that happiness can be good asan end and not as a means and yet be good extrinsically as long asit derives its goodness from something else for example, a goodwill.

So something can be good as an end even though it is not in-trinsically good. I also maintain that something can be an intrinsic

19



good even though it is not good as an end, although Korsgaarddoes not discuss that possibility.12 For one thing, there are intrin-sic goods that ought not to be ends of action. Pleasure for oneselfis often in this category, as well as many good emotions. (Think,for example, how odd it is in most circumstances to aim at feelinglove.) There are also things that are both intrinsically good andgood as ends, but even then we should distinguish the two waysin which they are good. God is said to be a final end, not desirableas a means to something else, but to say that God is intrinsicallygood is not to say that. It is to say that God is not made good byanything else. In addition, there may be intrinsic goods that arenot the sort of thing that can coherently be considered somethingwe choose or at which we aim. The beauty of nature may be anintrinsic aesthetic good, but ordinarily it cannot be good as anend because it cannot be an end, since we cannot choose meansto bring it about.13 The goods of fortune (such as health, long life,and the absence of war and strife) may also be intrinsically good,even though it is very difficult, and in some cases impossible, forthem to be the objects of our aims although they can be ends inthe related sense of objects of desire, and according to Aristotlethey are components of eudaimonia, our natural end. I would alsopropose that the value of personhood is intrinsic, but it does notmake sense to say that it is desirable as an end in itself, becausepersonhood is not the kind of thing that can be either chosen ordesired, and because nothing is a means to it, nor is it a naturaltelos in the Aristotelian sense.14

12 Korsgaard has said in correspondence that she does not think that therecan be intrinsic goods that are not good as ends; in fact, she now thinksthat there are no intrinsic goods at all. Not even a good will is intrinsi-cally good. All value is conferred by human choice. But as we will see,Korsgaard means something different from what I mean by intrinsic good.See Korsgaard (1996b) for the development of her view on intrinsic goodssince the publication of Two Distinctions of Goodness.

13 Actually, we can, up to a point, bring about the beauty of nature by cul-tivating gardens. But the result is no longer nature, but nature improvedby art.

14 Kant, of course, makes the idea that persons are ends in themselves a cor-nerstone of his theory. But in making that claim, Kant is referring to theway persons ought to be treated. Since it is possible to treat persons as

20



The mistake of confusing intrinsic good and the good of ends istypically made by assimilating the former to the latter. Korsgaarddiscusses the way in which this error has arisen in recent philos-ophy, but it can also be found in Plato and Aristotle. Plato saysthat the good is what every soul pursues (Republic, Book 6), andAristotle begins the Nicomachean Ethics with the pronouncement,Every craft and every investigation, and likewise every actionand decision, seems to aim at some good; hence the good has beenwell described as that at which everything aims (emphasis added). Byclaiming that the good just is the good of ends, Aristotle and Platoignore intrinsic good. Many contemporary philosophers make thesame mistake.

The conflation of intrinsic good with the good of ends is partlyresponsible for J. L. Mackies famous complaint about the oddityof morality. Mackie says:

The Form of the Good is such that knowledge of it providesthe knower with both a direction and an overriding motive;somethings being good both tells the person who knows thisto pursue it and makes him pursue it. An objective good wouldbe sought by anyone who was acquainted with it, not becauseof any contingent fact that this person, or every person, is soconstituted that he desires this end, but just because the end hasto-be-pursuedness somehow built into it. (Mackie 1977, p. 40)

It does seem odd that a property in the world could be essentiallymagnetic to any perceiver of it, but the peculiarity arises largelybecause Mackie assumes that objective value is the value of ends,

means, to say that a person is an end in Kants sense is just to say that heought not to be treated that way. The good of ends in the Kantian sense, inthe Aristotelian sense, and in the sense of aims must all be distinguishedfrom intrinsic good, since the former have to do with the way we valuethings rather than with the source of the value. This may be less clear inthe Aristotelian sense of end than in the other two. That is because anAristotelian telos is given to us by nature and we need not be explicitlyconscious of it, although Aristotle thought that we do consciously aim atour telos, eudaimonia, but with only a dim awareness of our target. Peo-ple all agree that were aiming at eudaimonia, but disagree about what itconsists in.

