A World Without Values: Essays on John Mackie's Moral Error Theory (Philosophical Studies Series)

A World Without Values

PHILOSOPHICAL STUDIES SERIES VOLUME 114

Founded by Wilfrid S. Sellars and Keith Lehrer Editor Stephen Hetherington, The University of New South Wales, Sydney, Australia Senior Advisory Editor Keith Lehrer, University of Arizona, Tucson, AZ, U.S.A. Associate Editor Stewart Cohen, Arizona State University, Tempe, AZ, U.S.A. Board of Consulting Editors Lynne Rudder Baker, University of Massachusetts, Amherst, MA U.S.A. Radu Bogdan, Tulane University, New Orleans, LA, U.S.A. Marian David, University of Notre Dame, Notre Dame, IN, U.S.A. John M. Fischer, University of California, Riverside, CA, U.S.A. Allan Gibbard, University of Michigan, Ann Arbor, MI, U.S.A. Denise Meyerson, Macquarie University, NSW, Australia François Recanati, Institut Jean-Nicod, EHESS, Paris, France Mark Sainsbury, University of Texas, Austin, TX, U.S.A. Stuart Silvers, Clemson University, Clemson, SC, U.S.A. Barry Smith, State University of New York, Buffalo, NY, U.S.A. Nicholas D. Smith, Lewis & Clark College, Portland, OR, U.S.A. Linda Zagzebski, University of Oklahoma, Norman, OK, U.S.A. For further volumes: http://www.springer.com/series/6459

Richard Joyce · Simon Kirchin Editors

A World Without Values Essays on John Mackie’s Moral Error Theory

123

Editors Richard Joyce University of Sydney Department of Philosophy Sydney NSW 2006 Australia [email protected]

Simon Kirchin University of Kent Department of Philosophy School of European Culture & Languages Cornwallis North West Canterbury, Kent United Kingdom CT2 7NF [email protected]

ISBN 978-90-481-3338-3 e-ISBN 978-90-481-3339-0 DOI 10.1007/978-90-481-3339-0 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2009939617 © Springer Science+Business Media B.V. 2010 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

Contents

Against Ethics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . John P. Burgess

1

Nihilism, Nietzsche, and the Doppelganger Problem . . . . . . . . . . . Charles R. Pigden

17

Patterns of Objectification Richard Joyce

. . . . . . . . . . . . . . . . . . . . . . . . .

35

Mackie’s Internalisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . Walter Sinnott-Armstrong

55

Mackie’s Realism: Queer Pigs and the Web of Belief . . . . . . . . . . . Jamie Dreier

71

Mackie on Practical Reason . . . . . . . . . . . . . . . . . . . . . . . . . David Phillips

87

The Argument from Moral Experience . . . . . . . . . . . . . . . . . . Don Loeb

101

Beyond the Error Theory . . . . . . . . . . . . . . . . . . . . . . . . . . Michael Smith

119

Normativity, Deliberation, and Queerness . . . . . . . . . . . . . . . . . David Copp

141

A Tension in the Moral Error Theory . . . . . . . . . . . . . . . . . . . Simon Kirchin

167

Business as Usual? The Error Theory, Internalism, and the Function of Morality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Caroline West

183

The Fictionalist’s Attitude Problem . . . . . . . . . . . . . . . . . . . . Graham Oddie and Daniel Demetriou

199

Abolishing Morality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Richard Garner

217

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

235 v

Contributors

John P. Burgess Department of Philosophy, Princeton University, USA. David Copp Department of Philosophy, UC Davis, USA. Daniel Demetriou Department of Philosophy, University of Minnestota, USA. Jamie Dreier Department of Philosophy, Brown University, USA. Richard Garner Department of Philosophy, Ohio State University, USA. Richard Joyce Department of Philosophy, University of Sydney, Australia. Simon Kirchin Department of Philosophy, University of Kent, UK. Don Loeb Department of Philosophy, University of Vermont, USA. Graham Oddie Department of Philosophy, University of Colorado, USA. David Phillips Department of Philosophy, University of Houston, USA. Charles R. Pigden Department of Philosophy, University of Otago, New Zealand. Walter Sinnott-Armstrong Department of Philosophy, Duke University, USA. Michael Smith Department of Philosophy, Princeton University, USA. Caroline West Department of Philosophy, University of Sydney, Australia.

vii

Introduction Richard Joyce and Simon Kirchin

Just as history is written by the winners, so too is moral philosophy written largely by the believers. Although moral skepticism has been a theoretical presence in Western philosophy for as long as anyone can discern, the position has nearly always been presented by its opponents. Callicles was probably a historical figure, and Thrasymachus certainly was, but it is unlikely that the lines that Plato placed in their mouths are remotely close to a sympathetic transcript of anything they ever asserted; their role in the dialogue is to fall silent as Socrates bullies his way to inevitable victory. This pattern repeats through the centuries: Moral skepticism is wheeled on to the stage for the sole purpose of the audience witnessing its crushing defeat. However, unlike the explanation for the paucity of historians from losing sides, the absence of the skeptic’s voice from the dialectic of moral philosophy is not due to his having been defeated (either militarily or intellectually). Indeed, the very fact that moral skepticism needs to be countered again and again – centuries of novel stratagems and ingenious arguments – indicates a foe that cannot be defeated easily, implying that there must exist significant considerations in its favor. The real explanation for the dearth of real-life moral skeptics plying their wares in the philosophical marketplace may be nothing more insidious than a natural process of self-filtration: Those who are drawn to moral philosophy sufficiently to publish works on the topic are more likely than not to be antecedently hostile towards moral skepticism. By analogy, consider theology. One need not believe in God in order to be a capable theologian, but how many atheistic theologians does one really expect to find in the profession? The average atheist, as a matter of contingent fact, simply has little interest in the practice. Similarly, perhaps, the average moral skeptic tends to expend her intellectual energies elsewhere. We suspect that moral skepticism enjoys a higher proportion of support among philosophers in general than it does among moral philosophers in particular.

R. Joyce (B) University of Sydney, Sydney, NSW, Australia S. Kirchin (B) University of Kent Cantesbury, Kent, UK

ix

x

Introduction

Of the handful of exceptions to this trend – of the handful of able-minded moral skeptics publicly championing the position – John Mackie must figure high on the list. His 1977 book, Ethics: Inventing Right and Wrong, continues to challenge and motivate metaethicists in a way to which few books from that era can claim. Despite the fact that it is a slim book with modest production values, despite the fact that its contents have provoked far more disagreement than concurrence, and despite the fact that few readers explore beyond Chapter 3, it must surely be acknowledged as one of the landmark texts of 20th-century metaethics. If it accomplished nothing more than staking out an extreme position that moral philosophers can use as an effective foil in the course of more “positive” theory-building, it would be an important book. Yet we think that it is much more than just a useful foil; it advocates a metaethical position to be taken very seriously, and employs arguments in favor of that position that, for all the intense critical scrutiny to which they have been subject, remain alive. The past 30 years have seen a great deal written about Mackie’s book, but more often than not the treatment reflects the history of moral skepticism mentioned above: His view is acknowledged only to be summarily dispatched, allowing the dialectic to proceed to other matters. However, Mackie’s metaethical standpoint deserves discussion of a more focused and sustained sort – an end to which this anthology aims to contribute. Readers will note that we have not selected papers that eulogize or even concur with Mackie’s views; some of the contributions barely even explicitly mention them, and, of those that do, a good many are robustly opposed. It is a testament to the fecundity of Ethics that it leads to philosophical encounters that Mackie did not explore, and that it continues to inspire lively and noisy critical debate. Although we have been referring to Mackie as a “moral skeptic,” it should be noted that he is not a skeptic in the classical sense of the term: He does not maintain that we cannot know whether moral claims are true and therefore ought to withhold passing judgment on the matter (as Pyrrho and his coterie held as a global position). Of the ancient trio of views – dogmatist (believer), nihilist (disbeliever), and skeptic (in a perpetual state of uncertainty) – Mackie’s so-called skepticism actually has much more of the flavor of nihilism. He does not merely doubt morality, he denies it. This denial takes the form of arguing for a moral “error theory” – the view that (A) moral discourse has the aim of securing the truth, but that (B) it systematically fails to do so. In arguing for (A), the error theorist contrasts with the noncognitivist, who claims that moral discourse is not even in the market for truth (because, for example, it consists of commands veiled in the indicative mood); in arguing for (B), the error theorist contrasts with the “success theorist,” in whose ranks appear all advocates of moral realism. There is nothing terribly complicated, esoteric, or unfamiliar in the idea of taking the error theoretic stance towards a problematic subject matter. It is, after all, simply the attitude that atheists take towards religion, and that devotees of one religion take towards any other non-equivalent religion. It is the attitude that sensible persons take towards phlogiston, astrology, the Loch Ness monster, and the existence of reliable causal relations between severed rabbits’ feet and episodes of good luck. Whenever people talk about a range of objects, relations, or properties for which a temptation

Introduction

xi

arises to declare that the world doesn’t contain the elements necessary to render their assertions true, we face the option of endorsing an error theory. Given that such temptations arise for an enormous range of perennial philosophical puzzles, the relevance of the error theoretic option is ubiquitous. We find it, thus, somewhat surprising that it was not until 1977 that the possibility of a moral error theory became a well-defined metaethical contender. Of course, moral nihilism antedates Mackie’s book. Mackie himself had advocated the view in his little-read 1946 article in the Australasian Journal of Psychology and Philosophy (soon to drop the “. . .Psychology and. . .”). Bertrand Russell articulated the idea in the 1920s (see Pigden, this volume). And one can trace its origins back as far as one pleases in Western moral philosophy, at least to Anaxarchus of the 4th century BC.1 Nevertheless, it seems fair to say that it was not until Ethics that the option became clearly delineated in the minds of modern metaethicists.

Understanding the Moral Error Theory Mackie introduces the term “error theory” by saying that most people’s moral judgments involve an implicit claim “to be pointing to something objectively prescriptive,” but “these claims are all false” (1977, p. 35). As far as trying to understand the error theoretic stance in general terms goes, the part about “objective prescriptions” can be put aside as optional; one would just as surely be advocating an error theory if one asserted, say, that most people’s moral judgments involve an implicit claim to the existence of pure autonomy, and all such claims are false. In light of this, let us say, as a first approximation, that one way of being an error theorist about morality is to hold that most people’s moral judgments involve implicit claim X, and all such claims are false. This needs to be refined in various ways. First of all, it is presupposed by nearly all subsequent discussions of the moral error theory that the reference to “most people” is a mistake. What is distinctive about the error theoretic stance is that there is something faulty about moral judgments per se, not merely most people’s moral judgments. Were it just a matter of most people being at fault, then there would be room for these people to revise their ways into line with the faultless moral judgments of the minority, and thus the error theorist’s key claim that there is something wrong with morality per se would evaporate. Indeed, it is often put forward as an objection to Mackie’s view that even if it is true that most people imbue their moral judgments with a problematic kind of objective force, this is a dispensable aspect of moral discourse; a morality stripped of this flawed element would still deserve to be called a “morality.” (See Dreier, Kirchin, West, this volume.) Mackie expresses the stronger and more metaethically interesting view when he asserts that the claims of objective prescriptivity are “part of what our ordinary statements mean: the traditional moral concepts of the ordinary 1 See

Warren (2002, p. 81).

xii

Introduction

man as well as of the main line of western philosophers are concepts of objective value” (1977, p. 35). Second, it may be prudent for the error theorist to say that the problematic claims in question are all untrue, rather than all false. To see why, consider that other paradigm of an error theoretic stance: atheism. The atheist thinks that the predicate “. . .is a god” has an empty extension (at least at the actual world), hence all sentences of the form “X is a god” are false. So far so good. But the word “god” functions not only as a count noun, but is often capitalized into the proper noun “God,” and what does the atheist say of the sentence “God is omnipotent”? It is not so obvious that this should be considered false. Peter Strawson’s well-known (though rather old-fashioned) view is that the atheist should consider the sentence neither true nor false (Strawson 1950). Sometimes noncognitivism is described as the view that sentences of the target discourse are neither true nor false – a characterization that leads to the embarrassing implication that, if Strawson is right, the atheist is a noncognitivist about (some) religious discourse. It is preferable to distinguish the noncognitivist from the error theorist by saying that the former denies, and the latter affirms, that moral utterances are assertions. Thus the atheist will not be a noncognitivist about “God is omnipotent,” since the sentence is typically used assertorically, even if Strawson is correct about its truth value. We then distinguish the error theorist from the success theorist (e.g., the moral realist) by saying that the former, but not the latter, claims that moral utterances are all either false or (hedging her bets on Strawson’s views) neither true nor false – a disjunction that can be captured by the claim that moral utterances are all untrue. (See Dreier, this volume, for possible problems with this taxonomy.) This way of distinguishing the error theorist from the noncognitivist also reveals what is mistaken about the widespread (though, in truth, fairly innocent) tendency to express the error theory in ontological terms: as the view that X doesn’t exist. The problem with this expedient is that noncognitivist also holds that the items in question do not exist (discounting the linguistic permissions that may be achieved via the quasi-realist program). The problem is circumvented if we insist that, when speaking carefully, the error theoretic claim is made of a discourse, not any cluster of objects, properties, or relations. We are not error theorists about ghosts, we are error theorists about ghost discourse. The common expression “an error theorist about morality” harmlessly fudges this distinction. To hold that moral judgments are all untrue is not to hold that every sentence containing a moral term is untrue. The atheist doesn’t think that every sentence involving the count noun “god” or the proper noun “God” is untrue. Consider “No gods exist” or “Augustine believed in God.” Rather, the error theorist focuses on a proper subset of sentences containing the problematic terms: those that seem to imply or presuppose the instantiation of a moral property. “Stealing pears is morally wrong” will be such a sentence; “Augustine believed that stealing pears is morally wrong” will not be. (See Pigden, this volume.) Not only is endorsing a moral error theory consistent with the continued use of moral terms (as in “Nothing is morally wrong”), it is even consistent with the continued use of atomic moral claims (such as “Stealing pears is morally wrong”).

Introduction

xiii

It is typically assumed that the moral error theorist must be a moral eliminativist: advocating the abolition of all atomic moral sentences. But in fact what the error theorist decides to do with the erroneous moral language is a matter logically independent of the truth of the moral error theory. Perhaps the moral error theorist will carry on asserting moral judgments although she believes none of them – in which case she will be lying to her audience (assuming her audience consists of moral believers). But if lying is a fault only in a moral sense, then the moral error theorist may remain unperturbed by this state of affairs. Or perhaps the moral error theorist carries on uttering moral sentences but finds some way of removing assertoric force from these utterances, in which case she need not be committing a moral or epistemological sin any more than does an actor reciting the lines of a play. The error theorist who advocates maintaining moral language in this way is endorsing a kind of fictionalism (a view which is further described towards the end of this introduction). Although one could be a moral error theorist by implication – either because one endorses a radical global error theory (thus being skeptical of morality along with modality, colors, other minds, cats and dogs, etc.), or because one endorses an error theory about all normative phenomena – typically the moral error theorist thinks that there is something especially problematic about morality, and does not harbor the same doubts about normativity in general. The moral error theorist usually allows that we can still deliberate about how to act, she thinks that we can still make sense of actions harming or advancing our own welfare (and others’ welfare), and hence she thinks that we can continue to make sense of various kinds of nonmoral “ought”s, such as prudential ones. (See Joyce 2007.) Thus we may observe the moral error theorist assert a claim like “One ought not harm others,” adding that it is not a moral “ought” that is being employed. (In the same way, an atheist can assert that one ought not covet one’s neighbor’s wife, so long as it clear that this isn’t an “. . .according to God” prescription.) There is a certain kind of opponent to the error theory who finds this dubious. If so much normativity is conceded to be non-erroneous (the opponent maintains) – if one can make good sense of practical reasons, prudential norms, altruistic desires, and so forth – then the error theoretic position simply collapses, for these things are the very warp and weft of morality. The opponent might go so far as to say that so long as it is allowed that there are decisions to be made about how we shall live together, then the error theoretic position fails, for to engage in such a decision procedure is to endorse a moral point of view. This dispute concerns the definition of “morality.” An opponent may complain that the error theorist’s definition is far too restrictive; the error theorist will respond that the opponent’s definition is too liberal. We will discuss Mackie’s particular way of defining morality below, but it is worth foreshadowing here. When discussing the point of morality, Mackie suggests that it is a mechanism for counteracting our limited sympathies (1977, p. 107). However, it seems unlikely that he intends to define “morality” by reference to a distinctive subject matter. Rather, what Mackie thinks is distinctive about morality is not its content but rather the unusual nature of its norms: an authoritative normativity that purports to bind agents “from the

xiv

Introduction

outside,” irrespective of their desires, projects, or interests. (This will be described in more detail later in this introduction, and it is of course a theme that arises repeatedly in the papers in this volume.) Since there is no reason to doubt that a given action might be valued or prescribed by two or more different normative systems at once, it follows that we might not be able to “read off” whether a judgment counts as moral from its subject matter or surface form. Consider a token of the sentence “You must not take the rook.” It might express a norm of chess (because taking the rook is an impermissible move), or a norm of prudence (because taking the rook is a poor move and you’d like to win the game), or a norm of etiquette (because taking the rook would be impolite), or even a moral norm (because you’ve made a nontrivial promise not to take the rook). It depends on the force with which it is uttered, of which the intentions of the speaker are surely an important component. (And, indeed, the speaker may be somewhat undecided, leaving the matter indeterminate.) In any case, because Mackie, like most other moral error theorists, wishes to damn morality while tolerating other forms of normativity, then he must not only convince us that this special kind of practical authority is irredeemably flawed, but also show us that it is not also a characteristic of the kinds of normativity that he endorses. Even in advance of the outcome of that debate, however, one related point is clear and worth emphasizing: There are no grounds for assuming that the moral error theorist must be a sociopathic cad. The point is particularly worth stressing since much of the opposition to the moral error theory is motivated (we suspect) by an inchoate practical fear of what might happen should moral skepticism be widely adopted. Two thousand years ago, Aristocles of Messina asked “What evil deeds would he not dare, who held that nothing is really evil, or disgraceful, or just or unjust?”2 Paraphrasing Dostoyevsky, one might declare “If there is no moral truth, then everything is permitted.” And Dr Johnson memorably said of the moral skeptic: “If he does really think that there is no distinction between virtue and vice, why, sir, when he leaves our houses let us count our spoons.” Such rhetoric, however, does not withstand careful scrutiny. There are no grounds for assuming that the moral error theorist must be tolerant of those actions that would usually be opposed on moral grounds; her skepticism does not (in any obvious way) exclude the possibility of her being motivated by compassion, love, and altruism. There are many entirely nonmoral considerations that speak against sneaky nastiness against one’s fellows, even when the chips are down.3 Thus the moral error theorist need not be embarrassed by the common rhetorical trick of asking (in a tone of outraged wonderment) something along the lines of: “So you don’t think that strangling babies is morally wrong?!” The best response is: “No, but nor do I think that strangling babies is morally good or morally acceptable either. I don’t think it is morally anything. Putting aside that 2 Quoted

by Eusebius of Caesarea in Praeparatio Evangelica, book 14, c. 313–324/1903. incidentally, indicates what is confused in the pseudo-Dostoyevskyan slogan. What kind of permission does the paraphraser of Dostoyevsky intend to invoke? It cannot be moral permission, for if there are no moral facts then nothing is morally permitted. But if it is some other kind of permission, then one needs a reason for thinking that the non-existence of moral facts will affect it. (An analogous dilemma faces the real Dostoyevskyan dictum, concerning the death of God.)

3 This,

Introduction

xv

whole bankrupt conceptual scheme, the thought of someone strangling a baby sickens me and I oppose such behavior with every fiber of my being.” In the context of everyday conversation, to deny that there is anything morally wrong with X implies that one thinks that X is morally good or morally permissible – but a conversational context in which radical moral skepticism is live contender is precisely one where that implication breaks down. Attempts to discomfit the error theorist by forcing him to admit that he doesn’t think that there is anything morally wrong with [insert your favorite horrible transgression here] trade on equivocating between these contexts. The fear that moral skepticism will lead to a breakdown of social cohesion presupposes that our moral commitments are the only thing keeping us well-behaved (and the only thing that could keep us well-behaved), which, ironically, reveals just the kind of pessimism about the human spirit that the moral skeptic is often accused of indulging in. On the other hand, it seems reasonable to assume that morality plays some kind of role in motivating prosocial behavior, and thus becoming a moral skeptic can be predicted to have some kind of proportionate impact on one’s motivations. (See West, Garner, this volume.) We think that the starting point for making progress on this matter is the recognition that the cluster of relevant questions – what is the role of moral belief in motivation? what would be the practical impact of removing those beliefs? might some alternative available psychological mechanism(s) serve these ends just as well? – are all empirical issues, not to be pronounced upon with any confidence from one’s armchair.

Arguing for a Moral Error Theory It is important to keep separate in one’s mind the error theoretic view and the arguments that Mackie employs in his attempt to establish that view. Yet again, it is useful to compare the atheist: What it takes to be an atheist is one thing (disbelief in any god); one’s grounds for being an atheist are something else. There are many such grounds possible – some more plausible than others – and it is clear that defeating one kind of pro-atheist argument hardly proves the existence of a god. Yet we find in much of the metaethical literature a tendency to assume that exposing the flaws of Mackie’s arguments in favor of the moral error theory provides a sufficient basis for the wholesale rejection of moral skepticism. This is not so; Mackie articulates just some of the considerations that might speak in favor of a moral error theory. Nevertheless, it is understandable that opposition to the moral error theoretic position has focused heavily on Mackie’s two arguments in favor of that viewpoint: the Argument from Relativity and the Argument from Queerness. Before introducing these arguments, let us consider the strategy of arguing for an error theory in more general terms. Such an argument typically has two steps: the conceptual and the ontological. First the error theorist may establish that moral discourse is centrally committed to

xvi

Introduction

some thesis X (or some theses X, Y, . . . , n 4 ). The phrase “centrally committed” is supposed to indicate that to deny X would be to cease to participate competently in that discourse. Imagine a phlogiston theorist who, upon hearing of the success of oxygen theory, claims that his theory has been vindicated; he asserts that he has been talking about oxygen all along but just by a different name. When the important differences between the two substances are pointed out to him (that phlogiston is stored in flammable materials and released during combustion, while oxygen combines from the atmosphere with flammable materials and is destroyed during combustion), he admits that he’s had some false beliefs about the nature of the substance in question, but remains adamant that he was still talking about oxygen all along. This seems unacceptable, roughly because the thesis about being stored and released is a “central commitment” of phlogiston talk; to deny this thesis with respect to some substance is to cease to talk about phlogiston. The ontological step of the error theorist’s argument is to establish that thesis X (whatever it may be) is false. This may be achieved either through a priori means (demonstrating X to be incoherent, say), or through a posteriori methods (investigating the world and coming to the conclusion that nothing satisfies X). Which method is appropriate depends on the nature of the error that has been attributed to moral discourse. Sometimes the moral error theorist will hold that there is something impossible or incoherent about moral properties, such that the error theory is necessarily true. But it suffices for being an error theorist to hold that the non-instantiation of moral properties is a merely contingent affair. Mackie is often interpreted in the former way, but in certain moods (at least) he prefers the latter. In an oftenoverlooked passage he concedes that if theism were true, then “a kind of objective ethical prescriptivity could be introduced” (1977, p. 48). Though an avowed atheist, Mackie did not, apparently, maintain that theism is necessarily false, and thus on the basis of this passage we must conclude that he took the moral error theory to be only contingently true. The two steps of the error theoretic argument produce two kinds of opposition: the “concessive” response and the “head-on” response. The challenger may concede that the putatively defective attribute that the error theorist assigns to morality really is irredeemably defective, but deny that this attribute is an essential component of morality; a normative framework stripped of the troublesome element will still count as a morality. Alternatively, opponents may accept that the putatively defective attribute is a non-negotiable component of anything deserving the name “morality,” but deny that it really is problematic; they embrace the challenge headon of making sense of the prima facie puzzling aspect. So, for example, if the error theorist is claiming that moral properties require a kind of pure autonomy which the universe does not supply, then one type of opponent will insist that morality requires

4 We

will drop this qualification in what follows, simply for brevity. Nevertheless, it is important to bear in mind that the moral error theorist might accuse moral discourse of a number of faults – any one of which would be insufficient to ground an error theory, but the sum total of which is sufficient.

Introduction

xvii

nothing of the sort, while another will insist that the universe does indeed contain such autonomy. The error theorist must be prepared to defend herself on both fronts. Mackie offers two arguments for his metaethical position. The first, the Argument from Relativity (often more perspicaciously referred to as “the Argument from Disagreement”), begins with an empirical observation: that there is an enormous amount of variation in moral views, and that moral disagreements often seem unusually intractable. Mackie argues that the best explanation of these phenomena is that moral judgments “reflect adherence to and participation in different ways of life” (1977, p. 36). This, at least, is a better explanation than the hypothesis that there is a realm of objective moral facts to which some cultures have inferior epistemic access than others. The example he uses is of two cultures’ divergent moral views regarding monogamy. Is it really plausible, he asks, that one culture enjoys access to the moral facts regarding marital arrangements whereas the other lacks that access? Isn’t it much more likely that monogamy happened to develop in one culture but not in the other (for whatever cultural or anthropological reasons), and that the respective moral views emerged as a result? Opposition to the Argument from Relativity can, broadly speaking, take two forms. First, one might deny the empirical premise, arguing that moral disagreement is not really as widespread as it is often made out to be, or at least arguing that much of the conspicuous disagreement masks extensive moral agreement at a deeper level pertaining to more fundamental moral principles. Mackie makes some brief remarks in response to this argument (1977, p. 37). Second, one might accept the phenomenon of moral disagreement at face value but deny that the best explanation of this favors the error theory. Often both strategies are deployed side by side. Mackie’s second skeptical argument, from Queerness, has two strands: one metaphysical and one epistemological. The first states that our conception of a moral property is essentially one of a very unusual kind of property, such that countenancing its instantiation requires us to posit in the world “qualities or relations of a very strange sort, utterly different from anything else in the universe” (1977, p. 38). The second states that in order to track such weird properties we would need “some special faculty of moral perception or intuition, utterly different from our ordinary ways of knowing everything else” (ibid.). These are not independent arguments, since we are forced to posit weird epistemological equipment only if it has already been established that the properties in question are weird. Thus really it is the metaphysical strand of the Argument from Queerness that is load bearing. The Argument from Queerness may be taken to refer to Mackie’s specific version or may be considered in a generic sense. In the generic sense, whenever one argues (A) that morality is centrally committed to some thesis X, and (B) that X is bizarre, ontologically profligate, or just too far-fetched to be taken seriously, etc., then one has presented a kind of Argument from Queerness. (Arguments for the moral error theory need not take this form; one might, for example, simply discover that X is empirically false.) This is generic since “X” could denote any of an open-ended

xviii

Introduction

range of options. But even understanding the Argument from Queerness in a nongeneric sense is no straightforward matter, since it is not entirely clear what Mackie puts in place of “X.” Mackie says that for moral properties to exist would require the existence of “objective values” and “objective prescriptions,” and it is evidently these values and prescriptions that he finds metaphysically queer. He claims that in denying the existence of such values he is denying that any “categorically imperative element is objectively valid” (1977, p. 29). A categorical imperative is an imperative (“Do φ”) that is applied to a subject irrespective of that person’s ends. It is to be contrasted with a hypothetical imperative, which does depend on a person’s ends. Thus “Go to bed now” is usually understood to be tacitly conditional, depending on something like “. . .if you want to get a decent night’s sleep.” If it turns out that the person lacks this desire (and any other desire that promises to be satisfied by following the advice), then the imperative should be withdrawn. By contrast, the categorical imperative “Don’t murder children” cannot be begged off by the addressee explaining that he really enjoys murdering children, that he lacks any desires that will be satisfied if the imperative is obeyed; it is not a piece of advice at all. Note that it does not appear to be categorical imperatives per se that trouble Mackie, but categorical imperatives that purport to be “objectively valid.” Quite what he means by this restriction, however, remains the subject of debate. He gives two concrete illustrations of what he has in mind – of what the world would have to be like in order for these putatively weird moral properties to be instantiated. First, he mentions Plato’s account of the Form of the Good, which is such that the mere comprehension of the fact that something participates in the Form (i.e., is good) somehow automatically engages the motivation to seek that thing. The Good, for Plato, has a kind of magical magnetism built into it. Second, Mackie mentions Samuel Clarke, who in the early 18th century argued for (in Mackie’s words) “necessary relations of fitness between situations and actions, so that a situation would have a demand for such-and-such an action somehow built into it” (1977, p. 40). The fact that these two illustrations are subtly but importantly different is responsible for at least some of the confusion surrounding the putative source of queerness. The Plato example suggests that the weirdness resides in properties the recognition of which compels motivation; the Clarke example suggests that the weirdness resides in properties that demand action (and thus motivation). The latter is arguably the more charitable interpretation, and also seems to fit better with comments made elsewhere by Mackie concerning the role of practical reasons in the Argument from Queerness. (See Garner 1990; though compare Sinnott-Armstrong, this volume.) He writes that “to say that [objective prescriptions] are intrinsically action-guiding [which is one way Mackie sometimes describes the queerness whose existence he is denying] is to say that the reasons that they give for doing or for not doing something are independent of that agent’s desires or purposes” (Mackie 1982, p. 115). It would make sense if Mackie were, then, simply to deny the existence of such “desire-transcendent” reasons (in the vein of Williams 1981); but his position is characteristically more nuanced than this. He allows that we often legitimately

Introduction

xix

employ talk of reasons regarding persons who have no desires that will be satisfied by performing the action in question. If another individual is suffering, for example, and there is some course of action I can take to relieve that suffering, then “it would be natural,” Mackie says, to claim that these sufferings “constitute some reason. . . independent of any desire that I now have to help these other people” (1977, pp. 78–79). Though Mackie doesn’t attempt to discredit appeals to such desire-transcendent reasons, what he does insist on is that such reasons talk is made legitimate only by the presence of an institution: What allows the transition from “There is a stranger writhing in agony before me” to “I have a reason to help” is a cluster of institutional facts, not brute facts. Examples of institutions, given by Mackie, include the rules of chess, social practices such as promising, and the thoughts and behaviors associated with the idea of a person’s identity persisting through time. Such institutions have rules of conduct which guide the behavior and speech of adherents, and transgressions of which are condemned. Importantly, such requirements “are constituted by human thought, behaviour, feelings, and attitudes” (1977, p. 81), and thus any such requirements are, in a central sense, mind-dependent. This, perhaps, provides insight into why Mackie objects not to categorical imperatives per se, but to objective categorical imperatives: It is categorical imperatives that profess to transcend all institutions, that purport to depend for their legitimacy on “requirements which simply are there, in the nature of things” (1977, p. 59), that are singled out as erroneous. As with categorical imperatives, so with reasons: It may not be false to claim “Anyone has a reason to ease the suffering of others,” but its truth is guaranteed only by invoking an institutional way of speaking – an institution of which one may or may not be an adherent. (Mackie writes that one is never “logically committed” to offer allegiance to an institution.) It is only when such a reason claim purports to transcend all institutions – when it is imbued with ambitions of objectivity – that it oversteps the mark. In light of these observations, the error theory arises because (Mackie thinks) moral discourse is pervaded through and through with aspirations to robust, institution-transcendent objectivity.

Preview of This Book In 2007, to mark the 30th anniversary of the publication of Mackie’s Ethics: Inventing Right and Wrong, we edited a special issue of the journal Ethical Theory and Moral Practice in which six papers were presented. The one element of that project with which we were disappointed was its modest length. We therefore decided to have another run at the project, this time as a full-length anthology. We here reproduce the six original chapters (Burgess, Demetriou and Oddie, Garner, Loeb, Phillips, and Pigden), plus five more invited contributions, and two additional papers from the editors. We believe that the rich and varied result indicates the enormously fruitful influence that Mackie’s work continues to have on contemporary metaethics.

xx

Introduction

At more or less the same time as Mackie was putting the finishing touches on Ethics, John Burgess (better known for his prolific work in logic and philosophy of mathematics) was penning an essay arguing for a position called “anethicism” – which turns out to be equivalent to the moral error theory. Burgess’ chapter was not published at that time – partly because Mackie beat him to the punch – but it was informally circulated for decades, especially among Princetonians with an interest in metaethics. Publication had to wait until the 2007 special issue of Ethical Theory and Moral Practice, by which time, of course, certain anachronisms of terminology had become apparent (e.g., a seemingly odd use of “cognitivism”) and many illustrative examples (e.g., references to Nixon and Jackie Onassis) had become charmingly dated. Despite these extrinsic qualities – which, in our opinion, add to the “time capsule” appeal of the paper – the case that Burgess develops remains highly relevant to the modern debate, so much so that we judge that whatever made it fitting to open that special issue with Burgess’ chapter provides warrant for doing so again. The second chapter begins by examining one of the historical antecedents of Mackie’s metaethical view – namely, Nietzsche’s convoluted moral philosophy. Charles Pigden argues that Nietzsche was indeed an error theorist, and in doing so he counters an objection that may be made to moral nihilism in general – an objection that Pigden finds articulated in the work of Crispin Wright and Simon Blackburn. The final sections of Pigden’s chapter help us to formulate the error theory more carefully. Mackie described the view as holding that all moral claims are false, which, Pigden notes, leads to trouble (the “Doppelganger Problem”). If “X is P” is a false moral claim, then “It is not the case that X is P” must be true (by classical standards); but if the latter claim also counts as moral, then it cannot be that all moral claims are false. Pigden provides a tidy solution to the problem. There follow three chapters that can be loosely grouped as attempts to analyze Mackie’s argument or the status of its conclusion. (Of course, all chapters in this collection do this to a certain extent; it’s a matter of focus.) Richard Joyce is interested in the logical relation between Mackie’s skepticism and Mackie’s “objectification thesis”: the idea that moral judgments are the result of our having “projected” our affective attitudes onto our experience of the world. Rather than just being an explanation for where this huge flaw in human moral thinking comes from (for the curious, as it were), the objectification thesis plays an important live role in Mackie’s case for moral skepticism. Though he has defended moral skepticism on several occasions before, Joyce’s position here is officially neutral; his objective is to understand the argument rather than advocate it. So too with the next chapter, from Walter Sinnott-Armstrong. Most modern commentators share the view that some thesis or other going by the name “internalism” must play some kind of important role in Mackie’s argument for moral skepticism; yet opinions differ as to how most properly to formulate this thesis – both with regard to Mackie’s intentions and concerning the construction of the best argument for moral skepticism. Sinnott-Armstrong draws attention to the many variables hidden in “internalism” and proceeds to tease apart an almost-bewildering array of internalist theses. Many of these are rejected as far-fetched or as not serving

Introduction

xxi

Mackie’s ambitions, and in doing so Sinnott-Armstrong homes in on those theses that might be both defensible and useful for Mackie. Jamie Dreier’s opinion is considerably less sympathetic towards the moral error theory, but nevertheless the principal goals of his chapter are largely neutral. He first presents the jaw-dropping claim that Mackie was in fact a moral realist, and then spends most of the paper trying to convince us of this startling contention. Dreier’s case begins with the recognition that the rise of quasi-realism and minimalist accounts of such notions as truth, property, and assertion have made it increasingly difficult, in recent decades, to distinguish the moral antirealist from the realist. Dreier offers a new way of drawing the boundary, pertaining to whether, in our attempts to explain philosophical puzzles surrounding morality, we look to properties in the world (as the realists think) or look to moral concepts and the relations among them (as the anti-realists think). The surprising outcome is that when the boundary is drawn using this criterion, Mackie’s view seems to come out as realist (though this need not be true of the moral error theorist in general). The fifth and sixth chapters of the anthology return to the actively supportive attitude towards the moral error theory that was evident in Burgess and Pigden. David Phillips provides a penetrating analysis of Mackie’s neglected view of practical reason, and illuminates the role of practical reason in the Argument from Queerness. Phillips contrasts Mackie’s position on practical reason with that of Bernard Williams, and argues that the former is superior in several respects. (Incidentally, one of the editors of the present volume is the target of some of Phillips’ admonition, and hereby acknowledges that he finds the critique generally convincing.) Don Loeb critically confronts an argument that is frequently leveled against the moral skeptic, which he calls “the Argument from Moral Experience.” It consists of two premises: (1) that the phenomenology of moral experience is in line with the moral realist’s perspective, and (2) that this creates a burden of proof that the moral skeptic has to work harder to overcome. In other words, the world’s seeming to be a moral way creates a presumption in favor of the world’s being a moral way. Loeb questions both premises. Should our moral phenomenology be characterized in the way that moral realists typically paint it? And, even if it should, is the moral realist justified in thinking that this favors her case? Following Phillips’ and Loeb’s essentially pro-Mackie stance come four contributions from the opposition – from Michael Smith, David Copp, Simon Kirchin, and Caroline West. Both Smith and Copp respond to Mackie’s Argument from Queerness in what was described earlier as a “head-on” manner: They attempt to show that objective prescriptions are not mysterious within a naturalistic framework, after all. Smith argues that Mackie’s moral skepticism would be defused if we were able to make sense of “ends that are absolutely prescribed by reason.” He outlines several attempts to provide the goods, settling ultimately (though tentatively) on the view that there may be certain desires the having of which is constitutive of being rational, and, thus, Mackie’s challenge is answerable.

xxii

Introduction

Like Smith, Copp investigates the possibility of opposing Mackie “head-on” by making sense of actions prescribed by reason, actions that it would be irrational to refrain from performing. Copp maintains that it is only if one comes to the puzzle with an instrumentalist theory of practical rationality in hand that such actions seem so problematic; but such a theory is strictly optional. He uses his “society-centered theory” of moral properties as an illustration of a kind of moral naturalism that (A) provides reasons for actions that it would be irrational to ignore, but (B) does not involve any dubious metaphysical excesses. Kirchin’s opposition to the Argument from Queerness can, by contrast, be classified as of the “concessive” type. Perhaps objective prescriptions of the kind outlined by Mackie would be too weird to countenance, but who says that ordinary moral discourse is committed to that? Kirchin argues that moral discourse is in reality a messy business, with speakers’ commitments far more nebulous, inchoate, and varied than the error theorist typically acknowledges. But this very messiness counts against the plausibility of the error theory, since it counts against the likelihood of our locating “non-negotiable moral commitments” which can then be successfully charged with error. West’s position also makes limited concessions to Mackie. She agrees with Mackie’s (apparent) claim that moral discourse is committed to some thesis of internalism, and she thinks that the internalism in question is untenable. As evidence for this commitment, she explores several pragmatic uses to which we put our morality (why do we have it at all?), explaining how a belief in internalism would be crucial for each practical purpose. But if we really are in such a predicament, then how should we proceed? West outlines several options, without firmly taking sides. At least one of these options concedes the whole game to the error theorist (fictionalism), whereas others hope to resurrect a kind of moral realism from the situation. It is on the strength of her sympathetic description of a “revisionary moral realism” that we categorize West’s chapter as charting a route for the concessive opponent to the error theory. Though morality may heretofore have committed us to some strong internalist thesis – a thesis that is false – it is possible that this commitment was never “non-negotiable”: perhaps a weakened internalism can be embraced in good faith while still underwriting enough of the pragmatic purposes to count literally as a (revised) morality. While the latter sections of West’s chapter address the question of how we might proceed if the error theorist turns out to be largely correct, the final two contributions of this collection deal entirely with the aftermath of a triumphant moral error theory. What should we do with our moral talk and thought if we were to become convinced that Mackie is correct? There are broadly two camps. Until recently the assumption was that the error theorist should also be an abolitionist (a.k.a. eliminativist): eschewing moral talk in the same way as we all now eschew positive talk of, say, witches. Thus, it has been supposed, if we catch a professed moral error theorist employing moral language, we can gleefully cry “Aha!” (Any such accusation would be an argument not against the moral error theory but against the theorist – showing her to be a hypocrite, disingenuous, in bad faith, or vacillating between belief and disbelief. To suppose that this somehow undermines the possibility of the

Introduction

xxiii

moral error theory being true is to commit an ad hominem fallacy.) But the increasingly popular alternative (or, at least, the increasingly discussed alternative) is that the error theorist might adopt the fictionalist stance: a commitment to continue to make moral utterances and have moral thoughts, while withholding assertoric force from the utterances and withholding doxastic assent from the thoughts. On the assumption that morality is in various respects useful when it is asserted and believed, eliminativism will (ceteris paribus) constitute a practical cost. The fictionalist option, therefore, becomes attractive if and only if it promises to recoup some of these costs. The advocate of fictionalism holds that some of these losses may be recovered by adopting a policy of employing moral language, engaging in moral deliberation, and being moved by moral emotions, but throughout it all remaining disposed to deny the truth of any moral proposition if pressed in an appropriately serious manner (e.g., when in the philosophy classroom), thus not really believing any of it and thus not violating any epistemic norms.5 The debate between the fictionalist and the abolitionist is one of the more intriguing recent legacies of Ethics, yet it is hard to discern Mackie’s own thoughts on the topic. On the final page of the book he hints that morality might continue with the status of a “useful fiction” – but one searches in vain for any discussion of how he thinks this might be supposed to work. Taking the moral fictionalist Mark Kalderon as their principal target, Graham Oddie and Dan Demetriou (in a joint chapter) raise a serious challenge for the fictionalist (and not just the moral fictionalist), which they call the “acceptancetransfer problem.” They argue that, as with the well-known Frege-Geach challenge to noncognitivism, there is a worry about how there can be any rational transfer between various claims in a piece of everyday moral reasoning (say, moral modus ponens) if the attitude taken by the speaker towards one of the premises (the moral claim) is something unlike belief and more akin to make-belief. In the thirteenth and final chapter of this issue, Richard Garner – a firm supporter of moral skepticism – puts the case for abolitionism most forcefully, reminding us of the reservations that Mackie himself had about the institution of morality (voiced not in Ethics but in Hume’s Moral Theory of a few years later). Garner has grave doubts about the fictionalist program: doubts about its psychological viability, doubts about its touted pragmatic pay-offs. He makes a strong case that abolitionism may be the most honest and practical attitude for the moral error theorist to adopt.

5 The

kind of fictionalism being described here is revolutionary and noncognitivist. In being revolutionary it contrasts with hermeneutic fictionalism, according to which we are already doing something like engaging with a fiction when we make moral judgments. In being noncognitivist it contrasts with cognitivist fictionalism, according to which moral judgments are truth-evaluable assertions whose content concerns some fiction (e.g., “X is morally wrong” = “According to fiction F, X is morally wrong”).

xxiv

Introduction

References Eusebius of Caesarea. 1903. Praeparatio evangelica. Trans. E. H. Gifford, 313–324. Oxford: Clarendon. Garner, Richard. 1990. On the genuine queerness of moral properties and facts. Australasian Journal of Philosophy 68: 137–146. Joyce, Richard. 2007. Morality, schmorality. In Morality and self-interest, ed. Paul Bloomfield, 51–75. Oxford: Oxford University Press. Mackie, John L. 1946. A refutation of morals. Australasian Journal of Psychology and Philosophy 24: 77–90. Mackie, John L. 1977. Ethics: Inventing right and wrong. London: Penguin. Mackie, John L. 1980. Hume’s moral theory. London: Routledge and Kegan Paul. Mackie, John L. 1982. The miracle of theism. Oxford: Oxford University Press. Strawson, P. F. 1950. On referring. Mind 59: 320–344. Warren, James. 2002. Epicurus and Democritean ethics: An archaeology of ataraxia. Cambridge: Cambridge University Press. Williams, Bernard. 1981. Internal and external reasons. In his Moral luck, 101–113. Cambridge: Cambridge University Press.

Against Ethics John P. Burgess

’ε´ γ νων, α’ ν ε´ γ νων,

κατ ε´ γ νων – Julian

The familiar theories of ethics can be divided into two classes. Cognitivist theories (for example, utilitarianism and intuitionism) maintain that there is a body of fact about right and wrong, good and evil, which it is the business of moral philosophy to uncover. According to linguistic theories (for example, emotivism and prescriptivism), proper analysis shows moral judgments to be meant not as assertions about an alleged realm of objective values, but as some quite different type of “speech act.” Could it be that both types of theory are wrong? Anethicism (or moral skepticism) maintains that they are: Ordinary people’s moral judgments are meant as statements of impersonal fact about absolute values, but there are no such objective values, so moral thinking involves a fundamental mistake and illusion. Anethicism is to ethics as atheism is to theology. Many philosophers are aware of the existence of the anethicist position. Indeed Thomas Nagel has gone so far as to say (personal communication) that “fear of it has I think been the unspoken motive of most of the recent contortions in metaethics.” Yet until the recent appearance of J. L. Mackie’s Ethics (1977), it has been hard to find even one expression of the anethicist position in print. (It has been hard to find an argument against linguistic theories of ethics that did not end by endorsing cognitivism, and vice versa.) I hope the discussion that follows may help bring further attention to this neglected position.

1 When Archbishop Lefebvre says, “To Hell with the Pope!” he gives vent to emotion. When he says, “Vatican II was inspired by Baalzebub,” he means to assert something. When Muhammad Ali says, “Eating pork is against my religion,” he J.P. Burgess (B) Department of Philosophy, Princeton University, Princeton, NJ, USA

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_1, 

1

2

J.P. Burgess

says something he knows would be false coming from Oral Roberts or Reverend Ike. When he says, “Allah forbids pork,” that is supposed to be impersonally true. If Baruch Korff says, “Zen is an old religion,” we have to pick up from context whether he means something like “by American standards” or “compared to Mormonism and Christian Science,” which would make his statement true, or something like “by Far Eastern standards” or “compared to Taoism and Shinto,” which would make it false. If he says, “Buddhism is a false religion,” his statement is not relative in this way. Billy Graham takes comfort in the thought that the angels of the Lord encamp round about all them that fear Him. But he does not believe this thought is true because it is comforting. Its truth is supposed to derive from the actual carryings-on of supernatural beings, and to be in a strong sense a matter of objective fact. Slippery philosophers of religion, super-subtle theologians, William Jameses and Paul Tillichs, may imagine that in reciting traditional creeds they are performing some quite extraordinary speech act. The ordinary belief has always been that, however poorly even King James English reflects celestial mysteries, church dogma embodies objective, absolute, and supremely important truths. I hold that ordinary people, leaving aside a few hyper-sophisticated moral philosophers, similarly mean their moral judgments as objective, non-relative assertions. That moral judgments are meant as any kind of assertion at all is denied by the once-influential theories of emotivism and prescriptivism. According to the former, “Abortion is immoral” is just an expression of emotion, akin to “Down with abortion!” According to the latter, it is a species of imperative, amounting to something like, “Let no one ever procure or perform an abortion!” The by-now familiar difficulties these theories encounter are worth reviewing. We grew up, most of us, acquiring a tangle of ethical concepts whose complexity philosophers are just beginning to appreciate. Among them were those of an immoral or wicked act, of an act one morally shouldn’t or oughtn’t commit, of a course of action less good than some alternative, of an act deserving blame or punishment, of a violation of rights or obligations, of a morally depraved or reprobate character, of a vice.We learned to deploy these concepts not merely in categorizing acts as right or wrong, but also in such dicta as: “Conscience is a very fallible guide to right,” “It is easier to denounce the shortcomings of others than to amend one’s own vices,” “Good people seldom keep bad company,” “Nice guys finish last.” We learned to reason with these concepts in arguments like the following (the first of which I borrow from Peter Geach): If it is wrong to do a thing, it is equally wrong to induce someone else to do it. But tormenting the cat is wrong. Ergo, getting your little brother to torment the cat is wrong.

Uncle Sam believes arson for profit is morally permissible. But it isn’t. Therefore, you can’t always take Uncle Sam’s word for it that a thing is morally permissible.

If moral judgments are seriously taken to be in any halfway literal sense expressions of emotion, or imperatives, it is impossible to account for the sorts of things that are actually said by perfectly ordinary people who have never opened a volume of moral philosophy in their lives. “If let no one engage in a practice, then let no

Against Ethics

3

one ever induce anyone else to engage in it!” is gibberish. Neither R. M. Hare nor anyone else has ever explained how the antecedent of a conditional could sensibly be regarded as affected by the force of an imperative. “Even though Uncle Sam does not agree with down with a practice, it may still be that down with it!” is likewise ungrammatical. One can’t even begin to formulate an emotivist or prescriptivist reading of “Virtue is its own reward.” Emotivism and prescriptivism deny things ordinary speakers tend to say about moral judgments. Talk of knowledge of good and evil, familiar from Genesis and the M’Naughton Rules, occurs every day when parents scold, “Since you knew it was wrong, why did you do it?” One who says, “Abortion is as wicked as murder,” may meet with the response, “I doubt that very much!” or “Do you sincerely believe that?” or “Do you realize the implications of such a stand?” or “You’re completely off the wall!” Ordinary people in ordinary situations speak of moral judgments as things that can be known, believed, doubted; that can be certain, likely, or improbable; that can contradict or imply each other; that can be true or false. In matters of meaning, you can fool some of the people (philosophers) all of the time, and all of the people some of the time (in idioms), but you can’t fool all of the people all of the time. Emotivism and prescriptivism, by denying that what ordinary speakers regularly treat as assertions are genuine assertions, merely deprive the term “assertion” of its sense. Now, if moral judgments are meant as assertions, might they not still be meant as subjective, relative ones, mere expressions of personal preference? This was not the view of the philosopher who wrote: When a man denominates another his enemy, his rival, his adversary, he is understood to speak the language of self-love. . . . But when he bestows on any man the epithets vicious or odious or depraved, he then speaks another language. (Hume, Enquiry concerning the principles of morals, IX, 6)

Calling Nixon evil means more than adding him to a personal or parochial or partisan enemies list. At least, that is my “linguistic intuition.” One indication that I’m right is the fact that one can speak of, say, Hirohito and Paul IV as having contradictory opinions as to the morality of suicide. In a case of mere subjective tastes, as when the Emperor says, “I like raw fish,” and the Pope replies, “I don’t,” one could not speak of contradiction. Another indication is that people laugh at the relativist analysis of moral judgments. The Devil’s Dictionary of Ambrose Bierce contains the following entry: MORAL, adj. Conforming to the local and mutable standard of right.