21



of objects of pursuit. If value is objective, he assumes, it mustexercise its motivational force from outside the mind. But thereis also the possibility that there are things that are intrinsically and hence objectively good, but that are not ends. They may alsomotivate, but as forces inside the mind moving the agent to actrather than as magnets attracting the agent to them. I see no reasonto deny that there are states in the mind that are intrinsically good.Many philosophers already think that there are states in the mindthat have intrinsic epistemic value for example, true beliefs. Whynot think that there are other psychic states with intrinsic valueas well, even moral value? I will argue in Chapter 2 that emotionscan have intrinsic value. I do not expect readers to be convincedyet of the truth of this claim, but only of its possibility. However,its possibility will not occur to anyone who does not distinguishintrinsic good from the good of ends.

3 Motive versus aim

A related confusion is that between motive and aim. It is oftenconvenient to identify a motive by its end, since we name a motivein the context of giving an explanation for an act, and an easy wayto explain an act is to show how that act constitutes the specificmeans to some specific end of the agent. For example, the agentssearching behavior is explained by pointing out that she aimsat finding her keys, and this can be expressed by saying that hermotive is to find her keys. Or her behavior of preparing her resumeand reading the classified ads is explained by saying that her endis getting a job, and so it is said that her motive is to get a job. Butit confuses the nature of motives to identify them by their ends,for several reasons. For one thing, more than one motive can havethe same end; there is more than one motive aimed at getting ajob. More importantly, even though motives have characteristicends, the intentional structure of a motive is quite different fromthe structure of a psychic state that aims at a particular end. Wecan see this by comparing motive explanation with means-endexplanation of human action. Means-end explanation tells us whyan agent acts under the assumption that he has a certain end, buta further explanation is needed for the fact that he has the end

22



that he has. Motive explanation tells us why the agent does x,not because x leads to y and the agent aims at y, but becausex-behavior exhibits motive M (love, jealousy, compassion, etc.),we understand what it is like to be in state M, and we see thatx-behavior is part of a pattern of behavior exhibiting M. Motiveexplanation explains a much wider range of an agents behaviorthan means-end explanation. This is why insight into the motivesof characters in novels explains so much of their behavior, andwhy insight of this kind allows us to predict the future behaviorof others.

Motives are essentially pushing states, not pulling states, andthey push the agent to perform a variety of different acts in dif-ferent circumstances and to adopt a variety of different ends. Ajealous lover sometimes tries to harms his rival and sometimesharms his beloved or himself instead. We understand this becausewe understand that the intentional object of jealousy (he is jealousof her) is distinct from the end of jealous action (to harm the rival,or her, or himself). So more than one motive can have the sameend, and the same motive can have different ends.

The value of the aim of an act and the value of its motivecan differ, so the difference between motive and aim can makea moral difference. In the scene in which Elizabeth Bennett meetsMr. Wickham in Pride and Prejudice, he tells her that Mr. Darcysgood acts indeed, good aims are motivated by pride: It hasoften led him to be liberal and generous to give his moneyfreely, to display hospitality, to assist his tenants, and relieve thepoor. Family pride, and filial pride, for he is very proud of whathis father was, have done this (Chapter 16). We find out laterthat Wickham has been grossly unfair in his description of Darcy,but the person he describes is a possible one. There are variousmotives for the aim to give to the poor one might give out ofpride, or fear, or simply because their suffering sickens one andnot all of these are morally laudable. In fact, Elizabeth is horri-fied at Darcys alleged motive. In each case, the motive can bethe cause of the act without being the aim. A man can have thesame aim as the compassionate person to alleviate suffering without acting out of the motive of compassion. His motive isnot as good, and arguably his act is not as good either. But

23



that difference cannot be explained if motives are confused withaims.