This can’t be an accurate transcription of the literal sense of the word, else how could it amuse? Rather, moral judgments are in theory supposed to make objective, absolute claims. The wit of the “definition” lies in its dramatic revelation of the gap between theory and the practice of Bierce’s Victorian contemporaries, who had no firmer basis for their (double) standards than superstition and prejudice. It is rather like another definition in the same work: SELF-EVIDENT, adj. Evident to one’s self and to nobody else.

4

J.P. Burgess

The few positive arguments that have been advanced in favor of relativism are far from conclusive. To take a recent example, consider Gilbert Harman’s distinction between “inner” and “outer” judgments (Harman 1975). He notes there is something odd in saying that Hitler or Attila morally oughtn’t to have killed so many people, while there’s no oddity in calling them evil men. It’s not just that “ought not” is too weak in view of the enormous numbers slain. Rather, it’s somehow out of place in view of the inconceivability of either of these gangsters being influenced by the kind of moral considerations that influence, say, readers of the Philosophical Review. Harman accounts for such puzzling restrictions on the sensible use of “morally ought not” by postulating that moral judgments are meant to be relative to some set of principles of conduct which must be specified, however tacitly and implicitly, in the context of their utterance, and that normally when X tells Y that Z ought to have done A, the principles to which X means his judgment to be relative to are those common to X, Y, and Z. If this relativist analysis is accepted, it is clear why “ought not” judgments are out of place when Z shares few principles with X and Y. Yet one who, like me, holds that moral judgments are meant absolutely, can equally well explain the restrictions on “morally ought not” by, say, attributing to “Z ought not to have done A” the logical form, “Doing A was wrong according to principles whose truth Z himself recognizes.” (Here “recognize” is used in the strong sense in which recognizing the truth of p presupposes p is true.) True, this does not explain why “ought” judgments, unlike “evil” judgments, should be restricted to cases where Z has some correct moral beliefs. But neither does Harman explain why “ought” judgments, unlike “evil” judgments, should be relative to principles shared by X, Y, and Z, and not just those common to X and Y. If there is any explanation, it must be sought in features of “ought” and “should” common to both moral and non-moral uses. One feature, noted by many writers, is the strong advice-giving component of “ought” and “should” judgments. Such judgments answer “shall I?” questions. Saying what Hitler morally ought to do might be likened to vicariously advising him what to do in order to be moral. One doesn’t have to be a relativist to see why such advice would be out of place in the case of someone who is not interested in being moral (or who considers the adviser’s moral opinions so erroneous as to be not worth listening to). Even a Thomist could understand that. “Hitler morally oughtn’t to have killed so many people” is odd, and “Hitler was very evil” quite in order, for the same reason that “Capone legally should not have killed so many people” is odd, and “Capone was a great criminal” in order. In short, the restrictions on the sensible use of “morally ought not” judgments discovered by Harman have more to do with “ought” than with morality, and so are explicable without any special assumptions (relativist, Thomist, anethicist, or what have you) about the nature of morality. I know of no other linguistic arguments for relativism more compelling than Harman’s. The main negative argument that has been raised by relativists against the view that moral judgments advance objective, absolute assertions is that if accepting a moral principle is “merely agreeing about a fact,” then it is impossible to explain the power of moral judgments to motivate. This objection requires careful consideration, and a section to itself.

Against Ethics

5

2 Many philosophers have insisted that some link with emotion and motivation is part of the very meaning of ethical concepts. Consider a patient who, after a normal upbringing, suffers something like a lobotomy. Imagine that the patient says, in all apparent sincerity, that various practices are morally wrong, and yet habitually engages in them without hesitation or regret, without signs of inner struggle or violent passion. The place of “right” and “wrong” in such a patient’s mental life would be far from the norm. Many philosophers would say the patient no longer possessed the standard moral concepts at all. According to prescriptivism, moral judgments are imperatives, so “accepting” a moral principle can only amount to resolving to obey the imperative it represents. According to the relativist theory developed by Harman, belief that an act is morally wrong amounts to belief that it goes against the way people in one’s group are trying to live. Thus these linguistic theories provide a transparent link between moral judgment and motivation, between “reasons” for a course of action in the two senses of “evidence of its moral rightness” and “motivation considerations in its favor.” But is it necessary to accept a theory according to which accepting a moral judgment reduces to having a certain intention, in order to explain how moral judgments motivate? I will argue that it is not – that a strong, even an analytic, link between moral judgment and motivation is compatible with the thesis I have been defending, that moral judgments advance claims about what are supposed to be matters of absolute, objective fact. It may help to consider the concept of religious duty and of violation of that duty, or sin. Surely belief that an act is a sin has a strong inhibitory effect. Indeed, we would hardly know what to make of someone who in all apparent sincerity insisted that contraception was a sin, and yet habitually engaged in it without hesitation or regret. One might be tempted to say such a person did not possess the standard concept of “sin,” just as one was tempted to say the lobotomized patient did not possess the standard concept of “wrong.” And yet belief that an act is a sin surely does not simply amount to intending not to commit it. All sorts of beliefs about what are supposed to be matters of objective fact are involved: belief that God opposes such acts, that He has revealed His will in the matter, and so on. Actually, there should be no mystery about how religious and moral judgments can acquire a double role, both motivating action and making assertions. A child who scribbles on the wall is normally both told that it is naughty, and punished for its naughtiness. The child simultaneously learns to assent to the statement that wallscribbling is bad, and to refrain from scribbling on the wall. It is rather like one of those Suzuki children learning to play the violin: the child simultaneously learns to recognize which notes are properly called “in tune” and learns to play in tune. Nor is it surprising that moral judgments should be clothed in the form of assertions, rather than, say, imperatives, even if, as many philosophers have claimed, the whole point of moral discourse is to guide action. It is not a practical method of conduct-training to wait for children to exhibit each form of undesirable behavior

6

J.P. Burgess

and then “condition” them against repeating it. The usual strategy is quite different. “Negative reinforcement” of such acts as wall-scribbling establishes a tendency for the categorization of an act as “wrong” to induce an inhibition against performing it. Then children can be trained not to, say, commit tax fraud, by telling them it is wrong. Moreover, it is not just particular categorizations such as “Tax fraud is wrong,” but also general admonitions such as “If it is wrong to do a thing, it is equally wrong to induce someone else to do it” and “It is more important to correct one’s own faults than to point out the shortcomings of others,” that have their place in normal conduct-training. And such forms, as we have seen, are only available when moral judgments take the form of assertions. The fact that moral judgments take the form of assertions makes it possible for gaps to develop between even the sincerest belief that an act is a duty and an actual intention to carry it out. As W. V. Quine puts it: It is one thing to learn the difference between right and wrong, and another to suit the action to the word. The one is a matter of learning . . . to talk; the other is a matter of learning . . . to behave. The similarity basis for learning the word “good” is reward, or impunity, and the basis of moral training is the same, but still the lessons learned are different. The child may go on properly applying the word “good” even after the training has stopped. . . . The individual stands to gain nothing from private linguistic deviation. . . . But he may stand to gain much from wicked behavior, once the sanctions are dropped. (Quine 1974, p. 51)

We are all familiar with the existence of such gaps between moral belief and action. A good many people can, like Oscar Wilde, resist anything but temptation. They constantly find themselves doing things “they just know they shouldn’t.” Again, there are episodes, common in adolescence and not unheard of in later life, when angry, rebellious people do things they believe to be wrong because they are wrong. It is bad psychology to deny that people can deliberately and knowingly do things which they believe at the very moment they are doing them to be without qualification wrong. It is a contradiction in terms to say that people can deliberately and knowingly do things which at the very moment they are doing them they unqualifiedly intend not to do. Thus any linguistic theory of ethics which identifies moral judgment too closely with intention must be rejected. The motivating power of moral judgment must not be exaggerated. On the other hand, it would be quite abnormal for the motivating power of moral judgments established in earliest childhood ever to fade away completely. As Quine continues: It is remarkable how successful we often are in training the young to police themselves against their own selfish interests; surely there is a native amenability at the bottom of it all, and it has been favored by natural selection because of its survival value for the race taken collectively. (ibid.)

For the connection between moral belief and action to be entirely erased would be so unusual that I have no quarrel with philosophers who claim it would imply a departure from the conventional “meaning” of “right” and “wrong.”

Against Ethics

7

I have argued that moral judgments can both motivate and be meant as assertions. But why are these assertions taken absolutely and not relatively? Well, all assertions that do not explicitly involve personal pronouns will be taken impersonally and objectively unless it is specifically taught that they are meant otherwise. Even aesthetic judgments are at first taken absolutely, and beauty taken for an objective quality. In sophisticated societies one eventually learns (often not until fairly late in one’s education) that beauty is in the eye of the beholder, that chacun a son goût. There is simply no parallel in moral education to this last phase in aesthetic education. Society has much to gain from encouraging people to tolerate harmless differences of taste, but little to gain from encouraging people to condone wrongdoing in others. Hence the relativity of aesthetic standards becomes part of the standard concept of “taste,” while the relativity of moral standards is no part of the standard concept of “right.” Switching from behaviorist to Freudian jargon, there is another strong reason why moral judgments will be taken objectively. The moral judgments of children, especially, are to a large extent based on the dictates of conscience. Conscience, or “superego,” is a specialized part of the mind, created by “internalization” of parental scolding. It is quite alien to the “conscious ego”; it is not recognized as part of the self. To children, “superego” and “id” are rather like those little angels and devils that whisper in the ears of cartoon characters. All unsophisticated persons tend to “project” the products of their own minds on to the universe. The tribeswoman who dreams of a dead ancestor believes she has had a communication from a ghost; the neurotic relieved of some of his psychosomatic symptoms by suggestion at a faith healing service thinks heavenly powers have cured him. In the same way, hearing the voice of conscience is mistaken for perceiving an external reality through a “moral sense.” As one of the few philosophers who have defended anethicism puts it: We have only moral feelings, but objectify these and think we are recognising objective facts and qualities. (Mackie 1946, p. 90)

Many philosophers have claimed that the whole point of moral discourse is to guide action; and I suppose in some sense this is true. What the foregoing psychological speculations are meant to show is that, even if this is the point of moral discourse, a natural, almost inevitable, process will lead to that discourse taking the form of assertions that are understood absolutely and objectively.

3 When Madelyn Murray O’Hair hears someone say that there are two natures in the incarnation, or that Jackie Onassis is in a state of grace, she doesn’t agree. Yet it would be misleading at best to say that she believes there are not two natures in the incarnation, or that Jackie is not in a state of grace. The famous opponent of school prayer is not a monophysite heretic like Haile Selassie, nor does she harbor a secret suspicion that Jackie has committed a mortal sin. Rather, her belief is that

8

J.P. Burgess

there never was any incarnation, and she rejects the concepts of sin and grace. If the theosophists or the Jehovah’s Witnesses are right, the Hassidim and the Jesuits have many erroneous ideas about God. But if the atheists are right, many of those groups’ beliefs cannot even be called false. I similarly maintain that moral judgments are without truth-value. I regard the question of whether abortion is immoral in the same way that Ms. Murray regards the question of whether it is against God’s law. I have no more use for the concept of moral turpitude than she has for the concept of sin. My reason is this: It is part of the normal understanding of moral judgments that what truth-value they have is supposed to be independent of the person by whom, and the circumstances in which, they are uttered. Yet though each speaker may have quite definite (albeit unconscious) criteria for applying the term “moral,” there is not enough common to all speakers’ criteria to provide “Abortion is immoral” with a speaker-independent truth-value. The relativity of morals is not unlike the relativity of simultaneity. No one, I hope, will pretend that even back in Newton’s day there was lurking in the “deep structure” of statements about the temporal order of events a hidden personal pronoun, a relativization to somebody’s frame-of-reference, that does not appear on the surface. No, such statements were always understood impersonally. Only a close analysis of the actual criteria applied in arriving at such judgments, an analysis first carried out by Einstein, revealed that the criteria of observers in relative motion disagree in their consequences. The discrepancies are too small to be noticed in everyday life; they don’t make people miss appointments. They would become obvious if motion at speeds approaching that of light were feasible. Similarly, I reject the suggestion that even back in Calvin’s day moral judgments carried the connotation of being merely personal and subjective. It was those moral Einsteins, the philosophes of the eighteenth century, who discovered, by analyzing the sources of moral beliefs, that they are commonly derived from conscience and moral sentiment, and that conscience and moral sentiment, alas, dictate quite different things to different people. This could not be noticed by people living in small, isolated, homogeneous, stable communities. Travel to distant lands and life in big cities have enabled a minority, consisting of philosophers, anthropologists, and some other unbiased observers, to detect the relativity of morals. The spectacle of the diverse moral codes left us by stern Moses, gentle Buddha, and a thousand others, makes it apparent that there is not enough common to different people’s criteria for what counts as right and wrong to make any answer to the question, “Is abortion immoral?” the right one. About virtue and justice I would say what Karen Blixen’s character says about God: “Thus,” he said, “with the general human idea of virtue, justice, or, if you will, of God. If anybody were to ask me what was the truth about these things, I should answer: ‘My friend, your question is without meaning. The Hebrews conceived their God like this; the Aztecs of America, about whom I have just read a book, like that; the Jansenites, again, like that. If you want the details of the various views I shall be pleased to give them, having devoted a certain amount of my time to this study. But let me advise you not to repeat your question in intelligent company.’ But at the same time I should be, for this superior view of mine, in debt

Against Ethics

9

to the naive people who have believed in the possibility of obtaining a direct and absolutely truth idea of God, and who were mistaken. For had they made it their object only to create a special Hebrew, Aztec, or Christian idea of God, where would the presupposition of the observer have been found?. . . Indeed, my friend, while the fools could have done without us, we are dependent on the fools for our better knowledge.” (Blixen 1934, p. 383)

It is not hard to understand why people’s criteria of right and wrong differ. Let us imagine that at some remote period the moral principles of each society were calculated to promote some intelligible goal – the survival of the society, say. What would happen as time went on and conditions changed? Values that were once conducive to survival would no longer be so; for example, sexual practices that produced stability in times of low population pressure and high infant mortality could produce disaster under altered circumstances. Yet there is a strong conservative tendency in morals. Even “swinging” parents, we read, often bring up their children the way they themselves were brought up, with old-fashioned, traditional values. Like all human traits, this conservativeness will run stronger in some family lines than in others. After some time it could be expected that a rift would develop between reactionaries bent on clinging to the old ways at any cost, and progressives willing to change their customs in order to promote the old goal of survival under changed conditions. Moreover, different societies would always have had different values, for the character traits that sustain a farming community will not do for a tribe that lives by plundering its neighbors, and the values appropriate for a nomadic existence won’t work in big cities. By the time complex societies developed, absorbing people of the most diverse origins, it would not be the least bit surprising to find people who believe divorce-and-remarriage is as wicked as murder living next door to people who believe monogamy is obsolete. This is what we do find in the modern world. Yet there are some philosophers who insist that au fond everyone from Angela Davis to Phyllis Schlafly shares the same basic criteria of right and wrong. Differences of opinion over abortion, meat-eating, homosexuality, usury, vivisection, censorship, and war are attributed by these philosophers to the difficulty of working out the consequences of shared criteria in concrete cases, and to differences of opinion on matters of non-ethical fact. On this view, we do all share enough common criteria for what counts as right and wrong to make one side in each of these debates objectively the right side. Yet the application of our common criteria may require settling many difficult questions (about, say, the long-range effect of euthanasia or pornography on human survival and happiness) on which at present there is little reliable evidence. If we all do share the same ultimate criteria of right and wrong, it ought to be possible to say what those criteria are; and some philosophers have tried to do just that. The analytic utilitarians, for instance, hold that calling an act morally right means no more than that it promotes the greatest happiness for the greatest number. Unfortunately, objections going back to G. E. Moore show the principle of utility is not a very plausible candidate for the ultimate criterion of morality. For consider the following views:

10

J.P. Burgess Even if it’s true that most people are always going to be unhappy in this world, still it wouldn’t be right for the politicians to put us out of our misery by blowing up the planet. Animals have just as much right as human beings not to be caused suffering. A society in which those who contribute most to human happiness are honored and rewarded is better than one in which the most benefits go to lazy idlers who contribute nothing to their fellow creatures, even if people on the whole are as happy in the one society as in the other. However content the citizens of the Brave New World may be, it would be evil to create such a society.

If analytic utilitarianism is right, then these views are not just mistaken, but are in a class with belief in married bachelors and round squares. I doubt whether even a philosopher would seriously want to maintain that. Other proposed definitive criteria of “good” – for example, “what is most highly evolved,” or “what we desire to desire” – are subject to similar objections. Indeed, there are two very general reasons for believing no universally shared ultimate defining criteria of morality exist. The first is that many people take some quite particular moral judgments as axiomatic or self-evident, and don’t in their own minds deduce them from anything more basic, whatever rationalizations they may offer others. For example, many people find certain sexual practices aesthetically disgusting, and some of them believe that they are, besides, morally wrong, though there is no way to deduce this from general principles. True, such people sometimes say they are against these practices “because they cause blindness,” or “because God is opposed.” Still it would be a misunderstanding of the psychology of these people to suppose that overwhelming medical evidence has convinced them of the deleterious effects of the practices in question on vision, and that therefore they oppose them, or that close textual analysis of the scriptures has convinced them that God opposes these acts, and hence they do. Rather it is their conviction that the practices are evil and deserve to be punished that leads them to give credence to quite unsubstantiated reports about their medical effects. They believe God opposes these practices not because the Bible tells them so – it is shockingly lacking in condemnations of, for example, polygamy – but because they themselves do so. A great many moral judgments seem to have the status of primitive principles in the minds of those who accept them. It is hard to view disagreement over such judgment as disagreement over how to apply some general criterion. All this was noted long ago by Bertrand Russell: The appeal to reason [in Aquinas] is, in a sense, insincere, since the conclusion to be reached is fixed in advance. Take, for example, the indissolubility of marriage. This is advocated on the ground that the father is useful in the education of the children, (a) because he is more rational than the mother, (b) because, being stronger, he is better able to inflict physical punishment. A modern educator might retort that (a) there is no reason to suppose men in general more rational than women, (b) that the sort of punishment that requires great physical strength is not desirable in education. He might go on to point out that fathers, in the modern world, have scarcely any part in education. But no follower of St. Thomas would, on that account, cease to believe in lifelong monogamy. (Russell 1945, p. 462)

Against Ethics

11

A second indication is that both advocates and opponents can be found, right here in our modern, developed, Western, industrial, capitalist society, for some extremely general principles. Jeremy Bentham in all apparent sincerity defended the principle of utility, which nearly every other important moral philosopher has rejected. Lewis Carroll held that people have an absolute right to take the lives of animals, but not to cause them pain. Animal liberationists dissent from the first part of Carroll’s thesis, and all those who are not anti-vivisectionists dissent from the second part. Russell attached supreme importance to the preservation of the human species. Patriots of the better-dead-than-red school think otherwise. It is hard to picture these fundamental disagreements as disagreements over how to apply some still more sweeping, universal criteria that everybody shares. Surely the burden of proof is on those who believe there are such cosmic moral principles to come up with some plausible examples; to show how basic disagreements over whether, say, plants or animals or zygotes have rights, can be viewed as disagreements over how to implement some still more basic agreements; to show how kinky views about, say, sexual matters, can be regarded as derived from general principles. Mere optimistic talk by philosophical Pollyannas cannot conjure away the very real differences in ideals and standards that exist in our present divided world and society.

4 Some French philosophers of the last century, we read, were so appalled by the idea of a person without a religion, that they would say to all the atheists they met, l’athéisme c’est donc votre religion. Imbeciles have confused atheism with paganism and Satanism. People who ought to know better sometimes label whatever beliefs an atheist holds most dear his or her “religion”; thus we hear Marxism, psychoanalysis, surrealism described as “religions.” Atheists themselves sometimes resort to quasi-religious terminology in describing their deepest feelings, trying to give secular senses to “mystic” and “sacred.” The correct thing to say about those who reject every form of supernatural belief from Anabaptism to Zoroastrianism is that they have no religion. In the same way, I have no moral beliefs. There are some people, called sociopaths, who lack all conscience and community feeling. They are delinquents and criminals and are generally held in low esteem by law-abiding citizens. Yet one school of moral philosophers think we should all be like them. These are the nihilists, whose motto is “Fais ce que voudras!” A related school, the immoralists, urge doing whatever prigs and prudes find shocking. “Épatez le bourgeois!” is their slogan. Those who adopt anethicism, the view I have been defending, by no means agree with nihilists and immoralists. Indeed, only the sort of people who confuse atheism with Satanism could mix up anethicism with nihilism or immoralism. There is no reason why anethicists should not have personal ideals and standards without the intellectual baggage of moral belief that usually accompanies them. The atheist Russell, after all, could share a number of attitudes and character traits with

12

J.P. Burgess

Quakers and Buddhists, without sharing their theology. Indeed, he approached much more closely to the ideals of these religions than the ostensible believers Nixon and Lon Nol. What anethicists like me object to is the language of morals. I am willing to speak of personal ideals and standards, and of group mores and ethos, but not of morality and ethical conduct. The ethical vocabulary, I believe, carries an implication, or at least an inevitable suggestion, of a claim to objective and absolute, or at least intersubjectively valid and impersonal, backing for one’s standards. Nobody has such backing. If some day people all over the world came to share the same values and goals, we could begin to assign “right” and “wrong” speaker-independent, intersubjectively valid extensions. Until that day, those of us who have appreciated the relativity of morals have several options: (1) We might use “right” and “wrong” in a least common denominator sense, calling “immoral” what nearly everybody’s standards forbid, and “permissible” what nearly everybody’s standards allow. (2) We might, each of us, use “right” and “wrong” relative to his or her personal standards, while explicitly repudiating any intention to universal, impersonal claims when using these words. (3) We might drop the terms “right” and “wrong” from our vocabularies. Similarly, after Einstein one might use “earlier” and “later” either for relativistic causal time, or use them relative to one’s own proper time, or drop them entirely, speaking always of “intervals” rather than distances or durations. In ethics, course (1) would require us to grant that abortion, for instance, is neither immoral nor morally permissible. We might be able to say that it would be immoral to induce abortion in a case where the woman involved did not desire it and would be medically damaged by it, and the fetus if allowed to come to term would become a normal, healthy infant, and there was no danger of overpopulation. We might be able to say that if abortion were morally permissible in one case, it would be equally permissible in every similar case. For that much might be common to everybody’s notions of morality. But we certainly could not hope to get a verdict of “immoral” or “permissible” on abortion in general. In physics, the course of using “earlier” and “later” relative to one’s own timeframe can be safely adopted outside the laboratory, because the differences between frames of reference for people moving at normal speeds on the surface of the earth are negligible. In ethics, the situation is rather different. On course (2) we would have to be constantly reminding ourselves that those of us who say abortion is wrong, and those who say it is all right, are not contradicting each other. Either course (1) or (2) would be bound to cause great confusion so long as the majority has not appreciated the relativity of morals. And I doubt whether today more people are aware of that relativity than understand Einstein’s theory. Moreover, it is really rather incongruous to go on speaking of “right” and “wrong” once one realizes there are no absolute standards. It is rather like a convert to atheism continuing to speak of those who die as “going to meet their Maker.” Therefore I advocate course (3), giving up ethical vocabulary. In any case, I emphasize that adopting either course (1) or (2) would be a major departure from the way in which the ethical vocabulary has conventionally been understood.

Against Ethics

13

Music critics don’t write for the sort of people who prefer muzak to music, Bacharach to Bach. Nor need moral philosophers write for those who take no interest in human survival and happiness. For those of us who do, adding the epithets “vicious” or “depraved” to a factual account of Nixon’s or Hitler’s career adds nothing, or nothing but a misleading suggestion that the universe disapproves of these thugs as much as we do. The human race has no agreed-upon purposes, aims, goals – no, not even its own survival – no common values and ideals, no shared principles and standards. Nor does the universe take sides in human affairs. There’s no use kidding ourselves.

Postscript (2007) I had, while a student, read the standard writers on metaethics from the first half of the last century. Personal experience as an adolescent in the 1960 s left me disposed to find the cognitivist theories very dubious, and the linguistic theories quite absurd. I knew that those making moral judgments meant them as statements of fact, but after having changed my own mind so many times, and having encountered so many so passionately convinced of so widely divergent views, I very strongly doubted that there were any facts to be stated. It seemed to me, however, that the linguistic theorists had sufficiently refuted the cognitivists, and vice versa, so that there was nothing left to add. Shortly after I arrived in Princeton in the fall of 1975, Gil Harman received and distributed to his colleagues reprints of his 1974 article on moral relativism. I at once recognized that his position, though one I would class as a sophisticated member of the linguistic family, was one member of that class that escaped, among others, Geach’s objections to Hare. I still could not accept the view, but I did not think the existing literature from the other side contained a refutation of it, so I concluded that it might be worthwhile for me to write out a statement of my own reservations. A semester’s leave in the fall of 1977 gave me an opportunity to complete a draft. But as I did so, youthful indiscretion carried me beyond my original intention of writing just a response to one recent article, and led me to produce a position paper or manifesto vigorously stating, and sketchily indicating arguments for, a view I thought unrepresented in the literature, one rejecting of cognitivist and linguistic theories alike. I had no sooner completed a draft that satisfied me than I came across Mackie’s Ethics in a book store. So the view was represented in the literature after all, and actually had been so, as I learned, since before my birth, with Mackie’s 1946 article “A Refutation of Morals.” I at first felt Mackie’s work made my own efforts entirely superfluous, but in the end I concluded that there were enough differences between Mackie’s more thoroughly worked-out position and my own sketchy one to leave some role for my paper, though naturally it had to be revised to include at least some citation and quotation of Mackie. My senior colleagues Gil Harman and Tom Nagel both (being perhaps more amused by the now-dated topical humor than edified by the philosophy) encouraged

14

J.P. Burgess

me to publish the revised version. Over the next few years I made several attempts to do so, and accumulated several rejection notices before I gave up. (I became the more content to do so when I came into communication with Ian Hinckfuss, and knew that he was working to get his To Hell with Morality into print. He eventually succeeded, though the work appeared only after much delay and under a tamer title.) I also accumulated third and fourth and nth drafts of “Against Ethics,” since I revised between each rejection and the next submission. I returned to metaethics briefly about 10 years after my original effort. In the late 1980s I gave a talk under the title “Ethics as an Intellectual Discipline” at Princeton and a couple of other places. I will not outline the talk here, except to say that I began with the claim that the important issue is not the existence of moral facts. This already represented a shift, at least in angle of attack, since “Against Ethics,” and I have shifted more since. To protect myself from the temptation to waste time trying to publish, I never reduced the notes from which I spoke to the form of a paper. By now I have lost even the notes, and retain only a fading outline in memory. By now I have lost also all versions of “Against Ethics” except that reproduced above, and even for it the last page, with the footnotes identifying the sources of quotations, was missing and had to be reconstructed. I believe it to be my second draft, the first produced after Mackie’s book came to my attention and the first to be submitted to and rejected by a journal. If so, it should date from 1978. But my aging memory is very fallible. Despite changes of view on my part, I have never refused requests for copies of the paper, so other versions may be in circulation. I remember one with a different epigraph, an untranslatable French pun, in place of the Apostate’s “I read, I understood, I condemned”; I remember one, possibly the same one, with a three part thesis-antithesis-synthesis structure; and I remember one, again possibly the same one, with an explicit statement of the Puritan syllogism, to illustrate how moral predicates may be used in intermediate steps of deductions from non-moral premises to non-moral conclusions: “Whatever causes pleasure without hard work is evil. Whatever is evil in the end is injurious to health. Ergo, whatever causes pleasure without hard work in the end is injurious to health.” I would appreciate hearing from anyone who may have a copy of any of these alternative versions. Again despite changes of view on my part, I am content to have the paper appear in print at last in the present venue. Though I would probably have declined any other offer to publish had I received one, how could I refuse to contribute to a volume in honor of Mackie? My own occasional historical research has convinced me of the badness of the practice of altering papers when reprinting, and even though the present appearance of my old paper is not exactly a reprinting, I have made no changes in transcribing my one remaining old manuscript, except to correct typos, eliminate the rather eccentric capitalization, and regularize the punctuation. As to the doctrines the paper contains, I would today still defend the principle, illustrated in the paper with a quotation from Bierce, that a definition that strikes people as a joke is unlikely to be a correct analysis.

Against Ethics

15

References Blixen, K. (pseud. Isak Dinesen). 1934. The poet. In her Seven gothic tales. New York: Random House. Harman, G. 1975. Moral relativism defended. Philosophical Review 84:3–22. Mackie, J. L. 1946. A refutation of morals. Australasian Journal of Psychology and Philosophy 24:77–90. Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin. Quine, W. V. 1974. The roots of reference. La Salle: Open Court. Russell, B. 1945. A history of western philosophy. New York: Simon and Schuster.

Nihilism, Nietzsche, and the Doppelganger Problem Charles R. Pigden

1 Introduction Let me start with two claims: (1) I am a moral nihilist, (2) so was Nietzsche. The first claim is not particularly controversial. Absent brain injury or massive selfdeception, I am the best authority on what I believe, and I can assure you that I believe something that can reasonably be described as “moral nihilism,” namely a minor variant of the error theory of J. L. Mackie. It is otherwise with the second claim. Some say Nietzsche was a nihilist, indeed a perfect, complete or uninhibited nihilist.1 Some say that this is all a horrible misunderstanding and that Nietzsche was nothing of the kind.2 I shall argue that he was a nihilist in much the same sense as I am, but variously a diagnostician, an opponent, and a survivor of certain other kinds of nihilism.3 Then, with Nietzsche’s aid, I shall defend the moral nihilism that we both believe (metaethical nihilism or the error theory) against a common line of criticism that nihilism can’t be true because if it were we would have to give up morality or, at least, moralizing. I then raise a problem (the Doppelganger Problem) for metaethical nihilism, reinforce the problem, and solve it by reformulating the doctrine. Thus although I think that trying to get Nietzsche right is a worthwhile intellectual enterprise, the real point of the paper is to vindicate the error theory (of which Mackie was the foremost defender) against certain kinds of criticism. For the record, I agree with (what I take to be) Nietzsche’s metaethic but disagree with his ethic of Calliclean self-assertion. I also think that there are lots of historical, psychological and philosophical insights scattered through Nietzsche’s works (mixed with a good deal of silliness) though they don’t always repay the effort of putting up with his big-noting, his button-holing, and his “Hi Ma! Look at Me!” style of writing.

C.R. Pigden (B) Department of Philosophy, University of Otago, New Zealand 1 See

Ansell-Pearson (1994), especially Chapter 10; Danto (1965), pp. 22, 30, 31. (1995, Chapters 2 and 3). 3 Thus my interpretation is broadly in accord with that of Brian Leiter (2002). 2 Schacht

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_2, 

17

18

C.R. Pigden

2 Metaethical Nihilism So, in what sense am I a moral nihilist? I think (as a first approximation) that moral judgments, specifically moral judgments concerning the thin moral concepts (“good,” “right,” “ought,” “wrong,” etc.), are propositions, that they are (in the current jargon) truth apt. And I think that they are all false.4 For there are no such properties as goodness, badness, wrongness or obligatoriness. You can’t do genuinely good deeds since there is no such property as goodness for your deeds to instantiate: at best they can be good in some watered down or ersatz sense. With the thick moral concepts (“honest,” “kind,” “spiteful” or “loyal”) the situation is more complex. Judgments such as “Abigail is honest” can be true so long as they are construed factually as describing Abigail’s propensity to tell the truth, refrain from falsehood, and stick to her word. If Abigail does indeed have these characteristics, then “Abigail is honest” is true just as “George Bush is honest” is not (since the illustrious President has a free and easy way with the facts). But “Abigail is honest” is false if it is taken to imply that being honest is a good thing, virtuous, or how one ought to be, and hence that Abigail is good, virtuous, or that she does what she should. In other words, judgments involving the thick moral concepts can be true so long as the thin coating of evaluative content is scraped away. But if such judgments are freighted with thin evaluative content they too are condemned to falsehood. Moreover, judgments involving the thin and the thick concepts can both be true, so long as they are construed sociologically or in an “inverted commas” sense. It is true (I hope) that Abigail is a good girl since she conforms (on the whole) to my paternal ethic. It is true that Achilles is agathos, even though he procures the defeat of his own side because of a quarrel about a slave girl, since his actions still conform to the heroic code. But Abigail is not absolutely good nor is Achilles absolutely agathos: Abigail is good-according-to-her-Dad and Achilles is agathos-according-to-the-heroic-code. These pedantic and R. M. Hare-ish distinctions are quite useful when it comes to deciphering Nietzsche’s thought. For example, when he says that to become moral is not itself moral (Nietzsche 1881, §97/1982, p. 59), what he means is that the motives and characteristics which induce people to subscribe to a morality (like the methods people use to propagate a morality) may qualify as moral failings according to that very morality. One’s commitment to virtue may be due to an “inverted commas” vice. To the best of my knowledge, the first clear formulation of metaethical nihilism or the error theory in the twentieth century was due to Bertrand Russell who propounded it at a meeting of the Cambridge Apostles in 1922: “There seems to me no doubt that our ethical judgments all claim objectivity; but this claim, to my mind, makes them all false” (Russell 1922/1999, p. 123). Since “the Society” (as the Apostles were known) was a secret society, Russell’s paper did not have much of an

4 Though

as we shall see, this has to be carefully qualified to avoid self-refutation.

Nihilism, Nietzsche, and the Doppelganger Problem

19

impact at the time. The true champion of metaethical nihilism in the twentieth century was J. L. Mackie. “Although most people in making moral judgments implicitly claim . . . to be pointing to something objectively prescriptive, these claims are all false” (Mackie 1977, p. 35). His “Refutation of Morals” (1946) put nihilism on the agenda, and his Ethics: Inventing Right and Wrong (1977) reminded people that it had not been dealt with. Since then, numerous efforts have been made, so much so that Michael Smith’s recent anthology Metaethics (1995) largely consists of desperate attempts by various philosophers to find facts that will make moral judgments true. (In the circumstances, Smith’s omission of Mackie makes the book not so much Hamlet without the Prince of Denmark, as Hamlet without the ghost.)

3 Nietzsche and Nihilism But was Nietzsche a nihilist? Well, absent brain injury or massive self-deception, he was the best authority on what he believed, and he certainly said he was a nihilist – indeed, “the first perfect nihilist in Europe” (Nietzsche 1883–1888/1968, p. 3). But this does not prove the point. To begin with, in his case, we cannot altogether discount the possibility of brain injury. He wrote these words about a year before his descent into madness, and the syphilis was beginning to undermine his intellect (the conceit and the megalomania were becoming more pronounced, the style less ironic and more abusive). Moreover, he often says things that suggest that he is not a nihilist. In Twilight of the Idols, “Maxims and Arrows,” he quotes Flaubert’s “One cannot think or write except when seated,” and exclaims “There I have caught you nihilist! Sitting still is the very sin against the Holy Spirit!” (Nietzsche 1889, II.34/2005, p. 160). Since Nietzsche had his best ideas whilst wandering lonely as a cloud through Alpine holiday resorts, the conversational implicatum would appear to be that, unlike Flaubert, Nietzsche was not a nihilist. Even Nietzsche’s claim to be the perfect nihilist is rather equivocal since his “perfection” consists in the fact that he has lived through nihilism and come out on the other side, “leaving it behind, outside himself.” In other words, the reason he is the perfect nihilist is that he is not a nihilist any more. Against this, Nietzsche’s writings are peppered with passages that suggest, imply or express metaethical nihilism. Daybreak §103 is explicit: “To deny morality” . . . can mean to deny that moral judgments are based on truths. Here it is admitted that they really are motives of action but that in this way it is errors which, as the basis of all moral judgments, impel men to their moral actions. This is my point of view.

The Genealogy of Morals is, in part, an attempt to demonstrate that since the current “slave” morality would have been believed even if it were false, the fact that it is widely believed gives us no reason to think it true. But the same argument quite obviously applies to the aristocratic morality that Nietzsche evidently prefers. That too was believed because it suited the needs of the aristocracy; hence the fact that it was believed affords no argument for supposing it to be true. But since moral claims purport to be truth apt, it follows that if they are not true (which is what Nietzsche

20

C.R. Pigden

seems to suggest) then they are false. Nietzsche’s Twilight of the Idols expresses the same idea in less metaphorical language: There are absolutely no moral facts. What moral and religious judgments have in common is the belief in things that are not real. Morality is just an interpretation of certain phenomena or (more accurately) a misinterpretation. (1889, VIII.1/2005, p. 182)

So how do we reconcile the rejection of nihilism with these nihilistic sentiments? Nietzsche was in my sense a metaethical nihilist – that is, an error theorist in the style of John Mackie – but the nihilism he claimed to have survived, the nihilism he regarded as a menace, the nihilism he hoped to transcend, was not just the belief that moral judgments are all false but the psycho-social malaise caused by this belief. “Scepticism regarding morality is what is decisive,” says Nietzsche, adding “The end of the moral interpretation of the world which no longer has any sanction after it has tried to escape into some beyond, leads to nihilism. ‘Everything lacks meaning’” (1883–1888/1968, p. 7. my emphasis). Obviously, if moral skepticism leads to nihilism it cannot be identical with the thing that it leads to. To be a nihilist, then, is not just to believe in the error theory but to believe in the error theory and to feel bad about it. Thus Nietzsche has ceased to be a nihilist (or has perfected his nihilism) not because he has ceased to be an error theorist but because he has ceased to feel bad. His project is the revaluation of all values, the reconstruction of a new morality, the morality of the overman, which, though equally false, will be more bracing, more life-enhancing and more conducive to the “higher” type of man. It will have a higher utility (given Nietzsche’s elitist and eccentric ends), but, since there are pernicious truths and useful (even life-enhancing) falsehoods, the fact that it is useful won’t make it true. Morality for Nietzsche is like mathematics for Hartry Field (1989): it does not have to be true to be good – in the sense of being good-for-something (though of course he believed that current moralities are neither true nor good, at least not good for anything which Nietzsche himself valued). If we believe (or make-believe) in this new morality, the world will be re-enchanted; it will become meaningful again. For as Zarathrustra makes plain, meaning is not something to be discovered but something to be imposed, and it is imposed by our moral beliefs: Only man placed value in things to preserve himself – he alone created a meaning for things, a human meaning. Therefore he calls himself “man” which means “the esteemer” . . . without esteeming the nut of existence would be hollow. (1883, I “On the thousand and one goals”/1954, p. 171)

So if the nut of existence seems to be hollow, if we suffer from a malaise of meaninglessness induced by moral skepticism, the solution is to forget our doubts (except perhaps when we are doing metaethics) and to create a new morality by “esteeming” a novel collection of goods. Of course the new morality will be immoral according to current norms, which is why Nietzsche calls himself an immoralist. But a new morality is not no morality, nor is it any less of a fiction because it conduces to the (rather vague) ends that Nietzsche has set himself.

Nihilism, Nietzsche, and the Doppelganger Problem

21

4 Schacht’s Objections Richard Schacht (1995, pp. 49–61) will have none of this. In his view, Nietzsche can’t have been an “axiological [or metaethical] nihilist.” Nietzsche may have been an “immoralist” but he was also a moralist, the advocate of a new, more healthy ethic based on the “will to power.” He felt that unless this new “interpretation” were adopted, European civilization was moribund. He “held that this must not happen and that life [presumably the life of the European elite] ought to flourish, ought to be enhanced and ought to continue to develop” (Schacht 1995, p. 53). Nietzsche could not have “held” this if he thought all moral judgments (including his own) were false. Schacht’s argument is interesting since it is closely related to a family of arguments designed to prove, not that Nietzsche was not a nihilist, but that nihilism in its metaethical form is false. As I understand it, it goes something like this: S1 If anyone thinks that all moral judgments are false he must (a) give up moralizing (i.e., making and defending moral claims, adopting moral beliefs) and (b) give up morality (i.e., acting on the basis of moral beliefs). S2 Nietzsche did not give up either moralizing or morality. On the contrary, he was a dedicated advocate of the ethic of the overman. Therefore: S3 Nietzsche was not an axiological nihilist: he did not believe that all moral judgments are false. Now, the problem with this argument is that premise S1 is clearly untrue. Some metaethical nihilists – such as Richard Garner (1994 and this volume) and Ian Hinckfuss (1987) – give up (or profess to give up) both moralizing and morality, and some – such as Mackie, Richard Joyce (2001), and myself – do not. (Strictly speaking, what Joyce recommends is that we continue to act morally and to think in moral terms whilst abandoning belief in the propositions of morality.) The latter part of Mackie’s Ethics: Inventing Right and Wrong is largely devoted to moralizing, and Mackie is described by his erstwhile neighbor, George Cawkwell, as one of the most duteous persons he had ever met (Cawkwell 1985, pp. 219–220). And though Mackie missed out on the Chair at the University of Tasmania because of his nihilistic metaethical opinions, it was his successful rival, the moralistic Sydney Orr, who was later dismissed from the post for “gross moral turpitude” (Pybus 1993, pp. 206–207). As for me, I am not only addicted to moralizing but I sometimes even act on my principles, which means that I too am a counterexample to S1. But perhaps I have misconstrued Schacht’s argument. The “must” in S1 is not an alethic but a deontic operator. It is not that the nihilist must give up both moralizing and morality (that he cannot help himself, as it were) but that in some sense he ought to do so. Really the argument should start from:

22

C.R. Pigden

S1 If all moral judgments are false then we ought to (a) give up moralizing and (b) give up morality. But of course this premise is at a considerable distance from Schacht’s desired conclusion. For even if S1 is true, Nietzsche may not have realized it, and even if he did realize it, the fact that he did not give up moralizing does not prove he was not a nihilist. For he may not have done what he thought he ought to do. Still, if we assume that S1 is obvious to all philosophically sensitive persons we might arrive at the following: S2 Nietzsche thought that if moral judgments were all false, he ought to (a) give up moralizing and (b) give up morality. Then, if we assume that Nietzsche generally did what he thought he ought to do, the fact that he did not give up either moralizing or morality indicates that he did not think he ought to do so. In which case he did not believe that all moral judgments were false. The trouble is that S2 itself is palpably false. For Nietzsche did not believe that, in general, people should give up or cease acting upon false beliefs: The falseness of a judgment is for us not necessarily an objection to a judgment. . . . The question is to what extent it is life preserving, species preserving, perhaps even species cultivating. And we are fundamentally inclined to think that the falsest judgments (which include synthetic judgments a priori) are the most indispensable for us. . . . To recognize untruth as a condition for life – that certainly means resisting accustomed value feelings in a dangerous way. (1886, §4/1968, pp. 201–202)

But if untruth can be a condition for life – if falsehoods can be indispensable – then Nietzsche would probably have rejected S1 . For if falsehood is not necessarily an objection to beliefs in general – if indispensability (for certain purposes) is an excuse for falsehood – then it may be a sufficient excuse for moral beliefs. So long as they are “species preserving” and all the rest of it – which Nietzsche’s overman ethic was supposed to be – moral beliefs need not be given up. Hence Schacht’s objection collapses.

5 Wright, Blackburn, and the Hoi Polloi But even if Schacht fails to prove that Nietzsche was not a nihilist he may provide the makings of an argument against nihilism proper. For many people think that if nihilism is true, then we ought to give up either moralizing or morality. And this is somehow supposed to be an objection to nihilism, i.e., a reason to think that it is false. Sophisticated philosophers such as Simon Blackburn and Crispin Wright focus on the alleged duty to give up moralizing, whereas simple folk focus on the duty to give up morality. To give up moralizing is to give up the practice of moral discourse as currently constituted: to give up making, defending and arguing for moral claims and to give up our distinctively moral beliefs. To give up morality is to give

Nihilism, Nietzsche, and the Doppelganger Problem

23

up our moral beliefs and to give up acting (and getting other people to act) on the basis of moral beliefs. It is not or not necessarily to give up the practices sanctioned by morality. There may be other reasons for not coveting my neighbor’s ox besides the alleged fact that it is morally wrong. (For instance: coveting leads to theft, and theft is illegal, and illegal actions risk punishment, thus coveting is imprudent.) But the simple folk are surely right in supposing that if our beliefs suffered such a seachange our practices would not remain unaltered. If considerations of duty ceased to motivate I would not do quite the same things as I do now, even if I refrained from the excesses of a Sid Vicious. Morality is not a redundant institution. However, Blackburn seems to suppose that we could give up moralizing without giving up morality. For he seems to think that we might substitute “shmoral” beliefs for moral beliefs and “shmoral” claims for moral claims, and that this “shmorality” might play much the same role as morality does at present (though, as we shall see, he regards such a substitution as silly). Thus giving up moralizing does not entail giving up morality but only altering it by replacing moral beliefs with shmoral beliefs. According to Blackburn: If a vocabulary embodies an error [especially, he seems to think, an error which infects it with falsehood] it would be better if it were replaced by one that avoids the error. . . . Surely it would be better if we avoided moral (erroneous) views altogether and contented ourselves with some lesser, purged commitments that can be held without making metaphysical mistakes. . . . The puzzle is why, in the light of the error theory, Mackie did not at least indicate how a moral vocabulary would look and why he did not himself go on to shmoralize not to moralize. And in my view this is enough of a puzzle to cast doubt back on to the original diagnosis of error. In other words, it would be a silly thing to do, to try to substitute some allegedly hygienic concepts for the moral ones; but that in itself suggests that no error can be incorporated in mere use of those concepts. (1993, pp. 149–150)

Blackburn’s argument then is this: B1 If moral judgments are all false, we ought to give up moralizing (and shmoralize instead). B2 But it is not the case that we ought to give up moralizing (and shmoralize instead). This would be “silly.” Therefore: B3 It is not the case that all moral judgments are false. Blackburn goes on to contend that the reason moral judgments are not all false is that strictly speaking they are neither true nor false (since their true purpose is to express attitudes) even though (for various subtle and complicated reasons) it is OK to call them true or false in common parlance. However, I am not concerned with Blackburn’s defense of quasi-realism but only with his critique of nihilism. Crispin Wright has something similar in mind: The great discomfort with such an [error theoretic] account [either about morality, maths or “the comic”] is that, unless more is said, it relegates discourse about the comic to bad faith. [Not such a calamity in the case of the comic, one is inclined to say!] . . . [A]s soon as

24

C.R. Pigden philosophy has taught us that the world is unsuited to confer truth on any of our claims about the funny the reasonable response ought surely to be to forgo making any such claims. This would not be to forgo the right to laugh [That’s a relief!] . . . But it would, apparently be to forgo the point of reasoned appreciation and debate about what is funny and of criticism of others’ opinions about it. And these consequences are most calamitous . . . within moral discourse. If it is of the essence of moral judgment to aim at the truth, and if philosophy teaches that there is no moral truth to hit, how do I take myself seriously in thinking the way I do about any issue which I regard as of major moral importance? (1992, pp. 9–10)

The obvious and unkind comment is that a man who doesn’t have a problem taking himself seriously when engaged in reasoned appreciation and debate about the comic need not fear that even nihilism will induce a lapse of seriousness when it comes to moral issues. But satire aside, Wright does have an argument, though it would appear to be one rung below Blackburn’s on the logical ladder. For Blackburn’s argument, whatever its faults, is at least formally valid. The same cannot be said for the argument that Wright propounds. Wright’s first premise is much like Blackburn’s: W1 If moral judgments are all false, we ought to give up moralizing (and take our moral commitments a lot less seriously). But his second premise is simply this: W2 Giving up moralizing would be a real calamity. We could not take our moral commitments seriously. And from this nothing in particular seems to follow. At best W1 and W2 when taken together provide us with a reason for wishing that nihilism were not true but not with a reason to think that it is not true. However, we can patch up Wright’s argument by substituting W2 for W2: W2 It is not the case that we ought to give up moralizing (for it would lead to a loss of moral commitment). And W1 and W2 do indeed entail the desired conclusion: W3 It is not the case that all moral judgments are false.

6 Nihilism and its Consequences Having carefully distinguished between the two philosophers and the hoi polloi, and between one philosopher and the other, in what follows I am going to run them all together. What is wrong with this line of argument does not depend on the details we have distinguished, so the premise I start with combines three lines

Nihilism, Nietzsche, and the Doppelganger Problem

25

of thought: that of Blackburn, that of Wright, and that of the simple folk. It is this: 1 If all moral judgments are false then we ought to (a) give up moralizing and (b) give up morality. The rest of the argument runs as follows: 2 It is not the case that we ought to (a) give up moralizing and (b) give up morality. Therefore: 3 It is not the case that all moral judgments are false. This argument is formally valid. But is it sound? This depends upon which “ought”s we have in mind. Let us take premise 1. Is it true? Not if the “ought” is moral, such that 1 becomes: 1 If all moral judgments are false then we morally ought to (a) give up moralizing and (b) give up morality. For if all moral judgments are false, then there is nothing that we morally ought to do, including giving up moralizing or giving up morality. So on this interpretation the argument fails. But not all “ought”s are moral “ought”s so this does not settle the matter. Let’s try again. Suppose we treat the “ought” in 1 as a hypothetical “ought” so that it expresses what we ought to do if we want to achieve some end. And let us suppose to that this end has something to do with rationality. (When Blackburn says that if Mackie were right, it would be “better” to give up moralizing in favor of shmoralizing, he seems to mean rationally better. Similarly his justification for the second premise is that shmoralizing would be silly – i.e., irrational. As for Wright, he explicitly says that “the reasonable response” to an error theory is to give up making the erroneous claims.) One such end is truth. Rationality is often defined with respect to truth so that a rule or procedure is rational if it tends to result in true beliefs either in fact or under the appropriate conditions. So perhaps the idea behind premise 1 is that if all moral judgments are false, we rationally ought to give up both moralizing and morality, i.e., that we ought to give them up if we want to have true, as opposed to false, beliefs. This gives us: 1 If all moral judgments are false, then if we want our beliefs to be true (and not false) we ought to (a) give up moralizing and (b) give up morality. Even if, like Nietzsche, we do not regard truth as the supremely rational end, if nihilism is true, we must still give up moralizing and give up morality if we are to achieve truth. For if truth demands that we give up moralizing – which means giving up our moral beliefs – it also demands that we give up acting on our moral beliefs

26

C.R. Pigden

since we won’t have any moral beliefs to act on. But of course, the truth of 1 does not settle the matter. The second premise – revised so as to maintain validity – has to be true too: 2 It is not the case that if we want our beliefs to be true (and not false) we ought to (a) give up moralizing and (b) give up morality. Now, 2 might be true. It might be that it is not the case that if we want our beliefs to be true (and not false) we ought to give up both moralizing and morality. But 2 will only be true if nihilism is false. For if nihilism is true then we can’t have moral beliefs without having false beliefs. Conversely, if we can have moral beliefs without having false beliefs then nihilism is not true. Indeed, so long as moral judgments are truth apt, the truth of 2 is tantamount to the falsehood of nihilism and vice versa. Thus 2 begs the question. It cannot provide an independent reason for supposing nihilism to be false since it more or less amounts to the negation of nihilism. Hence, this version of the argument is valid and may even be sound, but it is not rationally persuasive since it begs the question against nihilism. Perhaps we should try another tack. It is not that if moral judgments are false we ought rationally give up moralizing, where rationality has something to do with truth or consistency. Rather, if all moral judgments are false we ought pragmatically to give up moralizing. The idea is that there are ends, perhaps human ends, that are best achieved by giving up both moralizing and morality. This is in fact the view of Hinckfuss and Garner (and briefly, perhaps, of Bertrand Russell (1922/1999, pp. 184–188)). They are what I call “humanistic amoralists.” They don’t just think that moral judgments are false – they think that they are pernicious falsehoods which serve as a prop to tyranny and an excuse to torturers. Whether the humanistic amoralists are right and giving up morality would lead to a bonanza of tolerance, freedom and equality is a decidedly moot point. But there are presumably some ends that would be furthered by giving up morality and moralizing. Never mind what they are, let us just designate them as X. Then 1 will be true (indeed true by fiat): 1 If all moral judgments are false, then, if we are to achieve ends X, we ought to (a) give up moralizing and (b) give up morality. But the trouble is that this is trivially true, not just in the sense that we have defined ends X as the ones that would be achieved by giving up morality, but because the conditional is true in virtue of its consequent alone. The falsehood of all moral judgments has nothing to do with the fact that abandoning morality would be conducive to ends X. The “if” expresses no dependency of the consequent on the antecedent. Which means that 2 is simply false: 2 It is not the case that, if we are to achieve ends X, we ought to (a) give up moralizing and (b) give up morality.

Nihilism, Nietzsche, and the Doppelganger Problem

27

Hence, this version of the argument is valid but unsound. My conclusion is that the 1, 2, 3 line of argument cannot be made to work. If the “ought” is moral, the first premise is false. If the “ought” is rational, the first premise may be true, but the second premise presupposes the falsehood of nihilism which means that the argument begs the question. If the “ought” is pragmatic then the first premise may be true, but if so, it is true in virtue of its consequent, which means that the second premise, which consists in denying the consequent, is false. In my view, the argument derives its appeal from an unconscious equivocation. Premise 1 is read with a rational or truth-related “ought” but premise 2 is read with either a moral or a pragmatic “ought” (thus if all moral judgments are false we ought rationally to shmoralize rather than moralize, but shmoralizing would be pedantic and inconvenient and hence silly, i.e., something we are not pragmatically obliged to do). Appealing as it is, this line of argument does not show that nihilism is false.

7 The Doppelganger Problem So far, so good. Metaethical nihilism has survived the objections of Blackburn and Wright. But we now come to another problem: the Doppelganger Problem and its supercharged variant, the Reinforced Doppelganger. The Doppelganger Problem in its simple form is this: It seems that not all moral judgments can be false, for (in many cases at least) the negation of a moral judgment, X, is itself a moral judgment. And if X is false, its negation not-X must be true. But the error theory is precisely the thesis that all moral judgments are false (at least with respect to their core moral contents). So the error theory or metaethical nihilism is false; indeed, incoherent. The problem is a general one, which afflicts error theories of all sorts. According to Geoffrey Sayre-McCord’s famous taxonomy (1986), a realist about a domain of discourse K is someone who believes two things: (a) that K-statements express propositions (that is, are truth-apt) and (b) that some of them are literally true (that is, true when construed literally). Conversely, an antirealist about a domain K is someone who either (a) denies that K-statements are really propositions, truth-apt, true-or-false (at least with respect to their core meanings) or (b) insists that all of them are false. Thus with respect to ethics, emotivists and other noncognitivists are type-a antirealists, and nihilists or error theorists are type-b antirealists. SayreMcCord rather hoped that his taxonomy would be not only neutral but fair in a certain sense; it was designed to map out a series of positions that could be consistently (if not sanely) held. But if type-b antirealism is not such a position then his taxonomy is in deep trouble. And the Doppelganger Problem suggests precisely this. For it seems to show that there are not, or at least, that there should not be, any type-b antirealists about any domain (including ethics). And the reason is that typeb antirealism is incoherent, and thus collapses into straightforward Sayre-McCord realism.

28

C.R. Pigden

Here’s why: If the membership of a domain of discourse K is closed under negation, then, if X is a statement of kind K, its doppelganger ~X is a statement of kind K. But if X is false, ~X is true. So it cannot be the case that all judgments of kind K are false. If we were not dealing with potentially infinite domains we would have to say that at most half the statements of kind K are false. The rest of them – their negations, their doppelgangers – must be true. But the thesis that judgments of kind K are truth apt and that some of them are true is realism, not antirealism, according to Sayre-McCord’s taxonomy. Thus type-b antirealism does not mark out a coherent class of theories. Error theories as such seem to be self-refuting.

8 Solutions to the Doppelganger Can the Doppelganger Problem be solved? Perhaps. But it is not clear that the problem can always be solved or that the same solution will work for every domain. When Field says that mathematics is false he means that every mathematical statement that quantifies over abstract objects is false. Now, the negation of a statement that quantifies over abstract objects does not quantify over abstract objects. Hence the Doppelganger Principle – that if X is a statement of kind K, ~X is a statement of kind K – does not apply within this domain. And if the Doppelganger Principle does not apply within a domain the Doppelganger Problem is dissolved. Thus Field’s type-b antirealism does not collapse into realism. Can we make the same move within metaethics? I think not. For though the negation of a moral proposition is not always a moral proposition, it seems to me that sometimes and in some contexts it is. (Consider “We ought to keep our marriage vows,” and its doppelganger “It is not the case that we ought to keep our marriage vows.” The latter, like the former, might have a considerable impact on our conduct if it came to be widely believed.) In other words, the Doppelganger Principle applies but intermittently. Even so, it scuppers metaethical nihilism as I have described it. For if the negations of some moral judgments are moral judgments, then it cannot be the case that all moral judgments are false. Thus metaethical nihilism needs to be reformulated. I suggest the following: All non-negative atomic moral judgments are false. This requires elucidation. First we specify a range of primitive “thin” moral predicates: “good” (morally good), “bad,” “right,” “wrong” “ought to” etc. (there may be a problem about this, as some of them are interdefinable). We then define an atomic moral judgment as a proposition ascribing an n-place moral predicate to n specific items. As defined these are nonnegative, i.e., not governed by the negation operator, but we redundantly specify that they are non-negative for the sake of clarity. Nihilism now amounts to the claim that all non-negative atomic moral propositions are false. And the argument is the standard nihilistic argument that there are no moral properties or relations corresponding to the moral predicates and thus no moral facts. Although this new formulation of nihilism is much more restricted than the original doctrine, it captures the spirit, though not the letter of the original thesis. It captures the spirit, since moral facts are

Nihilism, Nietzsche, and the Doppelganger Problem

29

denied and error ramifies through the great systems of morality, rendering them systematically false. But it rejects the letter, since some items that might reasonably be described as moral judgments will come out true (material conditionals with atomic moral judgments for their antecedents, disjunctions in which one disjunct is moral and the other not, etc.). But a myth does not cease to be a myth because it contains a few random truths, and what might be called “restricted nihilism” converts morality into a collection of myths. Does this deal with the Doppelganger Problem? Apparently yes. For the negation of a non-negative atomic moral proposition is never a non-negative atomic moral proposition. And where the Doppelganger Principle does not apply, the Doppelganger Problem does not arise. There is a general lesson here. Error theories as characterized by Sayre-McCord are only viable on one condition. We can only say that all propositions of kind K are false if the negation of a kind K proposition is never itself a proposition of kind K. Thus the error theorist must be very careful about defining his kind K if he is to escape self-refutation. Nietzsche omits this precaution. In The Will to Power 15, he flirts with what might be called “Global Metaphysical Nihilism” or “the Global Error Theory”: “The most extreme form of nihilism would be the view that every belief, every considering-something-true, is necessarily false because there simply is no true world. [This form of nihilism] would be a divine way of thinking.” Divine or not, it is absurd. For it amounts to the thesis that all propositions are false. Here the kind K of propositions is the kind of propositions as such. This kind is closed under negation since the negation of a proposition is itself a proposition. Hence it cannot be the case that all propositions are false. We may be able to save Nietzsche’s bacon as a metaethicist by restricting his thesis to non-negative atomic moral propositions, but the “divine way of thinking” seems to be beyond redemption.

9 The Reinforced Doppelganger I come now to the Reinforced Doppelganger, a particular problem for metaethical nihilists. Let us take a specific act (say the slaying of Caesar on the Ides of March, 44 BC) performed by specific actors (Brutus and Cassius for convenience, though of course other conspirators were involved). Then, according to revised metaethical nihilism, proposition B Brutus and Cassius’ slaying of Caesar was wrong is false. And this, in turn, entails: ~B It is not the case that Brutus and Cassius’ slaying of Caesar was wrong. But, given that the slaying of Caesar was a deliberate action, ~B would appear to entail:

30

C.R. Pigden

Br Brutus and Cassius’ slaying of Caesar was right (in the sense of “morally permissible”). Generally speaking, “Action X is not wrong” appears to entail “Action X is right” and “Action X is not right” appears to entail “Action X is wrong.” Call these the “RD” (for “reinforced Doppelganger”) principles: RD1 “It is not the case that action X is wrong” entails “Action X is right.” RD2 “It is not the case that action X is right” entails “Action X is wrong.” But if either of these principles is correct, my solution to the Doppelganger Problem fails. For the falsehood of a non-negative atomic moral proposition entails its negation and, in some cases at least, the negation of one non-negative atomic moral proposition entails the truth of another. So it can’t be the case that all non-negative atomic moral propositions are false. If all actions are not wrong then all actions (at least, those which exist) are right or morally permissible. And if all actions are not right or morally permissible then all actions (at least, those which exist) are wrong. Either way we have moral truths – non-negative atomic moral truths – in abundance. It seems that the only alternative to moral realism (in Sayre-McCord’s sense) is noncognitivism, after all. There is only one way out for the nihilist. He has to deny the RD principles. “Action X is not wrong” does not entail that action X is right (in the sense of morally permissible) nor does “Action X is not right” entail that action X is wrong. But is this bold and blunt assertion anything more than the desperate response of the cornered nihilistic rat? No, because (I think) it can be motivated. A entails B if it cannot be the case that A is true and B false. Or A entails B if there is no conceivable situation (possible world) in which A is true and B false. Is there a conceivable situation in which ~B is true and Br false? Yes: the situation in which there are no moral properties or relations, and specifically no properties of rightness, wrongness or obligatoriness, which attach themselves to acts. In such a situation, Brutus and Cassius’ slaying of Caesar won’t have the property of wrongness, but it won’t have the property of rightness (moral permissibility) either. It won’t have any moral properties at all. Now this situation is precisely the situation that nihilists think obtains. Thus the RD principles cannot provide independent evidence against nihilism. For they rest on the thesis not only that nihilism is false but that it is necessarily false. The point can be expressed with the aid of the following two diagrams (Figs. 1 and 2). Figure 1 represents logical space (for deliberate actions) as presupposed by the RD principles. Figure 2 represents logical space (for deliberate actions) as represented by nihilists. (Of course, nihilists believe that all actual acts are in the doubly shaded area.) Nihilists believe, as RD theorists do not, that it could be that actions are neither right nor wrong (indeed, they argue that this is not just the way it could be but the way it is). RD theorists effectively deny even the possibility of nihilism. But to say that nihilism is impossible – that it is absolutely inconceivable

Nihilism, Nietzsche, and the Doppelganger Problem

31

Fig. 1 Logical space (for deliberate actions) as presupposed by the RD principles

Fig. 2 Logical space (for deliberate actions) as represented by nihilists

that neither moral rightness nor wrongness attach to actions – is to make a large and implausible claim. Yet if this claim is not true, the RD principles are both false. Thus metaethical nihilism (somewhat revised) can survive both the Doppelganger and the Reinforced Doppelganger Problems. It is perhaps worth stressing that the RD principles are not analytic since another famous attempt to refute the error theory (along with many other forms of moral antirealism) rests on the claim that they are. In his famous paper “Objectivity and Truth: You’d Better Believe it” (1996), Ronald Dworkin argues that wholesale or Archimedean moral skepticism of the kind advanced by Mackie (and in my view by Nietzsche) is fundamentally incoherent. You can’t be a skeptic about all moral claims, since if you think that abortion is not wrong – or if you think that it is not full-bloodedly true that abortion is wrong – you are committed to the first-order view that abortion is morally permissible. But that only holds if you subscribe to something like RD1 – that the claim that actions of kind X are not wrong entails that actions of kind X are right (in the sense of morally permissible). But nihilists (if they have any sense) reject such claims. Dworkin might reply that this is like people who believe both that Oscar is round and that Oscar is square but absolve themselves from the charge of inconsistency by rejecting the thesis that what is round is not square (and vice versa). The problem with this is that it really is analytic (or at least necessarily true) that what is round is not square, but it is not analytic that actions

32

C.R. Pigden

that are not wrong are morally permissible. For it does not hold in worlds where there are no moral properties, which is precisely that kind of world that nihilists think we inhabit. Dworkin is like the citizen of a mighty empire in which everything is legal unless the emperor forbids it; taking this to be an analytic truth, he concludes that outside in the Badlands, where the writ of the emperor does not run and nothing is forbidden, everything is legal. “Those who adopt the second-order view that the Badlands are lawless are in fact committed to the first-order view that everything in the badlands is legal! For they admit – nay, they insist – that in the Badlands nothing is forbidden by the emperor!” But where there is nobody with the authority to permit or forbid, the fact that something is not forbidden does not entail that it is permitted.

10 Unfinished Business In §§5–6, I discussed a family of arguments derived, in part, from Blackburn and Wright, which criticize the nihilistic view that moral judgments are all false. These arguments fail, but perhaps they do better against the amended form of the error theory that I have been defending in §§7–9. Premise 1 transforms into premise 1∗ : 1∗ If all non-negative atomic moral judgments are false then we morally ought to (a) give up moralizing and (b) give up morality. Is 1∗ true? Surely not. It is not quite clear how the consequent of 1∗ should be analyzed, but it is most naturally rendered as a universal quantification: For any person x, x morally ought to (a) give up moralizing and (b) give up morality. Thus, for the consequent of 1∗ to be true there must be a relation of obligation between each individual and the act-types of (a) giving up moralizing and (b) giving up morality. But if all non-negative atomic moral propositions are false, because there are no such things as obligations, then it will not be true of each individual that he/she ought to (a) give up moralizing and (b) give up morality. Thus, if the antecedent is true, the consequent will be false, which means that the conditional itself is false. Thus the first variant of the argument (with the “ought”s interpreted as moral) is unsound. What about the other two? In the second variant of the argument in which the “ought”s are read as hypothetical imperatives indexed to some truth-seeking project, premise 1 transforms into premise 1∗ : 1∗ If all non-negative atomic moral judgments are false, then if we want our beliefs to be true (and not false) we ought to (a) give up moralizing and (b) give up morality. 1∗ appears to be true, but premise 2 , which does not need to be amended, is just as question begging in this version as it was in the earlier argument, since it is tantamount (in context) to the claim that not all non-negative atomic moral judgments

Nihilism, Nietzsche, and the Doppelganger Problem

33

are false. As for the third variant (pragmatic “ought”), the uniform substitution of “non-negative atomic moral judgments” for “moral judgments” makes no difference to its status. The arguments fail against the older version of the error theory and they are just as unsuccessful against the amended version.

11 Conclusion In this chapter I have argued that Nietzsche was a certain sort of nihilist, i.e., an error theorist about ethics, defending my interpretation against Schacht. I then defended this more or less Mackian position against Blackburn, Wright, and the hoi polloi. Morality does not have to be true to be good (in the sense of good for something), thus if it is good for something and even worth persisting with, this does not show that it is composed of truths. I raised and solved both the Doppelganger Problem and the Reinforced Doppelganger Problem, amending the error theory along the way. In order to escape self-refutation, error theorists like Nietzsche and Mackie must pull in their horns. The claim should not be that all moral judgments are false but only that non-negative atomic moral judgments are all false. Thus we can move from the non-existence of moral properties to the systematic falsehood of morality without adopting the incoherent idea that everything that might reasonably be regarded as a moral judgment is condemned to error. But my aim has been to vindicate the error theory against certain objections, not to establish its truth. The error theory may be in error, but I hope I have shown that it is a lot less silly than some have supposed.

References Ansell-Pearson, K. 1994. An introduction to Nietzsche as a political thinker: The perfect nihilist. Cambridge: Cambridge University Press. Blackburn, S. 1993. Errors and the phenomenology of value. In his Essays in quasi-realism, 149–165. New York: Oxford University Press. Cawkwell, G. 1985. A memorial address. In Morality and objectivity: A tribute to J. L. Mackie, ed. T. Honderich, 219–222. London: Routledge and Kegan Paul. Danto, A. 1965. Nietzsche as philosopher. New York: Macmillan. Dworkin, R. 1996. Objectivity and truth: You’d better believe it. Philosophy and Public Affairs 25: 87–139. Field, H. 1989. Realism, mathematics and modality. Oxford: Blackwell. Garner, R. 1994. Beyond morality. Philadelphia: Temple University Press. Hinckfuss, I. 1987. The moral society: Its structure and effects. Canberra: Australian National University. Joyce, R. 2001. The myth of morality. Cambridge: Cambridge University Press. Leiter, B. 2002. Nietzsche on morality. London: Routledge. Mackie, J. L. 1946. A refutation of morals. Australasian Journal of Psychology and Philosophy 24: 77–90. Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin. Nietzsche, F. 1881. Daybreak, Trans. R. Hollingdale. 1982. Cambridge: Cambridge University Press.

34

C.R. Pigden

Nietzsche, F. 1883. Thus spake Zarathrustra. In The portable Nietzsche, Trans. and ed. W. Kaufmann. 1954. New York: Viking Penguin. Nietzsche, F. 1883–1888. The will to power, Trans. R. Hollingdale and W. Kaufmann; ed. W. Kaufmann. 1968. New York: Vintage. Nietzsche, F. 1886. Beyond good and evil. In The basic writings of Nietzsche, Trans. and ed. W. Kaufmann. 1968. New York: Modern Library. Nietzsche, F. 1887. The genealogy of morals, Trans. C. Diethe; ed. K. Ansell-Pearson. 1994. Cambridge: Cambridge University Press. Nietzsche, F. 1889. Twilight of the idols. In The Anti-Christ, Ecce Homo, Twilight of the Idols: And other writings, Trans. J. Norman; eds. A. Ridley and J. Norman. 2005. Cambridge: Cambridge University Press. Pybus, C. 1993. Gross moral turpitude: The Orr case reconsidered. Melbourne: William Heinemann. Russell, B. 1922. Is there an absolute good? In Russell on ethics, ed. C. Pigden. 1999, 119–124. London: Routledge. Sayre-McCord, G. 1986. The many moral realisms. Southern Journal of Philosophy 24(suppl.): 1–22. Schacht, R. 1995. Making sense of Nietzsche. Urbana: University of Illinois Press. Smith, M. ed. 1995. Metaethics. Aldershot: Dartmouth. Wright, C. 1992. Truth and objectivity. Cambridge, MA: Harvard University Press.

Patterns of Objectification Richard Joyce

John Mackie thinks that the “objective prescriptivity” with which our moral discourse is essentially but so fatally imbued is the result of our “tendency to read our feelings into their objects” (1977, p. 42). He invokes Hume’s famous projectivist image of the human mind’s “great propensity to spread itself on external objects,” and, indeed, it is in his book-length analysis of Hume’s moral theory (Mackie 1980) that the topic receives a more careful discussion than in Ethics: Inventing Right and Wrong. In both books he musters some considerations in favor of the objectification thesis, and reveals to us that he thinks that “it is very largely correct” (1980, p. 72). But what is the relation between Mackie’s objectification thesis and his thesis of moral skepticism? Is the error theory a premise in an argument to establish the objectification thesis? Or vice versa? Or are they logically unrelated? On the face of it, the objectification thesis appears to entail the error theory, but if this is so then one is left wondering why, in his 1977 book, this thesis is described straight after the argument from queerness, for if Mackie took himself to have some arguments in favor of moral objectification, then might it not have been strategically viable for him to establish the thesis of objectification first and then by implication argue for the moral error theory? Wondering whether Mackie might have chosen to establish his moral error theory on the basis of the thesis of objectification is just my dialectical point of departure. My principal goal in this paper is to try to get a firmer handle on just what the thesis of objectification really is, and to investigate what evidence might support it and what conclusions may follow from it. I will disambiguate two forms of the thesis. One does trivially imply moral skepticism but cannot be established independently of that skepticism, whereas the other may well be substantiated on independent grounds but is neutral on the matter of moral skepticism. First, some terminological clarifications. I will use “moral skepticism” and “moral error theory” interchangeably, reflecting Mackie’s own practice. I will also interchangeably use the verbs “objectify” and “project” – and the associated nouns “objectification” and “projection,” and “the thesis of moral objectification” and R. Joyce (B) Department of Philosophy, University of Sydney, Sydney, NSW, Australia

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_3, 

35

36

R. Joyce

“moral projectivism.” Mackie does this himself, in both his 1977 and 1980 books.1 In this terminological vein, let us start by distinguishing “objectification” from some similar notions. Hypostatization (a.k.a. reification) is the practice of taking something abstract and speaking of it (or thinking of it, or treating it) as if it were concrete (e.g., “Religion was his guiding light,” “Justice is blind”). As a literary device or simply as a metaphorical manner of speaking, there is obviously nothing objectionable about the tendency; we do it constantly. If, however, one began to accept such propositions literally, at face value, then that would be a kind of error. Sometimes, though, it may be the subject of dispute whether an instance of this is an error. When philosophers try to provide a concrete explication of a seemingly abstract concept – like set or number – then some will treat this as an admirable extension of the naturalistic program while others will consider it a misguided blunder. Anthropomorphism is often said to be a special type of hypostatization, wherein aspects of the inanimate world are imbued with human qualities. This taxonomy cannot be quite correct, however, since the subject of anthropomorphism need not be abstract. “Nature abhors a vacuum” attributes a human quality (abhorrence) to an abstract entity (nature), whereas “That stretch of road is treacherous” attributes a human quality (treachery) to a concrete entity (the road). As before, there is nothing wrong with anthropomorphism as a literary device or figure of speech, but one commits a kind of straightforward error if one really believes, of something incapable of human qualities, that it has such qualities. And, as before, there are areas of dispute, such as what kind of mental attributions can be made to animals or computers. In Modern Painters, John Ruskin gave the name the pathetic fallacy (from “pathos”) to a certain anthropomorphic tendency in writers and poets. He derided tired and uninspired anthropomorphic devices (“it is only the basest writer who cannot speak of the sea without talking of ‘raging waves,’ ‘remorseless floods,’ ‘ravenous billows,’ etc.” (Ruskin 1856/1908, p. 65)), but thought that some anthropomorphisms may be aesthetically justified when they reveal something genuine about the emotional life of the poet, despite being packaged with a false surface expression. (Still, Ruskin thought that the very best poets should be able to do away with such devices altogether.) Ruskin’s anachronistic ruminations on aesthetics need not detain us; I mention the pathetic fallacy here because Mackie explicitly refers to it himself. He says that his thesis of moral objectification is analogous to the pathetic fallacy (1977, p. 42). It is not obvious in what sense, exactly, the two are supposed to be analogous, but it is certainly important to observe a disanalogy. Witness what Mackie goes on to say immediately following, supposedly explaining the pathetic fallacy: “If a fungus, say, fills us with disgust, we may be inclined to ascribe to the fungus itself a nonnatural quality of foulness” (ibid.; he uses the same example in his 1946 article). This is actually not a good example of the pathetic fallacy (though it may be a good

1 The

disjunctive phrase “projection or objectification” appears on p. 42 of Mackie (1977), and twice on p. 72 of Mackie (1980).

Patterns of Objectification

37

example of what Mackie is really driving at). The pathetic fallacy proper would occur if one ascribed to the fungus the human trait of being disgusted – which one would do, obviously, only in the context of joking or speaking metaphorically. By contrast, to attribute to the fungus the quality of foulness (irrespective of whether that property is non-natural or non-existent) is not a piece of anthropomorphism at all. In this example, a certain feeling – disgust – has caused the person to experience the world as containing a certain quality – foulness. It seems pretty clear from context that Mackie will say that this foulness is not a real quality of the fungus; we think that it is only because our disgust somehow leads us to experience matters this way. Thus, to the extent that the fungus-foulness example exemplifies Mackie’s notion of objectification (which I think it does), objectification is not an instance of anthropomorphism. Contrary to the quote given in the opening paragraph, the process of objectification is not one by which we “read our feelings into their objects” – we do not see the inanimate world as being disgusted, being angry, being sad, (the waves as raging, the floods as ravenous), and so forth.2 Rather, the process is one by which our feelings cause us to read into their objects qualities that we would not otherwise judge them to have – that (it is tempting to say) they do not really have at all, that nothing has. In his 1980 book, Mackie uses the vague but more accurate phrase: that we “read some sort of image of [our] sentiments” into the actions and characters that arouse them (p. 71): the foulness is “some sort of image” of our disgust; the wrongness is “some sort of image” of our disapproval; and so on. In his much earlier 1946 article, he writes that “in objectifying our feelings we are also turning them inside out. . . . The feeling and the supposed quality are related as a seal or stamp and its impression” (pp. 81–82). Note that the characterization of objectification just given has several components. First, there’s a complicated claim about our experience of the world and what has caused that experience (a claim I will attempt to make more precise in due course); second, there’s a metaphysical claim that the quality we are experiencing the world to have is not really instantiated at all. One might think that if this is objectification, then the error theory would indeed follow by implication. But that is not quite so, for the characterization just given does not exclude a noncognitivist interpretation of the matter. Let noncognitivism be the metaethical view that moral utterances are not assertions. The characterization of objectification says nothing, nor implies anything, about the nature of the speech acts that one might use to communicate one’s experience. Suppose, for example, that we experience Xs as having the property Q, but in fact nothing instantiates Q – our experience is brought about by some emotional mental state E. This characterization is neutral regarding what is going on when we utter the sentence “X is Q.”

2 Cf.

Barry Stroud’s discussion of Humean projectivism: “We do not think that the sequence of events on the billiards table – the one ball’s striking the others and the second ball’s moving – itself has a feeling or impression like the feeling Hume says we humans get when we observe it. . . . Nor do we think that an act of willful murder itself has a feeling of disgust or disapprobation, any more than we think that a painting on a wall has a sentiment of pleasure or awe” (Stroud 2000, p. 22).

38

R. Joyce

Perhaps we are asserting that X is Q (in which case an error theory holds) or perhaps we are merely expressing the emotion E (in which case noncognitivism holds). Thus, in order to have a characterization of objectification that implies an error theory, we would have to add a further clause: that our utterances about the subject matter in question are assertions (or that they are false, or that they are erroneous, or something else along such lines). It is possible that Mackie did have in mind such a complex account of objectification, though it is difficult to say with confidence. He several times states that the qualities that are projected “are fictitious” (1980, p. 71), but one cannot tell whether he intends this to be a defining component of objectification itself or just an additional point that follows from other things that he has argued. (I suspect the matter was not precisely demarcated in his thinking.) The point to which I should like to draw attention is this: If one does have in mind the complex account of moral objectification – the one that does imply the moral error theory – the one that consists of (i) a complicated claim about our experience of the world and what causes that experience, (ii) the metaphysical claim that the quality we are experiencing the world as having is not really present at all, and (iii) the cognitivist claim that our associated utterances are assertions – then whatever arguments one employed to establish this thesis would already be entirely sufficient to establish moral skepticism. The moral error theoretic position, after all, just is the conjunction of components (ii) and (iii). Therefore, if we are using this complex account of objectification, then the argumentative strategy mentioned at the start of this paper – of establishing the moral error theory on the basis of the thesis of objectification – turns out not to be dialectically viable, after all. Suppose, alternatively, that we worked with a less complicated version of objectification. Suppose we stripped away both the cognitivist subthesis (iii) and the metaphysical claim (ii), leaving just (i) a claim about our experience of the world and what causes that experience. Elsewhere (Joyce 2009a) I have called this remaining theory “minimal projectivism”; here, in line with Mackie’s preferred terminology, let us call it “minimal objectification.” Would it be viable to argue for moral skepticism on the basis of minimal objectification? In order for a positive answer, minimal objectification must imply moral skepticism, but it must do so non-trivially – it cannot be that making the case for minimal objectification would require first making the case for moral skepticism. Before proceeding, we must be more precise regarding this “claim about our experience of the world and what causes that experience.” This claim is the heart of objectification; it is the part that is supposed to capture, in literal terms, the whole idea of something mental being projected onto the world. It is not my intention here to offer a general account of objectification, for that poses a number of complications that are surprisingly challenging to overcome and the effort is not necessary on this occasion. I am satisfied to sketch an account of a pertinent proper subset of objectification: minimal affective objectification.3 3 For

the curious, I will quickly outline the difficulties of providing a general account of minimal objectification (or “projectivism” as I will call it in this footnote, since I will relate it to subjects for

Patterns of Objectification

39

S’s experience of X as P is an instance of minimal affective objectification if and only if: (1) S experiences P as an objective feature of X, and (2) this experience has its immediate causal origin in some affective attitude (e.g., the emotion of disapproval) rather than a perceptual faculty.

Call (1) the phenomenological thesis and (2) the causal thesis.4 We experience the fungus as objectively foul, but this experience is the immediate result not of perceiving the property of foulness, but rather of our emotion of disgust. Naturally, perception is involved prior to our feeling disgust: we see the fungus, we smell the fungus, and so forth. It is in acknowledgement of this prior engagement of perceptual faculties that the causal thesis claims that the experience (e.g., of foulness) has its immediate origin in an affective attitude rather than a perceptual faculty. (That this is a vague and potentially problematic qualification does not escape me.) This appearance of “immediate” is designed also to exclude non-projection cases where some affective state, such as emotional arousal, guides subsequent perceptual processes, such as focusing one’s attention on aspects of the world that might otherwise escape notice. An intensely jealous person, for example, might notice something in the body language of another that she otherwise would not have seen. Let us say that the body language is real: The other person really is (say) sitting with uncrossed legs. And the jealous party experiences this body language as an objective aspect of the world (thus satisfying the phenomenological thesis). And were she not jealous, she would not have noticed it. However, her experience is not an instance

which the latter term is more familiar). An adequate general account of projectivism should cover at least the following cases: (1) moral projectivism, (2) causal projectivism, (3) psychopathological projectivism, and (4) color projectivism. In the interests of illustrating the difficulties of achieving a general account, let me sketch, in the most provisional terms, what these four theories might look like. In moral projectivism, something in the world prompts one to feel disapproval (say), which leads one to experience the thing in the world as forbidden. In causal projectivism, a regularity in the world prompts one to have an expectation, which leads one to experience the world as containing a causal relation. I am not confident that I can give a general account of all forms of psychopathological projectivism, but examples are not hard to come by: A person’s poor selfimage leads her to interpret her parents as being overly critical or demanding. So far we have three mental states that serve as “intermediaries” between the world and one’s experience of the world: disapproval, expectation, and a poor self-image. Yet it is hard to come up with an overarching category for these three that will not end up capturing too much. And in the case of color projectivism, it is challenging even to come up with an analogous intermediary mental state. The idea is (roughly) that one’s visual experience of color owes its quality to the nature of one’s sensory apparatus rather than to the real nature of objects’ surfaces (even though the sensory apparatus is sensitive to real properties of surfaces). It is not obvious what mental activity is supposed to be getting projected in the creation of one’s visual color field. (Visual qualia?) My hunch is that color projectivism can be articulated only as a metaphysical (non-minimal) thesis. It is in light of these kinds of complications that I have sidestepped the delicate task of trying to unify this family of stock examples (and others besides) with a general account, though I confess to harboring the hope of yet doing so in the future. Perhaps in the end there is no entirely satisfactory general account of minimal projection in the offing. 4 See Joyce (2009a), where some features of this account (e.g., what might be meant by “objective”) are discussed in more detail.

40

R. Joyce

of minimal affective projection because it does not flow immediately from the affective attitude of jealousy: rather, the jealousy has guided the subsequent engagement of her perceptual faculties. In the case of foul fungus, by comparison, the disgust does not lead one to see the fungus as foul via channeling one’s perceptual attention to aspects of the fungus (though it may incidentally direct perceptual attention as well); the relation between the disgust and the foulness – the relation that Mackie described as the latter being “some sort of image of” the former – is more direct than that. In order to articulate the thesis of minimal moral affective objectification we need do nothing more than stipulate that “P” (as in “. . .experiences X as P”) stands for a moral adjective. I cannot think of a theory of moral projectivism for which the mental state putatively “projected” in the creation of moral experience is supposed to be anything other than an affective attitude (e.g., disapproval, subscription to a normative framework, etc.), so in what follows I shall drop the “affective” qualification. Clarity requires, however, that we keep the “minimal.” I do not suppose that every philosopher who has spoken of “moral projectivism” or “moral objectification” over the years has really had in mind, even tacitly, the minimal variety. Both the phenomenological thesis and the causal thesis appear to be psychological claims. Quite how we would go about empirically testing them for the case of morality (or any other case) is a nice question (into which I won’t delve on this occasion), but it seems pretty clear that, on the face of it, whatever methods of empirical psychology we employed would not require us first to establish that moral properties do not exist, or that moral judgments are uttered with assertoric force. Therefore one desideratum of the strategy of arguing for moral skepticism on the basis of minimal objectification appears to be satisfied: The latter does not imply the former trivially. The other desideratum, however, is not satisfied. Minimal moral objectification does not imply moral skepticism at all, for it is metaethically neutral. Not only is it silent on the cognitivist/noncognitivist debate, it is also silent on the metaphysical debate over the existence or non-existence of moral properties. It is, therefore, compatible with moral realism and thus does not imply moral skepticism. Let me explain. Note, first, that there is a prima facie pressure in favor of maintaining this compatibility, based simply on the observation that were the phenomenological and causal theses to be confirmed by empirical inquiry (as they very well might be), it would surely be astonishing if the moral realist were to roll over and concede the game to the skeptic. It may nevertheless clarify matters to explore this compatibility in more specific terms. A simple example will suffice to get the compatibility on the table, and then I will develop a more satisfactory example for the moral case. Consider an everyday usage of the notion of projection. Suppose that a person tends to experience others she encounters in social situations as critical and reproachful; but in fact this is due to her own meek and self-doubting nature. It would not be at all peculiar for us to describe this person as projecting her sense of her own inadequacy onto others. Now suppose that on a given occasion this person encounters a man who really is unusually critical and reproachful. Let it be stipulated that he has not yet indicated to the woman, even in the most subtle fashion,

Patterns of Objectification

41

that he has such a personality; she simply assumes that he does, just as she does with everyone else that she meets. The woman’s judgment “This man is critical and reproachful of me” is just as much a projection on this occasion as ever it is; and this conclusion is not undermined by the fact that the judgment happens this time to be true. Even if the woman often encountered critical and reproachful persons, her subsequent negative judgments about them (which will now often be true) might nevertheless remain instances of her projecting her sense of inadequacy, so long as we specify that she would have formed these same judgments even if these critical and reproachful persons’ personalities had been otherwise. An analogous situation in the case of moral judgment would suffice to show the compatibility of minimal moral objectification and moral realism. We can imagine a scenario in which, on the one hand, moral judgments are acts of minimal objectification/projection, while, on the other hand, these judgments (when uttered) are assertions that are often true. According to one influential view, satisfying the latter conditions suffices for moral realism (Sayre-McCord 1986). Some would prefer to add a further clause to moral realism: that the assertions in question are true in virtue of some objective state of affairs. (See Joyce 2007 for discussion of the definition of “moral realism.”) Let it be so; imagining the scenario in such a way that this additional clause is also satisfied will not spoil the example by undermining the projectivist stipulation. But one may remain unsatisfied with this demonstration of compatibility, for the example had the projection-derived judgments turning out true by coincidence. The example shows minimal objectification and moral realism to be compatible according to the letter of the law, but perhaps not the spirit. Can we eliminate this aspect of accidentality from the demonstration of compatibility? I believe that we can. Return to Mackie’s example of foul fungus. Assume the minimal affective objectification account holds true: A person experiences the fungus as having an objective quality of foulness, and this experience has its immediate origin in the person’s disgust. Now let us see whether realism about foulness might also hold true. There is certainly nothing to exclude us holding that when the person makes public her judgment via the utterance “That fungus is foul!” she is making an assertion. It is the other realist elements that might be deemed troublesome: that the assertion (A) is true, (B) is true in virtue of an objective fact, and (C) is non-accidentally true – i.e., is made in a way dependent on the truth-rendering fact. (Whether (B) is really a necessary aspect of realism is moot, and that (C) is a necessary part of realism seems doubtful, but let us add these components for the sake of argument.) Consider a response-dependent account of the property of foulness. (See Casati and Tappolet 1998; Johnston 1989, 1992, 1993; Wright 1988.) Foulness, on this account, is a disposition to produce a certain kind of psychological response R in a certain kind of subject S in a certain kind of circumstance C. It would be no challenge to specify these variables R, S, and C in such a way that certain items in the actual world – such as bits of fungus – have this disposition, thus rendering assertions of the form “X is foul” true. The account can also claim to satisfy an important kind of objectivity. (See Pettit 1991.) The disposition in question will be understood

42

R. Joyce

in modal terms, such that the fungus would prompt R if apprehended by S in C. Thus the fungus may have the disposition even if no Ss have ever been in C, even if no Ss have ever experienced R, even if no Ss ever have existed or ever will exist. Thus the disposition does not depend on the existence of any particular psychological state, or even the existence of psychological states generally. (One might point out that the disposition is nevertheless conceptually mind-dependent – since it cannot be adequately described without reference being made to psychological state R – but it is not at all obvious that realism requires “conceptual mind-independence.” It certainly doesn’t require every form of mind-independence. One would like to be a realist about domesticated dogs, despite the fact that they are causally minddependent entities: existing only because of generations of intentional behavior on the part of humans.) In order to emphasize the fact that satisfying these realist conditions has not undermined the projectivist assumptions that we started with, let us specify R in such a way that the psychological response in question is in fact an act of objectification. In other words, suppose that foulness is the disposition to prompt Ss (when in C) to feel disgust and to objectify this disgust in their experience of the world. This appears to build the satisfaction of the minimal objectification thesis (i.e., the satisfaction of the phenomenological and the causal theses) into the response dependent account. We are thus in a position to see that the judgment that is derived from a process of objectification – “That fungus is foul!” – does indeed manifest a dependence relation on the truth-supplying fact; it is not merely coincidentally true: Were the fungus not to have been foul, the observer would not have made the judgment.5 One might worry that in striving to satisfy the dependence relation the account has undermined the causal thesis. If there is this dependence relation in place – if the counterfactual that ends the last paragraph holds true – then isn’t the person’s experience of the foulness (the foulness that we are here supposing to objectively exist) a case of perception, after all? And if so, then the causal thesis is undermined, in which case my attempt to show the compatibility of minimal moral objectification and realism will have failed. I respond not by offering a full account of perception (which I have neither the space nor the expertise to do), but by observing that the counterfactual dependence mentioned is certainly not a sufficient condition for perception. A couple of simple examples will suffice. Consider learning something from reading a book: Suppose 5 Needless

to say, Mackie himself won’t buy the response-dependent account of moral properties. He criticizes such views in general terms in his 1980 book (Chapter 5); and in his 1973 book (Chapter 4) he doubts even the existence of dispositional properties. It is also doubtful that Hume’s multifarious uses of the projectivist metaphor are supposed to be compatible with realism. Stroud (2000) emphasizes how, in Hume’s account, the content of projectivist experience – be it causal connection, beauty, color, or virtue – is something that could not even be intelligibly predicated of items in the world. Immediately following the famous Treatise projectivist image of the mind’s “great propensity to spread itself on external objects,” Hume declares that sounds and smells “really exist no where” (Hume 1739–1740, 1.3.14.24/1978, p. 167) – and context makes it reasonable to think that he will say the same of color and necessary connection. I discuss the error theoretic commitments lying behind Hume’s views in Joyce (forthcoming).

Patterns of Objectification

43

one learns for the first time that Napoleon lost Waterloo. The book, we are supposing, is an accurate one, in the sense that had Napoleon not lost Waterloo, the book would not have contained the claim that he did lose. And we will also suppose that the reader judges that Napoleon lost Waterloo solely on the basis of reading this reliable history book. This judgment thus manifests the appropriate counterfactual relationship with the relevant fact, but we would not on this basis conclude that the reader perceives Napoleon losing Waterloo. The second example concerns someone judging that a certain object instantiates a dispositional property. Suppose someone works out that a vase is fragile by some means other than breaking it. Perhaps she smashes a lot of similar vases, or perhaps she asks some authoritative people. We will suppose that whatever means she employs is reliable, in the sense that had the vase not had the disposition, she would not have come to make the judgment about its fragility that she does make. But, though her true judgment (“The vase is fragile”) manifests a dependence relation on the truth-rendering fact (the vase’s being fragile), we would not say that she has literally perceived this vase’s fragility. Thus, in striving to come up with a realistic account of foulness that satisfies the desideratum of the judgment being dependent on the truth-supplying fact (along with satisfying all the other realist criteria, too), we have not undercut our starting projectivist assumption that the experience has its immediate origin in an affective attitude rather than a perceptual faculty. And what goes for foulness here can go for moral qualities, too. I conclude, therefore, that minimal moral objectification is compatible with moral realism – even a fairly robust version of moral realism.6 Whatever may seem surprising or counter-intuitive about this conclusion probably stems from the fact that non-minimal versions of objectification are more familiar to us, both in vernacular settings and in the philosophy classroom. We are more likely to describe the case of psychopathological projection in terms such as: “Her sense of her own inadequacy makes her see others as overly critical when really they’re not.”7 We are more likely to describe moral objectification in terms such as: “Our feelings of disapproval and aversion lead us to see the world as containing moral qualities that it does not really contain.” It is not my intention to condemn such non-minimal, metaphysically-committed uses of objectification; they may, indeed, be the more natural and useful formulations in most circumstances. My point has been, rather, to show that in order to establish the truth of any such metaphysically-committed objectification thesis one would need to have already shown that the quality in question does not exist, and so the strategy of supporting moral skepticism by means of first establishing a metaphysically-committed version of moral objectification is unworkable. We have also seen that the strategy of supporting moral skepticism solely on the basis of establishing a minimal version

6 There

is precedence for seeing projectivism and realism as compatible in the Humean literature. See Craig (2000) and Sainsbury (1998). 7 Freud, remember, categorized this sort of projection as a kind of delusion – indicating an antirealist construal.

44

R. Joyce

of moral objectification is also unworkable, for the minimal version is silent on the debate between the moral realist and the moral skeptic. As to this last claim, one might object that of course minimal objectification should be silent about these metaethical claims, for the whole point, after all, is that it should not imply moral skepticism trivially; it should imply it only in conjunction with some other premises. Minimal objectification may well be consistent with moral realism, but perhaps these other premises will serve to narrow the range of possibilities so as to exclude realism along with every other theoretical possibility bar the error theory. One or more of these additional premises may function to knock noncognitivism out of the running. (Mackie does, recall, develop several arguments against the noncognitivist in both his 1977 and 1980 books.) Other additional premises may serve to establish the non-existence of moral properties. Perhaps when we add up all of these premises we end up with an argumentative route from minimal objectification to the moral error theory. At first blush, the problem with this objection appears obvious. The additional premises adverted to would appear to be sufficient to establish the moral error theory; they would be doing all the work, and the thesis of minimal objectification would be entirely superfluous in this argument. Therefore, we seem to see once more that the strategy of establishing the moral error theory on the basis of the thesis of minimal objectification turns out not to be viable. On closer inspection, however, the objection has more merit. Consider the passage with which Mackie introduces the topic of objectification, immediately after presenting his arguments from relativity and queerness: Considerations of these kinds suggest that it is in the end less paradoxical to reject than to retain the common-sense belief in the objectivity of moral values, provided that we can explain how this belief, if it is false, has become established and is so resistant to criticisms. (1977, p. 42)

The thesis of objectification is supposed to satisfy this “proviso.” In other words, Mackie has exposed a theoretical option: We can either “reject. . . the commonsense belief in the objectivity of moral values” or we can “retain” it. The thesis of objectification is evidently supposed to function as the tie-breaker, making rejection the reasonable choice. We are forced to conclude that (Mackie thinks that) without the thesis of objectification, retaining the commonsense belief would remain a live option. Objectification functions as an explanation of where the massive error embodied in morality comes from, in such a way that without that explanation there remains doubt that it is an error at all. There is a tempting alternative reading of Mackie, which accords the thesis of objectification a lesser role. According to this alternative, by the end of the section in which he presents the argument from queerness (Section 9, Chapter 1) Mackie has established the moral error theory to his own satisfaction. A reader might at that point accept the moral error theory but then be naturally curious to know where this widespread systematic human error has come from, and the thesis of moral objectification is supposed to satisfy this curiosity. According to this reading, the thesis

Patterns of Objectification

45

of moral objectification does not function to satisfy a proviso, but is, rather, a supplement to the completed skeptical argument: an interesting explanation but strictly dispensable in establishing the case for moral skepticism. The above-quoted passage, however, indicates that this alternative reading is incorrect. The case for moral skepticism (embodied in the arguments from relativity and queerness) is plainly considered incomplete until the thesis of objectification is put forward. But what role precisely does the thesis of objectification play in Mackie’s case for moral skepticism? The answer, I believe, lies in Mackie’s almost-tacit acceptance of some kind of epistemological conservatism. Just prior to presenting the arguments from relativity and queerness, he has admitted that since moral skepticism “goes against assumptions ingrained in our thought and built into some of the ways in which language is used, since it conflicts with what is sometimes called common sense, it needs very solid support” (1977, p. 35). Mackie appears to be acknowledging that the counterintuitiveness of moral skepticism in itself represents a burden of proof that the error theorist must strive to overcome – that the error theorist’s arguments need to be even more convincing than those of his opponent if he is to win the day. That Mackie is an epistemological conservative may seem an unexpected conclusion, considering how accustomed we have grown to seeing the moral realist reach for the principle of epistemological conservatism as one the main weapons in the anti-skepticism arsenal. (See, e.g., Huemer 2005; Brink 1989, pp. 23–24; Dancy 1986, p. 172.8 See Loeb, this volume, for discussion.) But it is apparent that Mackie does indeed consider the “intuitiveness” of a philosophical thesis a valid consideration in deciding whether to endorse it. All else being equal, an intuitive theory is to be preferred over a counter-intuitive one; in other words, the very fact that a belief is held supplies it with a certain prima facie epistemological justification. And that there are intuitions in favor of morality is hardly to be denied; any error theory worth arguing about is, ex hypothesi, counter-intuitive. Mackie’s arguments in favor of moral skepticism, then, must overcome these standing intuitions. The vital role of the thesis of objectification is to explain away the content of these pro-morality intuitions by providing an account of their origin that does not imply or presuppose their truth. Such a genealogical explanation serves to defeat or block whatever prima facie justification these intuitions might otherwise have been granted. The skeptic does not deny or doubt the principle of conservatism; he takes it seriously. In particular, the skeptic attends to the principle’s “all else being equal” clause. “When are things not equal?” the skeptic wonders. When are intuitions defeated, and under what conditions might they not even be accorded prima facie epistemic status? One answer (among many, no doubt) is that things are not equal if one has a plausible, or even empirically confirmed, theory of where the intuitions in question come from that is consistent with their being false.

8I

offer some criticisms of this line of argument – and of Michael Huemer’s version of it in particular – in Joyce (2009b).