The confusion of motives with aims leads to a problem in ex-plaining the source of the value of a motive. It is common and notunreasonable to think that the value of aiming at some end derivesfrom the value of the state of affairs at which the agent aims. It isgood to aim at relieving suffering, because the relief of sufferingis a good thing. It is bad to aim at punishing the innocent, becausethe punishment of the innocent is a bad thing. But if a motive isnot defined by its end, there is no temptation to think that thevalue of a motive derives from the value of an end. A given mo-tive is not identical to whatever state it is that has a given end.The value of the motive, therefore, cannot be explained by thevalue of the end but must be explained in some other way. I havealready mentioned in passing the possibility that a motivationalstate such as love is an end itself. It is possible to aim to producelove or compassion in oneself. In that case, a motive would havethe kind of value that ends have. Its value would not come froman end; it would be an end, and its value would be the value ofends.15 A moments reflection, however, reveals that this is nottypical. We do not often act in order to produce certain motivesin either ourselves or others, and in any case, the evaluation ofmotives is not limited to those cases in which we do. Ordinarily, amotive is neither identical to whatever state it is that aims at someend, nor an end itself. If intrinsic good is confused with the goodof ends, the apparent conclusion is that motives have no intrinsicvalue. In Chapter 2, I will argue that this is a mistake, but in orderto identify the mistake we need to clear up the confusion betweenmotive and aim as well as that between intrinsic good and thegood of ends.

15 The idea that virtuous motives should be the ends of the moral life is aform of perfectionism. An interesting and rarely discussed example ofa theory of this kind appears in Josiah Royces The Philosophy of Loyalty.Royce makes the motivational state of being loyal central to his theory. AsI read Royces account, he sees loyalty as both intrinsically valuable andvaluable as an end. Loyalty is the only ultimately proper object of loyalty.Good causes and bad causes are distinguishable in that the former are allforms of loyalty to loyalty. See Royce (1916), pp. 118ff.

24



To summarize, a motive is not the same as the state of aim-ing to bring about a certain end, nor is it typically an end itself.Its value does not derive from the value of a state of affairs atwhich it aims, nor is its value typically the kind of value thatends have. A motive is a psychological spring of action with anintentional structure, and the kind of value it has needs to be deter-mined. I will argue in Chapter 2 that emotions can have intrinsicvalue.

4 The myth of value bipolarity

The fourth confusion is related to the previous two, but it is themistake not of confusing two things for one but of confusing manythings for two. The mistake is in thinking that value is bipolar.Almost everyone assumes that there are two values, good andbad, which fall on opposite ends of a spectrum, and it is almostas common to think that the appropriate response to the former isattraction and to the latter repulsion. I want to insist not that thegood/bad bipolarity is a mistake, but only that it might be. I am,however, convinced that the attraction/repulsion bipolarity is amistake, both as a way of explaining the meaning of value and asa way of explaining the appropriate response to it, and I have notfound any other bipolar response that applies to all value.

Consider the following ways of having value or disvalue: pe-culiar, humorous, awesome, enviable, ugly, contemptible, unjust,pitiful. If the only values are degrees of good and bad, and if thegood is the attractive and the bad the repellant, these propertiesshould differ only in their descriptive components and in the de-gree to which they attract or repel. I doubt that that is true of anyof them, but perhaps the clearest cases are the last two, the unjustand the pitiful. I would not deny that there is something unattrac-tive about both, although there is probably also some feature ofboth of them that is attractive. In the case of the unjust, we maybe attracted to the victim with whom we identify; in the case ofthe pitiful, the quality that makes a person pitiful rather than con-temptible might attract us. And we probably also feel repelled bypersons we perceive to be pitiful or unjust. But turning away isnot the appropriate response or even the natural one; at least, it is

25



not the dominant response. We do not always turn our backs onthe pitiful and the unjust, nor should we.16

Perhaps there is some other bipolarity of appropriate responseto value. Another candidate is promotion and elimination. Per-haps we aim to promote the good and to eliminate the bad. Thisdoes seem to apply to the unjust and the pitiful, since the appro-priate response is to eliminate them, usually by eliminating theircauses. But while promotion/elimination applies to some of thevalues on the list, it is doubtful that it applies to all of them. Isthe appropriate response to the awesome to promote it or to theenviable to eliminate it? I do not think so.