46

R. Joyce

The moral skeptic, then, needs two lines of reasoning. The first attempts to show that there is something fishy about moral facts. (For Mackie, this is achieved by the conjunction of the arguments from relativity and queerness.) But the conclusion of this reasoning is, all parties agree, counter-intuitive. A principle of epistemological conservatism threatens to allow our pro-morality intuitions to override the skeptic’s case. So the skeptic offers a second body of evidence: explaining away the content of those pro-morality intuitions by revealing them to be the product of an unreliable process of formation. The second move aims to show “how even if there were no such objective values people not only might have come to suppose that there are but also persist firmly in that belief” (1977, p. 49). The skeptic thus discharges the burden of proof with which he is lumbered – not by bolstering his initial argument (presenting more evidence in support of premises, devising new argumentative moves, etc.), but by casting into a doubtful light those very intuitions that promised to give the conservative principle traction. (Moreover, nor can these same pro-morality intuitions be raised as a consideration against the argument for objectification, since it is a corollary of the objectification thesis that such intuitions will be in place.) My claim that the intent of Mackie’s two skeptical arguments (relativity and queerness) is merely to show that there is something “fishy” about morality might itself seem fishy; surely, one might object, these arguments are supposed to be more decisive than that. If these are sound arguments (the objection continues), then they need no extra argument to act as a tie-breaker; and if they are not sound arguments, then why should we pay them any attention? However, I think this dilemma does not succeed in undermining the interpretation being offered. Let us briefly consider Mackie’s two skeptical arguments in turn. The argument from relativity takes the form of a competition between two hypotheses: the phenomenon of moral disagreement may be explained either by the supposition that some parties have privileged epistemic access to the realm of moral facts (the realist’s hypothesis) or by the supposition that there are no moral facts at all (the skeptic’s hypothesis). Among the many criteria that we might employ in deciding between these two hypotheses, a comparison of their levels of mesh with our intuitions may well figure. Certainly the epistemological conservative allows it to figure. Therefore putting forward evidence (such as the thesis of objectification) that explains away the content of intuitions in favor of one hypothesis is both strategically permissible and potentially determinative. The same point is slightly less obvious in the case of the argument from queerness, for here, it might seem, we have an argument that purports to stand soundly on its own: Premise 1 is a piece of conceptual analysis (that moral discourse is centrally committed to the existence of objective prescriptions) and premise 2 is an ontological claim (that there exist no objective prescriptions). However, it is, I think, slightly naive to suppose that Mackie considers the argument from queerness to be a sound argument with demonstrably true premises. It is more realistic to think of it as providing a firm consideration in favor of moral skepticism, its premises having the status of hypotheses on whose acceptability many factors may have a

Patterns of Objectification

47

bearing. There may, for example, be considerations to be raised in support of the claim that there exist no objective prescriptions, but also considerations to be raised in support of the claim that there do exist objective prescriptions. No impartial spectator who has been paying attention to the debate could seriously doubt this; it is a complex and nuanced discussion that leaves plenty of wriggle room for competing interpretations of key elements (as some of the contributions to this volume demonstrate). The epistemological conservative allows that if there are standing intuitions in favor of view that objective prescriptions exist (as it seems reasonable to suspect), then these may be accorded a role in weighting the debate against premise 2 and thus against the argument from queerness. Therefore, again, the strategy of raising evidence (such as the thesis of objectification) that casts into doubt those very intuitions that speak against the argument from queerness – showing that they arise from a potentially unreliable source – is entirely legitimate and may very well swing the argument the skeptic’s way. We have seen, then, that the two lines of reasoning need each other. The case for moral skepticism is not achieved by the arguments from relativity and queerness alone; evidence to explain away the counter-intuitiveness of the conclusion (or individual premises) is also required in order to overcome the challenge from epistemological conservatism. And the thesis of objectification alone will not provide a skeptical conclusion. The minimal version of the thesis is metaethically neutral, and to employ a metaphysically-committed version to this end would simply beg the question. Thus far I have been concerned entirely with the role that the objectification thesis plays in Mackie’s overall strategy for establishing moral skepticism, but we have not yet examined any of the arguments he provides to convince us that the thesis is actually plausible. I will close by running through the considerations in support of the objectification thesis that Mackie offers in his 1977 and 1980 books.9 In the interests of clarifying the structure of the argument we should first acknowledge that it was open to Mackie to eschew the task of mustering evidence in favor of the objectification thesis, and instead simply present the thesis as a coherent and possible hypothesis of the genealogy of moral judgment according to which these judgments are not, or might not be, true. But it is evident that this is not Mackie’s attitude towards the thesis – and it is as well that it is not, for this strategy would place the objectification thesis in the same category as a host of other skeptical hypotheses that lack any real plausibility but which have the (dubious) virtue of thwarting all attempts at falsification. (Brains in vats and deceiving demons spring to mind.) The objectification thesis plainly isn’t supposed to function in this disappointing way – merely as a skeptical hypothesis that might, for all we can prove to

9 Space

does not permit an examination of Mackie’s arguments for moral objectification found in his 1946 paper. I do not think any of the arguments found there are superior to those problematic ones which I shall discuss.

48

R. Joyce

the contrary, hold true. It is supposed, rather, to be a serious hypothesis for which we can marshal solid evidence. Nevertheless, in my opinion, Mackie misses much of the opportunity to establish the plausibility of the thesis by looking for supportive evidence in the wrong places. The few pages devoted to this in his 1977 book (pp. 43–48) are uncharacteristically somewhat obscure. In particular, he does not do an adequate job, in my opinion, of distinguishing between the thesis that moral experience is the result of our having objectified affective attitudes (feelings, wants, and demands are mentioned by him) and the thesis that we simply have false beliefs about the objective status of moral properties. That we should recognize such a distinction seems highly desirable. Consider something that we would ordinarily consider a non-objective matter – say, what counts as polite behavior at meal times. A person might foolishly believe that the propriety of keeping one’s elbows off the table is an utterly mindindependent affair, and that all those other cultures that allow alternative rules of etiquette are simply mistaken. This person has a false belief about the objectivity of something. But must it be an instance of objectification? It might be, of course, but it also seems reasonable to suppose that it might not be. Objectification essentially involves a certain sort of psychological operation that leads to belief (or, speaking more carefully, that leads to a certain quality of experience). “Objectify” is a transitive verb; there must be something that gets objectified. But one can have a (false) belief about the objectivity of something without any operation of objectification having occurred. Mackie’s failure to attend to this distinction undermines the force of many of the considerations he raises in support of the objectification thesis. For example, Mackie begins by pointing out that a widespread belief in objective moral properties might fulfill certain human needs: Such properties (if they existed) would have a kind of practical authority over human affairs such that a widespread belief in their instantiation would regulate interpersonal relations in an effective way. This is a complex but broadly plausible claim. Yet it doesn’t obviously provide any evidence in favor of the objectification thesis as opposed to the “false-beliefs-aboutobjectivity” thesis. And even as evidence for the latter it is weak: The consideration merely shows that we might have a motive for believing in objective moral properties, which falls short of demonstrating that we do so believe. One might try to wring from this some (proportionally weak) support for the objectification thesis by pointing out that if we have a motive for believing in an objective morality, and this motive does lead us to belief, then we have objectified that motive, thus satisfying the criteria of the objectification thesis. But this is a problematic line of argument. What it is to “have a motive” is a complicated and indeterminate matter. On one reading it means that there is a reason to do something, even if one is unaware of this fact. (“The Romans had a motive for ceasing to line their aqueducts with lead.”) A more robust reading requires making reference to an agent’s desire. (“Romeo had a motive for climbing to Juliet’s balcony.”) This indeterminacy creates a fatal dilemma for the argument under consideration. Suppose, first, that when we say that humans “have a motive” for believing in an objective morality, we are using “motive” in something like the former sense, to mean that, as a matter of fact, things will go better for us (each of us, let’s say) if we

Patterns of Objectification

49

all have this belief. This, however, may be true while we all remain utterly ignorant of the fact, showing that such a motive might not exist in our psychological profile in any sense, and therefore is not there to be objectified. The item which we are calling “the motive” might nevertheless have causal powers. That fact that things go better for organisms with X than organisms without X can lead to change – through a process of natural selection, for example. I admitted that it is broadly plausible that things will go better for groups of humans if they believe in an objective morality, and we can thus imagine this belief becoming fixed in a population through some process of cultural (or biological) evolution. All this would make it permissible to claim: “Humans have a motive for believing in an objective morality, and this motive brings it about that they do so believe.” But at no point need we maintain that the “motive” in question is psychologically represented by/to humans – let alone as an affective attitude – and therefore the criteria for objectification are not satisfied. Alternatively, we could understand “motive” in the latter sense, to mean that humans (generally) do desire to believe in an objective morality, and this desire brings it about that they do so believe. This would satisfy the criteria of the objectification thesis, but only at the cost of being a fantastic empirical claim. Bearing in mind that we are using “desire” in the sense of occurrent, affective attitude (something that has causal powers within an individual’s psychological economy), it is wildly implausible that humans typically desire to believe that moral properties are objective. I conclude, therefore, that while it might be true that humans “have a motive” for believing in an objective morality, and that this fact might have had (and continue to have) an important causal role in bringing it about that we do so believe, this does not provide grounds for claiming that this belief is the product of our having objectified that motive. The next strategy explored by Mackie (1977, pp. 43–44) in seeking support for the objectification thesis is to give some specific examples of moral objectification, which he designates “patterns of objectification.” Sometimes we desire something for perfectly sound (non-moral) reasons, but then we “confuse” this basis for our desire (the item’s “subjective value”) with the idea that the item in question has objective value. Sometimes we think that someone ought to do something for instrumental reasons, but then we suppress the instrumental conditional clause and claim that she ought do it simpliciter. We might do this because expressing it in this way is more likely to lead to compliance. Later (p. 47), another kind of confusion is mentioned: when we muddle a descriptive and objective sense of the goal of humans (as in what we in fact pursue, or what posited goal will confer sense upon our actual actions) with a normative but subjective sense of that goal (as in what we ought to be pursuing). These might very well be cases of objectification, and it is not at all implausible to suppose that Mackie is accurately describing some real human phenomena. But if his goal is to provide evidence that human moral judgment is typically (always?) the product of a process of objectification, then these examples hardly count as strong evidence. After all, I doubt that anyone (apart from some philosophers who worry in

50

R. Joyce

their characteristic manner about the details of the thesis of projectivism) will deny that moral judgment is sometimes the product of projected desires, emotions, and moods. Therefore, exposing some cases of moral objectification does not suffice to explain how the (putatively) false intuitions in favor of morality have “become established” and are “so resistant to criticisms” (p. 42). By analogy, suppose one were to doubt the existence of human character traits, and sought to explain away those strong intuitions we seem to have that such traits do indeed exist. One potential explanation would be the projectivist one: Our tendency to see others as instantiating stable character traits is the result of our projecting aspects of our own mental lives onto them. In providing evidence for this projectivist thesis, it would clearly not suffice to point out some instances of this kind of projectivism – such as the example I employed earlier of someone’s sense of her own inadequacy leading her to see others as critical and reproachful. All parties (we can suppose) accept the existence of the phenomenon of psychopathological projection, but much more evidence would be needed to show that this phenomenon somehow generalizes in unexpected ways. By referring in the plural to “patterns of objectification,” Mackie might hope to give the impression of a systematic and widespread tendency here (as in “There is a pattern emerging!”), but in reality he has not provided evidence sufficient for his purposes. A further strand in Mackie’s case for the objectification thesis is broadly historical. Once upon a time a lot of our moral language was embedded in an ontological framework that included an all-powerful, all-seeing, loving deity doling out rewards and punishments. Mackie concedes that if “this theological doctrine could be defended, a kind of objective ethical prescriptivity could be introduced” (p. 48).10 He looks with some sympathy upon Elizabeth Anscombe’s conjecture that “modern moral philosophy” (à la mid-twentieth century) consists of trying to make sense of a family of normative concepts “outside the framework of thought that made [them] really intelligible” (Anscombe 1958, p. 6). However plausible this “conceptual residue theory” may be (and I agree with Mackie that there is surely something to it), it nevertheless seems to have little to do with the psychological process of objectification. Anscombe may have explained the origin of a widespread but false belief in the objectivity of morality (and therefore does provide resources upon which Mackie can draw), but the hypothesis does not fit with the desired model of this belief being the result of our having “spread” our wants and demands onto “external objects.” I conclude, then, that the case for the objectification thesis in Mackie’s 1977 book is very weak. We find a more structured and clear argument for the objectification thesis in 1980’s Hume’s Moral Theory. Here Mackie claims that it is the only theory that can properly make sense of three phenomena: (i) that the evidence seems to favor metaethical cognitivism, (ii) that moral statements are taken to be intrinsically 10 Incidentally, this comment reveals that Mackie believes that so long as there exists some possible

world at which the requisite kind of supernatural being is real, then the error in morality is but a contingent matter. In his forceful case for atheism in his 1982 book, Mackie repeatedly declares the existence of God to be “improbable” (pp. 100, 130, 252–253) – not impossible.

Patterns of Objectification

51

action-guiding (i.e., not to rely on subjects’ ends and goals for their legitimacy), and (iii) that “the essential fact of the matter, which underlies moral judgments. . ., is that people have various sentiments” (1980, p. 72). Though the last is stated rather clumsily, it is apparent to what Mackie is referring. In the preceding chapter he has identified nine arguments used by Hume to press the negative view that moral judgments are not the product of reason, and then he has provided a number of wellknown passages from Hume variously stating the positive “plain hypothesis”: “that morality is determined by sentiment” (Hume 1751, appendix 1.10/2006 p. 160). Mackie assesses various specifications of this positive hypothesis – one of which is the objectification thesis – and it is precisely because he wants to characterize the sentimentalist hypothesis in a way that leaves open a range of theoretical options that he words (iii) so formlessly. Mackie thinks that (i)–(iii) jointly knock out all rivals to the objectification thesis as follows (very roughly): Noncognitivist offerings fail to satisfy (i). Versions of “subjectivism” according to which moral utterances make reference descriptively to some real or hypothetical agent’s emotions (such as the ideal observer theory) fail to satisfy (ii). Various forms of moral rationalism, intuitionism, and naturalistic realism – theories that do not accord emotion a central role in our apprehension of moral truths – fail to satisfy (iii). By contrast, the objectification thesis is supposed to pass the test. According to this thesis, moral judgment begins with humans responding to certain actions and characters in the world with affective attitudes (thus satisfying (iii)), which we then project onto our experience of the world, reading “some sort of image” of the attitude into the item that prompted it, seeing (for example) the action as categorically required (thus satisfying (ii)); and we are, by and large, fooled by this operation into thinking that the normative property really is instantiated, in which case our language for discussing it is, naturally, assertoric and propositional (thus satisfying (i)). To assess this argument for the objectification thesis would require a comprehensive metaethical investigation. We would have to evaluate whether all these rivals do indeed fail to satisfy the desiderata that Mackie claims they fail; we would have to assess whether the objectification thesis really does satisfy the three desiderata; we would have to consider whether any theory other than the objectification thesis promises to satisfy the three desiderata; we would need to investigate whether these three desiderata really are that (that a theory’s failure to satisfy one of these criteria really does represent a reason for rejecting the theory); and we would need to reflect on whether there might exist additional desiderata (ones, perhaps, that the objectification thesis fails to satisfy). Needless to say, such an assessment is not going to be attempted here. The point I want to observe is the general one that here Mackie is seeking to establish the objectification thesis via a metaethical route – and a long and controversial metaethical route at that. It is clear that this argument for objectification is not independent of pivotal elements of his other arguments for moral skepticism. For example, it is a central plank of the argument from queerness that moral judgments are imbued with “objective prescriptivity” – “something that involves a call for action or for the refraining from action, and one that is absolute, not contingent

52

R. Joyce

on any desire or preference or policy or choice” (1977, p. 33). But note that this premise in the argument from queerness is really nothing more or less than desideratum (ii) employed in the argument for the objectification thesis. Many critics of the argument from queerness complain that moral discourse is committed to nothing so extravagant, and if they are correct then the argument will clearly collapse. Whatever grounds these critics of the argument from queerness have for their view are grounds for denying that (ii) is a criterion of theory acceptance. Another class of critic of the argument from queerness will maintain that this quality of “objective prescriptivity” can indeed be satisfied by (clusters of) naturalistic properties. But saying this is nothing more or less than denying Mackie’s claim that moral naturalism cannot satisfy (ii). That Mackie’s argument for the objectification thesis shares central premises with the argument from queerness is not a problem in the context of his 1980 book, for the objectification thesis is not there functioning to satisfy a proviso to another argument. But it does mean that we cannot lift the argument for the objectification thesis found in the 1980 book and use it to help establish the objectification thesis in the context of the 1977 argument. The interpretation I have offered of the 1977 dialectic has acceptance of the arguments from relativity and queerness held in abeyance until the objectivity thesis steps in as a tie-breaker, rendering it “less paradoxical to reject than to retain the common-sense belief in the objectivity of moral values” (1977, p. 42). But this strategy plainly will not work if whatever doubt hangs over the arguments for moral skepticism also hangs over the argument that would convince us of the thesis that would satisfy the proviso. I conclude that Mackie’s use of the objectification thesis in Ethics: Inventing Right and Wrong is a reasonable idea that is poorly executed. The general strategy of overcoming epistemic conservatism by showing moral intuitions to be the product of an unreliable process of formation is a sound one. Seeing moral experience as the product of an operation of projection or objectification is one prominent example of this strategy (among others) that might very well succeed. But the objectification thesis necessary and sufficient for the job is the minimal psychological thesis, to be established (if at all) by empirical investigation independent of any metaethical arguments. By muddling up psychological hypotheses with metaphysical commitments (whereby objectification involves “false belief in the fictitious features” (1980, p. 72)), Mackie makes it impossible for himself to use the objectification thesis in support of moral skepticism in a non-question-begging way.

References Anscombe, G. E. M. 1958. Modern moral philosophy. Philosophy 33: 1–19. Brink, D. 1989. Moral realism and the foundations of ethics. Cambridge: Cambridge University Press. Casati, R. and C. Tappolet. eds. 1998. European review of philosophy 3: Response-dependence. Stanford, CA: CSLI Publications.

Patterns of Objectification

53

Craig, E. 2000. Hume on causality: Projectivist and realist? In The new Hume debate, eds. R. Read and K. Richman, 113–121. London: Routledge. Dancy, J. 1986. Two conceptions of moral realism. Proceedings of the Aristotelian Society 60(suppl.): 167–187. Huemer, M. 2005. Ethical intuitionism. London: Palgrave Macmillan. Hume, D. 1739–1740. A treatise of human nature, eds. L. A. Selby-Bigge and P. H. Nidditch, 2nd ed. 1978. Oxford: Clarendon Press. Hume, D. 1751. An enquiry concerning the principles of morals, ed. T. L. Beauchamp. 2006. Oxford: Oxford University Press. Johnston, M. 1989. Dispositional theories of value. Proceedings of the Aristotelian Society 62(suppl.): 139–174. Johnston, M. 1992. How to speak of the colors. Philosophical Studies 68: 221–263. Johnston, M. 1993. Objectivity refigured: Pragmatism without verificationism. In Reality, representation, and projection, eds. J. Haldane and C. Wright, 85–130. Oxford: Oxford University Press. Joyce, R. 2007. Moral anti-realism. Stanford Encyclopedia of Philosophy. http://plato.stanford.edu/archives/fall2008/entries/moral-anti-realism/. Accessed 17 October 2008. Joyce, R. 2009a. Is moral projectivism empirically tractable? Ethical Theory and Moral Practice 12: 53–75. Joyce, R. 2009b. The skeptick’s tale. Philosophy and Phenomenological Research 78: 213–221. Joyce, R. Forthcoming. Expressivism, motivation internalism, and Hume. In Hume, motivation, and virtue, ed. C. Pigden. London: Palgrave MacMillan. Mackie, J. L. 1946. A refutation of morals. Australasian Journal of Psychology and Philosophy 24: 77–90. Mackie, J. L. 1973. Truth, probability and paradox. Oxford: Oxford University Press. Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin. Mackie, J. L. 1980. Hume’s moral theory. London: Routledge. Mackie, J. L. 1982. The miracle of theism. Oxford: Clarendon Press. Pettit, P. 1991. Realism and response-dependence. Mind 100: 587–626. Ruskin, J. 1856. Modern painters, vol. III. Excerpt in Selections from the work of John Ruskin. 1908. Boston: Houghton Mifflin. Sainsbury, R. M. 1998. Projections and relations. Monist 81: 133–160. Sayre-McCord, G. 1986. The many moral realisms. Southern Journal of Philosophy 24(suppl.): 1–22. Stroud, B. 2000. “Gilding or staining” the world with “sentiments” and “phantasms.” In The new Hume debate, eds. R. Read and K. Richman, 16–30. London: Routledge. Wright, C. 1988. Moral values, projections, and secondary qualities. Proceedings of the Aristotelian Society 62(suppl.): 1–26.

Mackie’s Internalisms Walter Sinnott-Armstrong

1 Mackie’s Basic Argument When Mackie calls certain judgments “queer,” he seems to mean that they differ from other judgments in some way that makes them dubious. Each version of the argument from queerness then picks out a different way in which moral judgments are supposed to differ from other judgments. Since Mackie’s goal is an error theory, he also needs to show that the differences between moral judgments and the other judgments make moral judgments dubious in such a way that they are never true in the same way as the other judgments. The version of queerness based on internalism claims that moral judgments are queer because some special force is internal to moral judgments but not to other judgments. This argument can be presented in steps: 1. 2. 3. 4.

Moral judgments have some special force built into them. Other judgments do not have that special force built into them. Therefore, moral judgments are unlike other judgments. If moral judgments are unlike other judgments in this way, then they are never true (in the same way as the other judgments). 5. Therefore, moral judgments are never true (in the same way as the other judgments). This overall argument works only if Mackie can specify some special force that (a) moral judgments have, (b) other judgments lack, and (c) is important enough in some way that supports Mackie’s skeptical conclusions. Mackie, thus, needs to argue for some kind of internalism in his first premise, and he also needs the right kind of internalism to make the rest of his argument follow. Comparison to other judgments, though suggested by Mackie’s label “queer,” might not be essential to Mackie’s argument. Sometimes (e.g., 1977, p. 35) W. Sinnott-Armstrong (B) Department of Philosophy, Duke University, Durham, NC, USA

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_4, 

55

56

W. Sinnott-Armstrong

Mackie seems to argue simply that moral judgments would need to be both objective and prescriptive. To call them “prescriptive” is, presumably, to say that they have some kind of special force built into them. If this internal prescriptivity conflicts with some kind of objectivity that is also essential to moral judgments, then this conflict could make moral judgments dubious, regardless of whether other judgments resemble moral judgments in these respects. To make this argument work, Mackie would need to specify a special force that (a) moral judgments have and (b) conflicts with a kind of objectivity that (c) moral judgments also have. As with the first interpretation of Mackie’s argument, the trick is to specify and defend a kind of internalism that fits into the rest of the argument. My goal in this chapter is to distinguish several versions of internalism and then to ask which are defensible and which might fit into Mackie’s argument. I will not try to determine whether any version of internalism really does conflict with objectivity or whether objectivity really is built into moral judgments. Instead, I will focus exclusively on Mackie’s premise that some special force is built into moral judgments. The rest of Mackie’s argument will have to wait for another occasion. Mackie’s first premise deserves this special attention, because Mackie’s argument cannot even get started until he tells us precisely which special force is supposed to be built into moral judgments. It is not clear which kind of internalism Mackie has in mind, because he runs together several distinct claims. Here is his canonical presentation of this argument: Plato’s Forms give a dramatic picture of what objective values would have to be. The Form of the Good is such that knowledge of it provides the knower with both a direction and an overriding motive; something’s being good both tells the person who knows this to pursue it and makes him pursue it. An objective good would be sought by anyone who was acquainted with it, not because of any contingent fact that this person, or every person, is so constituted that he desires this end, but just because this end has to-be-pursuedness somehow built into it. Similarly, if there were objective principles of right and wrong, any wrong (possible) course of action would have not-to-be-pursuedness somehow built into it. Or we would have something like Clarke’s necessary conditions of fitness between situations and actions, so that a situation would have a demand for such-and-such an action somehow built into it. (1977, p. 40)

This quotation contains a lot to unpack. We need to distinguish reasons from motivations, overriding motivations from influences, facts from judgments, moral judgments from other normative judgments, and various relations between judgments and special forces. Only after drawing all of these distinctions can we ask whether any kind of internalism could make Mackie’s argument work.

2 Which Special Force? Motivation vs. Reason My phrase “special force” is supposed to be ambiguous enough to cover reasons as well as motivations. Which is Mackie discussing? In the quoted passage, Mackie sometimes refers to “to-be-pursuedness.” To say that something is “to be pursued”

Mackie’s Internalisms

57

is, presumably, not to predict that it will be pursued but, instead, to make a normative judgment that it ought to be pursued or that there is a reason to pursue it. Similarly, when Mackie (following Clarke) suggests that moral judgments are related to a “demand,” this claim seems to be about a certain kind of reason instead of motivation, since an act can be demanded even if the agent who is subject to the demand is not at all motivated to do the demanded act (perhaps because the agent is unaware of the demand). If Mackie’s notions of “to-be-pursued” and “demand” refer to reasons, then he would seem to be claiming some form of reason internalism: Reason internalism: Moral judgments have reasons built into them.

In other parts of the quoted passage, however, Mackie refers not to reasons but instead to a “motive,” and he suggests that judging something good “makes him pursue it.” He also mentions “influence” in the following paragraph (quoted below). At these points, Mackie seems to claim some form of motivation internalism: Motivation internalism: Moral judgments have motivations built into them.

Reason internalism and motivation internalism are distinct if reasons differ from motivations. (On these distinctions and more below, see Brink 1989, pp. 37–42.) Whether reasons differ from motivations depends on what counts as a reason. We need to distinguish reasons for agents to do things (called justificatory reasons or justifications) from reasons why people do things (called explanatory reasons or explanations). An act is justified by citing some fact about the act or the situation (such as that the act prevents some harm or fulfills a promise), whereas an act is explained by citing some fact about the agent (such as the agent’s beliefs and desires). A justificatory reason for an agent to do something (such as because he promised to do it) need not explain why the agent did it (since the agent might do the promised act out of some independent desire to do it). Conversely, an explanatory reason why someone did something (such as because he wanted to harm a hated rival) might not be any justificatory reason at all. Explanatory reasons are tied closely to motivation if an explanatory reason why an agent does an act is always some motive of the agent. This universal claim is questionable, but let that go for now. The crucial point here is that a justificatory reason for the agent to do an act need not be related in any way to the agent’s motives. An agent can have a justificatory reason to buy milk simply because he promised to buy milk, even if he has no motivation at all to buy milk, because he forgot that he promised to buy milk. Similarly, imagine that pushing on the brake pedal of my car is necessary to avoid running over Jim, and I want to harm Jim and have no motivation to avoid running over Jim. I can still have a justificatory reason to push on the brake pedal and to avoid running over Jim, because the fact that pushing on the pedal is necessary to avoid harming Jim would justify that act if I were to do it. Williams (1981) and his followers deny this point and claim that no reasons are external roughly in the sense of independent of desire and motivation. However,

58

W. Sinnott-Armstrong

Williams’ claim seems to depend on a demand that all reasons be explanatory. Once we distinguish explanatory reasons from justificatory reasons, it becomes clear that justificatory reasons need not always be based on desire or motivation (or any other mental state of the agent). In any case, Mackie admits that justificatory reasons can be independent of motivation, even though he goes on to suggest that such reasons depend on institutions (1977, pp. 78–79), so I will assume that independence here. Thus, if reason internalism claims only that justificatory reasons are built into moral judgments, this view is distinct from motivation internalism. In contrast, if reason internalism claims that explanatory reasons are built into moral judgments, this view might not be separable from motivation internalism. Hence, for the sake of clarity, the former, justificatory version is what I will mean by reason internalism from now on. This kind of reason internalism does not seem queer in any way that should raise doubts about objective values or moral facts. After all, Mackie analyzes “ought” in terms of reasons: A first attempt at a general equivalent of “a ought to G” might be “There is a reason for a’s G-ing.”. . . We could then say that different uses (not different senses) of “ought” introduce different kinds of reason. (1977, pp. 73–74)

Hence, it is analytic that, if a morally ought to G, then there is a ( justificatory) moral reason for a to G. Given this analysis, Mackie cannot claim that there is anything queer about moral judgments having justificatory reasons built into them internally. These reasons are simply what the judgments are about. There still might be a separate problem with objective justificatory reasons themselves (cf. Mackie 1977, p. 79). That problem might even be the basis for an independent argument that moral judgments are never true. However, such a problem for objective justificatory reasons could not be shown by pointing out that moral judgments imply objective justificatory reasons. Conversely, the fact that moral judgments imply objective justificatory reasons also cannot show that there is anything dubious about objective justificatory reasons. The reasons themselves are distinct from the relation between moral judgments and reasons. The argument by Mackie that is our topic here concerns the relation between moral judgments and a special force. Even if objective justificatory reasons were dubious on independent grounds, the relation between moral judgments and reasons still would not be what makes moral judgments queer or dubious. To appreciate this point, compare a non-normative case: “son” has “male” built into it, and nothing else does, unless it is also analytically connected to “son,” since every male is a son. This relation does not show anything queer about being a son that would raise any suspicions about whether sons exist. Of course, if the property of being male were suspect, that other suspicion would transfer to being a son, but the relation of “son” to “male” still would not be suspect and would not be the source of suspicion about being a son. Analogously, even if “ought” has “reason” built into it, and nothing else does, unless it is analytically connected to “ought,” that relation between moral judgments and reasons does not show anything queer

Mackie’s Internalisms

59

about oughtness that should raise doubts about whether oughtness exists or whether any moral judgments about what ought to be done can be true. Of course, if reasons are seen as necessarily motivating, as in the case of explanatory reasons, then reason internalism still might support queerness and skepticism. However, that is not because of reasons, but because of motivation. Thus, the only kind of internalism that could help Mackie’s argument here is motivation internalism. On this view, moral judgments would be queer, because one kind of mental state (a motivation) is built into an apparently very different kind of mental state (a judgment). Moral judgments seem to be cognitive appraisals of the moral status of acts or agents, whereas motivations involve a separate part of the brain and mind that concerns action rather than cognition. Cognition might be necessary for action, but it is not clear how action or motivation to act could be necessary for or built into cognition. We will discuss some possibilities below. The point for now is only that motivation internalism captures a kind of queerness that might make Mackie’s argument work.

3 How Much Motivation? Some vs. Overriding Motivation Motivations conflict and come in degrees. In the passage quoted at the start, Mackie refers to “an overriding motive.” He adds, “something’s being good. . . makes him pursue it,” and a motive does not make a person act unless it is overrides all conflicting motives. This suggests that Mackie has in mind something like this: Overriding-motivation internalism: Moral judgments have overriding motivations built into them.

In other passages, however, Mackie mentions less than overriding motivation. Consider the paragraph right after the one quoted initially: The need for an argument of this sort [that is, for motivation to be built into objective values] can be brought out by reflection on Hume’s argument that “reason”—in which at this stage he includes all sorts of knowing as well as reasoning—can never be an “influencing motive of the will.” Someone might object that Hume has argued unfairly from the lack of influencing power (not contingent on desires) in ordinary objects of knowledge and ordinary reasoning, and might maintain that values differ from natural objects precisely in their power, when known, automatically to influence the will. To this Hume could, and would need to, reply that this objection involves the postulating of value-entities or value-features of quite a different order from anything else with which we are acquainted, and of a corresponding faculty with which to detect them. That is, we would have to supplement his explicit argument with what I have called the argument from queerness. (1977, pp. 40–41)

The last sentence refers to the epistemological part of Mackie’s argument from queerness. The relevant point here is that an “influencing motive” need not always be overriding. Hence, this passage suggests that the problem of queerness arises even for a weaker view: Some-motivation internalism: Moral judgments have some motivation built into them.

60

W. Sinnott-Armstrong

Of course, a motivation need not be overriding, so some-motivation internalism is much weaker than overriding-motivation internalism. This weaker view is also more plausible. People normally act on their overriding motives, so overriding-motivation internalism makes it hard to see how anyone could ever do what they know is morally wrong or bad. Overriding-motivation internalism can also be refuted by common examples. Consider a conflict between self-interest and moral judgment. Even if I judge that it is wrong to steal some money, I might go ahead and steal the money because my desire for the money (and what it can buy) overrides the motivation from my moral judgment. My moral judgment provides some impartial motivation not to steal, but that motivation might not be as strong as my conflicting selfish or partial motivation to get the money. This kind of example refutes overriding-motivation internalism, but it is compatible with some-motivation internalism. Some-motivation internalism is also strong enough for Mackie’s argument if nonnormative judgments do not have any motivation at all built into them. Since an argument is better when its premises are weaker and more plausible, Mackie’s argument is better if we interpret him as claiming only some-motivation internalism. So that is what we will do, despite some passages to the contrary.

4 What has Special Force? Facts vs. Judgments So far internalism has been formulated in terms of judgments, but that word is ambiguous. People who talk about judgments sometimes refer to a mental state or act of making a judgment, but sometimes they refer instead to the content of the judgment, which is an abstract proposition. When that content is true, it represents a fact. It is not clear, then, whether it is facts or mental states that are supposed to have a special force, according to Mackie. Mackie sometimes says that “something’s being good” or a “course of action” or a “situation” has some special force built into it. These formulations suggest that what has the special force is separate from the mental state of the person who makes the moral judgment. At other times, in contrast, Mackie says that “knowledge” of the Good or being “acquainted” with a good provides such a special force. These formulations suggest that what has the special force is partly a mental state of the person who makes the moral judgment. These pictures point toward very different forms of internalism: Fact internalism: Moral facts have some special force built into them. Judgment internalism: Moral judgments (as mental states or acts) have some special force built into them.

To see the difference between these views, consider an agent who chooses between two possible actions in a situation where one of the actions is immoral, but the agent does not judge (or know) that the action is immoral. Fact internalism implies that not-to-be-pursuedness is built right into the situation or possible action, so not-to-be-pursuedness should be present in this example. In contrast, judgment

Mackie’s Internalisms

61

internalism does not imply any not-to-be-pursuedness in this case, because no judgment about immorality is present. Conversely, if an agent judges an action to be immoral when it really is not immoral at all, then judgment internalism implies notto-be-pursuedness, but fact internalism lacks that implication. As Mackie presents them in the initial quotation (1977, p. 40), Clarke is a fact internalist, whereas Plato is more of a judgment internalist (or knowledge internalist). Since these views are so different, we need to separate them. Mackie might seem to be unable to distinguish fact internalism from belief internalism, because he is an error theorist who denies normative facts. However, Mackie is here characterizing the commitments of his opponents, and they claim normative facts. Hence, he needs to distinguish these kinds of internalism in order to specify what his opponents are supposed to be committed to. The plausibility of these versions of internalism depends on whether the special force involves reasons or motivations. Fact internalism for motivations would claim that moral facts have motivations built into them. It is hard to see how a person’s motivation could be affected by a fact of which the person has no awareness at all. That really would be queer. A fact can still provide a justificatory reason to do an act even when the agent is not aware of the fact, such as when the agent forgot that he promised to do the act. Thus, fact internalism is plausible for reasons but not for motivations. In contrast, judgment internalism makes little sense if the claimed special force involves reasons, because the judgment might be false or even unreasonable, and then it cannot provide a justificatory reason. Judgment internalism seems more plausible if the special force involves motivation instead of reasons, since then it claims that people who accept moral judgments are motivated to act accordingly, and the truth or falsity of such judgments does not seem to affect their motivational force. Thus, judgment internalism fits better with motivations than with reasons. Since we already suggested that Mackie should have in mind motivation (as opposed to reason) internalism, he should also have in mind the judgment (as opposed to fact) version of motivation internalism. Mackie does not, however, explicitly mention judgments in the quotation at the start (from 1977, p. 40). Instead, he discusses knowledge in the second quoted sentence, and then his third sentence refers to being acquainted with an objective good. These topics are distinct, because knowledge does not require acquaintance, even if acquaintance guarantees knowledge (which is questionable, even if Mackie assumes it). If knowledge is justified true belief or more, then it is not clear whether the special force of moral knowledge is supposed to be built into the belief itself, into the justification for the belief, or into whatever it is that makes some beliefs knowledge. Thus, we need to distinguish at least these versions of internalism: Knowledge internalism: Moral knowledge has some special force built into it. Acquaintance internalism: Acquaintance with moral facts has some special force built into it. Justified-belief internalism: Justified moral beliefs have some special force built into them.

62

W. Sinnott-Armstrong

These views have distinct implications for the presence of the special force in cases where people have knowledge without acquaintance or acquaintance without knowledge, or where people have knowledge without justified belief or justified belief without knowledge. All of these views are distinct from judgment internalism, as defined above, insofar as people can make normative judgments that are not justified, do not count as knowledge, and are not based on acquaintance. Knowledge internalism, acquaintance internalism, and justified-belief internalism do not imply any special force in such cases, but judgment internalism does. What turns a judgment into knowledge or acquaintance or justified belief seems to be something external to the judgment itself, such as the truth (or facts), the cause of the judgment, the reliability of a method, or maybe coherence with other beliefs. It is again hard to see how such external factors could be what guarantees motivation. The more plausible claim, then, is that moral judgment includes motivation regardless of whether that judgment is justified or knowledge or based on acquaintance. Thus, Mackie’s argument is best interpreted as resting on judgment internalism rather than knowledge, acquaintance, or justified-belief internalism. Judgment internalism itself, however, still needs to be clarified. The term “judgment” is intended to cover both beliefs and thoughts (and possibly also attitudes). An occurrent thought, such as when I momentarily direct my attention to the fact that Mars is a planet, is more active and more temporally restricted than a standing belief state with the same propositional content. On one account, the belief state consists in a disposition to assent to the claim that Mars is a planet, if one thinks about it; and that dispositional state can exist even while I am not thinking at all about Mars or planets. This distinction between beliefs and thoughts applies as well to moral judgments: I can believe that torture is immoral even at a time when I am not thinking about torture or morality. This distinction yields two possibilities: Thought internalism: Normative thoughts have some special force built into them. Belief internalism: Normative beliefs have some special force built into them.

If the special force involves reasons, and if I can have a reason to do something even while I am not thinking about it, then thought internalism seems too narrow, and belief internalism (or, rather, justified-belief internalism or knowledge internalism) might be a more natural view. However, if the special force involves motivation, then belief internalism would imply that I am motivated to keep my promise even when I am asleep or not thinking about my promise at all, but that seems implausible, assuming that motivation is occurrent rather than dispositional. Thus, if the special force involves motivation, then the natural views would seem to be that moral thoughts contain motivation and that moral beliefs (as dispositions to have such thoughts) contain only dispositions to be motivated. It might not, however, be all moral thoughts that contain motivation, since not all moral thoughts are equally perspicuous. Some acts, such as rape, immediately strike me as immoral. In other cases, it takes a trusted authority or a long train of reasoning to convince me that something is wrong. Suppose I could get away with

Mackie’s Internalisms

63

claiming a deduction on my taxes, and I believe that I ought to be allowed to take this deduction, but I also know that it violates a silly tax rule that happens to be in force. Is it immoral for me to claim that deduction? I think it is, but taking this deduction still does not strike me as immoral in the immediate way that rape strikes me as immoral. After inferring or being told that the tax deduction is wrong, I might come to see and feel that it is wrong, but I might not. Thus, I can believe, think, and even know that some acts are wrong, on the basis of trusted testimony or long argument, even though they do not immediately strike me as wrong or appear or seem wrong to me. My situation then resembles the Müller-Lyer illusion. In that illusion, I believe and even know that the lines are the same length, but the lines still do not appear or seem to have the same length. In the analogous moral case, an act does not seem immoral to me even after I come to think that it is immoral. Although I can believe what does not seem to me to be true, and I can fail to believe what does seem to me to be true, sometimes I do endorse what seems true, and then I believe or think that it is true. The mental state that occurs when a beliefcontent seems or appears to be true can be called an appearance or seeming (see Tolhurst 1998). A belief or thought occurs when the seeming is endorsed. It is not clear whether the special force claimed by internalism is supposed to be built into the seeming or into the endorsement of the seeming, so an internalist has these options: Seeming internalism: Moral seemings have some special force built into them. Seeming-judgment internalism: Moral judgments that seem true have some special force built into them.

These versions have different implications when an act seems wrong but that seeming is not endorsed. Imagine that you were raised to believe that dancing is immoral, but now you have reached the conclusion that dancing is not immoral after all. Dancing still might seem immoral (just as the lines still seem different lengths in the Müller-Lyer illusion). Your mental state must then include the special force (or motivation) according to seeming internalism but not according to seeming-judgment internalism. On either version, moral thoughts that do not seem true need not have any special force, but all moral thoughts that do seem true to us must have that special force. One way to describe the difference between believing and seeming might be to say that when a wrong act seems wrong to me, then I am acquainted with its wrongness. In contrast, when a wrong act does not seem wrong to me, then I am not acquainted with its wrongness, even if I know its wrongness in some other way, such as by testimony or inference rather than by acquaintance. If Mackie means something like this by “acquainted” when he says, “An objective good would be sought by anyone who was acquainted with it” (1977, p. 40), then his premise in this argument is something like seeming internalism or seeming-judgment internalism. Seeming internalism and seeming-judgment internalism are weaker and, thus, easier to defend than judgment internalism, because some purported counterexamples to judgment internalism are not counterexamples to seeming internalism or to

64

W. Sinnott-Armstrong

seeming-judgment internalism so long as the judgment in the purported counterexample does not seem true to the person who lacks motivation to act accordingly. Nonetheless, if non-moral judgments do not have any special force built into them, even when they seem true, then seeming internalism and seeming-judgment internalism are still strong enough to show that moral seemings are unlike those non-moral seemings. Mackie still needs to show why this difference raises doubts about the existence of objective values or about the truth of moral judgments, but the best premise for getting Mackie’s argument started is some form of seeming internalism or seeming-judgment internalism.

5 Which Judgments? Normative vs. Moral The versions of internalism defined so far apply only to moral judgments, so they are all versions of: Moral internalism: Moral judgments have some special force built into them.

However, if what gives moral judgments their special force is something shared by other normative judgments, then internalists might endorse this broader claim: Normative internalism: Normative judgments have some special force built into them.

A more restricted version is suggested when Mackie introduces queerness while talking about normative judgments that claim objectivity: Objective-normative internalism: Normative judgments that claim objectivity have some special force built into them.

These last two internalisms imply moral internalism if moral judgments are normative and claim objectivity, as Mackie argues. He specifies that his view is supposed to apply not only to moral judgments but to all normative (and evaluative) judgments, or at least to those that claim objectivity (1977, p. 15), so he seems to endorse not only moral internalism but also either normative internalism or objective-normative internalism. It is hard to see any rationale for objective-normative internalism, however. Why would normative judgments that claim objectivity have motivation built into them if normative judgments that do not claim objectivity lack such special force? Without an answer to this question, anyone who endorses objective-normative internalism would have just as much reason to endorse the broader view of normative internalism. A claim to objectivity might conflict with having motivation built into a judgment, and that conflict might be what makes moral judgments dubious, but the claim to objectivity does not seem to be what generates the motivation. In any case, some normative judgments that claim objectivity do not seem tied to action or motivation in the same way as moral judgments. One example is epistemic judgments. The fact that I have a certain kind of visual experience in certain

Mackie’s Internalisms

65

circumstances can make me justified in believing that my car is red. This epistemic judgment is normative insofar as it affects what I ought to believe, but it need not motivate me to act in any particular way, such as to seek or avoid the car or its color or to say out loud that my car is red, even if asked. (I might want to keep its color secret.) Thus, if the special force claimed by internalism involves motivation to act, then it is hard to see why internalism should be extended to all normative judgments or even to all normative judgments that claim objectivity, since either of these extensions would include epistemic judgments. In contrast, if the special force claimed by internalism involves only reasons and not motivation, then it is easier to understand how all normative judgments that claim objectivity might seem to have such special force. After all, when my visual experience makes me justified in believing that my car is red, I have a reason for that belief, even if I have no motivation to do anything about it. Thus, normative epistemic judgments will be just as queer and suspect as normative moral judgments if what makes moral values queer and suspect is their relation to reasons. However, as we saw, their relation to reasons does not make moral values queer. In contrast, if it is their relation to motivations that makes moral values queer and suspect, then there might be nothing queer or metaphysically suspect about epistemic facts or judgments, even if those epistemic judgments are normative, and even if moral facts and judgments are queer and suspect, as Mackie claims. It is also not clear whether normative aesthetic judgments have any special motivation built into them (or whether they make the same claim to objectivity as moral judgments). Mackie (1977, pp. 15, 43) includes aesthetic judgments, but it is not clear why I cannot judge that a modern painting (or piece of music) has the positive aesthetic quality of being creative, even though I have no desire at all to see it (or hear it played). If moral judgments have motivation built into them, that would seem to be because they are practical – that is, because they are related to action in a certain way that needs to be specified. Then all other practical judgments, including some non-moral ones (such as judgments about what we ought to do for self-interested reasons), would also be implicated in Mackie’s claim. We cannot tell exactly which normative judgments have motivation built into them until we determine what relation to action is needed in order to make a judgment practical and which judgments are practical in that way. All we can say for now is that it would be natural for internalists to claim that all practical judgments have motivation built into them, where practical judgments include more than moral judgments but less than all normative judgments. Practical judgments include a special kind of overall judgment. When I say that an agent ought to do an act, this might sometimes mean that there is most reason overall for that agent to do that act, considering all reasons of all kinds (moral, selfinterested, aesthetic, etc.). If such overall practical judgments have motivation built into them, it might seem to be overriding motivation, because there is no reason that is not considered which could provide any other motivation to override the motivation to do what one judges one overall ought to do.

66

W. Sinnott-Armstrong

However, not all practical judgments are overall judgments. Practical judgments often consider a limited class of reasons. They also often claim only prima facie reasons or reasons that are overridden. That is why practical judgments in general have at most some motivation and not necessarily overriding motivation built into them. Besides, overall practical judgments refer to moral reasons among others, and sometimes these moral reasons determine what someone ought to do overall. Suppose someone overall ought to do something because doing it is morally required, even though it conflicts with that person’s self-interest. For example, as above, imagine that I judge that it is morally wrong for me to steal some money, but I steal the money anyway, because my desire for money (and what it can buy) overrides the motivation from my moral judgment. My impartial moral motivation is overridden by my selfish desire for the money. Then, even if I have some motivation to do what I judge that I overall ought to do, I do not have overriding motivation to do it. Thus, even overall practical judgments need not guarantee overriding motivation. At most some motivation is built into even overall practical judgments. Internalism might be defended by limiting it not only to practical judgments but, further, to a subset of practical judgments. After all, practical judgments vary along several dimensions that might affect their relation to motivation. First, practical judgments can be very general or very specific (or even particular), and the precise relation to motivation varies accordingly. People might make a general judgment that it is morally wrong to break promises even if they themselves never make any promise, so they are never motivated by their judgment to do anything in particular. They are at most conditionally motivated to keep a promise if they make one. The same point applies to particular judgments about hypothetical situations. People can judge that a certain act is morally wrong even if they never could do that act, such as when a Catholic male judges that it would be morally wrong for him to get an abortion in a specific set of circumstances, such as if he were in a female friend’s situation. Such a Catholic male is not motivated to avoid an abortion, except maybe conditionally, and he also might not be motivated to do anything to women who get abortions or to prevent anyone from getting an abortion. This point applies also to third-person moral judgments. I can judge that you morally ought to pay your debts without feeling any motivation to pay your debts for you or to do anything to you if you do not pay your debts. Perhaps I need to be motivated to pay my debts if I were in a similar situation, but that is another kind of conditional motivation. The conditional nature of such motivations matters because conditional motivation need not be detectable by introspection or by observation. The point is not about motivations that arise only in hypothetical situations but is instead about actual motivations with conditional content. Such conditional motivations need not produce any feeling, behavior, or skin conductance response, as far as we know. Hence, the absence of a feeling or skin conductance response cannot be used to show that no such motivation is present. (Pace Roskies 2003, 2008.)

Mackie’s Internalisms

67

The situation is different when someone makes a first-person “in situ” judgment about a concrete action that he or she is able to do at present, such as when I judge that right now I morally ought to go pick up my son, because I promised to do so at this time. (See Kennett and Fine 2008.) Then the motivation is not just conditional. This non-conditional motivation still might not be associated with any introspectible feeling or observable behavior or skin conductance response, but it seems more natural to expect some introspectible or observable difference when such non-conditional motivation is present than with only conditional motivation. What about first-person judgments of acts that I cannot do now but will be able to do in the future? Suppose I judge at 3:00 p.m. that I morally ought to pick up my son at 4:00 p.m. What is built into this judgment, according to internalism, is the judgment that I am motivated at 3:00 to pick up my son at 4:00. The judgment and the motivation have matching content (Smith 2008). Since I do not have any motivation right now to do anything right now, I need not currently feel any tug towards any particular behavior (unless I need to do something now in order to get ready to pick him up at 4:00). Thus, even if internalists are committed to some motivation at 3:00 in this case, they need not predict that this motivation will be detectable either by introspection or by any external means, such as behavior or skin conductance response. The weakest version of internalism limits its claim to first-person present-tense (in situ) moral thoughts that seem true. It is hard to see why internalists would limit their claim this much except to make it easier to defend. Nonetheless, this minimal internalism is strong enough to support Mackie’s skepticism about moral values if no motivation is contained in the corresponding non-normative judgments, and if Mackie or his supporters can show why this feature of moral judgments is relevant and important enough to cast doubt on the existence of objective moral values or truths.

6 What is it to be “Built in”? A final metaphor that needs to be unpacked is that of being “built in.” To claim that motivation (or reason) is built into normative (or moral) judgments (or facts) is to claim some strong relation between what is built in and what it is built into. One possibility is universal material implication: Universal internalism: Every person who makes a moral judgment has some motivation to act accordingly.

For judgments that an act is good or required, to act accordingly is to do the act. For judgments that an act is bad or wrong, to act accordingly is to avoid doing the act. Although universal, this view is not strong enough for Mackie’s purposes. One reason is that some non-normative judgments might also be related to motivation in this way. It might be true (that is, it is true in some possible worlds) that, whenever

68

W. Sinnott-Armstrong

a person judges that an act is such that whoever does it will die, then that person is at least somewhat motivated not to do that act. However, the judgment that an act is such that whoever does it will die is simply a prediction and not a normative or moral judgment. Hence, universal internalism would not show that moral judgments must be unlike all non-normative judgments, much less that they are queer in any way that would cast suspicion on the existence of objective normative values or truths. The same point applies if the relation to motivation holds by psychological or physical necessity based on the nature of our species or our brains. Perhaps all people who are not motivated to do what they judge to be good would have to be mutants whose bodies and brains are not wired in the way that is normal to the human species. This connection between judgment and motivation might be as much a part of human nature as the fact that eyes are normally connected to brains in the human species. However, this physical or psychological necessity would still not make moral, practical, or normative judgments unlike all non-normative judgments, since it also might be a part of human nature that all normal people have some motivation to avoid acts such that people who do them die. Mackie recognizes the need for a stronger relation when he writes: “An objective good would be sought by anyone who was acquainted with it, not because of any contingent fact that this person, or every person, is so constituted that he desires this end” (1977, p. 40). He refers in the next paragraph to the power of values “when known, automatically to influence the will.” The adverb “automatically” does not, however, help much here. My fire alarm goes off automatically when it detects smoke, but that is still just a matter of physical necessity based on the nature and structure of the alarm. To say that it is automatic is just to say that it happens without further steps such as me pushing a button or pulling a lever, as is required to set off some other fire alarms. Internalists are claiming more than that normative judgments lead to motivations automatically, if all that means is that the motivation arises without further steps such as conscious thought. They are claiming that motivation is somehow built into the judgment itself. Hence the name “internalism.” To see what this means, imagine a car coming at you fast. If you see it, that will influence your choice to move out of its way. It might even happen that every human is structured by human nature to have some desire to move out of the way. Nonetheless, it won’t be conceptually necessary that everyone who believes that a car is coming at them will be motivated to move, since there is no inconsistency in imagining a person who wants to be hit and hurt. Thus, moral beliefs still differ from such non-normative beliefs if the relation between moral judgments and motivation is conceptual necessity or something like that: Conceptual internalism: It is conceptually necessary that every person who makes a moral judgment has some motivation to act accordingly.