We now have two bipolarities, and neither appears to be re-ducible to the other. But perhaps both are reducible to a third.Attraction/repulsion and promotion/elimination have this incommon: We are glad that the object of the first response in eachpair exists, and we are sorry that the object of the second responsein each pair exists. That does seem to come close to a commonalityin the two pairs of value responses, and arguably it applies to allof the values on the list; but it does not apply to all values. Thereare forms of good and bad about which gladness and sorrow seemto be beside the point in fact, about which all responses seembeside the point. Would we say that the appropriate response to alife of Aristotelian eudaimonia, a life that fulfills human nature, isgladness? The problem is not so much that gladness is the wrong

16 Aristotle and other Greek writers classified pity as a form of pain, and sothey attempted to put it into the category of the repellant. In his studyof pity in ancient thought, David Konstan (2001, p. 11) speculates thatthe biological evidence might explain why the Greeks classified pity asa painful emotion. Joseph LeDoux (1996, p. 165) argues that basic emo-tions, such as fear, involve parallel transmissions to the amydala fromthe sensory thalmus and sensory cortex. The subcortical pathways arefaster, but more accurate representations of the external cause of the emo-tion come from the cortex. Konstan proposes that pity might also involvean initial aversive reaction followed by a more subtle and accurate as-sessment of the situation processed by the cerebral cortex. If so, pitywould be initially experienced as painful, and the Greeks would be partlyright.

26



response, since gladness is probably as good as any. The problemis that the whole question of response seems inappropriate whenapplied to certain kinds of value. A person who is eudaimon has agood life whether she is glad about it or not. She might not eventhink about her life as a whole. (Maybe when it is good enough,she has no reason to do so.) You will reply that if she were tothink about it, she would be glad. Perhaps so, but I doubt thatthat response is related in any significant way to what a good lifeis. Furthermore, there are value properties, such as the pitiful andthe contemptible, toward which the appropriate response is com-plex, much more so than any of the pairs of responses we haveconsidered.

These considerations affect the way in which a moral theoryought to be constructed. A satisfactory theory should make roomfor value of many different kinds. Some value properties are thosethat attract or repel us. Others are those we aim to produce or elim-inate. Still others are those we admire or find reprehensible. Somemay not be related to human responses in any interesting way,such as what is good for us. Some value properties do have appro-priate responses, but the responses are complex. Since a theoryought to be as simple as possible without unnecessarily distort-ing its subject matter, it would be helpful if it revealed both thedifferences between and the relations among these different typesof value. I will aim at a theory that has this feature.

5 Direct and indirect value of acts

Sometimes we evaluate an act not on the basis of its actual proper-ties, but on the basis of the properties of a class of acts to which itbelongs or the members of that class. So we may ascribe value toan act because acts of that type generally have good consequencesin circumstances of that kind, or because it follows a rule the gen-eral following of which has a good property that the act itself doesnot have, or because it is an act of a type a rational person woulddo or a virtuous person would do. In each case, the act gets valueindirectly. Either it gets value from the value of a property most ofthe acts of a certain class have, or it gets value from the value of a

27



property possessed by the class of acts itself. If it is the former, theact inherits value from a valuable property other individual actsthat are relevantly similar to it have (e.g., good consequences). If itis the latter, the act inherits value from a property that is valuablewhen possessed by a class of acts, such as the class of acts that fol-low some rule. It is interesting to consider the question of why weattribute value indirectly in this way. It may have something to dowith conceptual economy, although I doubt that that is the entireexplanation.17 In any case, my interest here is not in explainingwhy we attribute value to acts in this way; I merely want to callattention to the fact that we do so.

In addition to the indirect value we attribute to acts, we alsoattribute value in virtue of properties the act actually possesses.The rule consequentialist attributes value to an act either becauseit is a member of a class of acts most of which have good con-sequences, or because it follows a rule the general following ofwhich has good consequences. But the consequentialist cannot letit go at that. Surely, of two acts both of which are of the right con-sequential type, if one actually has good consequences and theother does not, the former is a better act. The actual consequencescan enhance the value of the act whether or not the act also hasvalue in the indirect way just described.

The same point applies to Kantian deontology, as Kant himselfrecognized. There is a difference between an act that follows theCategorical Imperative and an act that is actually motivated by theCategorical Imperative. An act follows the Categorical Imperativewhen it is a member of a class of acts that could be willed by agood will. Such an act is right (not wrong). But it is even better ifa given act is actually willed by a good will. The latter has whatKant considers distinctively moral value. The virtue theory that Iprefer also has this difference. An act has a form of value when itis the type of act a virtuous person would do, but it is even better

17 Indirect value is sometimes attributed to beliefs as well as to acts. Relia-bilists attribute value to a belief on the basis of the fact that it is a memberof a class of acts most of which produce true beliefs. On one form of co-herentism a belief gets value (often called justification) on the basis of thefact that it is a member of a coherent set of beliefs.