This version, of course, implies universal internalism, but it would be denied by externalists who claim that it is a contingent fact about human nature that all normal humans are motivated not to do what they judge to be wrong.

Mackie’s Internalisms

69

Conceptual internalism is descriptive in the sense that it does not imply any negative normative judgment about anyone (except the judgment that those who deny it are wrong). Some philosophers (such as Smith 1994), in contrast, formulate internalism as a normative claim like this: Rational internalism: It is conceptually necessary that every person who makes a moral (or practical) judgment has some motivation to act accordingly or else that person is irrational.

This view allows that people can judge that an act is morally required but still not be motivated at all to do that act. Instead of ruling out such people as impossible, rational internalism admits their possibility but deems them irrational. This version is very different from the kinds of internalism that we have discussed so far. It is also easier to defend, because it cannot be refuted by hypothetical or actual counterexamples without establishing controversial claims about rationality. It is not completely clear, however, why it must always be irrational to make an impartial moral judgment without having any desire to be impartial. Consider, as before, someone who judges that it is morally wrong to steal some money, perhaps because he recognizes that he would favor a public rule against stealing if he were impartial. He could still not care about public rules or about being impartial. Then he might not have any motivation at all not to do the act that he judges to be morally wrong. His only motivation is a selfish desire to get the money for himself. However, since that desire is not irrational, it is hard to see why we would have to count him as irrational. Defenders of rational internalism might respond by limiting their view to judgments about what ought to be done overall, not just morally. However, as mentioned above, moral reasons sometimes determine what one overall ought to do. In such cases, our agent could judge that stealing is morally wrong and conclude that it is overall morally wrong, while still not caring or being motivated to act accordingly. It is not clear that he could fail to care if the overall judgment seems or appears true to him, but that unclarity is largely because the notions of seeming and appearance are themselves not determined precisely enough for this issue. Thus, rational internalism might or might not be defensible if limited to overall normative judgments that seem true. In any case, rational internalism cannot help Mackie, because it need not show any essential difference between normative and non-normative judgments. Consider the judgment that anyone who does a certain act will suffer severe pain and then die. Anyone who makes this judgment but is not motivated to avoid doing that act could plausibly be deemed irrational. This might even seem conceptually necessary by virtue of our concept of rationality (see Gert 2005). However, this judgment is not normative. It simply predicts what will happen to anyone who does act. Hence, rational internalism is not enough to show that normative judgments are queer, much less that they are queer in any way that would support Mackie’s skepticism.

70

W. Sinnott-Armstrong

7 Conclusions We have arrived at the end of our discussion, but where is that? I argued first that reason internalism would not support Mackie’s argument, so he needs some form of motivation internalism. Then I argued that overriding-motivation internalism is too strong to be plausible, but Mackie could rest his argument on some-motivation internalism. Third, I argued that fact-internalism about motivation is implausible, as are knowledge internalism, acquaintance internalism, and justified-belief internalism about motivation, so Mackie is best read as claiming judgment internalism or, even better, seeming or seeming-judgment internalism. Next, I suggested that internalism could be made even easier to defend by limiting it to first-person present-tense (in situ) moral thoughts that seem true. Finally, I argued that internalists still need to claim a strong conceptual relation between these judgments and motivation in order to end up with a version that could support Mackie’s conclusion. I have not determined whether any version of internalism is true. Instead, I have only been trying to clarify what internalism is, or which internalisms there are. The eventual goal is to determine whether any version of internalism is both plausible and strong enough to support Mackie’s conclusion. Reaching that goal will have to wait for another time, but I hope that this chapter has clarified some of the issues and options. Acknowledgments Thanks to Adina Roskies, Richard Joyce, and Simon Kirchin for helpful comments on drafts of this chapter.

References Brink, D. 1989. Moral realism and the foundations of ethics. New York: Cambridge University Press. Gert, B. 2005. Morality: Its nature and justification, Rev. ed. New York: Oxford University Press. Kennett, J. and C. Fine. 2008. Internalism and the evidence from psychopaths and “acquired sociopaths.” In Moral psychology, volume 3: The neuroscience of morality, ed. W. SinnottArmstrong, 173–190. Cambridge MA: MIT Press. Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin. Roskies, A. 2003. Are ethical judgments intrinsically motivational? Lessons from “acquired sociopathy.” Philosophical Psychology 16: 51–56. Roskies, A. 2008. Internalism and the evidence from pathology. In Moral psychology, volume 3: The neuroscience of morality, ed. W. Sinnott-Armstrong, 191–206. Cambridge MA: MIT Press. Smith, M. 1994. The moral problem. Oxford: Blackwell. Smith, M. 2008. The truth about internalism. In Moral psychology, volume 3: The neuroscience of morality, ed. W. Sinnott-Armstrong, 207–216. Cambridge MA: MIT Press. Tolhurst, W. 1998. Seemings. American Philosophical Quarterly 35: 293–302. Williams, B. 1981. Internal and external reasons. In his Moral luck, 101–113. Cambridge: Cambridge University Press.

Mackie’s Realism: Queer Pigs and the Web of Belief Jamie Dreier

This chapter is about John Mackie’s moral realism, and so part of its task will be to convince the reader that Mackie was a moral realist. Even if it fails in this task, I hope the main idea will be plausible enough to make the discussion worthwhile. That idea is to present the metaethical issues of realism and antirealism in a certain way and think about how some familiar notions about Mackie’s view look different in the new light. It sounds perverse to call Mackie a realist, but I am not doing it just to be perverse. We could start with the thought that an atheist is not an antirealist about God, or better, about God-talk (“religious discourse,” let’s say).1 The antirealists are those who think that there is no fact of the matter of God’s existence, that attributing holiness, or sacredness, or blessedness, to things, people, actions is not a matter of positing special objects and properties but rather a matter of taking a certain stance, a spiritual stance as it were, toward the same objects we can and do take an ordinary stance toward.2 Atheists think that judgments about spiritual matters are plain judgments (only maybe not about plain things), which can be true or false. They are all false (at least the ones with existential import). But they are truth-apt. Then we might be quite happy with the implication that Mackie’s “error theory,” which is surely analogous to atheism, is also a realist theory, according to which moral judgments are plain judgments of unplain (queer) facts, all truth apt, and all false (again, it’s really just the moral judgments with existential import that are false). I’ve managed to convince myself that this implication is a happy one; maybe I can convince the reader.3 J. Dreier (B) Department of Philosophy, Brown University, Providence, RI, USA 1 John

Bigelow first explained to me why we might think of atheism as a kind of realism about religious discourse, in conversation in 1994. 2 Think of A. J. Ayer on religious discourse in Language, Truth and Logic (1946) Chapter 6 – the same chapter that offers his conception of ethical discourse. 3 Oddly enough, Mackie does on one occasion seem to endorse Bigelonian antirealism: “One way of stating the thesis that there are no objective values is to say that value statements cannot be either true or false” (1977, p. 25). He rejects this as misleading, but not on the grounds that moral

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_5, 

71

72

J. Dreier

In the first section, I’ll do some metametaethics, starting off with just a bit of metaphilosophy. Then I’ll work through some Mackie exegesis. In the third section, I’ll try to show why, according to my metametaethics, Mackie was a realist. The fourth section explains a mistake central to a well-known literary plot, and in the end I will argue that Mackie made a parallel mistake in plumping for realism.

1 Metametaethics Some people think of antirealism as a metaphysical thesis, and some think of it as a thesis about language.4 (I mean that each sort of antirealism, whether in ethics, mathematics, modality, and so on, can be thought of as either a thesis about the metaphysics of a certain domain of objects and properties or else as a thesis about the semantics of a certain section of language.) There’s no reason that there couldn’t be two realist/antirealist distinctions, presumably related – two ways of dividing up philosophical theories, perhaps, that don’t draw the same boundaries, and no serious question about which way is correct. More plausibly there are different classification schemes each more illuminating for some purposes. At a first pass, then, semantic theorists are more interested in classifying theories according to whether they count sentences (of a certain kind) as “truth apt,” while metaphysicians are more interested in classifying theories by their ontology. So I could finesse the difficult task of convincing you to think of Mackie as a realist simply by pointing out that philosophy of moral language is an important topic and noting that Mackie’s view does count moral sentences as true or false. But I’m a little more ambitious than that. I’m going to try to give a reason that we should concentrate on a certain way of drawing the line. If we draw it that way, Mackie falls on the realist side. This will show us something about Mackie, and something about realism.

1.1 The Problem of the Boundary Drawing the boundary between realist metaethics and antirealist metaethics is a problem of metametaethics. The metametaethical problem I’m particularly interested in is one that was raised a couple of decades ago and is becoming more familiar

statements are false; rather, he follows what was the standard view that “relative” statements of value can be straightforwardly true or false, where the relativization is to the speaker or agent’s desires. But later, in any case, Mackie puts the matter the way he should, the way one would expect given his arguments: “[T]he denial of objective values will have to be put forward not as the result of an analytic approach, but as an ‘error theory,’ a theory that although most people in making moral judgements implicitly claim, among other things, to be pointing to something objectively prescriptive, these claims are all false” (1977, p. 35). 4 This section recaps and explores my metametaethical suggestions in Dreier (2005, pp. 23–24) and Dreier (2006).

Mackie’s Realism: Queer Pigs and the Web of Belief

73

nowadays as well as more pressing. It’s the problem of distinguishing sophisticated antirealists, like Simon Blackburn (1984, 1993), Allan Gibbard (1990, 2003), and Terry Horgan and Mark Timmons (2006), from some of the more extravagant moral realists, the Mooreans, among whom I include Russ Shafer-Landau (2003), Robert Audi (1997), and there are probably others. The problem arises, we all know, because, the sophisticated antirealists take it as part of their project to vindicate all of the realist-sounding language of ordinary moral conversation and debate; they are quasi-realists, as Blackburn says. It is made more acute because of the attractiveness and recent successes of deflationary philosophy of language, especially minimalism about truth and allied semantic notions. I think this problem and the way that the two trends (quasi-realism and minimalism) contribute to it are fairly well known, so I will just explain briefly. Quasi-realism presumably needs no introduction, but in case it does, here is Blackburn’s 1984 statement. Mackie, Blackburn notes, held that the “expressive” origin of our moral judgments, when properly understood, shows ordinary moral thinking to be in error. But perhaps there is no mistake. I call the enterprise of showing that there is none—that even on antirealist grounds there is nothing improper, nothing “diseased” in projected predicates—the enterprise of quasi-realism. The point is that it tries to earn, on the slender basis, the features of moral language (or of other commitments to which a projective theory might apply) which tempt people to realism. (Blackburn 1984, p. 171)

On the other side, minimalism about truth helps make it easier to earn some realist-sounding features of moral language. Truth-talk comes free when expressive language has the right grammatical features, as it plainly does, since calling a sentence true is equivalent to asserting the sentence (or at least to what seems on the surface to be asserting; if we can adopt a minimalist conception of assertion then anything with the grammar of assertion will just be assertion). Talk of propositions and facts, and of properties and reference, also follow in the wake of proper grammatical, propositional structure. The result is a kind of fading away of the metaphysical (facts, properties) and semantic (truth, reference) distinctions that once seemed so starkly to divide expressivists from realists. Here is Gibbard’s recent mention of the elision: Does this mean that there are no facts of what I ought to do, no truths and falsehoods? Previously I thought so, but other philosophers challenged me to say what this denial could mean. In this book, I withdraw the denial and turn non-commital. In one sense there clearly are “facts” of what a person ought to do, and in a sense of the word “true” there is a truth of the matter. That’s a minimalist sense, in which “It’s true that pain is to be avoided” just amounts to saying that pain is to be avoided—and likewise for “It’s a fact that.” Perhaps, as I used to think, there are senses too in which we can sensibly debate whether ought conclusions are true or false. Nothing in this book, though, depends on whether there is any such sense. (2003, p. x)

Similarly, the line drawn between the kinds of mental states that moral judgments expressed has been erased by the minimalist, deflationary conception of belief. One hopeful way to explain why expressivism isn’t a kind of realism is to say that the former takes the state expressed to be something other than belief. But Crispin Wright

74

J. Dreier

notes: “assertion has the following analytical tie to belief: if someone makes an assertion, and is supposed sincere, it follows that she has a belief whose content can be captured by means of the sentence used” (1992, p. 14). Horgan and Timmons adopt this line, insisting that the state expressed must obviously be a belief since it has all logical and grammatical features of a belief. Their “cognitive expressivism” is supposed to have the same metaphysics and same explanatory scheme as the noncognitive versions, differing in that it counts (minimalistically) the state of mind expressed by moral judgment as a belief. Quasi-realism advances, since we do quite ordinarily speak of moral belief, and the prospects for spelling out why expressivism isn’t a kind of realism recede.

1.2 Drawing the Boundary by Type of Explanation All of this quasi-realism and creeping minimalism doesn’t mean that we should give up on the metaphysical division of theories. The semantic criteria are in about as much jeopardy as the metaphysical ones. For just as we can’t pick out antirealists for their denial of the existence of moral facts or properties (since they may be happy with the minimalist versions and doubtful about what the more inflated versions might amount to), we find also that truth, proposition, and reference are also all of them common ground (at least the minimalist versions, and again we’ll have a hard time saying what else there is to want). Instead, I want to say that the difference between the types of theories resides in how they take certain special, “puzzling” or philosophically problematic phenomena to be explained and accounted for. As I’ll try to make clear, this is a metaphysical distinction after all. There are some features of moral language and thought that make its objects seem ripe for philosophical worry. These include the supervenience of the moral on the non-moral, especially combined with some sort of irreducibility5 ; and famously the “internalist” feature or “practicality requirement” linking moral judgment to motivation (or perhaps to reasons for action), along with possibly independent worries about how moral expressions might get their reference.6 Different sorts of metaethical theories have different explanations of these features. I’ll take Gibbard’s 2003 Thinking How to Live as my example of how expressivists try to explain. “Begin with the slogans I’ve been proclaiming,” Gibbard says. “Questions of what we ought to do are questions of what to do. Finish your deliberation, conclude what to do, and you’ve concluded what you ought to do” (2003, p. 19). The practicality requirement, that moral (or more generally, normative) judgment should show some specially intimate connection to a decision to act, is explained by the

5 See

Schroeder (2005) for a recent, clear-headed discussion of what might be at issue here. Horgan and Timmons’ assortment of “Moral Twin Earth troubles” papers; e.g., (1991), (1992a), (1992b).

6 See

Mackie’s Realism: Queer Pigs and the Web of Belief

75

hypothesis that the most fundamental sort of judgment of what we ought to do is a decision – or, as Gibbard says, a plan – for what to do.7 He continues: “these crude sayings will, of course, need qualification, but the distinctive claim of an expressivist is that dicta like these, suitably worked out, account for the subject matter of ethics.” This is Gibbard’s own idea about what distinguishes a view like his from a realist metaethical view. Then he adds: “If a self-avowed realist agrees, the two of us may have no quarrel. I don’t, however, know of any ‘ethical realist’ who accepts slogans like these explicitly.” Nor do I. The realist view, I suggest, is that the explanation goes the other way around: it’s the special subject matter of ethics that explains why and in what way moral judgment is essentially practical. This may seem unlikely. Practicality, one might think, is a feature of judgment that is explained by its role in deliberation, and not by its content.8 So the antirealist mark of expressivism, it appears, is that it explains the practicality requirement by some functionalist account of the nature of moral (or normative) concepts9 and not by the special objects of moral judgment. My example was Gibbard’s theory, but I could just as well have used Blackburn’s (see, e.g., Blackburn 1984, Chapter 6). And, I boldly conjecture, a similar line could be drawn by examination of the explanations that metaethical theories give of supervenience. The main idea, again, is that realists will say something about what properties moral predicates pick out, and what natural properties underlie them, and then say how it is that the moral properties are tied to the natural ones. By contrast, antirealists (expressivists are my paradigms) will say something about how moral concepts work and then show how concepts like these must be applied alike to objects that are alike in the underlying respects. The commonality among realists is that they will think that it’s the properties, or facts, that do the explaining; expressivist explanations will make no use of moral properties or facts, though expressivists may be happy to speak of such things “at the end of the day” (as Blackburn says).10 Let me give a very brief and schematic sketch of why it turns out that atheists are realists about religious talk. I am not the best person to identify puzzling features of religious discourse, but a plodding example may serve well enough. Suppose we are wondering why it is that thoughts and talk about God simultaneously fill people

7 I’m

not defending or advocating Gibbard’s explanation here. This is a paper on metametaethics. Defending a particular metaethical view is no more a part of my aim than defending a particular first order normative moral scheme is Mackie’s aim in (the first part of) Ethics. 8 But compare Peter Railton’s views (1993, p. 282). And Michael Smith (1994) also argues that the right sort of content can explain a moral belief’s role in practical reasoning. 9 This idea is elaborated by John O’Leary-Hawthorne and Huw Price (1996). 10 Compare a currently popular view about conditionals, especially indicative conditionals, according to which they have no truth conditions. The motivation for this view has its source in doubts, supported by powerful technical arguments, that the cognitive and inferential role of the state of mind of accepting a conditional can be explained by finding a special kind of proposition or truth conditions that indicative conditionals state. The upshot is that there couldn’t be any special kind of conditional fact whose nature and structure explain why our conditional judgments behave the way they do. See Adams (1975), Edgington (1991), Bennett (2003, especially Chapter 7, 94–113).

76

J. Dreier

with a sense of comfort and a kind of fear. A realist explanation will appeal to the kind of being that God is: infinite in benevolence but omniscient and perfectly just, a judge from whose court none can escape. The atheist accepts this explanation, in the sense that she agrees that precisely that content (we may suppose) is responsible for the fearsome comfort of thinking and talking about God. What would an antirealist philosophy of religious matters be, then? Maybe a kind of Freudian expressivism, according to which by speaking of which actions are blessed or of what God wants, we are expressing our longing to regain both the authority and the security provided by our parents in our childhood.11 Now we can return to Mackie.

2 Mackie Exegesis When I teach Mackie in my metaethics course, I find that students are strangely attracted to the view. I say this is strange because so few professional philosophers find Mackie’s arguments convincing. Personally, I feel quite confident that Mackie made a mistake in his arguments.12 Most professional philosophers who bother to argue against Mackie are moral realists (see, e.g., Brink 1984); those who are more sympathetic toward his views tend to be antirealists (e.g., Horgan and Timmons). The interesting question always seemed to me to be: Why do metaethical philosophers of an antirealist bent not share Mackie’s conclusions? Speaking loosely, many who do, or would agree with Mackie that there are no objective values, are reluctant to follow him into an “error theory.”13 What is the further premise that Mackie believes and most contemporary antirealists deny? In one sense, the answer is fairly clear. As Mackie puts it, there are two questions one can ask about the objectivity of morality. One is conceptual or semantic, the other is metaphysical. Mackie agrees with hyperrealist philosophers like Moore and Plato (those are the realists he mentions repeatedly) about the conceptual and semantic questions, but he joins antirealists in answering the metaphysical, ontological questions. This is what he says, and it’s what I teach students, and it’s not wrong. But it’s not the most illuminating answer, I now think. Rather, Mackie joins the realists in his conception of central metaphysical questions, and in the way he thinks they have to be answered. For example, he accepts the practicality requirement, and he thinks it is to be explained by some feature of moral properties or facts. That, in a nutshell, is what he thinks is so “queer” about moral goodness: that its nature must explain the internal connection between moral judgment and action.

11 I

am about equally uncomfortable with serious philosophy of religion and serious Freudian psychology. 12 So, yes, here, and when I explain the mistake, I am abandoning my haughty metametaethical position, and stooping to a bit of actual metaethics. 13 Richard Joyce, an editor of this volume, is a notable exception. See Joyce (2001).

Mackie’s Realism: Queer Pigs and the Web of Belief

77

Before I develop this point, which seems to me to be the absolutely crucial one for understanding how Mackie turned out to be an error theorist rather than an expressivist (or some sort of contextualist or relativist), let me run through a more standard reading of why he doesn’t agree with expressivism.14 Mackie first puts his own metaethical thesis like this: There are no objective values. (1977, p. 15)

Alternatively, he notes, we could say that values “are not part of the fabric of the world” (ibid.). But Mackie is aware that that this “moral scepticism” is not enough to distinguish what he’s trying to say from what some of his contemporaries are saying, and he does want to distinguish: It is true that those who have accepted the moral subjectivism which is the doctrine that moral judgements are equivalent to reports of the speaker’s own feelings or attitudes have usually presupposed what I am calling moral scepticism. It is because they have assumed that there are no moral values that they have looked elsewhere for an analysis of what moral statements might mean, and have settled upon subjective reports. (1977, p. 18)

Mackie remarks on what he calls the “multiplicity of second order questions,” and in particular that there are metaphysical (“factual”) questions that are not equivalent to conceptual or linguistic questions. His main claim is the metaphysical one, but he also has a linguistic/conceptual view which differs from the “subjectivists” and expressivists. Answers to one of the second order questions don’t entail answers to the others, he insists, though he thinks others will disagree: Recent philosophy, biased as it has been toward various kinds of linguistic inquiry, has tended to doubt this, but the distinction between conceptual and factual analysis in ethics can be supported by analogies with other areas. (1977, p. 19)15

So according to Mackie, we understand the difference between subjective value and objective value well enough to distinguish them. And moral value concepts are objective. If second order ethics were confined, then, to linguistic and conceptual analysis, it ought to conclude that moral values at least are objective: that they are so is part of what our ordinary moral statements mean: the traditional moral concepts of the ordinary man as well as the main line of western philosophers are concepts of objective value. (Mackie 1977, p. 35)

14 Actually,

the “nonfactualist” interlocutor that most interests Mackie is R. M. Hare; for my purposes, Hare is close enough to an expressivist to make no difference. 15 Mackie has a difficult, uncharacteristically opaque dispute with R. M. Hare over the question of whether there are any factual questions remaining once the linguistic and conceptual ones are resolved. For Hare’s side, see his “Nothing Matters” (1972, p. 47). Hare writes: “Think of one world into whose fabric values are objectively built; and think of another in which those values have been annihilated. And remember that in both worlds the people in them go on being concerned about the same things – there is no difference in the ‘subjective’ concern which people have for things, only in their ‘objective’ value. Now I ask, ‘What is the difference between the states of affairs in these two worlds?’ Can any answer be given except ‘None whatever’?” Mackie gives his side in Ethics, pp. 21–22.

78

J. Dreier

This means that Mackie ends up with commitments that split across the battle lines of his contemporaries. On the conceptual question he is on the side of Moore (and, he thinks, Plato and Kant, though nowadays we’d have a very hard time thinking of Kant as holding that morality is part of the “fabric of the world”), but he is with Stevenson and Hare on the metaphysical question. [T]he denial of objective values will have to be put forward not as the result of an analytic approach, but as an “error theory,” a theory that although most people in making moral judgements implicitly claim, among other things, to be pointing to something objectively prescriptive, these claims are all false. (ibid.)

So far, so good. But why does he agree with Moore about the concept if he’s with the subjectivists on metaphysics? An error theory is extremely uncharitable as an interpretation of ordinary thought and talk. And notoriously Mackie goes on, in Ethics, to vindicate quite a lot of ordinary moral thinking, at least so far as to use sentences that any of us would use to state our own moral views, and to use them sincerely, in his own voice. When he says these things, he naturally thinks of himself as free of the incorrect metaphysics that he takes to be presupposed by the moralizing of hoi polloi (and by that of his more metaphysically extravagant colleagues). If he can argue for qualified deontological principles against both pure consequentialism and “absolutist” deontology without committing the Error, then why can’t we?16 The simple answer is that he just thinks the “claim to objectivity” is built into moral concepts as people actually use them: [O]bjectivism about values . . . has also a firm basis in ordinary thought, and even in the meanings of moral terms. No doubt it was an extravagance of Moore to say that “good” is the name of a non-natural quality, but it would not be so far wrong to say that in moral contexts it is used as if it were the name of a supposed non-natural quality, where the description “non-natural” leaves room for the peculiar evaluative, prescriptive, intrinsically action-guiding aspects of this supposed quality. (Mackie 1977, p. 32)

Now, Mackie does give some reasonably good argument for why ordinary thought includes the “claim to objectivity,” but these arguments do not seem to show that the “claim” figures in ordinary thought as a kind of presupposition or as part of the content. Granting his points, that is, it still seems open for a subjectivist, prescriptivist, or expressivist critic to reply that although ordinary thinking may somehow include whatever is the best way to capture objectivity, this extra thought is just an add-on. It could be false. And, obviously, Mackie thinks that in a way it is false: there are no such properties. So why does this mean that nothing at all is morally right or wrong, rather than meaning that plenty of things (those

16 I

think this criticism of Mackie is overblown. In both Ethics and his later work, Mackie is much more apt than most to speak of what follows from this or that moral premise, of what commitments are incurred by adopting a principle, and of what we might be willing to say or accept, than most philosophers writing about ethics. When occasionally he slides back into the natural idiom of first-order ethics with existential commitment, it seems fair to allow him the façon de parler.

Mackie’s Realism: Queer Pigs and the Web of Belief

79

things we ordinarily think are morally right and wrong) really are, only the fact that they are is partly constituted by our thinking so, or by our sentiments, or that these are only facts in the pleonastic, minimal, deflated sense in that saying that they are facts is merely repeating our own moral commitments? Why couldn’t we be correct in our first-order moral thinking but incorrect in the metaphysical interpretation we give this thinking? Compare ordinary thinking about physical properties and relations like solidity or simultaneity. A sophisticated understanding of the physics that underlies these features of the world shows that they are not all that we think they are. Solid objects do not uniformly fill the space they occupy with undifferentiated matter; rather, they are mostly space, they are (constituted by) crystalline lattices of atoms held together by electromagnetic forces. Simultaneous events, furthermore, take place “at the same time” only relative to a frame of reference; there is no absolute simultaneity. But sophisticated physics does not tell us that there are no solid objects and it does not say that everybody’s judgments about which events are simultaneous are false. The theoretic gloss on our ordinary judgments is not properly understood as infecting every first-order judgment with false presupposition. Why, then, should false metaphysics infect our moralizing? It’s true that plenty of ordinary concepts turn out to be empty – uninstantiated – and that we come to be error theorists about them because of increasing theoretic sophistication, or more mundanely by finding things out about the world. There aren’t any witches, we know now. It’s not that we discovered that witches aren’t magic, or that magic is really just sleight of hand. When we see that what we took to be magic is accomplished by prestidigitation, we conclude that it wasn’t magic after all, and when we discover that eccentric women cannot use incantations to cause people to become sick, we conclude that there are no witches. Sometimes the theoretical or commonsense baggage is so woven into the ordinary concept that it cannot be discharged without abandoning the expressions. It isn’t absurd to suppose that the claim to objectivity is to moral wrongness as magical power is to witchiness, so Mackie’s combination of positions on the conceptual and metaphysical questions certainly makes sense. My question is why we are supposed to think that wrongness is like witchiness, rather than like simultaneity. Mackie never explains.

3 Queerness I now, finally, turn to my own explanation of why Mackie is a realist. In my sense, expressivism is marked off from all kinds of realism, remember, by the type of explanation that it offers for the odd, puzzling features of an area of thinking and language – in the present case, moral thinking and language. Mackie’s explanation is of a piece with Moore’s and other paradigm realists’. My account also shows why Mackie doesn’t adopt the conceptual part of the “subjectivist” view. I don’t want to go so far as to say that my way of looking at things makes everything about Mackie’s

80

J. Dreier

big picture perfectly clear and comfortable; I just say that it illuminates some dark corners. Unsurprisingly, I think the main textual focus has to be on Mackie’s infamous Argument from Queerness. I’ll be concerned with the metaphysical version. Mackie says also that the epistemology of moral values would have to be queer. Insofar as this is really a separate argument, I believe it rests on the conclusion of another argument – namely, Mackie’s argument about the supervenience of the moral. I say “insofar as” it is separate because Mackie himself says that the supervenience of the moral on the non-moral is itself an element of the queerness of moral properties.17 In any case, the point that I want to make is about metaphysical queerness, so I’ll just be focusing on that, without committing myself here to the stronger claim that metaphysical queerness is what’s really behind the error theory. Here is the telegraphic version of the argument. Moral values, if they were to exist, would have to be very queer sorts of things. It is hard to believe that anything so queer, or queer in this way, exists; nothing really is that way, as far as we can make out when we reflect on things soberly and clearly. So moral values don’t exist. One obvious shortcoming of this version is that it fails to say what would be so queer about moral properties. Mackie does say a number of things about why they would be queer. They would have to have a certain queer feature, which he variously calls “objective prescriptivity,” “built in to-be-pursuedness,” “absolute action-guidingness.” In one nice passage, he admits that there is a sense of the word “justice” in which it is quite objective whether something is just or not; we know of paradigm cases of injustice, like when a judge fails to apply the law, or when a competitor is given a lower score even though he does better measure up to the actual standards. But the statement that a certain decision is thus just or unjust will not be objectively prescriptive: in so far as it can be simply true it leaves open the question whether there is any objective requirement to do what is just and to refrain from what is unjust, and equally leaves open the practical decision to act in either way. (Mackie 1977, pp. 26–27)

Some facts can be prescriptive, and some can be objective, but none could be objective and prescriptive at once. But moral facts would have to be both. That’s queer. Still, this does leave us with the question of just what exactly “objectively prescriptive” means. It does sound queer: the facts themselves telling us what to do. But maybe that’s just a peculiar description of something ordinary and explicable. We need a more explicit version.

17 “Another

way of bringing out this queerness is to ask, about anything that is supposed to have some objective moral quality, how this is linked with its natural features. . . . The wrongness must be ‘consequential’ or ‘supervenient’; it is wrong because it is a piece of deliberate cruelty. But just what in the world is signified by this ‘because’? And how do we know the relation that it signifies?” (Mackie 1977, p. 41).

Mackie’s Realism: Queer Pigs and the Web of Belief

81

[T]he main tradition of European philosophy from Plato onwards has combined the view that moral values are objective with the recognition that moral judgements are partly prescriptive or directive or action-guiding. . . . In Plato’s theory the Forms, and in particular the Form of the Good, are external, extra-mental, realities. . . . But it is held also that just knowing them or “seeing” them will not merely tell men what to do but will ensure that they do it, overruling any contrary inclinations. (Mackie 1977, p. 23)18

This passage provides one of the two most explicit developments of the nature of the queerness of moral properties. Here is the other: Plato’s Forms give a dramatic picture of what objective values would have to be. The Form of the Good is such that knowledge of it provides the knower with both a direction and an overriding motive; something’s being good both tells the person who knows this to pursue it and makes him pursue it. An objective good would be sought by anyone who was acquainted with it, not because of any contingent fact that this person, or every person, is so constituted that he desires this end, but just because the end has to-be-pursuedness built into it. (Mackie 1977, p. 40)

The queerness of moral properties, I think it’s clear, is a direct product of the “internalist” feature of moral judgment, the “practicality requirement.” Glossing his own argument, he writes: The considerations that favour moral skepticism are: first, the relativity or variability of some important starting points of moral thinking and their apparent dependence on actual ways of life; secondly, the metaphysical peculiarity of the supposed objective values, in that they would have to be intrinsically action-guiding and motivating. (Mackie 1977, p. 49)

Mackie endorses a very strong version of the requirement – much stronger, I think, than any metaethicist, even one strongly sympathetic to the thesis, would endorse today. I doubt that his argument needs such a strong version. There are all sorts of ways of weakening it without seriously damaging the argument, but I’ll stick with the strong version. I do need to clarify one thing, though, so as to disambiguate (or maybe correct) Mackie’s statement of the requirement. Mackie speaks of knowledge of and acquaintance with moral good; it is these that he says provide the agent automatically and by their own force, these that will not merely tell us what to do but ensure that we do it. One version of internalism does indeed hold that it is knowledge or acquaintance with moral facts that must motivate. But I don’t think that’s the one Mackie had in mind. Notice that he attributes to “the European tradition” “the recognition that moral judgements are partly prescriptive or directive or action-guiding” (my emphasis). It’s the act of judging that is bound to motivation, and not some further contact with the facts judged. And it’s belief internalism, or judgment internalism, that figures in the argument. Here’s how I interpret the argument when the queerness of moral judgment is made fully explicit.

18 Though Plato is his favorite, because Mackie thinks Plato faced up to queerness more forthrightly

than other philosophers, he adds Kant and Sidgwick in this passage as having held essentially the same view.

82

J. Dreier

Q1

Moral goodness would have to be a property, G, such that judging that something is G entails having an overriding motivation to pursue it. Q2 But for no property P does judging that something is P entail having a motivation to pursue it. (What a queer sort of property that would be!) Therefore: Q3

There is no such property as moral goodness.

This is the argument that makes Mackie a realist. He subscribes to the realist style of the explanation of the internalist feature of moral judgment. For morality to be vindicated, according to Mackie, the moral properties themselves would have to account for the internal, necessary connection between moral judgment and motivation. I did say that my account would not exactly make absolutely everything clear, and I have to admit that I find it odd that Mackie should have thought that it was the moral properties that would have to explain the internal connection. One reason I find it odd is that it just seems like a mistake to me. It seems to me that Q1 and Q2 are natural grounds for expressivism, although an expressivist would put the same premises somewhat differently. E1

Judging that something is morally good entails having an overriding motivation to pursue the thing. E2 But there is no belief the having of which entails having any motivation. (What a queer belief that would be!) Therefore: E3

Judging that something is morally good is not a belief.

This version strikes me as much better. It is, surely, the belief that is supposed to be queer, according to internalism, and not the property – the property gets its own queerness in a rather irrelevant and relational way, if we absolutely insist on the Q version of the argument. Mackie seems to have mislocated the queerness. When I lay things out this way, his mistake seems very strange. But Mackie is not alone, actually. There is a group of literary characters who are guilty of an analogous mistake.

4 Queer Pigs In E. B. White’s novel Charlotte’s Web (1952), Wilbur the Pig, saved from early slaughter as runt of the litter by the pleading of the tender hearted Fern Arable, is getting dangerously fat. The danger is rather different for him than it would be for us: Fern’s uncle Homer Zuckerman has visions of pork chops and bacon. Fortunately

Mackie’s Realism: Queer Pigs and the Web of Belief

83

for Wilbur, he has friends in high places, especially Charlotte A. Cavatica, the grey barn spider who has spun her web in the doorway of the barn, directly over Wilbur’s pen. Charlotte conceives a plan to save Wilbur’s life: “The way to save Wilbur’s life is to play a trick on Zuckerman. If I can fool a bug,” thought Charlotte, “I can surely fool a man. People are not as smart as bugs.” (White 1952, p. 67)

That night, she sets about executing her trick, and by morning the trap is set: The web glistened in the light and made a pattern of loveliness and mystery, like a delicate veil. Even Lurvy, who wasn’t particularly interested in beauty, noticed the web when he came with the pig’s breakfast. He noted how clearly it showed up, and he noted how carefully built it was. And then he took another look and he saw something that made him set his pail down. There, in the center of the web, neatly woven in block letters, was a message. It said: SOME PIG! Lurvy felt weak. He brushed his hand across his eyes and stared harder at Charlotte’s web. (White 1952, p. 78)

Charlotte’s trick is having the desired affect. Lurvy tells Mr. Zuckerman, owner of farm and pig: “I think you’d better come down to the pig pen,” he said. “What’s the trouble?” asked Mr. Zuckerman. “Anything wrong with the pig?” “N-not exactly,” said Lurvy. “Come and see for yourself.” The two men walked silently down to Wilbur’s yard. Lurvy pointed to the spider web. “Do you see what I see?” he asked? Zuckerman stared at the writing on the web. Then he murmured the words “Some Pig.” Then he looked at Lurvy. Then they both began to tremble. Charlotte, sleepy after her night’s exertions, smiled as she watched. (ibid.)

Lurvy and Mr. Zuckerman confirm Charlotte’s low opinion of men. Struck by the miracle of English writing in a spider’s web, they are in the process of mislocating the miraculous object: they think it’s Wilbur. They think it must be the subject matter of the incredible, inexplicable writing, rather than something about the writing (or author). E. B. White may have a better opinion of women than Charlotte has of men. Mrs. Zuckerman is not so easy to hoodwink: “Edith,” [Mr. Zuckerman] said, trying to keep his voice steady, “I think you had best be told that we have a very unusual pig.” A look of complete bewilderment came over Mrs. Zuckerman’s face. “Homer Zuckerman, what in the world are you talking about?” “This is a very serious thing, Edith,” he replied. “Our pig is completely out of the ordinary.” “What’s unusual about the pig?” asked Mrs. Zuckerman, who was beginning to recover from her scare. “Well, I don’t really know yet,” said Mr. Zuckerman. “But we have received a sign, Edith– a mysterious sign. . . . [Mr. Zuckerman described the scene in the barn.] A miracle has happened and a sign has occurred here on earth, right on our farm, and we have no ordinary pig.” “Well,” said Mrs. Zuckerman, “it seems to me you’re a little off. It seems to me we have no ordinary spider.” “Oh no,” said Zuckerman. “It’s the pig that’s unusual. It says so, right there in the middle of the web.”

84

J. Dreier “Maybe so,” said Mrs. Zuckerman. “Just the same, I intend to have a look at that spider.” (White 1952, pp. 80–81)

Zuckerman tells his minister that a miracle has occurred on the farm. The minister’s assessment of the situation is no better than the farmer’s: “Don’t tell anybody else,” said the minister. “We don’t know what it means yet, but perhaps if I give some thought to it, I can explain it in my sermon next Sunday. There can be no doubt that you have a most unusual pig. I intend to speak about it in my sermon and point out the fact that this community has been visited with a wondrous animal.” (White 1952, p. 82)

In the weeks to come, Charlotte repeats her scheme, using words she finds in magazine advertisements. She describes Wilbur in turn as TERRIFIC, RADIANT, and HUMBLE. The victory is unqualified: Wilbur lives though holiday dinners, wins a special prize at the state fair and in the end looks forward to a long, if swinish, existence. “Ladies and Gentlemen,” said the loudspeaker [at the fair], “we now present Mr. Homer L. Zuckerman’s distinguished pig. The fame of this unique animal has spread to the far corners of the earth, attracting many valuable tourists to our great State. Many of you will recall that never-to-be-forgotten day last summer when the writing appeared mysteriously on the spider’s web in Mr. Zuckerman’s barn, calling the attention of all and sundry to the fact that this pig was completely out of the ordinary. This miracle has never been fully explained, although learned men have visited the Zuckerman pigpen to study and observer the phenomenon. In the last analysis, we simply know that we are dealing with supernatural forces here, and we should all feel proud and grateful. In the words of the spider’s web, ladies and gentlemen, this is some pig.” (White 1952, p. 157)

5 Conclusion: The Web of Belief Zuckerman and cohorts are guilty of an amusing mistake, one that requires some good-natured suspension of disbelief for readers to swallow. But John Mackie, quite a lot smarter than Zuckerman, Lurvy, or any bug (pace Charlotte), was guilty of a parallel mistake. In the presence of words in a spider web, Lurvy turned his gaze in the wrong direction, toward the subject matter of the words instead of to the nature of the web and words themselves. “Some pig,” he thought, instead of “Some spider.” “What a queer pig that must be, to be the subject of spider remarks! No pig that I’ve ever seen has been of such a terrific, radiant, extraordinary nature as to have a spider web advertisement be about him. It seems hard to believe in such a pig! And yet, there it is.” I hope this makes it clear how Mackie’s mistake is parallel. Moral beliefs are like a spider’s web with words. They are very queer indeed. It is not their subject matter that we should wonder about, be skeptical about, consider queer; it’s the beliefs. Those are some beliefs you have there, you moralizers. Our community has been visited, to paraphrase the minister, not by a wondrous property or pig, but by a wondrous and suspicious belief. It seems odd that Mackie would make the mistake I attribute to him. Mackie was no Lurvy. Still, I am pretty confident that he did make that mistake. Mackie was

Mackie’s Realism: Queer Pigs and the Web of Belief

85

right, at least given his premise of internalism, to find queerness in the domain of moral talk and thought, but he located it in entirely the wrong spot. What would Edith Zuckerman say? He should have kept his eyes on the miraculous nature of moral belief and doubted that there could be any such thing, instead of attributing it to the subject matter of the beliefs. At least so it seems to me; and here I have shown my metaethical hand. For it’s realists who expect to explain the queer (or anyway, philosophically puzzling) aspects of moral judgment by way of the moral facts and properties themselves, antirealists who try to explain them by other, functional or non-representational aspects of our words. (Spiders, maybe?) I’ve been trying to convince the reader that Mackie was a realist, but I don’t have to put my conclusion in such a provocative way. Let me recap. In the first section, I argued that traditional ways of drawing the boundary between realist and antirealist metaethical theories are no longer adequate, because deflationary or minimalist accounts of some of the concepts employed in carving the distinction render them useless for the purpose. I suggested that what makes expressivism antirealist and Moorean non-naturalism realist is that expressivists explain the special, puzzling features of normative thought and talk by means of features of our concepts and our language, while Mooreans look for the explanation in the subject matter, that is, in some special nature of moral properties and ought-facts. Since Mackie agrees with Mooreans on this question, he counts as a realist. The key to categorizing Mackie is found in his ideas about the queerness of moral properties and facts: they would have to be such as to explain the odd, intrinsically motivational features of beliefs about them. Notice that it isn’t the error theory per se that makes Mackie a realist, according to my scheme, but his reasons for subscribing to it. Maybe other error theorists have different reasons, and if they do maybe those reasons are antirealist reasons.19 Of course, I shouldn’t insist on my terminology. “Realism” and “antirealism” are bits of philosophical jargon, not terms of ordinary language, and philosophers are entitled to use them however they like. If a metaphysician declines my invitation to draw the boundary by reference to the location of explanations of the special features of thought and talk, I won’t really complain. Instead I can put my point about Mackie in a terminologically neutral but still polemical way. Mackie shares with Mooreans and Lurvy the view that it’s the pig that’s queer; expressivists are on the side of Edith Zuckerman in thinking that it’s the spider. Acknowledgments I gave a version of this paper at the APA Central Division meetings in Chicago, April 2005. Thanks very much to Terence Cuneo for the critical comments he gave at the symposium, and also to Richard Joyce and Simon Kirchin for their advice, both philosophical and editorial.

19 Thanks

to Richard Joyce for getting me to see this point.

86

J. Dreier

References Adams, E. 1975. The logic of conditionals. Dordrecht: Reidel. Audi, R. 1997. Moral knowledge and ethical character. Oxford: Oxford University Press. Ayer, A. J. 1946. Language, truth and logic. 2nd ed. New York: Dover. Bennett, J. 2003. A philosophical guide to conditionals. Oxford: Oxford University Press. Blackburn, S. 1984. Spreading the word. New York: Oxford University Press. Blackburn, S. 1993. Essays in quasi-realism. New York: Oxford University Press. Brink, D. 1984. Moral realism and the skeptical arguments from disagreement and queerness. Australasian Journal of Philosophy 62: 111–125. Dreier, J. 2005. Metaethics and the problem of creeping minimalism. Philosophical Perspectives 18: 23–44. Dreier, J. 2006. Was Moore a Moorean? In Metaethics after Moore, eds. T. Horgan and M. Timmons, 191–207. Oxford: Oxford University Press. Edgington, D. 1991. Do conditionals have truth conditions? In Conditionals, ed. F. Jackson, 176– 201. Oxford: Oxford University Press. Gibbard, A. 1990. Wise choices, apt feelings. Cambridge, MA: Harvard University Press. Gibbard, A. 2003. Thinking how to live. Cambridge, MA: Harvard University Press. Hare, R. M. 1972. “Nothing matters”: Is “the annihilation of values” something that could happen? In his Applications of moral philosophy, 32–47. Berkeley, CA: University of California Press. Horgan, T. and M. Timmons. 1991. New wave moral realism meets moral twin earth. Journal of Philosophical Research 16: 447–465. Horgan, T. and M. Timmons. 1992a. Troubles for new wave moral semantics: The open question argument revived. Philosophical Papers 21: 153–175. Horgan, T. and M. Timmons. 1992b. Troubles on moral twin earth: Moral queerness revived. Synthese 92: 221–260. Horgan, T. and M. Timmons. 2006. Cognitivist expressivism. In Metaethics after Moore, eds. T. Horgan and M. Timmons, 255–298. Oxford: Oxford University Press. Joyce, R. 2001. The myth of morality. Cambridge: Cambridge University Press. Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin. O’Leary-Hawthorne, J. and H. Price. 1996. How to stand up for noncognitivists. Australasian Journal of Philosophy 74: 275–292. Railton, P. 1993. What the noncognitivist helps us to see, the naturalist must help us to explain. In Reality, representation, and projection, eds. J. Haldane and C. Wright, 279–300. New York: Oxford University Press. Schroeder, M. 2005. Realism and reduction: The quest for robustness. Philosopher’s Imprint 5: 1–18. Shafer-Landau, R. 2003. Moral realism: A defence. Oxford: Oxford University Press. Smith, M. 1994. The moral problem. London: Blackwell. White, E. B. 1952. Charlotte’s web. New York: Harper Collins. Wright, C. 1992. Truth and objectivity. Cambridge, MA: Harvard University Press.

Mackie on Practical Reason David Phillips

My topic in this chapter is John Mackie’s treatment of practical reason. I will argue that his approach is both attractive and unjustly neglected. In particular, I will argue that it is importantly different from, and much more plausible than, the kind of instrumentalist approach famously articulated by Bernard Williams.1 This matters for the interpretation of the arguments for Mackie’s most famous thesis: moral skepticism, the claim that there are no objective values. Richard Joyce has recently defended a version or variant of Mackie’s view by invoking an instrumentalist theory like Williams’. While Joyce does a nice job in a number of ways of framing Mackie’s argument, I shall argue that it is a serious mistake to invoke instrumentalism in its defense. I will begin with interpretive issues, arguing that Mackie develops a view of practical reason significantly different from Williams’. I will then turn to philosophical assessment, arguing that the view I attribute to Mackie is significantly more plausible than Williams’ view. Finally, I will ask what this shows about how to understand Mackie’s arguments for moral skepticism.

1 Mackie and Williams on Practical Reason Begin with Williams, who develops the most famous recent version of the view that all reasons are desire-based. He distinguishes internal from external reasons, and argues that the only true or genuine reasons are internal. The simplest account of internal reasons, as he notes, is one according to which “A has a reason to φ if A has some desire the satisfaction of which will be served by his φing” (Williams 1981, D. Phillips (B) Department of Philosophy, University of Houston, Houston, TX, USA 1 Williams’ views are developed in four places. “Internal and External Reasons” was first published

in 1979 and appeared in his Moral Luck (1981). He returned to the topic in “Internal reasons and the obscurity of blame” (which was first published in 1989 and appears in his collected papers (1995a)) and then in “Replies” (1995b). His final reflections on the topic are to be found in “Postscript: Some further notes on internal and external reasons” (2001).

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_6, 

87

88

D. Phillips

p. 101). Williams’ account of internal reasons goes beyond this simple “subHumean” model in two major respects: the agent’s “subjective motivational set,” S, can contain more than just desires, and deliberation is not restricted to straightforward means-ends calculation. He then considers whether there are any genuine or correct external reasons statements, and concludes that there are not: The sort of considerations offered here strongly suggest to me that external reason statements, when definitely isolated as such, are false, or incoherent, or really something else misleadingly expressed . . . The formulation in terms of reasons does have an effect, particularly in its suggestion that the agent is being irrational, and this suggestion, once the basis of an internal reason claim has been clearly laid aside, is bluff. (Williams 1981, p. 111)

In “Internal and External Reasons,” Williams argues for his view by focusing on a case from Henry James, that of Owen Wingrave. As Williams describes the case, Owen’s father claims that Owen has a reason to join the army, but there is nothing in Owen’s S such that if he deliberated he would be motivated to join the army. The argument begins with the following (internalism) constraint: In considering what an external reason statement might mean, we have to remember the dimension of possible explanation. If something can be a reason for action, then it could be someone’s reason for acting on a particular occasion, and it would then figure in an explanation of that action. (Williams 1981, p. 106)

The argument then goes as follows: No external reason statement could by itself offer an explanation of someone’s action (because, ex hypothesi, it can be true independently of the agent’s motivations). So external reasons statements must explain action in conjunction with some “psychological link” (Williams 1981, p. 107). The obvious candidate to be that psychological link is belief. One possibility is that the distinctive explanatory role of external reasons statements is played by explanations invoking a person’s belief that a particular consideration constitutes a reason. But this possibility must be rejected, because a person who had such a belief “appears to be one about whom . . . an internal reason statement could truly be made” (Williams 1981, p. 107). So the distinctive explanatory role of an external reason statement “will have to be revealed by considering what it is to come to believe such a statement” (Williams 1981, p. 108, italics in original). But Williams argues that this possibility too must be rejected: Given the agent’s earlier existing motivations, and this new motivation, what has to hold for external reason statements to be true, on this line of interpretation, is that the new motivation could be in some way rationally arrived at, granted the earlier motivations. Yet at the same time it must not bear to the earlier motivations the kind of rational relation which we considered in the earlier discussion of deliberation—for in that case an internal reason statement would have been true in the first place. I see no reason to suppose that these conditions could possibly be met. (Williams 1981, p. 109)

For current purposes, the key feature of Williams’ position is this: he gives a very special status to internal reasons claims. They can, he claims, be true or correct in a way that no other reasons claims can be.