28



if it is actually produced in the same way that a virtuous personproduces it, that is, with a virtuous motive.

I do not think that direct and indirect value are systematicallyconfused, but the relation between them is often unclear. In partic-ular, it may be unclear whether they pertain to different kinds ofvalue or just to different degrees of the same kind of value. Indirectvalue is most often used in discussions of the deontic propertiesof right, wrong, and duty. An act is right or wrong because offeatures of a class of acts to which it belongs. This seems to meto be correct, although I have no considered position on why itis correct. Direct value may be more appropriate in discussionsof the praiseworthiness or blameworthiness of an act, althoughthat needs investigation. One confusion that ought to be avoidedis that between right in the sense of not wrong (permissible) andright in the sense of commendable. The two senses of right needto be kept distinct whether or not the difference is connected tothe difference between indirect and direct value.

iv a taxonomy of ethical theories

A moral theory is an abstract structure that aims to increase ourunderstanding of the moral realm by systematizing the conceptswe use in moral evaluation. John Rawls (197475) once proposedthat we compare theories by the different ways in which theyrelate the basic concepts of a virtue, a good state of affairs, anda right act.18 This approach may be a good place to begin, butit is too simple to do justice to significant differences betweenkinds of theory. For one thing, it ignores the distinction betweena good state of affairs and a good outcome.19 For another, it ig-nores the kind of good that is basic in Aristotelian and Thomisticvirtue theories: the good for beings of a kind most especially,for human beings. It also ignores such goods as knowledge, love,

18 Rawls actually says moral worth instead of virtue, but the two con-cepts are obviously connected.

19 Rawls seems to have outcomes in mind when he says states of affairs,and this confuses matters. I will have more to say about this in the discus-sion of Watson.

29



happiness, and life, which do not fall easily under either traits ofpersons or states of affairs, although many of them probably fallunder the category of goods for human beings. There may also bea distinct category of goods that Nagel calls perfectionist values such things as aesthetic value, scientific discovery, and the explo-ration of outer space. These allegedly differ from goods such aspersonal knowledge and love in that they are good apart fromwhat is good for human persons.20 The issue of whether there aresuch goods is an interesting one, but since these goods are not con-nected to the evaluation of human agents, human action, or a goodhuman life, I will give them only passing attention. The good for abeing of a kind is an important one, however, and it should not beignored in any ethical theory that purports to be complete.21

Some philosophers have wondered whether there is any form ofvirtue ethics that is interestingly different from consequentialistand deontological theories that is, different in structure fromthe other two types of theory. This question is important for thepurpose of deciding among competing theories. If virtue theorydoes not offer a distinctive structure, then it does not competewith consequentialism and deontologism in the same way thatthe latter two compete with each other. As already noted, somevirtue ethicists reject the idea of theory altogether, and so theyare not concerned to show that virtue ethics is capable of offeringa different theoretical structure. But it seems to me that virtueethics is much more interesting if it does compete by offering adistinctive theoretical structure as well as a different (and richer)substantive content.

20 Nagel (1979) does not explicitly distinguish perfectionist values fromgoods for human persons, but he does identify them as those that aregood apart from their value to the individuals who experience or usethem. He says, They are thought to have an intrinsic value, so that it isimportant to achieve fundamental advances, for example, in mathematicsor astronomy even if very few people come to understand them and theyhave no practical effects. Apparently Nagel believes that these goods aregood apart from the way in which they are good for individual humanbeings, although he may not have ruled out the possibility that they aregood for human beings collectively.

21 I thank John Zeis for reminding me of the importance of this sense of good.

30



Gary Watson has made a helpful study of the structural differ-ences between virtue theory and other kinds of theory, particularlythe class of theories that he calls an ethics of outcome.22 His worryis that virtue ethics is more likely to collapse into consequential-ism than into deontologism. I doubt that that is true, but in anycase, Watson says some interesting things about the taxonomy ofethical theories in the course of identifying what it would takefor a virtue theory to be significantly different from an ethics ofoutcome. The theory I am pr

zagzebski, linda - divine motivation theory. cambridge university press, new york (2004).pdf

Documents