Mackie on Practical Reason

89

Mackie’s treatment, I shall suggest, contrasts with Williams’ in this key respect. For Mackie, internal reasons claims have no such special status and in particular are not specially authoritative. The story about their authority is ultimately the same as the story about the authority of other normative claims: they are authoritative only when a certain background is presupposed; they lack the special authority that, Mackie the moral skeptic thinks, nothing in the world has. To show that this is Mackie’s view, begin by noting the common features of an approach to moral terms and concepts found in a number of places in the first three chapters of Ethics. On this approach, moral claims (or, more generally, value claims) can be assessed as true or false only when certain normative items (standards, requirements, desires, ends, etc.) are presupposed. For this reason, employing a term Mackie does not employ, I will call it “the presupposition approach.” It makes an early appearance in Section 5 of Chapter 1: One way of stating that there are no objective values is to say that value statements cannot be either true or false. But this formulation, too, lends itself to misinterpretation. For there are certain kinds of value statements which undoubtedly can be true or false, even if, in the sense I intend, there are no objective values. Evaluations of many sorts are commonly made in relation to agreed and assumed standards. (Mackie 1977, pp. 25–26)

Consider next what Mackie says in introducing the (familiar) notion of a hypothetical imperative: “If you want to do X, do Y” (or “You ought to do Y”) will be a hypothetical imperative if it is based on the supposed fact that Y is, in the circumstances, the only (or the best) available means to X, that is, on a causal relation between Y and X. The reason for doing Y lies in its causal connection with the desired end, X; the oughtness is contingent upon the desire. (Mackie 1977, pp. 27–28)

Then notice Mackie’s proposed “general definition of ‘good’”: Such as to satisfy requirements (etc.) of the kind in question. (Mackie 1977, pp. 55–56, italics in original)

Mackie goes on to suggest that context tells us what kind of requirements are in question. Finally, notice Mackie’s account of the meaning of “ought” and of the different kinds of “ought”-claims. His “first attempt at a general equivalent of ‘a ought to G’” is: There is a reason for a’s Ging. (Mackie 1977, pp. 73–74)

Then he allows for several different sorts of reasons with which to fill in this formula: hypothetically imperative reasons, institutional reasons, epistemic reasons, prudential reasons, reasons provided by the wants and needs of others. He also mentions the kind of “intrinsic requirements” in which, of course, he disbelieves. I suggest that in these different passages there is a common general idea: that value claims can be true or correct only when some normative item is presupposed. Care is needed in seeing just what follows from Mackie’s development of this presupposition approach. One of Mackie’s themes is that too much of the moral philosophy contemporary with the writing of Ethics wrongly focuses on linguistic and semantic issues, excluding important and independent metaphysical issues. Thus

90

D. Phillips

we should not expect Mackie directly to infer anything metaphysical from these linguistic or conceptual claims. He thinks that the truth of these linguistic or conceptual claims is compatible with the falsity of moral skepticism. Indeed, the structure of his case for his error theory requires that he distinguish, in Michael Smith’s terminology (Smith 1994, pp. 63–66), rationalism as a conceptual claim (which Mackie accepts) from rationalism as a substantive claim (which he rejects). Nonetheless, there is an important lesson suggested by Mackie’s embrace of the presupposition approach: that there is nothing special about hypothetical imperatives or internal reasons. Hypothetical “ought”-claims or internal reasons claims are true contingently on certain desires, in just the same way as institutional “ought”-claims (e.g., involving promising) are true speaking from within the institution of promising, and in just the same way as claims about the goodness of apples are true given presupposed standards or requirements that apples must meet. In none of these cases are normative claims true or correct in a presupposition-independent way. Mackie will go on to argue, in effect, that no normative claims are true or correct in a straightforward, or presupposition-independent way. Again, though, this further claim is the product of independent arguments against the existence of categorical imperatives (specifically, the arguments from relativity and from queerness); Mackie certainly does not take it to follow just from the presupposition approach itself. Read this way, Mackie’s view is importantly different from Williams’. Mackie does not pick out hypothetical imperative or internal reasons claims as having a special kind of authority or validity, as Williams does. They are in central respects just like other familiar evaluative claims; and in no such cases (he will go on to argue) is there this kind of authority.

2 Why Mackie’s Treatment is Better The presupposition approach is philosophically superior to Williams’ approach. The problem with Williams’ approach can be developed as a dilemma. To articulate the dilemma, begin by distinguishing between debunking and non-debunking accounts of normative practical reasons claims. Non-debunking accounts seek to capture the ordinary meaning of such claims. Debunking accounts take it instead that the ordinary meaning of such claims involves some metaphysical error, and so seek to offer instead the best non-erroneous reconstruction of such claims. In outline, the dilemma is this: Either Williams offers a non-debunking account of normative practical reasons claims, or he does not. If, as most interpreters suppose,2 2 Scanlon

is quite explicit in interpreting Williams as developing a non-debunking view. In What We Owe to Each Other, he writes: “I will assume that [Williams’] claim that there are only internal reasons does not reflect skepticism about reasons in the standard normative sense” (p. 365). Parfit is equally clear that he sees Williams as developing a reductive, but not a debunking or eliminativist, account of normative practical reasons (Parfit 1997). Korsgaard, however, may take Williams to be developing a debunking view; her celebrated critique of his work comes, after all, in an article entitled “Skepticism about Practical Reason.”

Mackie on Practical Reason

91

he offers a non-debunking account of practical reasons claims, then Williams faces three problems: first, that each person’s own (contingent, idiosyncratic) subjective motivational set is made a source of authority, a role it is ill-suited to play; second, that his view conflicts with firm intuitions about the reasons people have; and third, that his view conflicts with the phenomenology of deliberation. If, contrary to most interpreters, Williams is read as offering a debunking account of practical reasons claims, then the problem is that there are better debunking accounts available, including Mackie’s account, which allows us to make sense, in a way Williams’ account does not, of external reasons judgments. To develop the argument more fully, begin with the first horn of the dilemma. Suppose that we read Williams as offering a non-debunking account of practical reasons claims. The first problem then is that Williams makes the agent’s subjective motivational set the source of authority; and, on reflection, that is a role that subjective motivational sets are, in themselves, ill-equipped to play. The idea of a normative practical reason is the idea of an authoritative consideration applying to an agent. Williams’ view has no problem with application to an agent. But it does have a problem with authority. What someone’s desires are is, to a significant extent, an arbitrary, contingent fact about her. Why are her desires therefore a source, or the only source, of authority? The second problem is that Williams’ view conflicts with firm intuitions about the reasons we have. Any number of possible examples can be used to illustrate this point. Consider two. First, suppose that Alice wants to smoke cigarettes. She does not enjoy smoking cigarettes. She knows about the likely deleterious consequences for her health. But her desire to smoke cigarettes would, for all that, survive the kind of procedurally rational reflection for which Williams’ view allows. He is then compelled to say that Alice has a reason to smoke cigarettes. Her desire is itself a source of authority. But surely the mere fact that Alice has this desire, and that procedurally rational reflection would not cause her to lose the desire, does not show that she has a good reason to smoke cigarettes. Second, consider a famous example from Derek Parfit: the example of the person who is future-Tuesday-indifferent (Parfit 1984, pp. 123–124). She cares equally about pains and pleasure in the future, except that she doesn’t care at all about pain or pleasure on future Tuesdays. So she would prefer agony on a future Tuesday to mild discomfort on any other day of the week. She understands that the days of the week are just conventional calendar divisions, and has no other bizarre metaphysical beliefs in the light of which her attitude to Tuesdays would be more comprehensible. Her indifference to pain on future Tuesdays would survive procedurally rational reflection. Again, Williams’ view compels us to say that this person has no reason to care about pains on future Tuesdays. Again that conflicts with our firm intuitions about the reasons this person has. Someone might respond that almost all normal people will turn out to have elements in their subjective motivational sets such that, on Williams’ view, they do have reason to care about pain on future Tuesdays. But Parfit’s point is that we think that even people who (by hypothesis) lack any such element nonetheless do have reason to care about pain on future Tuesdays.

92

D. Phillips

The third problem is that Williams’ view conflicts with the phenomenology of deliberation. When we deliberate, we typically do not take the value of the options about which we deliberate to depend on our desires. Instead, we take the desires to be responses to the value. Let me quote two recent characterizations of this problem. The first is from Stephen Darwall, who invokes (and criticizes) terminology introduced by Philip Pettit and Michael Smith (1990): What a deliberating agent has in view is rarely his own will or desires as such, even if taking something as a reason is intimately tied to desire. Someone who wants to escape a burning building doesn’t evaluate her options by considering which is likeliest to realize what she wants or wills. She is focused, rather, on her desire’s object: getting out alive. The fact that a successful route would realize something she wants is apt to strike her as beside the point or, at best, as a trivial bonus. . . . This point is sometimes put by saying that desires are in the “background” rather than the foreground, of the practical scene a deliberating agent faces. The metaphor is somewhat misleading, however, since an agent’s desires are normally not so much in the background of her deliberative field as outside of it altogether. If we must locate them spatially, a better place might be within or behind the standpoint from which the agent views her alternatives rather than towards the back of the scene she views. (Darwall 2001, pp. 129–130)

The problem is also articulated by Scanlon: If I take myself to have reason to do something because it is worthwhile—to work to alleviate some people’s suffering, for example, or to prevent the destruction of some great building—this reason does not seem to depend on my seeing it as a reason. Rather, I think that I would be mistaken not to see that it is worthwhile or excellent, and mistaken not to care about such things . . . . Consequently, the claim that the reasons they give me derive solely from my desires, or from what I care about, seems deflationary. (Scanlon 1998, p. 42)

It is important to be clear about the exact character of this problem. The problem is not that Williams’ view requires deliberation to have the wrong focus. An agent who accepts Williams’ view need not focus on her subjective motivational set when deliberating. She can properly focus on the options between which she has to choose and the values of those options. On her view these values are ultimately the product of her subjective motivational set; but that does not mean that, while deliberating, she must focus on that fact or on her subjective motivational set. The problem is, rather, that (as the quotation from Scanlon suggests) a serious commitment to deliberation cannot easily coexist, in many normal agents, with acceptance of Williams’ view. Deliberation need not focus on the source of value. But some accounts of the source of value, including Williams’, seem to undermine much normal deliberation, in the sense that normal agents, in deliberating, assume their falsity, and would think such deliberation pointless if they came to accept these accounts. Given these very serious problems, there would have to be very powerful arguments on the other side to persuade us to accept Williams’ view. But his arguments have no such power. The argument explicitly offered in “Internal and External Reasons” trades on the internalism constraint identified earlier: If something can be a reason for action, then it could be someone’s reason for acting on a particular occasion, and it would then figure in an explanation of that action. (Williams 1981, p. 106)

Mackie on Practical Reason

93

Williams takes this to show that the only genuine reasons are internal in his favored sense, that they derive from the agent’s subjective motivational set. But in doing so he fails to see that the internalism constraint might just as well be formulated instead as Christine Korsgaard formulates it: Practical-reason claims, if they are really to present us with reasons for action, must be capable of motivating rational persons. (Korsgaard 1986, p. 11)

On this reformulation, it does not follow from the internalism constraint that someone’s failing to be motivated by a consideration shows that it is not a reason for her. For there is always an alternative diagnosis available: the person is irrational. Hence no argument from the existence of an internalism constraint supports the idea that all genuine reasons are internal in Williams’ favored sense that they depend on elements of the agent’s subjective motivational set. Relatedly, Williams’ writing sometimes suggests that he supposes that the only alternative to his broad assimilation of normative to motivating reasons is the implausible supposition that there is no connection between the term “reason” as used of motivating reasons, and the term “reason” as used of normative reasons. In “Internal reasons and the obscurity of blame,” Williams makes more explicit than in “Internal and external reasons” this general reason for accepting the internalism constraint: the internalism constraint should be accepted because there must be a connection between normative and explanatory reasons, and the internalism constraint characterizes such a connection. As he puts it: The first point is the interrelation of explanatory and normative reasons. It must be a mistake simply to separate explanatory and normative reasons. (Williams 1995a, pp. 38–39)

But this again is a mistake. It is perfectly possible to explain the connections between the two senses of “reason” without supposing that normative reasons are constructible out of the agent’s current or appropriately corrected motivations. Roughly, the connection is this: to say that I have a motivating reason to φ is to say that I have some belief and desire that will together motivate me to φ; to say that I have a normative reason to φ is to say that there is a belief-desire pair which would, if I had them, motivate me to φ, and where both belief and desire are correct. Such an account makes the connection between normative and motivating reasons appropriately non-accidental, but does not entail that normative reasons are constructible out of current motivations. At this point, despite the general view among Williams’ interpreters that he intends to give a non-debunking account of normative reasons, we might be tempted to try the other option. What if Williams does not really suppose that there are any normative reasons? What if he is really trying just to make the best sense he can of normative reasons claims on the assumption that none are ever straightforwardly true or correct? Here, though, we encounter a different problem. The problem is that Williams’ proposed account is not the best debunking account of normative practical reasons claims. It is not the best debunking account for a reason Williams’ argument emphasizes: it forces us to understand external reasons statements as false, or incoherent,

94

D. Phillips

or something of the sort. But, given that we really make external reasons statements, it is an undesirable feature of a debunking account that it make no sense of them. If we are anyway in the business of interpreting normative reasons claims on the assumption that there are no valid norms of practical reason, then there are better ways to treat external reasons statements. To see this more concretely, consider again the Owen Wingrave case. Instead of interpreting Owen’s father’s claim that there is a reason for Owen to join the army as problematically or incoherently invoking Owen’s subjective motivational state, we can better interpret it as invoking the values of family honor and military duty to which Owen’s father is indeed committed. Doing this gives a unitary semantics to reasons statements. The difference between internal and external reasons claims is then a matter of what gets invoked: standards or aims constructed out of the agent’s own subjective motivational states in the case of internal reasons statements, other contextually indicated standards and aims in the case of external reasons statements. Williams’ account of a subjective motivational set can then, indeed, supply a model for the kind of set of standards, or, as we may say, “normative set,” that can be used to interpret external reasons statements. Much of what Williams says about subjective motivational sets can usefully be transmuted into parallel and more general claims about normative sets. He says that S will typically contain much more than desires; correspondingly, normative sets may contain much more than norms having the very straightforward structure of desires. He says that it may be somewhat vague just what S does contain and that S will be revisable in the light of deliberation. Much the same can be said of normative sets like the standards of family honor and military duty invoked by Wingrave senior’s external reasons claim. And there will be a parallel between the kind of verdicts that subjective motivational sets and normative sets give about reason claims. According to Williams’ account of an internal reason claim, an agent has an internal reason to φ if she would be motivated to φ given full information. According to the parallel account of external reasons claims, an agent has an external reason to φ if the normative set invoked, after full deliberation, recommends φing. But the crucial point is that it is not sensible to employ the model only when in interpreting internal reasons statements. In an obvious modified form, the model can be applied to external reasons statements too. So supplemented, Williams’ view starts to sound much more like Mackie’s, and becomes much more attractive. But to interpret and supplement Williams in this way is to lose what is distinctive in his position. We can now explicitly make the comparison between Williams’ view and the presupposition view which (I argued) is to be found in Mackie. The comparison is very much to Mackie’s advantage. On the standard, non-debunking, reading, Williams admits that there are true normative reasons claims, but takes the only source of authority to be the agent’s subjective motivational set. His position then faces the three serious problems we identified: that it gives an inadequate account of the authority of normative reasons, that it conflicts with intuitions about the reasons we have, and that it conflicts with the phenomenology of deliberation. On the debunking reading, Williams’ treatment of external reasons statements, whereby no

Mackie on Practical Reason

95

straightforward sense can be made of them, is again unnecessarily costly. By contrast, Mackie’s position is a consistent debunking view. It treats all practical reasons claims in a unitary way. They can be true or correct only relative to certain assumed aims, purposes, ends, or norms, which cannot themselves be straightforwardly true or correct. The prominence of Williams’ view, and the relative inconspicuousness of Mackie’s view, in discussions of skeptical views of practical reason, reflects a quite mistaken estimate of their relative plausibility.

3 How to Argue for Moral Skepticism Mackie’s error theory is generated by the combination of two claims: (1) that ordinary moral judgments presuppose the existence of objective values, and (2) that there are no objective values. Focus for now only on the argument for (2), the thesis Mackie calls “moral skepticism.” Mackie distinguishes two different arguments for moral skepticism, the argument from relativity and the argument from queerness. In introducing the argument from queerness, he says it is “even more important” and “more generally applicable” than the argument from relativity. Mackie gives an illuminating summary of the case for moral skepticism as he sees it. In the summary the first consideration is the argument from relativity; the second, third, and fourth are elements of the argument from queerness: The considerations that favour moral scepticism are: first, the relativity or variability of some important starting points of moral thinking and their apparent dependence on actual ways of life; secondly, the metaphysical peculiarity of the supposed objective values, in that they would have to be intrinsically action-guiding and motivating; thirdly, the problem of how such values could be consequential or supervenient upon natural features; fourthly, the corresponding epistemological difficulty of accounting for our knowledge of value entities or features and of their links with the features on which they would be consequential; fifthly, the possibility of explaining, in terms of several different patterns of objectification . . . how even if there were no such objective values people not only might have come to suppose that there are but also might firmly persist in that belief. (Mackie 1977, p. 49)

There are important ways in which Mackie’s presentation of the case for moral skepticism can be improved. About two of these ways, I take myself to agree with Joyce. First, I think that the argument from queerness and the argument from relativity have a common target, a view about practical reason. According to this view, there are categorical imperatives, authoritative norms of practical reason. The argument from queerness and the argument from relativity are supposed to give us reason to suppose that there are no such authoritative norms. Thus it is a mistake to suggest, as Mackie’s writing sometimes does, that it is distinctively moral norms, rather than authoritative norms of practical reason more generally, that are his target. Secondly, I think it is a mistake to think (as Mackie sometimes suggests, for instance in the second item in the quotation above) that the problem with categorical imperatives is a motivational problem. As David Brink (1984, 1989) and others have argued, it is

96

D. Phillips

possible to make moral judgments objective without thinking that they necessarily motivate; and as Korsgaard and Smith have argued, the most sensible construal of the necessity in motivational internalism is as rational necessity: that such judgments will motivate you if you are rational. When motivational internalism is understood in this way, it is queer only if rationality itself is queer. It is a mistake to suppose that motivational necessity would have to be automatic in a way that would bypass an agent’s rationality (see Garner 1990). But in a crucial way, I think Joyce’s reconstruction of the case for moral skepticism is mistaken. Joyce devotes Chapter 5 of The Myth of Morality, entitled “Internal and External Reasons,” to a defense of a variant of Williams’ view. At the end of the chapter, summarizing the ways in which he differs from Williams, Joyce writes: I would not in fact go as far as to say bluntly that any use of an external reason is false. We must recall the discussion of institutional reasons from Chapter 2: reason claims that are legitimated by the rules of an institution, and which may well be external in nature. My objection is only with external reason claims that do not know their place–that overstep themselves by claiming to transcend all institutions. Such, I have argued, are moral reasons. Williams’ constraint on reason claims–that a reason must be a potential motivator–is best construed as a necessary condition of normative reasons. (Joyce 2001, p. 133; italics in original)

In this passage Joyce recognizes, in effect, the problems on the second horn of the dilemma we presented for Williams. But he shows no recognition of the problems on the first horn. But it is not just that, as we have already seen, Williams gives a very problematic account of normative practical reasons claims. Using Williams’ account of normative reasons to support Mackie is a grave strategic error. Mackie wants to offer naturalistic objections to the existence of genuine requirements of reason. If, like Williams, you allow that there are genuine requirements of reason (albeit requirements derived from an agent’s desires), the whole strategy of principled naturalistic objection to straightforwardly true normative practical reasons claims is undermined. The critic is perfectly placed to respond: you too admit that there are authoritative reasons. So you cannot launch a naturalistic objection to the notion of authoritative reasons. And once you have admitted that there are authoritative reasons, we can show you (by rehearsing the kind of problems discussed in Section 2) why you should reject the view that such reasons are derived, or solely derived, from the agent’s desires.

4 Objections A first objection is that there is really no difference between holding (as Williams and Joyce do) that the only real or genuine reasons are internal, and holding (as I suggest Mackie does) that categorical imperatives are problematic while hypothetical imperatives are not. But there is an important difference. There are two very different ways to develop the idea that hypothetical imperatives are unproblematic. One way is Williams’ and

Mackie on Practical Reason

97

Joyce’s way. It is to take there to be genuine reasons grounded in agent’s desires. The desires are the source, the only source, of genuine practical reasons. This line of thought faces the three problems we have already identified. The second way to develop the idea that hypothetical imperatives are unproblematic – which, I suggested, is Mackie’s way and is much more sensible – is to take it that the key idea in hypothetical imperatives is the idea of hypothesis, or presupposition. Such imperatives bind only on a certain hypothesis, the hypothesis involved in having a desire, that a certain end is good. But then we are in a position to see that an agent’s desires are not the only possible source of such hypotheses. Such hypotheses may derive from presupposed frameworks, institutions, or contexts, which are not the product of the desires of the agent. This latter way of understanding hypothetical imperatives is clearly superior, I suggest. For it does not involve an arbitrary privileging of the agent’s own current desires, and does not require the supporter of hypothetical imperatives to admit the problematic idea that there are framework- or presupposition-independent good practical reasons. Darwall develops the contrast I have in mind in an illuminating way: It is a familiar thought that instrumental reasoning issues “hypothetical” rather than “categorical” imperatives. However, care is required in interpreting this idea . . . . What we should say is that the imperatives or prescriptions that follow from instrumental reasoning are imperatives of practical consistency. They tell the agent either to take the means or to give up the end or the belief that the means in question is the only means. They do not tell the agent, If A is your end, and B the only means, then you should do B. There is a sense in which they do recommend B hypothetically, but in this sense B is recommended conditionally on a “hypothesis” that the agent assumes in having A as her end, namely, that A is to be done. The imperatives, then, do not recommend doing B simply on the condition that the agent has A as end. In this way, they are just like hypothetical theoretical reasoning. (Darwall 2001, pp. 141–142)

A second objection is that my comparison between Williams’ view and the view I attribute to Mackie is unfair, because, in developing the first horn of the dilemma for Williams, I take Williams to be in a more demanding and difficult business than Mackie: the business of analyzing or offering a non-debunking account, rather than revising or offering a debunking account, of normative practical reasons claims. In response to this objection, I would argue first that the comparison is not unfair. What I have argued is that Williams account fails whichever way you read it, while Mackie’s is a plausible debunking account. This matters in two ways. First, it matters in the general context of debates about practical reason. Williams’ aim appears in part to be to develop an alternative to the view of the “external reasons theorist,” a character Williams introduces towards the end of “Internal and External Reasons.” Williams characterizes the external reasons theorist initially as someone who “wants to oppose Hume’s general conclusion and to make a lot out of external reasons statements” (Williams 1981, p. 108). Mackie’s alternative to the view of the external reasons theorist is clearly superior to Williams’ alternative, and ought to be taken much more seriously in debates about practical reason. Second, the superiority of Mackie’s view over Williams’ view matters in the particular context of arguments for moral skepticism, the thesis that there are no objective values. Given

98

D. Phillips

that superiority, it is a mistake to burden arguments for moral skepticism with the peculiar costs of Williams’ view. There is, however, a further possibility. Mackie famously embraces an error theory. He thinks that the (mistaken) idea that there are objective values is built not only into philosophical theory but also into ordinary language. As he puts it: Objectivism about values is not only a feature of the philosophical tradition. It also has a firm basis in ordinary thought, and even in the meanings of moral terms. (Mackie 1977, p. 31)

But we might reject this idea. We could then ask how Mackie’s view of normative practical reasons claims would fare as a non-debunking view. In particular, we could ask how well it would fare against the three objections brought against Williams’ view on the first horn of the dilemma. Here again, Mackie’s presupposition approach is clearly superior to Williams’ approach. Consider in turn the three problems for Williams. First there is the problem that desires seem too arbitrary and contingent to be sources of authority. The presupposition approach would take the sources of authority to be assumed or presupposed standards, ends, criteria, etc. These are better sources of authority than desires. On the presupposition approach, none of them is straightforwardly or presupposition-independently correct. In that sense their authority is not complete. But they can be whatever standards, ends, criteria, etc., we most standardly take to be relevant and appropriate in particular normative contexts. In this way they are better sources of authority than desires. Second, there is the problem of conflict between Williams’ view and our intuitive judgments about what we have reason to do, as exemplified by the cases of the irrational smoker and the person who is future-Tuesday-indifferent. Again, the presupposition approach does better than Williams’ view. It does not force us to deny that the future-Tuesday-indifferent person has reason to care about pain on future Tuesdays. Rather, it tells us that this claim will be true when interpreted presupposing the norms we are most likely to invoke here – norms of prudence. If common sense tells us that these norms of prudence are themselves presupposition-independently correct, then common sense in this respect conflicts with the presupposition approach. But unlike Williams’ approach, the presupposition approach can do a reasonable job of accommodating the ordinary intuition that future-Tuesday-indifference is irrational, by understanding this intuition as invoking norms of prudence. Third, there is the problem of conflict between Williams’ view and the phenomenology of deliberation, the problem that much normal deliberation would be undermined by accepting Williams’ view. Again, Mackie’s presupposition approach does better. It is not in the same way incompatible with serious deliberation. In deliberating we often assume that the values of options are not the product of our subjective motivational sets; we assume that the options have value independent of our subjective motivational sets. It is not nearly so clear that we assume a more specific positive view: that that value is built into the fabric of the world, rather than built into normative systems which we accept. So, again, Mackie’s looks a

Mackie on Practical Reason

99

more promising non-debunking approach to normative practical reasons claims than Williams’. Normal deliberation does not obviously assume that the presupposition approach is mistaken, while much normal deliberation does assume that Williams’ view is mistaken. Thus, I suggest, Mackie’s approach to practical reason is not only superior to Williams’ as a debunking approach, it is also superior as a non-debunking approach. But this result, while interesting, is not directly relevant to the question how best to reconstruct Mackie’s case for moral skepticism. For Mackie holds that ordinary moral thought involves a commitment to objective values, and hence has to develop an error theory.

5 Conclusions Mackie’s approach to practical reason is distinctive and unjustly neglected. It is much superior to Williams’ more famous view. And it is a serious mistake, for both general and more particular strategic reasons, to reconstruct Mackie’s argument for moral skepticism with Williams’ view as a premise. Acknowledgments Many thanks to Carl Feierabend, Bredo Johnsen, Bill Nelson, an audience at the University of Houston, and the editors of this volume for helpful comments on earlier versions of this chapter.

References Brink, D. 1984. Moral realism and the sceptical arguments from disagreement and queerness. Australasian Journal of Philosophy 62: 111–125. Brink, D. 1989. Moral realism and the foundations of ethics. Cambridge: Cambridge University Press. Darwall, S. 2001. Because I want it. Social Philosophy and Policy 18: 129–153. Garner, R. 1990. On the genuine queerness of moral properties and facts. Australasian Journal of Philosophy 68: 137–146. Joyce, R. 2001. The myth of morality. Cambridge: Cambridge University Press. Korsgaard, C. 1986. Skepticism about practical reason. Journal of Philosophy 83: 5–25. Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin. Parfit, D. 1984. Reasons and persons. Oxford: Clarendon Press. Parfit, D. 1997. Reasons and motivation. Proceedings of the Aristotelian Society 71(suppl.): 99–130. Pettit, P. and M. Smith. 1990. Backgrounding desire. Philosophical Review 99: 565–592. Scanlon, T. M. 1998. What we owe to each other. Cambridge, MA: Harvard University Press. Smith, M. 1994. The moral problem. Oxford: Blackwell. Williams, B. 1981. Internal and external reasons. In his Moral luck, 101–113. Cambridge: Cambridge University Press. Williams, B. 1995a. Internal reasons and the obscurity of blame. In his Making sense of humanity, 35–45. Cambridge: Cambridge University Press. Williams, B. 1995b. Replies. In World, mind, and ethics, eds. J. E. J. Altham and R. Harrison, 186–194. Cambridge: Cambridge University Press. Williams, B. 2001. Postscript: Some further notes on internal and external reasons. In Varieties of practical reasoning, ed. E. Millgram, 91–97. Cambridge, MA: MIT Press.

The Argument from Moral Experience Don Loeb

1 Introduction: The Pervasiveness of the Argument It is often said that our moral experience, broadly construed to include our ways of thinking and talking about morality, has a certain objective-seeming character to it. There are two ways in which experience of this sort has been thought relevant to the central questions of metaethics. The first involves the traditional idea that moral experience – especially our dispositions to use the moral vocabulary in various ways – is among the best evidence we have about what it is we are (thinking and) talking about when we talk about morality: We are, it is claimed, talking about a realm of (putative) fact.1 The second way is not inconsistent with the first. It involves an inference from morality seeming a certain way (or our practices somehow presupposing it to be that way) to the reasonableness of a presumption that it is that way. It is widely thought that the objective-seeming nature of our moral experience supports a presumption in favor of objectivist theories (according to which morality is a realm of non-relative facts or truths) and against anti-objectivist theories such as Mackie’s error theory (according to which it is not). The presumption can be defeated, it is claimed, only if arguments against objectivist theories prove successful. I will call the argument that our moral experience supports objectivist theories the Argument from Moral Experience, or the AME for short. In some form or other, the AME (or at least the presumption it is said to ground) has played an important role in the debate over moral realism.2 Many of those accepting it consider themselves to be moral realists and take the AME to sup-

D. Loeb (B) Department of Philosophy, University of Vermont, Burlington, VT, USA 1 Mackie

(1977, pp. 30–35, 50–63) embraced this traditional idea. See also Smith (1993) and Jackson (1998). 2 For convenience, I’ll omit “or at least the presumption. . .” in what follows. It seems fair to assume that those employing the presumption are at least implicitly relying on the AME. Otherwise they owe us an alternative defense of the presumption, and it is hard to imagine what that could be.

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_7, 

101

102

D. Loeb

port their realism.3 Thus, Jonathan Dancy claims that a version of the AME is “the main argument for moral realism” and even “perhaps the only argument for realism, remaining thoughts being used for defence/offence” (Dancy 1986, pp. 172, 175). He describes “the simple form of the argument” as follows: [W]e take moral value to be part of the fabric of the world; taking our experience at face value, we judge it to be the experience of the moral properties of actions and agents in the world. And . . . we should take it in the absence of contrary considerations that actions and agents do have the sorts of moral properties we experience in them. This is an argument about the nature of moral experience, which moves from that nature to the probable nature of the world. (op. cit., p. 172)

According to David Brink, the “presumptive case in favor of moral realism. . . shift[s] the burden of proof to the moral antirealist” (1989, p. 36). David McNaughton puts it even more forcefully, saying “The realist’s contention is that he has only to rebut the arguments designed to persuade us that moral realism is philosophically untenable in order to have made out his case” (1988, pp. 40–41).4 This assignment of the burden of proof helps to explain why so much of the debate over moral realism has centered on arguments against it, and far less has focused on arguments in its favor. If the negative arguments can be defeated, it is widely thought, then moral realism wins by default. But those who view themselves as moral realists cannot appropriate the AME as readily as they have often claimed, for it can be argued that our moral experience does not presuppose that moral realism is correct, but at most that morality is objective in the sense that there are non-relative moral facts, truths, properties, or correct answers to our moral questions.5 All moral realists are objectivists, but many objectivists do not consider themselves to be moral realists. For example, most Kantians and Kantian constructivists would agree that a claim of objectivity is implicit in our moral thinking, and that this supports objectivism about ethics. But they would not agree that ordinary moral experience thereby supports moral realism. On their views, certain moral judgments are practically correct even though moral properties are not real and the correctness of moral judgments does not consist in correspondence between our statements and the world.6

say “. . . consider themselves to be moral realists” because it has become less and less obvious that we can find a satisfactory account of what it is to be a moral realist. (See Dreier, this volume.) Among the many in this category who appear to accept some version of the AME are: Bloomfield (2001), Brink (1989), Dancy (1986), Lovibond (1983), McNaughton (1988), Nagel (1986), ShaferLandau (2003), Smith (1994), and Wiggins (1988). 4 See also Nagel (1986): “[I]t is very difficult to argue for such a possibility [as value realism] except by refuting certain arguments against it” (p. 143). 5 For simplicity’s sake, I’ll speak simply of morality as a (possible) realm of fact. 6 See, for example, Korsgaard (1996, pp. 35–37, 44–48, 112). Although Korsgaard denies that there are moral facts, she nevertheless holds that there are correct answers to moral questions. 3I

The Argument from Moral Experience

103

Even those in the broadly noncognitivist7 camp, who hold that moral utterances are (primarily) something other than straightforward assertions of putative fact, often claim it an advantage that their theories can accommodate (or so they maintain) the most important features of our objective-seeming moral experience. For example, Simon Blackburn says that “the most forceful attack” faced by the moral projectivist “is that he cannot accommodate the rich phenomena of moral life,” but goes on to argue that his “quasi-realism” can indeed accommodate these phenomena (Blackburn 1993, p. 158).8 The near universal acceptance of the AME helps to explain why the error theory has had so few adherents in recent years. The error theory, made famous by Mackie (1977), combines cognitivism, the view that the central function of moral utterances is to make factual assertions, with the view that all (positive) assertions of this sort are false.9 Thus it is a prime example of an anti-objectivist theory. A weak, burden-shifting, presumption in favor of objectivist theories would not be sufficient to explain why the error theory is held in such low regard, however. Instead, a much more ambitious version of the AME seems to be in play, one which holds that there is an overwhelming presumption in favor of objectivist theories. If the error theory is true then it is not a fact that torturing infants for fun is morally wrong. Antiobjectivist implications like this are not merely taken to place the burden of proof on error theorists, but to make that burden extremely strong. Indeed, Mackie himself was in the grip of an ambitious version of the AME. Mackie was convinced that “ordinary moral judgments include a claim to objectivity, an assumption that there are objective values” (Mackie 1977, p. 35). This was the basis for the cognitivist component of his error theory. But he also believed that in virtue of this claim views like his faced an uphill climb: But since this is an error theory, since it goes against assumptions ingrained in our thought and built into some of the ways in which language is used, since it conflicts with what is sometimes called common sense, it needs very solid support. It is not something we can accept lightly or casually and then quietly pass on. (Mackie 1977, p. 35)

I agree with Mackie that objectivism should not be rejected lightly or casually. If objectivism is needed in order to validate common sense, then it might be very disturbing to find out that it is wrong. But that in itself gives us no reason to believe that objectivism is correct. In what follows, I ask, first: In what sense is it true that we experience morality as a realm of fact? Although this is largely an empirical question, I argue that evidence we already have suggests that things are much messier than objectivists 7 Some

prefer the etymologically more revealing “descriptivism/non-descriptivism” to the more common “cognitivism/noncognitivism.” See Timmons (1999, p. 19). 8 For related views, see Timmons (1999) and Wright (1992). See also Gibbard (1990, especially Chapters 8–13), who employs a number of ingenious strategies for accommodating what he calls the “objective pretensions” of moral thought and language. For earlier noncognitivist attempts to capture some of this seeming objectivity, see Stevenson (1950) and Hare (1981). 9 More recent error theorists include Garner (1994), Joyce (2001), Greene (2002), and Lillehammer (2004). For a related view, see Schiffer (1990).

104

D. Loeb

have supposed. Second: What support does our moral experience give to the claim that morality is such a realm? I argue that even if moral experience were to display morality as a realm of fact (or presuppose that it is), that would not in itself support objectivism. The stakes are high. If I am right, then, at a minimum, the playing field should be leveled. One might even have thought that those defending an existence claim should themselves accept the burden of proof. If so, and if, as Dancy claims, the AME is the only argument for objectivism, then its failure would mean that anti-objectivist views like the error theory have the upper hand.

2 Do We Experience Morality as a Realm of Fact? In what sense is it true that we experience morality as a realm of fact? To begin, “our moral experience” is not as uniform as that expression suggests.10 (See Kirchin, this volume.) Undoubtedly there are both cross-cultural and intra-cultural differences. Still, there are some obvious commonalities for many of us, and perhaps we can base a reasonable gloss on “our moral experience” on these. Brink (1989, pp. 25–29) presents one of the fullest cases for the claim that our moral experience has an objective-seeming character. Moral utterances, he points out, are often in the declarative mood, and thus appear to be statements of fact. Our beliefs that “one should not be held responsible for actions one could not have known were wrong,” that “goodness deserves reward,” and so on, contain implicit references to moral properties, facts, or knowledge, he argues. Furthermore we often wonder what morality requires, and when we do, “we often deliberate as if there is a correct answer to the question before us.” We disagree with those whose moral views seem mistaken. We recognize that we have been mistaken in the past (and might be now), but we also believe that we can be right.11 McNaughton (1988, pp. 19, 48, 56) adds other elements, centered on the phenomenology of our moral experience. Morality seems to be in the world apart from our happening to encounter it, not something that depends for its existence on our subjective inputs. We seem to have moral perceptions, which, although not strictly analogous to visual perceptions, nevertheless are like them in that they seem to be of something outside of us. We seem to see conduct as right or wrong. And when we are moved to act morally, it seems to be in virtue of our recognizing morality’s authority over us.12 This appears to be an impressive case for the claim that we experience morality as a realm of fact. Of course, few non-philosophers have thought about the issue in terms that make explicit reference to moral facts, and it seems illegitimate to

10 Indeed,

even what counts as “experience” will be disputed. Here I use the term very broadly. distinguishes features of our moral experience he thinks count against noncognitivism from those he thinks count against constructivism. 12 For another litany of objective-seeming features, followed by an extended discussion of one of them, see Timmons (1999, pp. 74–106). 11 Brink

The Argument from Moral Experience

105

attribute much metaethical theory to most people. But most people do seem to recognize a distinction between realms of fact and others (such as fashion) in which normative facts are not thought to play a central role. Perhaps they would be disposed on reflection to place morality in the former category. Or perhaps our moral practice and thought in some more inchoate way presuppose objectivism. For example, the claim that we can reason about moral questions can be true, it might be claimed, only if morality is objective. Our moral beliefs and practices might be appropriate if objectivism is correct, but not if it is incorrect. But is it clear that even this more limited presupposition is present? It is surprising how often philosophers are willing to generalize about complex, subtle, and largely empirical matters like this based merely on their own experience and intuitions. The evidence may well take us in more than one direction. People’s beliefs and presuppositions, especially about philosophical matters, are often intrapersonally inconsistent, and interpersonal and cross-cultural differences surely complicate things. Thus, even if we find substantial evidence of a commitment to objectivity, we may also find evidence against such a commitment. Indeed, certain features of common moral experience do suggest that we experience morality as something that is not objective.13 For example, just as we talk about moral beliefs, we often talk about moral feelings and attitudes as well, and in other contexts these words typically signify something other than beliefs. In fact, people often say things that seem quite incompatible with objectivism, such as that in ethics “it’s all relative,” or that what it is right for a person to do depends on that person’s own decisions. We cannot dismiss such statements as the products of confusion merely because they appear to conflict with views we think widely held.14 There are other problems for the claim that we experience morality as a realm of fact. In particular, anti-objectivists claim that their theories can accommodate many of the features often cited as pointing towards objectivism. Unless ordinary people have metaethical views inconsistent with that claim, we need a reason for thinking that their experience presupposes objectivity, and such a reason has not been supplied. Mackie, for example, thought it entirely appropriate to reason about questions of value, even though he denied that values are objective. Barring the implausible assumption that ordinary people think reasoning is only appropriate when it comes to objective matters, ordinary moral reasoning could be of just the sort Mackie endorses. Or it might fail to reflect either an objectivist or an anti-objectivist presupposition. Similarly, because Mackie thought of morality as something to be made and not discovered, it is not at all surprising to find him writing about values in the declarative mood. Similar things can be said about disagreement on questions of value, and even, perhaps, about the doubting and wondering mentioned by Brink. Once again, 13 Christopher Hookway (1986, p. 190) expresses doubts about the objective-seeming nature of our

moral experience. See also Nichols (2004, Chapter 8). (1988, pp. 3–4) himself begins his book with a discussion of “two contrasting feelings about our moral life that all of us share to some extent,” one of which “appears to lead to the view that there is nothing independent of our moral opinions that determines whether or not they are correct.”

14 McNaughton

106

D. Loeb

unless we have reason to believe that ordinary people are in the grip of a metaethic incompatible with Mackie’s, proponents of the AME need to say more if they are to make persuasive the claim that ordinary moral practices presuppose objectivity. In fact, much of the evidence cited in favor of the objectivist reading of moral experience – for example, that we tend to express our moral thoughts in the declarative mood – seems weak. We talk about matters of taste in the declarative mood (saying, for example, that chocolate is better than vanilla), but ordinary people would find “gastronomic objectivism” about such matters utterly implausible. Similarly, the fact that we seem to have perceptions of morality does not strongly support the claim that we experience it as objective, since we seem to have perceptions of the (comparative) goodness or badness of some foods’ tastes but many would find odd the claim that we experience the taste of food as in fact good or bad or that we typically presuppose objectivism about gastronomic value.15 Indeed, even if Mackie and other anti-objectivists are wrong about the reasonableness of ordinary people’s practices and beliefs (on the assumption that there are no moral facts), it may still be inappropriate to treat such people as presupposing moral objectivity, at least if by “presuppose” we mean to imply a sensitivity to the rightness or wrongness of these metaethical claims. The rightness or wrongness of those claims is unlikely to have had an impact on ordinary people’s thought. Still, if anti-objectivists are wrong about these matters, perhaps there is another sense in which ordinary people’s views and practices could be said to presuppose some form of objectivism: Their views would not be right and their practices would not make sense unless objectivism were correct.16 Presupposition in this second sense might be relevant to some (but not all) versions of the AME. But whether ordinary people’s views and practices in this second sense presuppose some form of objectivism is an important issue in the debate over moral objectivity. It is inappropriate to make a burden-setting presumption in that very debate rest on a particular resolution of a matter at issue between the two sides. Undoubtedly, much more could be said about these questions (see Loeb 2008). But it should already be clear that our moral experience is much more varied and complex than proponents of the AME have assumed. Rather than continue to investigate these matters here, I shall set them aside and ask whether experience of the sort gestured at above, even interpreted in a way reasonably favorable to objectivists, would by itself support objectivism. The answer, I will argue, is that it does not. To supplement it, objectivists would need a favorable resolution of certain issues already central to the debate over moral objectivity. That being so, a presumption in favor of objectivist theories is not warranted.

15 For

a sustained discussion of the parallels between gastronomic and moral realism, see Loeb (2003). 16 By analogy, if Kant’s arguments were sound, then human experience of objects would presuppose, in this second sense, that we have a priori concepts that correctly apply to things. The existence of these a priori concepts would be a necessary condition of our experience of objects; however, the claim that they exist is one very few people would even recognize to be true.

The Argument from Moral Experience

107

3 Is the Presumption Justified? Does a presumption even require a justification? It is by now a philosophical commonplace that we have to start somewhere. But we are not being asked to view the objectivist presumption as itself a philosophical starting point. We argue for the presumption on the basis of our moral experience, experience that is thought to be evidence for objectivism. That being the case, it seems fair to ask how the evidentiary support is supposed to work. Of course, we could make a further presumption here, that our moral experience is evidence for objectivism (even if we can’t explain how this evidence is supposed to work). However, that would seem rather ad hoc, to say the least. And, although those who appeal to moral experience often seem to be taking the evidentiary link for granted, there are hints of various attempts to support it. In what follows, I consider four related strategies for supporting the presumption. Although these appear to be the most promising, I argue that none of them is in fact successful.

3.1 The Best Explanation The most prominent approach suggests that objectivism allows us to best explain the nature of our moral experience. For example, Brink argues that we can most readily explain our moral thought and behavior on the hypothesis that moral inquiry is directed at discovering independently-holding moral facts, and that this supports moral realism: [G]eneral considerations about the nature of inquiry and considerations about moral inquiry in particular are most easily explained on the assumption that moral inquiry is directed at discovering moral facts that obtain independently of our moral beliefs and at arriving at evidence-independent true moral beliefs. I take this to establish a presumptive case in favor of moral realism and to shift the burden of proof to the moral antirealist. (Brink 1989, p. 36)

But the claim that moral inquiry is directed at discovering moral facts, even if true, does not by itself support the claim that such facts exist, any more than the fact that prayer is directed at God supports theism. The most obvious way to supplement it would be with a second explanatory claim – that the best explanation for the direction of moral inquiry involves the hypothesis that realism (or objectivism) is true. Brink may have something of this sort in mind when he suggests that “features of actual and possible moral inquiry are hard to understand on antirealist assumptions and much easier to understand on realist assumptions. Realism and realism alone,” he says, “provides a natural explanation or justification of the way we do and can conduct ourselves in moral thought and inquiry” (1989, p. 24). This argument, however, seems to run explanation and justification together. The fact that we need to understand our moral experience does not support the claim that we must show it to make sense. But conflating explanation and justification is fairly common, and this may misleadingly boost the appeal of the AME. It is often

108

D. Loeb

facilitated by speaking of the need to “make sense of” our moral experience.17 We do need to understand that experience. But the claim that our moral practices are most naturally justified on the hypothesis that moral realism (or objectivism) is true is controversial, and, even if correct, could not support moral realism (objectivism) – not, that is, without the added premise that our practices are typically justified, or something of that sort. Such a premise would be question-begging in this context since anti-objectivists believe that insofar as these practices presuppose that morality is a realm of fact they are not justified. What then of the claim that our moral practices can be explained more naturally or easily on the assumption that moral realism is true? Brink seems to be assuming that what he calls a more natural or easier explanation is a better explanation, and thus that we can best explain our moral practices, experiences, and beliefs on the assumption that moral realism is true. But he does not spell out how such an argument would proceed. One possibility (expressed in terms of objectivism) is as follows: If we experience morality as a realm of fact, there must be some explanation for that. The easier and more natural (and thus best) explanation is that morality is such a realm and that we are familiar with it as such. We think moral beliefs can be mistaken or correct because they can be mistaken or correct; we deliberate because there are answers. And so on. But if we are to accept the objectivist’s explanation as best, we need more than the mere claim that we experience morality as objective because it is objective. We need to know how its being objective can explain our having the experiences we do.18 Or at least we need some good reason for thinking that its being objective would explain why our experience turns out as it does. But such a reason has not been given in the context of this argument, and we cannot simply assume that one is available. More importantly, whether moral facts help to explain our moral experience is a special case of one of the central controversies in metaethics: whether moral facts figure in the best explanation of anything.19 It is inappropriate to make a presumption in favor of moral objectivity depend on the outcome of a major controversy in the debate over whether morality is objective. If the explanatory claim were to succeed then we would be beyond the point where we needed the presumption, and if it were to fail then the presumption could not be grounded on its success. It might be tempting to think that a demand that moral facts be shown to figure in best explanations sets the bar too high for a mere presumption. The presumption is only meant to set the burden of proof, not to establish that objectivism is correct. 17 For example, Dancy (1986) says that “we abandon moral realism at the cost of making our moral

experience unintelligible,” and that we can “make satisfactory sense of our experience of the moral properties of objects” only on the assumption that moral realism is correct (p. 173). See also Nagel (1986, p. 146), McNaughton (1988, pp. 16 and 52), and Smith (1994, pp. 5, 11). Timmons’ term, “accommodate,” is ambiguous in this way as well (Timmons 1999, p. 12). 18 Gilbert Harman (1986) makes a similar point. See also Blackburn (1993, p. 154). 19 The locus classicus of this debate is Harman (1977, Chapter 1), and Sturgeon (1984). For a recent attempt to put this debate into perspective, see Loeb (2005).

The Argument from Moral Experience

109

Shouldn’t a showing that a particular explanation is the most initially appealing be sufficient? Perhaps that is what Brink has in mind when he says that realism provides an easier and more natural explanation for our moral experience. The easier and more natural explanation, he could argue, is good enough to support a presumption in realism’s favor. But it could be overridden if a better explanation were ultimately to be discovered. The problem with this line of thinking is that whether or not realism provides the most initially appealing explanation is itself at issue between objectivists and anti-objectivists. First, it is not agreed that realism (or objectivism) does provide an easier and more natural explanation of our moral experience. Second, even assuming that easiness and naturalness are explanatory virtues, they are certainly not the only such virtues. Anti-objectivists are likely to point to others – such as parsimony, or fit with the best current psychological or evolutionary theory – as supporting their own explanations.20 Here again, the presumption cannot be grounded in a claim that reflects an unresolved issue in the debate.

3.2 Epistemic Conservatism Another possible ground for the presumption focuses mainly on one particular aspect of our moral experience – our beliefs. Many philosophers have felt that, other things being equal, we should favor the theory that requires us to give up as few of our current beliefs as possible.21 It is not clear, however, that the conservative strategy could work here, even in its own terms. If it is admitted that ordinary people are not explicit objectivists, then for them at least, a belief in objectivism itself is not available for conserving. Still, ordinary people do have beliefs objectivists might think cannot be true unless objectivism is true. And, it could be argued, objectivism allows us to conserve more of these beliefs than do anti-objectivist approaches such as the error theory. For example, most people believe that our moral thinking is fallible. Objectivists often allege that there is no room for this to be true within an anti-objectivist framework. If this allegation were correct, then a principle of conservatism might have us grant some credibility to objectivism, since objectivism would follow in some sense from things we already believe – things the conservative approach tells us to treat as having some credibility. Likewise, the objectivist could argue, our moral beliefs themselves cannot be correct unless objectivism is correct. If we are to conserve these beliefs, they could say, we must accept some form of objectivism. But the principle of conservatism is quite controversial, and for good reason. It is implausible that the mere fact that we already believe something should by

20 See

Mackie (1977, pp. 42–48). For a more recent, and much more detailed, suggestion along these lines, see Joyce (2006). 21 This is (roughly) Quine’s (1951) formulation of the principle of conservatism. Other formulations have been put forward, but the differences do not affect the argument given here.

110

D. Loeb

itself give that belief any warrant, quite apart from any other beliefs we might have about the reliability of the processes that led to the formation of that initial belief. Suppose I believe that the next Lotto number drawn will be even, but suppose further that I have no actual evidence for that hypothesis. I should believe that the odds are 50/50, in spite of the fact that I find myself with the ungrounded belief that they are not. That I do believe the number will be even gives me no warrant whatsoever for believing it. Indeed, even if I knew that most ordinary people had the same belief, this too would not by itself give me any evidence that it is true.22 Conservatives might try to restrict the range of beliefs to be conserved, so as to avoid counterexamples of this sort, for example by pointing out that we cannot see any way that my belief about the Lotto could be explanatorily connected to its own truth. In other cases, we can see such a connection. For example, I believe that the Battle of Hastings was fought in 1066, though I cannot remember my evidence for this belief. Still, I have reason to maintain the belief since it is likely that it was produced by a reliable process (such as my having read a history book). But such a reason is lacking in the case of my belief about the Lotto. Perhaps, then, we should conserve only those beliefs for which a plausible credibility-imparting explanatory story regarding their provenance can be told. Anti-conservatives will no doubt argue that to adopt an approach like this is to abandon conservatism entirely. But we need not enter into that debate here, for, returning to our previous examples, the restricted conservative strategy would require us to find a credibility-imparting explanation for the widespread belief in moral fallibility, or for our moral beliefs themselves – a showing that these beliefs are plausibly explained by the states of affairs that make them true. And that is just another instance of the question of whether moral facts figure in the best explanation of anything. As we have just seen, resolving this question cannot be a prerequisite for grounding an independent presumption in favor of objectivism. There may be other ways of restricting the range of beliefs to be conserved in order to avoid the implausible implications of wholesale conservatism. But if so, and if proponents of the AME wish to rely on them, they need to produce them. And they must show us that these approaches are not themselves disguised versions of the inference to the best explanation strategy.

3.3 The Principle of Credulity A third suggestion relies on another principle it is sometimes thought we must accept if we are to avoid falling into skepticism. This principle, often referred to as the Principle of Credulity, holds that we are entitled to presume that things are

22 For a persuasive critique of the principle of conservatism, see Christensen (1994). My discussion

here has benefited greatly from that paper and from conversations with its author.

The Argument from Moral Experience

111

pretty much as they seem to be, unless there is substantial evidence to the contrary.23 Just as the appearance of an external world is sometimes thought to entitle us to presume that there is such a world, the appearance that there are moral facts is thought to entitle us to believe that there are such facts.24 Even if we needed such a presumption in order to deflect external-world skepticism, however, this would not by itself give us reason to adopt a similar, but distinct, presumption to get us out of moral skepticism. Perhaps external-world skepticism is so implausible that we would never be satisfied unless we were to deny it. But that would not establish that moral skepticism involves a price high enough to make its avoidance imperative as well. (Indeed, most anti-objectivists think that the price is nowhere near as high, if there is any price to be paid at all.) Instead, the suggestion seems to be that a very general presumption that things are as they appear (barring substantial evidence to the contrary) is necessary for avoiding external-world skepticism, and that the very same presumption also has implications for the moral case. Now, given the implausibility of attributing a belief in objectivism to most nonphilosophers, it would be inappropriate to claim that morality seems objective to the average person. So here again we may feel compelled to rely on other prephilosophical beliefs thought by objectivists to be best accommodated within an objectivist framework. For example, since it seems as though we can reason about morality, we might be entitled to believe that we can reason about it, in the absence of substantial evidence to the contrary. If objectivism in fact does best accommodate this belief and others like it, a reasonable extension of the strategy might hold that we are entitled to accept it, at least provisionally. If morality is not a realm of fact, the argument would go, then things are not as they appear. Many people would find the kind of anti-skeptical strategy described above very unappealing, however, even in the external world case. If there are arguments against skepticism that are not based on the Principle, they would say, then we can examine them. But if they are unsuccessful, then we should let the chips fall where they may. Unless the Principle has some significant appeal beyond the avoidance of skepticism, they would argue, we are not entitled to adopt it for that purpose alone. By itself, the Principle seems question-begging as a response to skepticism. In fact, the Principle of Credulity is implausible for precisely the reason conservatism (likewise commonly defended as necessary for avoiding skepticism) is implausible. Seemings in and of themselves give us no more warrant than mere believings. And just as the scope of conservatism’s reach would have to be restricted in order to avoid counterexamples such as my belief about the next Lotto number, 23 Paul

Bloomfield reminded me of this version of the argument. McNaughton (1988, pp. 39–41, ff.) relies on it and cites the discussion of the Principle in Swinburne (1979, pp. 254–271). A classic source of the Principle is Reid (1785/1975). 24 Even before the publication of Reid’s “Essays” in 1785, Richard Price (1758/1948, p. 45) offered something like this version of the argument: “Is there nothing truly wrong in the. . . misery of an innocent being? – ‘It appears wrong to us.’ – And what reason can you have for doubting, whether it appears what it is?”

112

D. Loeb

so the Principle of Credulity would have to be restricted as well. Suppose it seems to me that the next Lotto number drawn will be even. Unless there is a plausible credibility-imparting explanation for things seeming this way – a showing that this seeming is explanatorily related to the truth of the beliefs it generates – there is no reason to treat it as providing any epistemic warrant whatsoever. In the moral case it would have to be shown that moral facts figure in the best explanations of these seemings – a matter, as before, at issue between objectivists and anti-objectivists. Such a showing cannot be a prerequisite for an independent presumption in favor of objectivism. The anti-objectivist about morality can take no comfort in the defeat of a more general anti-skeptical strategy, however. For, as objectivists often point out, moral skepticism is interesting only if it is a distinctive form of skepticism. It would be no victory at all for the anti-objectivist to show that there is no more reason for believing in moral facts than there is for believing in an external world! What is wanted is some justification for treating moral properties differently. This suggests a familiar constraint for anti-objectivists: They should not rely on arguments that provide as much support for external-world skepticism as they do for moral skepticism. But there is an asymmetry here. For it would not be unreasonable to hold that even if external-world skepticism turns out to be false there is still no good reason for believing in moral facts. For example, even if we believe that the best explanation for the appearance of an external world is that there is an external world, nothing about the moral case follows. Thus anti-objectivists can reject the unrestricted Principle of Credulity as a way out of either sort of skepticism, remaining agnostic (at worst) about external-world skepticism. It could be claimed that any route out of external-world skepticism would apply equally well to moral skepticism. But that claim would need to be defended, and such a defense has not been produced, nor does it seem obvious that it could be. In fact, the unrestricted Principle of Credulity is not the only anti-skeptical strategy worth considering. For example, a more modest presumption that our senses and memory are reasonably reliable would still allow us to avoid external-world skepticism but it would not have the implication that morality is as it appears to be, since the appearances relied upon in the moral case do not result (primarily) from the operation of these faculties.25

3.4 Wide Reflective Equilibrium A final strategy, also closely related to conservatism, relies on a more general epistemic principle having to do with considerations of theoretical coherence. It is hard to see how we could avoid appealing to such considerations in evaluating theories. Just as it is often thought that good moral thinking involves a search for “wide reflective equilibrium” (WRE), good reasoning in general seems to involve such a 25 For

an anti-skeptical suggestion along these lines, see Chisholm (1977).

The Argument from Moral Experience

113

search as well. We appear to have no choice but to start off with the beliefs we have, not all of which cohere with one another. Our aim, it is often thought, should be to bring our beliefs into a coherent balance after careful consideration. Arguably, such a balance provides the best sort of justification available to us. How might this approach to justification favor a presumption of objectivism, given the nature of our moral experience? Here again the focus is primarily on our beliefs. It might seem as though the approach commits us to the view that even our pre-reflective beliefs about the nature of morality (such as our belief in the possibility of moral reasoning) have some initial warrant, warrant they carry with them into the balancing process. After all, it could be asked, how could beliefs that have no warrant, even when made coherent upon reflection, produce beliefs that are warranted? Such an argument would not rest on the false assumption that nothing can have properties different than the properties of its parts. Instead, it would be an application of what is sometimes called the “garbage in, garbage out principle” (GIGO). Unless the beliefs we start out with have some warrant, the argument would go, they have nothing to pass on to the beliefs that emerge in WRE. Their initial warrant, on this view, might be outweighed. But all of our initial beliefs – including those thought by objectivists to be most compatible with objectivism – have some warrant, since the beliefs that emerge in WRE have warrant. This argument seems highly suspect, however, for it opens a back door to epistemic conservatism by once again treating mere belief as sufficient for warrant. According to this approach, I should again treat my belief that the next Lotto number picked will be even as having some warrant, because the WRE approach requires me to treat all of my initial beliefs as having some warrant. Yet the claim that belief is sufficient for some degree of warrant is just as implausible when put in terms of WRE as it was when put in terms of conservatism. Once again, the mere fact that many of us have beliefs about the possibility of moral reasoning (and the like) does not in itself show those beliefs to have any warrant. Is there any other way in which WRE might be brought in to support the presumption? A hint may be found in Ronald Dworkin’s response to Crispin Wright’s claim that realism about a particular domain has the burden of proof: Unlike an argument in a court of law, however, the course of a philosophical investigation is fixed not by any free-standing methodological postulates, like Occam’s razor which Wright cites, but by how opinion stands when the investigation begins. No skeptical argument can succeed, for anyone, unless it brings him skeptical conviction, and that means that none of us can accept such an argument unless we find its premises convincing even when we grasp their skeptical import. We must find these premises more plausible than what they require us to abandon. (Dworkin 1996, p. 117)26

Expanding on Dworkin, we might argue as follows: Either our current moral beliefs (some of them, at least) are justified (enough to be presumptively true) or they are not. If they are justified, then objectivism must be justified as well, since the beliefs

26 For

an argument similar to Dworkin’s (which acknowledges that connection), see Nagel (1997, Chapter 6, especially pp. 110–112, 115–118, and 125).

114

D. Loeb

can’t be right unless objectivism is right. If they are not justified now, then once we have gotten appropriately close to WRE, our web of belief will either contain those moral beliefs or not. If it doesn’t, that will be because there appeared to us to be some reason to take them out. But that is just to say that our moral beliefs (and hence objectivism) are presumptively true. We are entitled to accept them, the argument would go, unless there is some reason for rejecting them. Such a reason would indeed defeat the presumption in favor of objectivist theories, but in its absence, the presumption holds. Such an argument, however, would be in one sense too strong, and in another too weak. If it were meant to show our initial moral beliefs to have some (defeasible) epistemic warrant, it would fail. For it is true of all of our current beliefs that we would not abandon them, upon reflection, absent a (putative) reason for doing so. But that does not mean that each of our beliefs is presumptively true. To think otherwise is once again to allow conservatism to sneak in, since it is to hold that any belief is to some degree warranted merely in virtue of being held. Alternatively, if the argument is not meant to show that these beliefs have warrant, then it cannot show that the burden of proof is on the anti-objectivist, as proponents of the AME claim. On this reading, what looked like an argument is merely a description of what many people happen to believe (and perhaps a prediction that they are unlikely to give most of those beliefs up in response to philosophical arguments). It would be an overstatement to call such beliefs presumptively true. At best, we could say that they have been presumed to be true. Indeed, Dworkin himself denies that the considerations he points to show where the burden of proof lies or “that our convictions are right just because we find them irresistible, or that our inability to think anything else is a reason or ground or argument supporting our judgment” (Dworkin 1996, p. 118). No doubt most people do have moral beliefs, and it would take a reason (or at least a putative reason) for them to give these beliefs up. But if we reject conservatism, we will hold that this alone does not give those beliefs any positive epistemic status. Whatever else is needed in order for them to earn that status is needed by both our current beliefs and any competitors. There is no asymmetry. Still, while it is implausible to think that individual beliefs have warrant merely because they are held, perhaps we have no alternative but to treat as warranted certain theories that unify and accommodate those beliefs. Where else could we look for justification? If objectivism does indeed unify and accommodate the various theoretical and substantive beliefs about morality that in part constitute our moral experience, then perhaps we should provisionally accept it. Conservatism, however, is no more plausible at the level of theories than it is at the level of individual beliefs. Suppose, to take a silly example, the theory that this bus has been commandeered by Martians nicely unifies and accommodates my beliefs about each of the various passengers — that they are non-human and that they are from a nearby planet, for example. If the beliefs about the passengers are unwarranted (as they would be if their only warrant came from my holding them), then a theory that unifies and accommodates them is (without more) unwarranted as well.

The Argument from Moral Experience

115

But how, without lapsing into complete skepticism, can those employing the method of WRE avoid holding that sooner or later coherence with mere beliefs is warrant conferring? Is there a way to look at the WRE strategy that does not have conservatism’s implausible implications? Perhaps. Our initial beliefs represent the beginning of our search for WRE, not the end point. Often, these initial beliefs will have to be modified or rejected if the entire corpus of our beliefs is to be made coherent. Thus arguably we should treat only those beliefs that are able to survive significant reflection as having some degree of warrant, and not those that merely appeared plausible before we began to think things through more carefully. If this is correct, then acceptance of the WRE framework would not automatically allow us to treat an initially appealing theoretical gloss on our moral beliefs as warranted. The claim that it is warranted would depend, on this approach, on whether it could survive the minimally necessary process of reflection. How then should we envision this process of reflection? One answer, similar to a proposal gestured at above, begins with the claim that we have justified beliefs about the reliability of the processes that led to our beliefs. These beliefs about reliability might be thought to have a bearing on the question of how much credibility the initial beliefs should be given. For example, if a person’s commitment to the existence of angels stems entirely from dogmatically accepted religious beliefs he now sees reason to abandon, he may also have reason to avoid giving his belief in angels any weight in WRE. In contrast, arguably we should see our perceptual beliefs as warranted when they are made under normal circumstances, since we are justified in believing that perception is normally a reliable belief producer. If this is correct, then the WRE strategy cannot yet support objectivism, since it would be questionbegging to assume that our beliefs about morality (and our moral beliefs themselves) are the products of similarly reliable mechanisms. In any case, this approach to WRE seems problematic since it is not immediately clear how our beliefs about reliability themselves get justified. An alternative might be to look for justification in the way our beliefs fit into our larger explanatory picture. Thus, we might treat as warranted only those beliefs whose existence can plausibly be explained in a way that supports the claim that they are true. Finding a plausible explanatory theory is often thought to be the point of WRE. On such a view, the coherence of a set of beliefs is constituted in part by the beliefs’ fit into such an explanatory picture.27 How would the beliefs we have been discussing fare under this proposal? My belief that the next Lotto number drawn will be even cannot plausibly be thought to be best explained on the hypothesis that the number will in fact be even, nor can it be supported by any other aspect of a plausible explanation for my coming to have it. 27 On

most views of this sort, the justification would not depend upon anyone’s having actually gone through such a process, but on the availability of an explanatory picture of the kind gestured at here. Thus we might not currently be in a position to know (or be justified in believing) that certain of our beliefs are justified, even if they are. Still, if we are not in such a position, then we should not presume that those beliefs are justified until our explanatory theory improves.

116

D. Loeb

The same is true of my beliefs about my fellow bus passengers, and the theory that unifies and accommodates these beliefs. What about the everyday beliefs pointed to by those employing the AME? These would give objectivism some measure of warrant only if the best explanation for our holding them makes it more probable that objectivism is true. But, once again, whether such an explanation is available is merely a special case of the question of whether moral facts figure in the best explanation of anything, a question not appropriately considered in this context. Presumably, other ways of looking at the WRE strategy are available, and perhaps some of them can also avoid the implausible implications of conservatism, either at the level of beliefs or at the level of theories that purport to accommodate those beliefs. But however we interpret it, the WRE strategy should not treat our beliefs – and a fortiori should not treat controversial theories that attempt to explain those beliefs – as having warrant simply because the beliefs are held. Instead, the strategy should hold that the reflective process itself (or at least, the availability of such a process) is what confers warrant on some, but only some, of those initial beliefs. Whether it does so in the case under consideration here is a question that has not yet been answered.

4 Conclusion The AME has exerted an enormous influence on the debate between objectivists and anti-objectivists. Yet its underpinnings have never been carefully examined. If I am right, objectivists have unfairly attempted to shift the burden of proof to antiobjectivists, and anti-objectivists like Mackie have been all too quick to accept this burden. Unless someone does a better job of defending the AME, we should treat objectivism and anti-objectivism as, at a minimum, starting on an even footing. Acknowledgments For their comments, encouragement, criticism, patience, and generosity, I am grateful to Paul Bloomfield, Sin yee Chan, David Christensen, Stephen Darwall, Tyler Doggett, Richard Joyce, Simon Kirchin, Hilary Kornblith, Arthur Kuflik, William Mann, Mark Moyer, Derk Pereboom, Seth Shabo, and Walter Sinnott-Armstrong. I wish to thank the Dean’s Office, College of Arts and Sciences, The University of Vermont for providing modest grant support for an initial draft of this chapter. I am also grateful to the attendees at a colloquium at Brandeis University, a number of undergraduate students at Vermont (forced to read various drafts), and several anonymous referees (likewise). As always, my wife, Barbara Rachelson, suffered through many, many drafts and conversations about this chapter and was a tremendous help to me. I started work on this chapter so long ago that I am sure I have forgotten many others to whom thanks are due. You know who you are. Thanks.

References Blackburn, S. 1993. Errors and the phenomenology of value. In his Essays in quasi-realism, 149– 165. New York: Oxford University Press. Bloomfield, P. 2001. Moral reality. Oxford University Press: New York. Brink, D. 1989. Moral realism and the foundations of ethics. Cambridge: Cambridge University Press.

The Argument from Moral Experience

117

Chisholm, R. 1977. Theory of knowledge. 2nd ed. Englewood Cliffs: Prentice-Hall. Christensen, D. 1994. Conservatism in epistemology. Noûs 28: 69–89. Dancy, J. 1986. Two conceptions of moral realism. Proceedings of the Aristotelian Society 60(suppl.): 167–187. Dworkin, R. 1996. Objectivity and truth: You’d better believe it. Philosophy and Public Affairs 25: 87–139. Garner, R. 1994. Beyond morality. Philadelphia: Temple University Press. Gibbard, A. 1990. Wise choices, apt feelings. Cambridge, MA: Harvard University Press. Greene, J. 2002. The terrible, horrible, no good, very bad truth about morality and what to do about it. Dissertation, Princeton University. Hare, R. M. 1981. Moral thinking. Oxford: Clarendon Press. Harman, G. 1977. The nature of morality. Oxford: Oxford University Press. Harman, G. 1986. Moral explanations of natural facts: Can moral claims be tested against moral reality? Southern Journal of Philososphy 24(suppl.): 57–68. Hookway, C. 1986. Two conceptions of moral realism. Proceedings of the Aristotelian Society 60(suppl.): 167–187. Jackson, F. 1998. From metaphysics to ethics: A defence of conceptual analysis. Oxford: Oxford University Press. Joyce, R. 2001. The myth of morality. Cambridge: Cambridge University Press. Joyce, R. 2006. The evolution of morality. Cambridge MA: MIT Press. Korsgaard, C. 1996. The sources of normativity. Cambridge: Cambridge University Press. Lillehammer, H. 2004. Moral error theory. Proceedings of the Aristotelian Society 104: 95–111. Loeb, D. 2003. Gastronomic realism: A cautionary tale. Journal of Theoretical Philosophical Psychology 23: 30–49. Loeb, D. 2005. Moral explanations of moral beliefs. Philosophy and Phenomenological Research 70: 193–208. Loeb, D. 2008. Moral incoherentism: How to pull a metaphysical rabbit out of a semantic hat. In Moral psychology, vol. 2: The cognitive science of morality, ed. W. Sinnott-Armstrong. Cambridge, MA: MIT Press. Lovibond, S. 1983. Realism and imagination in ethics. Minneapolis: University of Minnesota Press. Mackie, J. L. 1977. Ethics: Inventing right and wrong. New York: Penguin. McNaughton, D. 1988. Moral vision. Oxford: Basil Blackwell. Nagel, T. 1986. The view from nowhere. New York: Oxford University Press. Nagel, T. 1997. The last word. New York: Oxford University Press. Nichols, S. 2004. Sentimental rules: On the natural foundations of moral judgment. New York: Oxford University Press. Price, R. 1758. A review of the principal questions in morals, ed. D. D. Raphael. 1948. Oxford: Clarendon Press. Quine, W. V. 1951. Two dogmas of empiricism. Philosophical Review 60: 20–43. Reid, T. 1785. Essays on the intellectual powers of man. In Inquiry and essays, eds. Ronald B. and Keith L. 1975. Indianapolis: Bobbs-Merrill. Schiffer, S. 1990. Meaning and value. Journal of Philosophy 87: 602–614. Shafer-Landau, R. 2003. Moral realism: A defence. Oxford: Oxford University Press. Smith, M. 1993. Objectivity and moral realism: On the significance of the phenomenology of moral experience. In Reality, representation, and projection, eds. J. Haldane and C. Wright, 235–255. New York: Oxford University Press. Smith, M. 1994. The moral problem. Oxford: Blackwell. Stevenson, C. L. 1950. The emotive conception of ethics and its cognitive implications. Philosophical Review 59: 291–304. Sturgeon, N. 1984. Moral explanations. In Morality, reason and truth, eds. D. Copp and D. Zimmerman, 49–78. Totowa: Rowman and Allanheld. Swinburne, R. 1979. The existence of God. Oxford: Clarendon Press.

118

D. Loeb

Timmons, M. 1999. Morality without foundations. New York: Oxford University Press. Wiggins, D. 1988. Truth, invention, and the meaning of life. In Essays on moral realism, ed. G. Sayre-McCord, 127–165. Ithaca: Cornell University Press. Wright, C. 1992. Truth and objectivity. Cambridge, MA: Harvard University Press.

Beyond the Error Theory Michael Smith

In Ethics: Inventing Right and Wrong, published just over 30 years ago, John Mackie famously argued for the error theory (Mackie 1977). Though the argument initially met with considerable skepticism (see, e.g., Blackburn 1993; McDowell 1985), in the years that followed many theorists came to the conclusion that Mackie had things more or less right (see e.g., Lewis 1989; Garner 1994; Joyce 2001). But which of these views is correct? Should we all be error theorists? Or is the renewed admiration for Mackie’s argument itself mistaken? To anticipate, my somewhat tentative suggestion will be that Mackie’s argument fails. Since I have been tempted to believe the error theory myself in the past (Smith 2006), this represents something of a shift in my own thinking.

1 An Outline of Mackie’s Argument Mackie’s argument for the error theory proceeds in two stages. He begins by pointing out that our concept of a moral value is the concept of a feature of things that is at once both objective and prescriptive. But, he then goes on to argue, general reflections of a metaphysical and epistemological kind show that nothing has such a feature: The concept of an objective and prescriptive feature isn’t instantiated. As we will see, the reasons he offers make it seem that the conclusion would have to be necessary, so the upshot, if the argument works, is not just that nothing has moral value, but that nothing could have such value. As even this briefest of outlines makes clear, the real power of Mackie’s argument is that it is addressed to people who are antecedently engaged in ordinary moral thought and talk in blissful ignorance of the error that it is the aim of his argument to lay bare. His strategy is to get those people to agree first to the conceptual claim – this fixes what it is that they are thinking and talking about – and then to the substantive reasons he offers for supposing that there could be nothing like that. M. Smith (B) Department of Philosophy, Princeton University, Princeton, NJ, USA

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_8, 

119

120

M. Smith

Mackie’s argument is thus an internal critique of morality: It purports to show that someone who is committed to morality shouldn’t be so committed, and it purports to show this on terms that that person can himself recognize. Let’s consider the two stages of the argument in more detail. Mackie’s defense of the conceptual claim consists in a rehearsal of what he sees as “the main tradition of European moral philosophy from Plato onwards,” a tradition which, he tells us, . . . has combined the view that moral values are objective with the recognition that moral judgments are partly prescriptive or directive or action-guiding. Values themselves have been seen as at once prescriptive and objective. In Plato’s theory the Forms, and in particular the Form of the Good, are eternal, extra-mental, realities. They are a very central structural element in the fabric of the world. But it is held also that just knowing them or “seeing” them will not merely tell men what to do but will ensure that they do it, overruling any contrary inclinations. . . . Similarly, Kant believes that pure reason can itself be practical, though he does not pretend to be able to explain how it can be. Again, Sidgwick argues that if there is to be a science of ethics—and he assumes that there can be, indeed he defines ethics as “the science of conduct”—what ought to be “must in another sense have objective existence: it must be an object of knowledge and as such the same for all minds”; but he says that the affirmations of this science “are also precepts,” and he speaks of happiness itself as “an end absolutely prescribed by reason.” (Mackie 1977, pp. 23–24)

Though Mackie seems to think that these diverse formulations are all in some sense equivalent, in what follows I will focus on just one of them, namely, Sidgwick’s. In Sidgwick’s terms, Mackie’s claim that our concept of moral value is the concept of an objectively prescriptive feature of things amounts to the claim that to conceive of (say) happiness as a moral value is to conceive of happiness itself as having the feature of being an end that is absolutely prescribed by reason. Mackie’s argument from this conceptual claim about moral value to the conclusion that there is no such value is brevity itself. He dubs it “the argument from queerness” and he tells us that it

. . . has two parts, one metaphysical, the other epistemological. If there were objective values, then they would be entities or qualities or relations of a very strange sort, utterly different from anything else in the universe. Correspondingly, if we were aware of them, it would have to be by some special faculty of moral perception or intuition, utterly different from our ordinary ways of knowing everything else. These points were recognized by Moore, who spoke of non-natural qualities, and by the intuitionists in their talk about a “faculty of moral intuition.” Intuitionism has long been out of favour, and it is indeed easy to point out its implausibilities. What is not so often stressed, but is more important, is that the central thesis of intuitionism is one to which any objectivist view of values is in the end committed: Intuitionism merely makes unpalatably plain what other forms of objectivism wrap up. Of course the suggestion that moral judgements are made or moral problems solved by just sitting down and having an ethical intuition is a travesty of actual moral thinking. But, however complex the real process, it will require (if it is to yield authoritatively prescriptive conclusions) some input of this distinctive sort, either premises or forms of argument or both. (Mackie 1977, p. 38)

Beyond the Error Theory

121

Though Mackie goes on to consider various replies to this argument, he doesn’t really say anything more to spell it out beyond what he says in this passage. So what exactly is the argument? Let’s begin by restating Mackie’s argument in Sidgwickian terms. Though we believe that happiness has the feature of being an end that is absolutely prescribed by reason, Mackie seems to be saying, the idea that happiness is such an end is both metaphysically and epistemologically queer. Why is it metaphysically queer? Mackie doesn’t explicitly say, but we can imagine what he is thinking. Ends are the sorts of things that each of us has, in so far as we aim at, or desire, different things. But while it is therefore true that some of us have happiness as an end, as some of us do desire happiness, the claim that happiness has the feature of being an end absolutely prescribed by reason, “the same for all minds,” would have to be made true by some further fact about happiness, a fact beyond this purely descriptive psychological fact. Yet what further fact is there? At this point, Mackie seems to just throw up his hands. He cannot see what further fact there could be. Or, more accurately, he cannot think of anything beyond its being a brute further fact, a Moorean non-natural fact the like of which he can make no sense (Moore 1903). This seems to be the metaphysical queerness he has in mind. Moreover, since he can make no sense of what kind of fact would make true the claim that happiness is an end absolutely prescribed by reason, he cannot think of any way in which we might come to know about such a further fact either. Or, more accurately, he cannot think of anything beyond our possessing a special faculty which enables us to detect non-natural features, a faculty the like of which he can make no sense either. This seems to be the epistemological queerness that he has in mind. I said earlier that the reasons Mackie gives for thinking that there are no moral values would, if they were successful, make the conclusion necessary. We can now see why this is so. The problem with Moorean non-natural qualities isn’t that there aren’t any such things as a matter of contingent fact. The problem is that we can literally make no sense of them: There are no possible worlds in which objects have such qualities (though see Shafer-Landau 2003). Since Mackie appears to think that the existence of objectively prescriptive features requires the existence of such non-natural qualities, it follows, if he is right, that there could be no objectively prescriptive features either. To sum up: Mackie’s argument from queerness consists in an analysis of the concept of moral value together with a pair of challenges which purport to show that we can make no real sense of how that concept could be instantiated. We are therefore left with the conclusion that there is nothing beyond facts like the purely descriptive psychological fact that some of us do indeed have happiness as an end. Suppose we were antecedently committed to morality. Mackie’s argument purports to show that we shouldn’t be. Moreover, it purports to show this on terms that we can ourselves recognize. If we are to respond to his argument then we must therefore either take issue with his analysis of value or else provide answers to the metaphysical and epistemological challenges.

122

M. Smith

2 Value or Obligation? Let’s begin with an initial worry about the analysis. Mackie’s conceptual claim, which focuses on the concept of moral value, is framed in rationalist terms. He tells us that to say that happiness is a moral value is to say that happiness has the feature of being an end absolutely prescribed by reason. But insofar as they go in for conceptual analysis at all, the concept most rationalists seek to analyze is not the concept of moral value, but rather the concept of moral obligation. Do things look any better if we restate Mackie’s argument in terms of the concept of moral obligation, rather than the concept of moral value? It might initially be thought that they do. Rationalists typically tell us that the concept of a moral obligation is the concept of a certain sort of reason for action. Thus, for example, if we assume that act utilitarianism is the correct theory of what we ought to do – and from here on, in deference to the earlier decision to focus on Sidgwick’s analysis of value, I will assume this for ease of exposition – then many rationalists claim it follows from this that there is a reason, perhaps conclusive perhaps non-conclusive (for there are stronger and weaker forms of rationalism), for each of us to maximize happiness and minimize suffering. One question to ask is whether rationalists mightn’t just stop at this point. Or is there meant to be something metaphysically and epistemologically queer about the concept of a reason for action as such? Do reasons for action presuppose the existence of Moorean non-natural qualities? (Relatedly, we might ask whether there is meant to be something especially queer about moral obligation, or moral reasons, or whether the queerness is supposed to attach to the idea of there being anything at all that we ought to do, or any reasons for action at all.) The answer is that Mackie’s argument from queerness, if it works at all, establishes that the concept of a reason for action invoked by the rationalists is metaphysically and epistemologically queer as well. There are various ways to bring this out, but for present purposes the following observation should suffice. If we add to rationalism – understood as a claim about the link between moral obligations and reasons for action – the following “Williams-Korsgaard” thesis about the nature of reasons for action (Williams 1981; Korsgaard 1986): WK: If an agent has a reason to φ then she would want herself to φ if she engaged in a suitable course of deliberation,

then it follows that claims about what we morally ought to do entail claims about what we would desire ourselves to do after a suitable course of deliberation. But since a suitable course of deliberation is simply one in which the deliberator is maximally informed and then forms his desires on the basis of that information in accordance with the requirements of rationality – this is something about which both Bernard Williams (a Humean) and Christine Korsgaard (a Kantian) agree – it further follows that claims about what we morally ought to do entail claims about what we would want ourselves to do after forming our desires in the light of full information and the requirements of rationality. (See also Smith 1994.) Mackie’s

Beyond the Error Theory

123

two challenges thus re-emerge, but this time about the possibility that our concept of a moral obligation is instantiated. Why do Mackie’s two challenges re-emerge? They re-emerge because, if he is right that there is something metaphysically and epistemologically queer about the idea that happiness is an end absolutely prescribed by reason, then there is something equally metaphysically and epistemologically queer about supposing that desiring to maximize happiness and minimize suffering is absolutely prescribed by reason. This, after all, is what we must be supposing if we think that we would each desire ourselves to maximize happiness and minimize suffering if we formed our desires in the light of full information and the requirements of rationality. If the former presupposes the existence of Moorean non-natural qualities, then so must the latter. Given WK, the argument from queerness is thus equally an objection to the rationalist’s traditional analysis of the concept of moral obligation in terms of the concept of a reason for action. (Indeed, this suggests that the argument from queerness equally calls into question not just the existence of moral obligations, but there being anything we ought to do and the existence of any reasons for action.) Nor, for the record, should this be in the least surprising. For a natural way to understand what it is to have a reason for action is in terms of the value produced by the action that we have reason to perform. As Davidson puts it, an action we have reason to perform must have certain “desirability characteristics” (Davidson 1963). The WK condition that must be satisfied by our moral obligations, if our moral obligations give rise to reasons for action, is thus much the same as the condition that must be satisfied by the states of affairs brought about by our actions if those states of affairs have moral value, given the Sidgwickian account of what it is for something to have moral value. In each case what’s required is that there is some desire, or end, absolutely prescribed by reason. In the case of moral value, the desire in question concerns a state of affairs. In the case of a reason for action, the desire in question concerns the action of bringing that state of affairs about. There’s no surprise here given that it follows from the fact that we have a reason to do something that our doing that thing will bring about something of value. This detour is, however, helpful, as it enables us to focus more clearly on how we might respond to Mackie’s two challenges. First of all, remembering once again our simplifying assumption about what we are morally obliged to do, we must explain either what the world would have to be like or what the requirements of rationality would have to be like for it to be the case that we would each desire ourselves to maximize happiness and minimize suffering if we formed our desires in the light of full information and the requirements of rationality. Second, we must show that it is plausible to suppose that the facts about the world, or the requirements of rationality, are like that. And third, on the assumption that we can know what’s of moral value or what we have reason to do, we must make explicit how it is that we are able to acquire knowledge of the relevant facts about the world and requirements of rationality. In what follows I will describe and assess what I take to be the four main approaches to responding to Mackie’s two challenges. The discussion will of necessity be incomplete, as each approach can be developed in different ways, some of

124

M. Smith

which are more promising than others. My main aim, however, is not to say the last word about any of these approaches, but rather to put all four on the table for further discussion. My own view, to anticipate, is that only the fourth has any chance of showing what’s wrong with Mackie’s argument. Though I am I tempted by the fourth approach, I must confess to a sense of unease about it.

3 The Instrumental Approach According to what I will call the “Instrumental Approach,” only one requirement of rationality governing the formation of desires is needed to underwrite the truth of the claim that we would each desire ourselves to maximize happiness and minimize suffering if we formed our desires in the light of full information and the requirements of rationality. The requirement is some variation on the following, fairly uncontroversial, requirement of means-ends rationality (Note that hereafter “RR” is short for “Reason requires that”): ME: RR (if a subject has an intrinsic desire that p and a belief that he can bring about p by bringing about q, then he has an instrumental desire that he brings about q).

This is because all of the work is done by a crucial empirical fact, one that we would each come to appreciate if we had full information, about the means by which we will get what we intrinsically desire, no matter what we intrinsically desire. (For a related argument see Gauthier 1986.) Before we get to that, however, let’s focus for a moment on ME itself. If the Instrumental Approach is to succeed, then we will have to explain what sort of fact ME is and we will also have to explain how it is possible for us to come by knowledge of this fact, in so far as we have knowledge of moral values and reasons for action. My own view is that ME is best understood in much the same way as we understand claims like “It ought to be the case that knives cut well.” This “ought”claim derives from the metaphysically mundane fact that knife is a functional kind. Since what a knife is is something whose function is to cut, it follows that knives can be ordered according to how well they serve that function: Knives that cut serve that function better than knives that don’t; knives that cut more efficiently serve that function better than knives that cut less efficiently; and so on. As I see things, the claim that it ought to be the case that knives cut well is simply an efficient way of saying that this raft of evaluative claims is true. No Moorean non-natural qualities are thus required to underwrite the truth of this claim. Similarly, it seems to me that ME derives from the metaphysically mundane fact that intrinsic desires and means-end beliefs are psychological states possessed by agents, where the psychology of an agent is also a functional kind. The psychology of an agent is something whose function is, inter alia, to produce action. Psychologies too can thus be ordered according to how well they serve their function: Those in which intrinsic desires combine with means-end beliefs in the way required to produce action – for this is what happens when they combine in such a way as to produce an instrumental desire (indeed, in my view, instrumental desires just are intrinsic desires that have suitably combined with means-end beliefs (Smith

Beyond the Error Theory

125

2004)) – better serve the function of the psychology of an agent than do those in which intrinsic desires and means-end beliefs do not combine in this way; psychologies in which intrinsic desires and means-end beliefs reliably combine in this way serve that function better than those in which they combine in this way albeit unreliably; and so on. ME, which is just an “ought”-claim about the relationship between desires and means-end beliefs, is simply an efficient way of saying that this raft of evaluative claims is true. Understood in this way, ME states a fact that is as metaphysically mundane as the claim that knives ought to cut well, and our knowledge of ME, much as our knowledge of the claim that knives ought to cut well, is mundane knowledge too. Just as the “ought”-claim about knives implied nothing about Moorean non-natural qualities, neither does ME. There is, however, one crucial difference between claims about the function of a knife and claims about the function of the psychology of an agent. Since knives are a human invention, the function of a knife can be traced in some loose way to human purposes. The psychology of an agent, by contrast, is not a human invention, so its function cannot be traced to human purposes. So what does fix the function of the psychology of an agent? Human psychologies are, of course, the causal product of a process of evolution, so it might be thought that the function of the psychology of an agent is fixed by the contingencies of that process. But the fact that a human psychology is the product of a process of evolution is not an essential feature of a human psychology as a psychology. ME purports to tell us something about the proper functioning of every possible psychology of an agent, including those that spring into existence willy-nilly, not just something about the contingencies of an evolved human psychology. Indeed, a human psychology might evolve in ways that make it flout ME. So, to repeat, what does fix the function of the psychology of an agent? My own view is that we must think of the psychology of an agent as a privileged kind in reality – a Lewisian elite property, if you like (Lewis 1984) – and that we must suppose that knowledge of the function of the psychology of an agent is purely speculative knowledge of a thing of that kind. We gain knowledge of the function of the psychology of an agent a priori by reflecting on the nature of psychology and agency, much as we gain knowledge of causation, persistence, freedom, and the like, by reflecting on their natures. At this point, my suggestion is, when we reflect on the nature of the psychology of an agent we learn that that is something that has a function captured, inter alia, by ME. A crucial question that will need to be addressed, in attempting to respond to Mackie’s argument, is whether we need to have a richer account of what it is for the psychology of an agent to function properly or whether this is all that needs to be said. With this account of the metaphysics and epistemology of ME in place, the Instrumental Approach holds that what’s been said is all that needs to be said. It holds that we need to appeal to just one further fact in order to explain why we would each desire ourselves to maximize happiness and minimize suffering – again, remember our simplifying assumption – if we formed our desires in the light of full information and the requirements of rationality. The further fact in question is a crucial empirical fact: namely, that acting so as to maximize happiness and minimize

126

M. Smith

suffering is an all-purpose means to the satisfaction of whatever desires agents happen to have. This is what full information would reveal to us, or so the Instrumental Approach tells us. Morality is an all-purpose means to our ends. In that case, no matter what people desire, so long as they have full information, and hence true beliefs about means, and so long as they go on to form their instrumental desires in accordance with the requirements of means-ends rationality in the light of these true beliefs, they will acquire an instrumental desire to maximize happiness and minimize suffering. What should we make of this response to Mackie’s argument? The Instrumental Approach certainly succeeds in showing that there is a lacuna in Mackie’s argument. Whereas Mackie says that the existence of desires that are absolutely prescribed by reason would require the existence of Moorean non-natural qualities, the Instrumental Approach shows that no such qualities are required. All that’s required is ME and a crucial empirical fact. Unfortunately, however, the alleged empirical fact – the fact, given our simplifying assumption, that maximizing happiness and minimizing suffering is an all-purpose means to the satisfaction of whatever desires anyone happens to have – seems to be no empirical fact at all. And nor would it help if we were to eschew the simplifying assumption and make different assumptions about what we are obliged to do, and hence different assumptions about what the all-purpose means to our ends are. For whatever we are in fact obliged to do, it seems not to be an empirical fact that our doing that is an all-purpose means to the satisfaction of whatever desires we happen to have. Some may balk at this. Couldn’t God see to it that doing what we are obliged to do is an all-purpose means to the satisfaction of our desires, whatever desires we happen to have? Couldn’t he institute a set of rewards for doing what we are obliged to do and punishments for our failing to do what we are obliged to do, where these rewards and punishments are in turn a matter of our getting whatever it is that we happen to want or be averse to? If so, and if God exists and sets up such a system of rewards and punishments, then the Instrumental Approach shows that he would thereby have seen to it that there exist features that are both objective and prescriptive. I said earlier that Mackie’s argument purports to show that there could only exist objectively prescriptive features if there were Moorean non-natural features, something we can literally make no sense of. But, it might be thought, we now see that his argument shows no such thing. For the existence of moral value would be equally secured by the existence of God, on the assumption that God can indeed set up a system of rewards and punishments as described. However I am not persuaded that this last crucial assumption is plausible. Suppose (for reductio) that God exists and sets up the required system of rewards and punishments in some possible world. So far, so good. Now let’s ask a question about that possible world. Had there been someone whose intrinsic desires are not satisfied by his doing what’s morally obligatory, would he have had moral obligations? This is a legitimate question to ask, because moral obligations, if they exist at all, are possessed not just by actual people, but also by those who would have existed had things been otherwise. And the answer is surely that he would still have

Beyond the Error Theory

127

had moral obligations. Imagine, for example, someone whose every intrinsic desire is for some X where X does not come about by means of whatever it is that we are imagining it is morally obligatory for him to do. That person, by hypothesis, has moral obligations. But, given our analysis of moral obligation, he too would have to be rationally required to desire to do whatever it is that we are imagining he has a moral obligation to do. Yet if ME is the only requirement of rationality governing the formation of desires, we have imagined his having desires that make this condition impossible to satisfy. If this is right then the upshot is that not even God can help the Instrumental Approach explain the existence of moral obligations. Let me sum up. If we take the Instrumental Approach, then, though we don’t commit ourselves to the existence of Moorean non-natural qualities, we do land ourselves with the error theory nonetheless. We land ourselves with the error theory because it is incoherent to suppose that our doing what we are morally obliged to do is an all-purpose means to the satisfaction of our desires whatever we happen to desire: That’s what the possibility of someone with desires like those just described shows. But the failure of the Instrumental Approach teaches us an important lesson. Imagine that everyone who is obliged to maximize happiness and minimize suffering would indeed desire themselves to so act if they formed their desires in the light of full information and the requirements of rationality. In that case there would have to be something irrational about someone whose every intrinsic desire is for some X where X doesn’t come about by means of whatever it is that we are imagining we have an obligation to do. We thus have no alternative but to suppose that such intrinsic desires themselves are subject to rational requirements. Thus, the question to which we require an answer, given our simplifying assumption, is whether we mightn’t be rationally required to intrinsically desire that we maximize happiness and minimize suffering. The remaining approaches all proceed by attempting to answer this question in the affirmative without presupposing the existence of Moorean non-natural qualities.

4 The Universalization Approach If our intrinsic desires themselves are subject to rational requirements, then there must be rational requirements beyond ME. But what might such requirements of rationality be like? One possibility is that, in order to be rational, our intrinsic desires must meet certain formal constraints, formal constraints that weed out all but the desire to maximize happiness and minimize suffering and those intrinsic desires whose satisfaction is consistent with the satisfaction of such a desire. The most obvious such constraint is some variation on the following requirement of universalization: U: RR (if a subject has an intrinsic desire that p, then either p itself is suitably universal or the satisfaction of the desire that p is consistent with the satisfaction of desires whose contents are themselves suitably universal).

128

M. Smith

The rationale for U is not hard to provide. (For a related argument see Kant 1786/1948.) If there are any norms of rationality governing desires at all, then, since no particulars have a privileged status in the rational order of things, it must be possible to state the norms in purely universal terms. There is an obvious analogy here with laws of nature. Because laws of nature assign no particulars a privileged role in the causal order, it follows that it must be possible to state them too in purely universal terms. According to U, this means that, to be rational, our intrinsic desires must have contents that are themselves suitably universal – they must mention no particulars – or, at any rate, their satisfaction must be consistent with the satisfaction of desires whose contents are themselves suitably universal. (From here on I will omit this qualification.) According to the Universalization Approach, however, there is a further argument that takes us from U to the conclusion that the only desires concerning particulars that are rational are those either derived from or consistent with the desire to maximize happiness and minimize suffering. But what exactly the further argument is supposed to be is very much a moot point. The best known argument for something like this conclusion is that given by R.M. Hare (1981). (Having said that, however, it must immediately be added that Hare himself didn’t accept U, which purports to state a requirement of rationality. According to Hare, though universalization is a condition of morality, it is not a condition of rationality. Hare therefore rejects the Sidgwickian analysis of value in terms of certain desires being absolutely prescribed by reason; this is where he parts company with Mackie. Let’s, however, leave Hare’s reasons for rejecting the Sidgwickian analysis of value, an analysis which Mackie accepts, to one side.) Hare did, however, argue that the only intrinsic desire that is suitably universal is a desire much like the intrinsic desire that there is as much happiness as possible. According to Hare, the only intrinsic desire that passes the universalization test is the desire that there is as much desire satisfaction as possible. But now look at what happens if we put U together with Hare’s views about the power of universalization arguments. We get the conclusion that, if we have any intrinsic desires at all, then, if we formed our intrinsic desires in the light of full information and the requirements of rationality, we would all desire that we maximize desire satisfaction. It might therefore be thought that the Universalization Approach provides us with a response to Mackie’s two challenges all by itself. U is hardly metaphysically queer, after all. Much like ME, it is a principle that tells us what must be the case for the psychology of an agent to function properly. According to U, the psychology of an agent functions properly only if the desires that issue in action are themselves suitably universal. No Moorean non-natural qualities there. Nor is there any epistemological queerness either – not, at any rate, if Hare’s arguments succeed in showing that the only intrinsic desire that passes the universalization test is the desire that there be as much desire satisfaction as possible. As I said, however, what the argument is supposed to be that establishes this conclusion is very much a moot point. Let me briefly explain why. Hare thinks, plausibly enough, that U would require us, in whatever situation we happen to find ourselves, to find something that we want to obtain in every possible situation identical in universal respects to this situation. Let’s suppose that we find

Beyond the Error Theory

129

ourselves in a situation, which we will call “S,” in which there are three people interacting in a certain way. In S: we desire that p, a second person desires that q, and a third person desires that r. In figuring out what we can want to obtain in every possible situation identical in universal respects to S, Hare suggests that we need to put together three distinct desires concerning three distinct ways things could be that are none the less identical in universal respects to S. The first is our desire that p and it concerns S itself. The second is the desire that we would acquire if we were fully to imagine ourselves in the possible situation identical in universal respects to S, but in which we ourselves occupy the position of the second person. According to Hare, this is the desire that, in the possible situation in which we occupy the second person’s position, q. And the third is the desire that we would acquire if we were to fully imagine ourselves in the possible situation identical in universal respects to S, but in which we occupy the position of the third person. According to Hare, this is the desire that, in the possible situation in which we occupy that position, r. The upshot, according to Hare, is that we have three conflicting desires concerning three possible situations all of which are identical to S in purely universal respects. In forming a desire for one thing to obtain in all of these situations, he thinks that there is therefore only one rational course, and that is to desire whatever will maximally satisfy our three conflicting desires concerning these situations. This in turn, he thinks, is equivalent to our desiring to maximally satisfy the desires of all three parties involved in S. There are many things to say about this argument, but for present purposes it will suffice to focus on just one. To repeat, Hare thinks that if we were fully to imagine ourselves occupying the position of (say) the second person, who desires that q, then we would acquire the desire that, in the possible situation in which we occupy that position, q. There are two ways to understand what he has in mind. One is that there is a necessary connection between belief and desire: When we form the belief that there is some possible situation in which we desire that q then that entails that we also form the desire that, in that possible situation, q. But since it is plainly false that there is such a necessary connection between belief and desire – it is at least possible for someone to believe that there is a possible situation in which she desires that p while being indifferent or averse to p’s being the case in that situation – the argument, so interpreted, isn’t very plausible. The other possibility is that Hare is positing a normative connection between belief and desire. He might be thinking that reason requires us, when we believe that there is some possible situation in which we desire that q, to desire that, in that possible situation, q. We may of course have the belief but lack the desire, but only at the cost of flouting the normative requirement. The trouble with this interpretation of Hare’s argument, however, is that it posits a rational principle independent of U. Whereas U tells us that there is a normative constraint on the form of our desires – our desires must have contents that are suitably universal – this rational principle tells us that the contents of our beliefs put normative constraints on the contents our desires. Understood in this way, Hare’s argument isn’t a version of the Universalization Approach at all. It is a version of a quite different approach, the one that I will consider next.

130

M. Smith

This brief consideration of the Universalization Approach hardly establishes that every variation on that approach is flawed. Perhaps a more plausible version can be found in Kant’s own much more extensively discussed version of the Universalization Approach (Kant 1786/1948), though the ink spilled explaining why Kant’s various arguments don’t work strongly suggest otherwise (see, most recently, Parfit forthcoming). My own view, however, is that this is unlikely to be so. When you look closely at them, all versions of the Universalization Approach seem to share the crucial feature of Hare’s just identified. Though it may be plausible to suppose that universalization is a condition of rationality, all versions of the Universalization Approach appeal to something beyond mere universalization in order to establish that some particular intrinsic desire is rationally required. Mackie’s two challenges thus remain. In these terms, his challenge is to dispel the sense that there is something metaphysically and epistemologically queer about the further thing to which appeal is made in such arguments. Perhaps universalization itself doesn’t presuppose the existence of Moorean non-natural qualities, but the further thing to which appeal is made does.

5 The Reasons Approach In What We Owe To Each Other, Thomas Scanlon explicitly rejects the idea that goodness is a Moorean non-natural quality. He claims that what it is for something to be of value is for there to be a reason to want it, or to appreciate it, or to have some other attitudinal response towards it, where the different attitudinal responses are markers of different kinds of value (Scanlon 1998). This is his well-known “buckpassing” account of value. On the plausible assumption that the existence of a reason to want something entails that wanting that thing is absolutely prescribed by reason, his account entails Sidgwick’s. Derek Parfit concurs and elaborates on the nature of the reasons that we have for wanting things when the value in question is intrinsic value (Parfit forthcoming). Parfit says that what it is for something to be intrinsically good is for that thing to have intrinsic features that provide us with reasons to want intrinsically that those very features be realized. This suggests the following alternative account of the principles of rationality governing our intrinsic desires. Remembering once again our simplifying assumption, the idea would have to be that the intrinsic nature of the states of affairs in which there is as much happiness as possible provides us with a reason to intrinsically desire that that state of affairs obtains, and hence a reason to desire that we bring that state of affairs about. If this claim about the reasons that exist for intrinsically desiring is correct, then that in turn suggests that our intrinsic desires are subject to the following rational principle: BD: RR (if a subject believes that a state of affairs has the intrinsic nature of that state of affairs in which there is as much happiness as possible, then he intrinsically desires that that state of affairs obtains).

Beyond the Error Theory

131

The idea, in other words, is that subjects are rationally required to form the beliefs and desires that there are reasons for forming when they believe that those reasons obtain; they are, in other words, rationally required to be sensitive to such reasons as they can appreciate. Let’s call this the “Reasons Approach.” How plausible is it to suppose that there are rational principles of the kind that the Reasons Approach posits? In answering this question we must remember that the question isn’t whether BD itself is plausible. The question is the more general one whether rational principles with BD’s form are plausible. In other words, the question is whether it is plausible to suppose that, for some p and some q, there is a basic rational principle of the form: RR (if a subject believes that p then she has an intrinsic desire that q).

Moreover, it is important to emphasize that the question concerns basic rational principles of this form because, on plausible assumptions, we can derive at least one principle with the same form as BD from a mere commitment to the existence of rational principles governing desires. This would not, however, be a vindication of the Reasons Approach. Let me explain why. Assuming that desires, like beliefs, are indeed subject to rational requirements, it follows that, no matter what form the rational requirements on desires take – whether the requirements are like those posited on the Instrumental Approach, the Universalization Approach, or the Reasons Approach – those who have the capacity to reflect on the rational standing of their desires may, as a result of their reflection, form beliefs about what they would desire if their desires conformed to such rational requirements. But now imagine someone who does indeed form such beliefs. Let’s suppose he forms the belief that he would intrinsically desire that q if his desires conformed to the rational requirements to which they are subject. It seems that we thereby imagine someone who is under rational pressure either to acquire the intrinsic desire that q or to give up his belief that that is indeed what he would desire if his desires conformed to rational requirements. In other words, we seem thereby committed to supposing that the following is a further requirement of rationality: RR (if a subject believes that she would intrinsically desire that q if her desires conformed to all rational requirements then she has an intrinsic desire that q),

where this principle has exactly the same form as BD; it is that instance where p is the proposition that she would intrinsically desire that q if her desires conformed to all rational requirements. Given that we think our desires are subject to some rational requirements or other, it therefore seems that we’re committed to their being subject to an additional rational requirement: a requirement that our desires match our beliefs about what our desires should be, given those rational requirements (Smith 2001). Nor is it surprising that our desires should be subject to such an additional requirement of rationality, for what we imagine, when we imagine someone who has the capacity to form beliefs about the rational standing of his desires, and indeed his beliefs too for that matter, is someone who doesn’t just have beliefs and desires that are subject to rational requirements, but someone who can bring about what

132

M. Smith

are, by his own lights, rational improvements in his beliefs and desires, taken as a whole. This capacity to reflectively manage one’s beliefs and desires is a distinctive rational capacity, one that presupposes rational requirements that govern the selfmanagement itself, rational requirements with the same form as BD. But, of course, these rational requirements are not themselves basic in the sense we’re after. They piggy-back on our commitment to more basic rational requirements that govern our beliefs and desires in the first place, those about which we form beliefs when we engage in the process of self-management. We now have a new way of putting the question we asked initially. How plausible is it to suppose that a principle of the kind to which we are committed on the Reasons Approach – BD – is both true and explanatorily basic in the sense of not piggybacking on our commitment to more basic rational requirements? It might be thought that our discussion of the Universalization Approach already suggests a positive answer to this question. After all, when we described Hare’s version of the argument from U to the conclusion that the only desire that is suitably universal is the desire that there is as much preference satisfaction as possible, we saw that his argument made crucial appeal to the following principle: RR (if someone believes that there is some possible situation in which she desires that q, then she desires that, in that possible situation, q).

This principle, which has the same form as BD, tells us in effect that a certain consideration – the nature of the possible situation in which we desire that q – provides us with a reason to form a certain desire, namely, the desire that, in that situation, q. The fact that Hare’s argument made crucial appeal to this principle was, I suggested, symptomatic of the failure of the Universalization Approach. That’s because, as we can now see, he was really offering a version of the Reasons Approach. To the extent that his argument has any appeal at all, its appeal rests entirely on the assumption that the principle just described expresses an explanatorily basic requirement of rationality. In fact, however, as we will shortly see, not only could this principle not be explanatorily basic, it seems doubtful that any such principle could be explanatorily basic. To see why this is so we need to reflect for a moment on what would have to be the case for such a principle to be explanatorily basic, so consider a slightly different case. Why should we suppose that the following is an explanatorily basic rational principle governing our beliefs? B: RR (if someone believes that p and believes that if p then q, then she believes that q).

The answer to this question may seem to run parallel to what we have already said about BD. We should believe that this principle is true and explanatorily basic, we might say, because the facts that p and that if p then q are reasons to believe that q, and rationality is a matter of sensitivity to reasons; that is, it is a matter of forming the beliefs that we have reason to believe when we believe that those reasons obtain. But in fact the answer in the case of B has an extra and important feature over and above anything that we can say about the principle to which we are committed on the Reasons Approach.

Beyond the Error Theory

133

Suppose someone asks why the facts that p and that if p then q are reasons for believing that q. I take it that this is a legitimate question and that we can answer it by saying something about the way in which the facts that p and that if p then q bear upon the truth of q. This is because what these considerations are reasons for is the attitude of believing, where believing is in turn an attitude whose nature we need to explain in terms of its having the aim of truth. The facts that p and that if p then q are reasons for believing q because, inter alia, belief aims at the truth and there are truth-supporting relations between p, if p then q, and q. If this is right, however, then the worry with BD can be stated rather simply. The worry is that we cannot say anything in defense of BD remotely similar to what we just said in defense of B. Suppose we ask why the fact that a state of affairs has the intrinsic nature of a state of affairs in which there is as much happiness as possible is a reason for intrinsically desiring that that state of affairs obtains. If the answer to this question were to run in parallel to the answer we just gave in the case of reasons for belief, then we would have to answer it by appealing to the aim of desire. But what is the aim of desire? The aim of desire obviously isn’t truth. That is the aim of belief. Nor would it help to suggest that the aim of desire is something like satisfaction, given our beliefs. That plainly won’t help us explain why the fact that a state of affairs has the intrinsic nature of a state of affairs in which there is as much happiness as possible is a reason for intrinsically desiring that that state of affairs obtains. This, at any rate, is what we learned from the failure of the Instrumental Approach. The only answer that seems likely to do the required work is that the aim of desire is the good. But the trouble with this answer is that the Reasons Approach is itself derived from the Scanlon/Parfit buck-passing account of the good, an account according to which the good is simply that which there is reason to desire. There therefore isn’t a good independent account of what there is reason to desire that could play the role of explaining what makes the considerations that are reasons to desire reasons to desire. It would be viciously circular to explain why a consideration is a reason for desiring in terms of the fact that the good is the aim of desire and then to immediately go on and explain the good in terms of what there is reason to desire. The upshot, it seems to me, is that if we adopt the buck-passing account of the good then we simply cannot explain why the fact that a state of affairs has the intrinsic nature of a state of affairs in which there is as much happiness as possible is a reason for intrinsically desiring that that state of affairs obtains in a way that parallels the explanation that we give of why the considerations that are reasons for beliefs are reasons for beliefs. But if this is right then it turns out that the basic relation out of which the Reasons Approach is constructed – the relation of a consideration’s being a reason for desiring – is a relation whose nature is (so far, at any rate) utterly opaque to us. We literally have no idea what would make one consideration rather than another a basic reason for desiring. This discussion of the Reasons Approach helps us better understand Mackie’s original metaphysical challenge to the existence of value. As I said earlier, Mackie says that if happiness is of value then it follows that happiness has the feature of being an end absolutely prescribed by reason, “the same for all minds.” We can now see that Mackie’s argument really poses a dilemma. If happiness is an end absolutely

134

M. Smith

prescribed by reason then, at least if we take the Reasons Approach, it follows that there must be some consideration that provides us all with a reason to desire happiness. But in order to understand what it would be for some consideration to be a reason to desire happiness we would have to have some independent conception of the good as the aim of desire. We might suppose that this is why Mackie thinks that we are led inevitably to Moorean non-naturalism about the good with its attendant metaphysical and epistemological difficulties. This is one horn of the dilemma. There is, however, another horn on which we deny the existence of Moorean nonnatural qualities and follow instead the buck-passers in conceiving of the good as that which there is reason to desire. On this horn, however, the problem is that we can give no explanation at all of what it is for certain considerations, as distinct from others, to be reasons for desiring at all. On this horn, the idea of there being reasons for desiring literally makes no sense. The Reasons Approach must therefore be rejected. This is not to say that we must reject BD, the principle to which we are committed on the Reasons Approach. BD may well be true, for all that’s been said. But if BD is true, then that will be because we have derived it in some way from something else that is itself explanatorily more basic. But what else is there to derive it from?

6 The Constitutivist Approach To avoid the error theory, it seems that we must explain why certain intrinsic desires are rationally required without presupposing that there are explanatorily basic reasons for desiring, since such reasons do presuppose the existence of Moorean non-natural qualities. But how might we do this? The only untried possibility I can imagine is that we might suppose that certain desires are constitutive of being fully rational. In other words, remembering once again our simplifying assumption about the substance of morality, we might suppose that the following is an explanatorily basic principle of rationality: C: RR (people intrinsically desire that there is as much happiness as possible).

Let’s call this the “Constitutivist Approach.” (Parfit considers views of this kind when he discusses the “critical” versions of the present aim theory (Parfit 1984).) To anticipate, it seems to me that we are better placed to respond to Mackie’s metaphysical and epistemological challenges to the existence of value if we take the Constitutivist Approach than if we take any of the others. As with the other approaches, C states a condition on the proper functioning of the psychology of an agent. It says, in effect, that if it is to function properly, an agent’s psychology must include the desire that there is as much happiness as possible. C is thus metaphysically innocent. It does not presuppose the existence of any Moorean non-natural qualities and it doesn’t presuppose unexplained reason relations either. C thus differs crucially from BD, the principle to which we are committed on the Reasons Approach. BD assumes that we can explain why a fully rational person would desire that there is as much happiness as possible in terms

Beyond the Error Theory

135

of reasons that exist for desiring. BD thus falls foul of the need to explain what it would be for a consideration to be a reason for desiring: This is what gives rise to the need to appeal to Moorean non-natural qualities and the attendant metaphysical and epistemological queerness. Since C makes no such assumption, it avoids this charge of metaphysical and epistemological queerness. Even so, note that C would, if it were true, explain the truth of BD. After all, if it is constitutive of being rational that we desire that there is as much happiness as possible then it follows that we are rationally required, when we believe that a state of affairs has the intrinsic nature of a state of affairs in which there is as much happiness as possible, to intrinsically desire that that state of affairs obtains. We might even express this by saying that the intrinsic nature of happiness provides us with a reason to intrinsically desire happiness. What’s crucial, however, is that BD, so understood, would not be explanatorily basic. What’s explanatorily basic is rather C itself: the claim that we are rationally required to desire happiness. But how plausible is it to suppose that this is so? In answering this question it is once again important to remember that the question isn’t whether this particular rational principle, C, is itself plausible. The question is the more general one whether it is plausible to suppose that, for some p, there is a rational principle of the form: RR (people intrinsically desire that p).

Moreover, it is also important to remember that this is a metaphysical question and that epistemological questions are therefore orthogonal. The principle to which we are committed on the Constitutivist Approach is a principle that captures what the psychology of an agent has to be like if it is to function properly. We thus mustn’t suppose that the principle purports to state some sort of obvious analytic truth about rationality. C itself may be no such thing. But the mere fact that it is far from obvious that C is true is neither here nor there given that C purports to be a metaphysical thesis, rather than an analytic truth about rationality. To be sure, if C is true then it is something that we can discover simply by thinking about what the psychology of an agent is like if it is to function properly, but it may be difficult and non-obvious for all that. So let’s now face the question fairly and squarely. How plausible is it to suppose that there are rational principles like those to which we are committed on the Constitutivist Approach? Here is where matters get tricky. Those who urge the Constitutivist Approach upon us will insist that, to the extent that we are convinced by Mackie’s conceptual claim, we just have to admit that each and every moral judgment we make commits us to a corresponding judgment that some desire or other is constitutive of being rational. The judgment that it is morally obligatory to keep some promise in certain specific circumstances, for example, commits us to the judgment that desiring to keep that promise in those specific circumstances is the product of some intrinsic desire that is constitutive of being rational plus facts about how keeping that promise leads to the satisfaction of that intrinsic desire; the judgment that it is morally obligatory to return a borrowed book in certain specific circumstances commits us to the judgment that the desire that we return the

136

M. Smith

borrowed book in those circumstances is the product of some intrinsic desire that is constitutive of being rational plus facts about how returning that book leads to the satisfaction of that intrinsic desire; and so we could go on. It thus seems that our moral beliefs commit us to the conclusion that there are rational principles of the kind to which we are committed on the Constitutivist Approach. The question, however, is whether we can rest content with this commitment. Mackie might say that we cannot. The commitment to these principles is, he might say, grist for his mill. Since we know, ex ante, that no desires are constitutive of being rational, the upshot is thus that we have to do a modus tollens and give up our moral beliefs. But the other possibility, of course, is that we can indeed rest content with the commitment. On this alternative way of thinking, the response we just imagined Mackie giving begs the question. For even if we were disposed to think that no desires are constitutive of being rational ex ante, after being convinced that our moral beliefs commit us to the conclusion that there are such desires, and after seeing that the supposition that there are such desires is metaphysically innocent, we should simply revise that belief. Our moral beliefs commit us to the conclusion that there are rational principles of the kind to which we are committed on the Constitutivist Approach, so that’s that. I must confess that I find it difficult to say which of these two responses is correct. Should we think that one way of figuring out what the psychology of an agent is like when it functions properly is by engaging in ordinary moral reflection? If so, then we should conclude that ordinary moral reflection provides us with insight into which desires are constitutive of being rational. Or should we instead suppose that our ex ante beliefs about the nature of rationality are themselves true? If so then, if we are indeed disposed, ex ante, to deny that there are any desires that are constitutive of being rational, then we should conclude that our moral beliefs are all false. My somewhat tentative suggestion is that the first supposition is more plausible than the second. But I say this mindful of the fact that that I thereby merely express a hunch rather than the conclusion of a reasoned argument. Once the suggestion that our moral views provide us with insight into the nature of rationality is on the table, an obvious way of figuring out which desires are constitutive of being rational presents itself. We figure out which desires are constitutive of being rational in exactly the same way as we figure out what the most fundamental moral principles are (compare Gilbert Harman on the autonomous approach to morality in Harman 1985). In other words, we should believe C, as opposed to some alternative claim about the desires constitutive of being rational, for much the same reason that we should believe that the principle of utility is the most fundamental moral principle, if indeed we should believe either of these things at all. We figure out what the most fundamental moral principles are by engaging in a certain canonical method of reasoning: the process of reflective equilibrium (Rawls 1951; Daniels 1979). We try to get our considered judgments about what our moral obligations are in specific cases into equilibrium with our reflective judgments about the most general moral principles that govern those specific cases. Similarly, the suggestion goes, we can figure out which desires are constitutive of being rational by getting our considered judgments about what reason requires us to desire in

Beyond the Error Theory

137

specific cases – where, remember, our commitments about this can simply be read off from our moral judgments about the specific cases, as the one entails the other – into equilibrium with our reflective judgments about what the most general intrinsic desires constitutive of being rational are that stand behind these more specific desires. Suppose that, via the process of reflective equilibrium, we are led to the conclusion that the intrinsic value of happiness and disvalue of suffering best explain and justify our more specific judgments of moral obligation. In that case it seems that that same process of reasoning will inevitably lead us to conclude that intrinsically desiring that there is as much happiness as possible is constitutive of being rational. The upshot is thus that, on the Constitutivist Approach, our ability to vindicate the truth of a candidate rational principle such as C, in the sense of being rationally justified in believing it to be true, goes hand in hand with our ability to provide a similar vindication of the principle of utility itself. Moral theorising and theorising about the nature of rationality are one and the same. What can we say to those who disagree with us about what our moral obligations are if we take the Constitutivist Approach? Let’s suppose that they have the beliefs and desires of someone with a firm commitment to commonsense morality, whereas we have the beliefs and desires of a committed utilitarian. One thing we can say is that, as we see things, they lack the desire that is constitutive of being rational, whereas we possess that desire. But it is worth adding that the account we have just given of how we come by knowledge of what our obligations are and which desires are constitutive of being rational shows that they may be liable to a different kind of charge as well. Those who do not believe the principle of utility may be such that, if they were to engage in the reflective equilibrium process, they would come to the conclusion that the principle of utility is the most fundamental moral principle, and in that case we can criticize them for being irrational in the further sense of having epistemically unjustified beliefs. Moreover, given that the belief they would have, if their beliefs were epistemically justified, commits them to the conclusion that they would desire to maximize happiness and minimize suffering if they had the desires constitutive of being rational, it follows that such agents are liable to a further charge of irrationality as well. For if they had epistemically justified beliefs, rationality would be on the side of their having a matching desire to maximize happiness and minimize suffering. In this way the Constitutivist Approach allows that there may be many different grounds on which we might criticize those with whom we have moral disagreements. And what can we say to someone who disagrees with us about the nature of rationality? Imagine someone who agrees that moral beliefs commit those who have them to the conclusion that certain desires are constitutive of being rational – to this extent they follow the Constitutivist Approach – but who then goes on to insist that, since there are no desires that are constitutive of being rational, it follows that all moral beliefs are false. Perhaps he goes on to add “That’s why I have never had any moral beliefs.” What are we to say to him? As I have already indicated, the only thing to say to such a person is that we quite reasonably take our moral beliefs to provide us with insights into the nature of rationality, insights that he evidently lacks.

138

M. Smith

Unfortunately, this means that we will be unable to convince him that his views about the nature of rationality are mistaken. But it could hardly be a requirement on any philosophical view that you have to be able to convince the arbitrary person that that view is correct. It is surely enough that we are able to convince ourselves of the reasonableness of our own view. Let me sum up. The Constitutivist Approach seems to me to offer the most promising way of responding to Mackie’s metaphysical and epistemological arguments for the error theory. Let’s grant that Mackie is right that something is of moral value just in case desiring that thing is, as Sidgwick says, “absolutely prescribed by reason.” It turns out that all that this requires is that desiring that thing is constitutive of being fully rational. Controversial though this claim is, the important point is that there is nothing metaphysically queer about it. It presupposes neither Moorean non-naturalism nor an unexplained reason relation. To be sure, we may not be able to convince everyone of the truth of this claim. But who would have thought that we could? Acknowledgments An earlier draft of this chapter was given at the Symposium on Moral Rationalism at the annual Australasian Association of Philosophy Conference held in Melbourne in 2008. I am grateful for helpful comments made by my co-symposiasts – Charles Pigden and Francois and Laura Schroeter – and by members of the audience, especially Daniel Cohen, Simon Keller, Norva Lo, and Geoffrey Sayre-McCord. I am also especially grateful for written comments I received from Richard Joyce and Simon Kirchin. Though they saved me from many errors, I fear that all too many remain. Finally, I would like to thank the anonymous referee who suggested that I fix my tortured prose.

References Blackburn, S. 1993. Errors and the phenomenology of value. In his Essays in quasi-realism, 149–165. New York: Oxford University Press. Daniels, N. 1979. Wide reflective equilibrium and theory acceptance in ethics. Journal of Philosophy 76: 256–282. Davidson, D. 1963. Actions, reasons, and causes. Journal of Philosophy 60: 685–700. Garner, R. 1994. Beyond morality. Philadelphia: Temple University Press. Gauthier, D. 1986. Morals by agreement. Oxford: Clarendon Press. Hare, R. M. 1981. Moral thinking. Oxford: Oxford University Press. Harman, G. 1985. Is There a single true morality? In Morality, reason, and truth, eds. D. Copp and D. Zimmerman, 77–99. Totowa: Rowman and Allanheld. Joyce, R. 2001. The myth of morality. Cambridge: Cambridge University Press. Kant, I. 1786. Groundwork of the metaphysics of morals, Trans. H. J. Paton. 1948. London: Hutchinson and Company. Korsgaard, C. 1986. Skepticism about practical reason. Journal of Philosophy 83: 5–25. Lewis, D. 1984. Putnam’s paradox. Australasian Journal of Philosophy 62: 221–223. Lewis, D. 1989. Dispositional theories of value. Proceedings of the Aristotelian Society 63(suppl.): 113–137. Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin. McDowell, J. 1985. Values and secondary qualities. In Morality and objectivity, ed. T. Honderich, 110–129. London: Routledge and Kegan Paul.

Beyond the Error Theory

139

Moore, G. E. 1903. Principia ethica. Cambridge: Cambridge University Press. Parfit, D. 1984. Reasons and persons. Oxford: Clarendon Press. Parfit, D. Forthcoming. On what matters. Rawls, J. 1951. Outline of a decision procedure for ethics. Philosophical Review 60: 177–197. Scanlon, T. M. 1998. What we owe to each other. Cambridge, MA: Harvard University Press. Shafer-Landau, R. 2003. Moral realism: A defence. Oxford: Oxford University Press. Smith, M. 1994. The moral problem. Oxford: Blackwell. Smith, M. 2001. The incoherence argument: Reply to Shafer-Landau. Analysis 61: 254–266. Smith, M. 2004. Instrumental desires, instrumental rationality. Proceedings of the Aristotelian Society 78(suppl.): 93–109. Smith, M. 2006. Is that all there is? The Journal of Ethics 10: 75–106. Williams, B. 1981. Internal and external reasons. In his Moral luck, 101–113. Cambridge: Cambridge University Press.

Normativity, Deliberation, and Queerness David Copp

According to an idea that is familiar from the Kantian tradition, moral obligations are a source of “authoritative reasons” – reasons it would be “irrational to ignore.” In this sense, moral obligations are “genuinely normative.” We can express the idea as follows: Authoritative Reasons Proposal: Necessarily, if a person has a moral obligation to do something, then (a) there is a moral reason for her to do it, and (b) if she is fully rational and believes she is obligated to do it or that there is a moral reason for her to do it, she takes this obligation or reason appropriately into account in deciding what to do.1

This proposal reflects a conception of normativity that I have elsewhere called “authoritative normativity” (Copp 2007b). It seeks to understand normativity in terms of reasons for action, and it views the content of rational decision-making as the litmus test of a genuine reason. It is of course also a familiar idea from the Kantian tradition that whether a person has a moral reason to do something depends only on the moral status of the action, chiefly on whether it is morally obligatory or forbidden.2 Crucially it does not depend on whether doing it would further the person’s desires, values, or interests, or whether it would be to her advantage. Non-Instrumental Reasons: An agent S has a “non-instrumental reason” to φ just in case there is a reason for her to φ and the fact that this is so does not depend on S’s having any particular desires, values, or interests, or on whether S’s φing would be to her advantage.

D. Copp (B) Department of Philosophy, UC, Davis, CA, USA 1I

use the term “moral obligation” to speak of what a person morally ought to do or is morally required to do as well as to speak of moral obligations in the strict sense. 2 I assume that a “moral reason” to φ is a fact, such as the fact that it is morally obligatory to φ or forbidden not to φ. In all such cases, there will be morally relevant facts in virtue of which, say, it is obligatory to φ. For example, it might be obligatory to φ because to do otherwise would be to mislead someone. Morally relevant facts of this kind might also be called “moral reasons” to φ, but I shall not refer to them as such. For my purposes, they are “morally relevant reasons.”

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_9, 

141

142

D. Copp

In this sense, morality is a source of non-instrumental reasons. Combining this idea with the authoritative reasons proposal, we reach the view that the reasons entailed by moral obligations are both non-instrumental and authoritative. This combination of ideas has fueled challenges both to moral naturalism and to moral realism more generally.3 It will help to consider an example. According to one kind of moral naturalism, the moral facts are, roughly, facts about the content of the “optimal” moral system – the system of moral rules the currency of which in society would maximize expected overall human welfare.4 If the optimal moral system calls for us not to lie, then, the theory says, it would be wrong to lie and there is a moral reason not to lie. According to the theory, the fact that there is a moral reason not to lie would be the fact that the optimal system calls for us not to lie. This theory is a kind of rule consequentialism, understood here as a metaethical proposal regarding the nature of the moral facts. Because of its similarity to a theory I have proposed elsewhere, which I have called the “society-centered theory,” I will use this name here as well (see Copp 1995, 2007a). The question to ask in this context is how a theory of this kind can account for the normativity of morality. The challenge is to explain why rational agents necessarily would take their beliefs about their moral reasons into account in deciding what to do if the society-centered theory is correct about what it is for there to be a moral reason. More generally, if moral reasons are non-instrumental, so that a person’s having a moral reason to do something does not depend on her desires, values, or interests, the challenge is to explain why rational agents necessarily would take their beliefs about their moral reasons into account in deciding what to do. Following J.L. Mackie, we might try to explain this by supposing that moral facts have a kind of “to-bepursuedness” built into them, but Mackie argues that no facts have any such builtin feature (Mackie 1977, pp. 38–42). Richard Joyce argues that there are no noninstrumental reasons, or, more cautiously, he thinks it is implausible to suppose that non-instrumental reasons are genuinely normative since it is not irrational to ignore them. The trouble is, he contends, there is a moral obligation to do something only if there is a genuinely normative, non-instrumental reason to do it (Joyce 2001, pp. 77, 101–102).5 These arguments lead Mackie and Joyce to an “error theory” according to which there are no moral obligations and nothing is morally wrong. I will discuss their arguments in what follows. The authoritative reasons proposal plays an important role in Joyce’s argument if not in Mackie’s. I do not accept this proposal, however. I find it counter-intuitive

3 Moral realism holds that there are moral “facts,” including facts about our obligations and reasons.

It holds there are such facts in whatever sense there are facts of other kinds, such as meteorological facts. See Copp (2007a). 4 I simplify by assuming that there is one such system, that there are no “ties.” 5 In some places, Joyce appears to say that (setting aside “motivating reasons”) there are only instrumental reasons (Joyce 2001, p. 70). In other places, he concedes that (still setting aside “motivating reasons”) there are “institutional” reasons, such as reasons of etiquette and moral reasons (Joyce 2001, pp. 39–42). He claims, however, that such reasons are not “normative” (Joyce 2001, pp. 101–102). I take the latter to be his official view. See below, Notes 14 and 15.

Normativity, Deliberation, and Queerness

143

to suppose that the wrongness of torture depends on whether it is a necessary truth that anyone who believed that torture is wrong would be irrational to fail to take this into account in deciding whether to torture. In this chapter, however, I will set this disagreement aside and assume for the sake of argument that the authoritative reasons proposal is true. I will contend that moral naturalism has nothing to fear from the proposal. The idea that moral obligations entail that there are reasons that are both authoritative and non-instrumental is compatible with moral naturalism if it is combined with a suitable account of practical rationality.6 To see this, it is important to bear in mind two points. First, by itself, the idea that moral reasons are authoritative implies nothing about the metaphysics of moral facts. It rather implies a constraint on the theory of practical rationality. It implies that a person does not qualify as rational unless she is “appropriately motivated” – unless she is motivated in such a way that if she believes she has a moral obligation or reason to do something, she takes this obligation or this reason appropriately into account in deciding what to do. Second, there is nothing metaphysically queer or otherwise problematic about the property of being appropriately motivated. The metaphysics is not problematic. The only issue is whether being appropriately motivated is a necessary condition of being rational. Moreover, I contend, this is a classificatory issue, not a substantive metaphysical issue. To be sure, there is a notion of instrumental rationality as a matter of, basically, the efficient pursuit of one’s goals while living in accord with one’s values. There is a substantive question whether being instrumentally rational in this sense entails being appropriately motivated. I think that it does not entail this. But we have the option of taking up a different conception of rationality, a conception according to which appropriate motivation is a necessary condition of being rational. That is, in other words, we have the option of using the term “rationality” differently, to refer to a psychological constitution that includes being appropriately motivated, rather than to the psychological constitution of merely being “instrumentally rational.” The choice between these different uses of “rationality” is not substantive. It cannot cut any ice in an argument against moral realism or moral naturalism. I return to this argument and respond to objections in Section 4 of this chapter. I begin, in Section 1, with Mackie’s argument from queerness. In Section 2, I examine Joyce’s argument for the moral error theory. In Section 3, I explain the authoritative reasons proposal and discuss five arguments in support of it. I contend that these arguments are unsuccessful. In Section 4, I argue that the authoritative reasons proposal is compatible with naturalistic moral realism. This returns us to the argument I sketched in the preceding paragraphs. In the brief concluding section, I explore the implications of the authoritative reasons proposal for the theory of rational agency.

6 In

Copp (2007b), I argued that Christine Korsgaard’s (1996) account of authoritative normativity is compatible with moral naturalism. In this chapter, I aim to show that any version of moral realism is compatible with the authoritative reasons proposal if it is combined with a suitable account of rationality.

144

D. Copp

1 Mackie’s Challenge In his famous argument from queerness, Mackie argues for the moral error theory from a thesis about “the authority of ethics” (Mackie 1977, p. 33). He holds that the “meanings of moral terms” include a “claim to objective, intrinsic, prescriptivity” (1977, p. 35). Given this, he thinks, ordinary moral judgments “presuppose” or, perhaps better, entail or “include” a claim “that there are objective values or intrinsically prescriptive entities or features of some kind” (1977, p. 40). He then claims that if objective values existed, they would be “entities or qualities or relations of a very strange sort, utterly different from anything else in the universe” (1977, p. 38). They would be metaphysically “queer.” In light of this, he concludes, objective values are “not part of the fabric of the world” (1977, p. 15). They do not exist (1977, p. 17). But since ordinary moral judgments include the claim that there are objective values, it follows that ordinary moral judgments are false. Nothing is morally right or wrong or good or bad. This is his “error theory” (1977, p. 35). As I understand it, the error theory is the thesis that there are no true “basic moral propositions” – where a basic moral proposition is one that entails or presupposes a proposition ascribing a moral property to something.7 An example is the proposition that torture is wrong. The fact that torture is wrong would be a “basic moral fact” – where a basic moral fact would consist at least in part in something’s having a moral property.8 Mackie’s thesis that there are no true basic moral propositions entails that there are no basic moral facts. It entails, among other things, that nothing is morally wrong.9 Mackie claims that there are no “objective moral values,” but it is not clear exactly what he means by this. I take him to intend to claim that, in his words, there are no “entities or qualities or relations” that are “intrinsically prescriptive” (1977, p. 40), or “intrinsically action-guiding and motivating”(1977, p. 49).10 He thinks that ordinary moral judgments entail or presuppose that there are such things, and it is this that he thinks is false. His idea seems to be that an “intrinsically prescriptive”

7I

use the term “moral property” to refer to the contribution made by a moral predicate, such as “right” and “wrong,” to the content of the beliefs we express in making moral assertions using that predicate. If we express a belief in asserting that lying is “morally wrong,” I say the predicate contributes the “property” wrongness to the content of the belief. 8 I assume a metaphysically robust construal of facts as states of affairs that are “truth-makers” for true propositions. But a moral realist needs to hold only that there are moral facts in whatever sense there are facts of other kinds, such as meteorological facts. 9 The proposition that nothing is wrong is non-basic. Compare Pigden, this volume, on “atomic moral judgments.” 10 In at least one passage, Mackie appears to have in mind a different conception of objective values. For he concedes that if God existed and if God “fixed the true purpose of human life,” then there would be a kind of “objective ethical prescriptivity” (Mackie 1977, p. 48). It would be true that we ought to act in accord with the true purpose of human life. Because human beings are imperfect, however, Mackie notes, this fact would not motivate everyone who was aware of it, and so it would not be “intrinsically action-guiding and motivating.” It would not be an objective value in the sense that concerns me in the text. Richard Joyce reminded me of this passage.

Normativity, Deliberation, and Queerness

145

entity would have mutually incompatible characteristics. On the one hand, intrinsically prescriptive facts would motivate those who were aware of them. Mackie says that an “objective value” would be “sought by anyone who was acquainted with it. . . just because the end has to-be-pursuedness somehow built into it” (1977, p. 40). For example, the belief that truth-telling is good would motivate us to some degree to be truthful. Mackie thinks that the normativity of goodness would consist at least in part in the fact that, necessarily, the belief that truth-telling is good would provide anyone with a motive to be truthful. On the other hand, Mackie thinks, a moral fact would be intrinsically prescriptive. The normativity of the fact that truthtelling is good would be intrinsic to what that fact is, and independent of facts about us. Thus, he says, a person making a moral judgment wants to say something “that involves a call for action or for the refraining from action, and one that is absolute, not contingent upon any desire or preference or policy or choice, his own or anyone else’s” (1977, p. 33). The problem is that it could not be intrinsic to a fact that our awareness of it necessarily motivates us. If there were a fact awareness of which necessarily motivated people, this would not be a truth about that fact as it is in itself. There could not be facts that, as a matter merely of the way they are in themselves, necessarily motivate people who are aware of them. At the very least, such facts would be metaphysically queer.11 This argument poses a challenge to moral realism. Some might think that non-naturalism can avoid the challenge. Obviously a non-naturalist could accept Mackie’s claim that objective values would be “utterly different from anything else in the universe” (1977, p. 38). Perhaps even a naturalist could accept this. But at the most fundamental level, Mackie’s challenge is grounded in a metaphysical puzzlement about the very idea of a normative fact or property. This puzzle is not solved by postulating the existence of normative properties and then simply declaring that they are “non-natural.” In order to meet the challenge, a realist might argue that moral facts are not normative, appearances to the contrary notwithstanding. I believe, however, that such an argument would be completely unpersuasive. We need to concede that moral facts are normative in some sense (Copp 2007b). One might propose a noncognitivist or expressivist account of normativity, but this option is not open to a realist.12 A better strategy for the realist would be to accept Mackie’s thesis that moral judgments presuppose or entail that there are normative facts or properties, but then to argue that this idea is not problematic. This is the strategy I pursue. I will argue that although

11 Even

if it does not follow that there cannot be intrinsically prescriptive entities, Mackie claims we should not suppose there are. We can explain why we have beliefs that entail or presuppose that there are such things without supposing that such things exist, and we have no other need to suppose them to exist (1977, pp. 42–46). But then, in light of the “paradoxical” characteristics that normative entities would have, if they existed, he concludes that we are best to deny that there are such things (1977, p. 42). 12 I have defended a position I call “realist-expressivism,” but I do not think it provides a satisfactory realist account of normativity. See Copp (2007d, 2007b).

146

D. Copp

Mackie is correct to think that moral facts would be normative, he is mistaken to think that normative facts would be “intrinsically prescriptive.” There are at least three errors in Mackie’s argument. First, Mackie argues that moral facts would be intrinsically normative on the basis that, as he says, a person making a moral judgment wants to say something “that involves a call for action or for the refraining from action. . . that is. . . not contingent upon any desire or preference or policy or choice, his own or anyone else’s” (1977, p. 33). But even if Mackie’s latter point is correct – we think, for instance, that the fact that torture is wrong (if it is) does not depend on anyone’s desires or preferences or policies or choices – it does not follow that the normativity of this fact is intrinsic to it. The most that follows is that the normativity of this fact (if there is such a fact) would not depend on certain kinds of psychological facts about persons. Second, Mackie is wrong to think that normative facts would be “prescriptive” – assuming he means by this that a fact p counts as “prescriptive” just in case, necessarily, any person who believed that p would be suitably motivated. I think that the normativity of the fact that torture is wrong is completely independent of whether it is necessarily the case that a person who believes that torture is wrong would be motivated appropriately. Notice that people are motivated to seek sweet things. As I have pointed out elsewhere (Copp 2007b), we could stipulate that a belief counts as the belief that something is “tweet” just in case (a) its content is the same as the content of a belief that the thing is sweet and (b) the person with the belief is motivated to seek the thing. Given this stipulation, it follows that it is necessarily the case that a person who believes that something is tweet would be motivated to seek it. But it should be obvious that the fact that sugar is tweet (i.e., sweet) is not normative. Mackie seems to have confused the motivational import of a basic moral belief with the normativity of a moral fact. Mackie’s third and crucial mistake is to suppose that ordinary moral judgments entail or presuppose that there are facts that are “prescriptive” in the above sense – facts that would motivate anyone who was aware of them (1977, p. 40). It is not incoherent to suppose that, say, torture is morally wrong, but that a person who was aware of its wrongness would not necessarily be motivated accordingly.13 I have discussed this issue elsewhere, and I will not pursue it here (Copp 2007b; see also Brink 1984). To be sure, it is plausible that an agent who deliberated rationally about what to do would take appropriately into account any relevant normative fact of which she was aware. This idea ties normativity to the motivation of rational agents rather than to the motivation of actual agents. Mackie’s mistake is to think that normativity can be reduced to facts about actual motivation. If I am correct, Mackie’s argument rests on an implausible account of normativity. It would be more plausible to tie the concept of normativity to the concept of a reason that it would be irrational to ignore in decision-making, as follows: 13 I am not merely denying that it is incoherent to suppose someone might believe torture is morally

wrong without being motivated accordingly. I am denying that it is incoherent to suppose that torture actually is morally wrong even though someone might believe it is wrong without being motivated accordingly.

Normativity, Deliberation, and Queerness

147

The Genuine Normativity Thesis: A reason to do something is “authoritative” or “genuinely normative” just in case a person who believed she had this reason would be irrational not to take it appropriately into account in deciding what to do.

If we combine this thesis with the idea that a moral obligation to do something entails that there is an authoritative reason to do it, we are led to the authoritative reasons proposal. And this proposal can be pressed into service in a new argument for the error theory, as we will now see.

2 Joyce and Authoritative Reasons Joyce thinks that there are moral obligations only if there are moral reasons to act accordingly and only if these reasons have “inescapable practical authority.” The key question is how this is to be understood. Joyce suggests that, on the most charitable interpretation, the idea of a reason with inescapable practical authority is the idea of a reason it would be irrational to ignore (Joyce 2001, pp. 102–103). And this suggestion leads him, tentatively, to the authoritative reasons proposal, the proposition that moral obligations entail reasons that it would be irrational to ignore. This proposition is a key premise in Joyce’s argument for the error theory. Taking some liberties with his official presentation (Joyce 2001, p. 77), we can formulate Joyce’s argument as follows: 1. Necessarily, a person has a moral obligation to φ only if she has a noninstrumental reason to φ, a reason to φ regardless of her desires, values and interests (Joyce 2001, p. 56). 2. Necessarily, this reason – the reason to φ that a person would have, of necessity, if she were morally obligated to φ – would have inescapable practical authority. It would be irrational to ignore it. 3. There are no authoritative, non-instrumental reasons. Instrumental reasons are the only reasons it would be irrational to ignore (Joyce 2001, pp. 100–103).14 4. Therefore, necessarily, no one has a moral obligation. Nothing is morally obligatory and nothing is morally wrong. In short, Joyce seems to be contending, ordinary moral thinking entails or presupposes that there are authoritative, non-instrumental reasons, but there are no such reasons (Joyce 2001, p. 104). I accept the first premise in Joyce’s argument, for I agree that moral reasons are non-instrumental. The second premise is the authoritative reasons proposal. I also

14 Joyce

allows that there are institutional reasons that are non-instrumental (Joyce 2001, pp. 39– 42). He contends they are not “normative” – it is not irrational to ignore them (Joyce 2001, pp. 100–105). In my terminology, they are not “authoritative.” See below, Note 15.

148

D. Copp

accept this premise, if only for the sake of argument. I will refer to the conjunction of these two premises as the “moral reasons doctrine.” Having granted to Joyce the moral reasons doctrine, I can avoid his conclusion only if I can reject his third premise, the negative premise that there are no authoritative, non-instrumental reasons. I contend that this premise and the moral reasons doctrine are not jointly plausible. That is, it would not be plausible to affirm all three premises of the argument. If we understand rationality in a way that makes the moral reasons doctrine defensible, then Joyce’s third, negative premise is implausible. That is, if we accept Joyce’s first two premises, we will rightly find it implausible that only instrumental reasons are authoritative. But if we think that only instrumental reasons are authoritative, then the moral reasons doctrine is indefensible. That is, if we accept Joyce’s third premise, we ought not to accept both of his first two premises. Let me explain. For convenience, let me stipulate that a “reason of practical rationality” is a reason to do something such that a person who believed she had this reason would be irrational not to take it appropriately into account in deciding what to do – it is a reason for acting that it would be “irrational to ignore,” as I shall say. Given this usage, and given the genuine normativity thesis, it follows that all and only reasons of practical rationality are authoritative or genuinely normative. Joyce’s argument for the third, negative premise in his argument turns on an instrumental theory of practical reason, a theory according to which reasons of practical rationality are instrumental to serving one’s desires, values, or interests. On such a theory, it would be surprising if any non-instrumental reasons were authoritative. For if a reason is authoritative, it is irrational to ignore, and if it is irrational to ignore, then, given an instrumental theory of practical reason, it must be that ignoring it would involve failing to serve one’s desires, values, or interests. But whether this is so must surely depend on one’s desires, values, and interests. If one is a gangster whose desires, values, and interests are well served by a life of crime, then on an instrumental theory of practical rationality, one might be irrational not to ignore any non-instrumental reasons regarding what to do. On an instrumental theory of practical rationality, then, it is at best doubtful that any non-instrumental reasons are authoritative. But then, if moral reasons are non-instrumental, as Joyce’s first premise maintains, it is at best doubtful that moral reasons are authoritative, as Joyce’s second premise maintains. Hence, on Joyce’s argument for the third premise, if we accept his first premise, his second premise is at best doubtful. His defense of the third premise cuts against the moral reasons doctrine of the first two premises. The issue is not whether there are any non-instrumental reasons, for Joyce agrees that there are. He thinks there are reasons of etiquette, for example, and reasons of etiquette are non-instrumental. There is a reason of etiquette for me not to belch at the table, and my having this reason does not depend on my interests, nor does it depend on what I happen to desire or to value. Joyce calls reasons of this kind “institutional,” since he takes them to be entailed by the requirements of “normative systems” or “institutions,” such as the rules of etiquette (Joyce 2001, p. 39–42).

Normativity, Deliberation, and Queerness

149

On reflection, there seem to be many kinds of non-instrumental reasons for action, including aesthetic reasons, legal reasons, and, perhaps, moral reasons.15 We surely do not want to say that any arbitrary normative system yields genuine reasons. As Joyce points out, merely by declaring a rule requiring everyone to purchase things made in Norway in the autumn, I cannot make it the case that there are genuine reasons of the Norwegian-autumnal-product kind (Joyce 2006, p. 203). Even where there is an ongoing institution, such as the Roman practice of gladiatorial combat, Joyce suggests, a participant might have no genuine reason to follow the rules. If a gladiator can save his life only by breaking one of the rules, Joyce suggests, the gladiator “may legitimately ignore” the rule (Joyce 2001, pp. 40–41). Institutional reasons are “escapable.” They are not “genuinely normative.” We take moral reasons to be inescapable, however, and the philosophical issue is what sense can be made of this (Joyce 2001, p. 102). As we saw, Joyce proposes understanding the issue to be whether it would be irrational to ignore moral reasons. If we understand the “inescapability” of moral reasons in this way, we are led to the authoritative reasons proposal, the idea that moral reasons are such that a rational person who believed she had a moral reason would necessarily take that reason appropriately into account in deciding what to do. The trouble is that it would be implausible to understand the “inescapability” of moral reasons in this way if we hold an instrumental theory of practical reason according to which practical rationality is a matter of serving one’s desires, values, and interests. We do not take moral reasons to be inescapable in the sense of being such that to ignore them is to fail to serve one’s desires, values, and interests, for we think that moral reasons are non-instrumental. In short, given the first premise in Joyce’s argument, the remaining two premises are not jointly plausible. Given the first premise, if we accept the authoritative reasons proposal as an account of the sense in which moral reasons are inescapable, then an instrumental theory of practical reason will seem implausible. But the third premise of Joyce’s argument rests on an instrumental theory of practical reason. The question, then, is whether to accept an instrumental theory of practical reason. Joyce appeals here to Bernard Williams’ argument that there are no “external reasons” (Williams 1981; Joyce 2001, pp. 106–134). I cannot here attempt a thorough discussion of the argument, but others have discussed it extensively (e.g., Korsgaard 1986; see also Phillips, this volume). Joyce himself thinks that Williams’ argument does not establish that there are no external reasons. He think it shows only that external reasons, such as institutional reasons, are not genuinely normative or authoritative (Joyce 2001, p. 102). I will therefore reformulate the argument so it speaks to the thesis that no external reason is authoritative. The key premise in Williams’ argument is that, as he says, “If there are [authoritative] reasons for action, it must be that people sometimes act for those reasons, 15 Joyce appears to agree that there are institutional moral reasons although he thinks these reasons

are not authoritative (Joyce 2001, p. 41). He holds, that ordinary moral discourse commits us to the thesis that moral obligations entail authoritative reasons to act accordingly. Institutional moral reasons, if there are any, would not be authoritative.

150

D. Copp

and if they do, their reasons must figure in some correct explanation of their action” (Williams 1981, p. 102). In short, (authoritative) reasons are potential motivators. As Joyce says, “If a normative [or authoritative] reason could not potentially motivate an agent, then, if presented with such a reason,” an agent could respond with a shrug of her shoulders and the query “So what?” This, Joyce claims, would be unacceptable (2001, p. 108). But what we can be motivated to do depends on the content of what Williams calls our “subjective motivational set” – our desires, projects, values, normative commitments and the like (Joyce 2001, p. 105). It follows that the nature of our (authoritative) reasons must depend on the content of our subjective motivational set. Our (authoritative) reasons must be “internal” to our subjective motivational set in some suitable sense so that it is guaranteed that these reasons are capable of motivating us to act. Unfortunately, it does not follow that all “internal” reasons are instrumental, so there is room to argue that moral reasons may be internal even given that moral reasons are non-instrumental. To see this, we need to be clear about the concept of an internal reason. For Williams, an internal reason is a reason for an agent to do something, where, given the agent’s subjective motivational set, there is a process of “sound reasoning” or “rational deliberation” that would lead her to be motivated to do the thing, assuming she had no false beliefs (Williams 1981, pp. 102–104). If we thought that only instrumental reasoning could count as “sound,” then we could perhaps move quickly to the conclusion that all of an agent’s internal reasons are instrumental. But it would be question-begging to assume that only instrumental reasoning could qualify as sound (Joyce 2001, p. 124), and Williams is very liberal about what can count (1981, pp. 104–105). He allows that an agent could become motivated to do something as a result of imagining a possibility she had not considered before. He allows that sound reasoning can remove desires from our initial subjective motivational set and that it can add desires (1981, p. 104). He leaves it open that an agent could become motivated to do something as a result of a kind of reasoning that does not rest on any antecedent desires, values or interests. Hence, even if all (authoritative) reasons are “internal,” it does not follow that all (authoritative) reasons are instrumental, so it does not follow that there are no (authoritative) moral reasons. Joyce agrees with this but maintains that a defender of moral reasons must develop an account of exactly what kind of reasoning would lead a person to become motivated to do what she is morally obligated to do, regardless of the content of her subjective motivational set. An evil gangster has a moral obligation to change his ways, but there may be nothing in his current subjective motivational set that would motivate him to do this, not even given sound deliberation and full information. If the defender of moral reasons is to show that the gangster’s reason to change his ways is an internal reason, she must explain exactly what kind of reasoning is supposed to be capable of motivating such a person to change his ways. The burden of argument, Joyce contends, rests with the defender of moral reasons (Joyce 2001, pp. 123–125). I will not attempt to take up this burden – I will not try to show that moral reasons may be internal. Instead, I will argue that Williams has not shown that there are

Normativity, Deliberation, and Queerness

151

no (authoritative) external reasons. Williams argues that “if there are [authoritative] reasons for action, it must be that people sometimes act for those reasons, and if they do, their reasons must figure in some correct explanation of their action” (1981, p. 102). I agree of course that people sometimes do act for reasons. But what Williams apparently intended to claim was rather that if a person has (authoritative) reason to do something, she can do it for that reason, and if she did, her reason would figure in a correct explanation of her action. I agree with this as well. We can see by inspection, however, that it does not follow that if a person has (authoritative) reason to do something, she can be motivated to do it simply by harnessing sound deliberation and full information to her subjective motivational set. What follows from Williams’ argument is that all (authoritative) reasons can be acted on, but not that they are all internal. Joyce appears to believe that this does follow provided that we assume a Humean theory of action, according to which action requires the cooperation of belief and desire. On the Humean theory, he says, “an external reason could never explain action,” for a person who believes she has an external reason must also have a relevant desire in order to act from that reason. And in that case, Joyce says, “the putative external reason collapses into an internal one” (2001, p. 110). There is a subtle confusion here. On the Humean theory, to act from a reason one must believe that one has the reason and one must also have a relevant desire. But of course it may be that one can act from a reason even if one won’t. Hence, even if one can act from a reason, it does not follow that some relevant desire already exists, nor that the agent could come to have such a desire simply by sound reasoning from her subjective motivational set, given full information. So it does not follow that all (authoritative) reasons are internal. Suppose that Allan is deeply irrational. When he believes he has a reason to do something, he loses all desire to do it and comes to desire very strongly not to do it, unless there is a full moon. Suppose there is not a full moon and Allan has a horrible headache. He believes his headache gives him a reason to take an aspirin, so he strongly desires not to take an aspirin. Given his subjective motivational set and his irrationality, there is no course of reasoning that would lead him to take an aspirin, not even if he were fully informed. But it does not follow that he cannot take an aspirin, nor that he cannot [take an aspirin for the reason that it will help reduce his pain]. All that follows is that he will not [take an aspirin for the reason that it will help reduce his pain] unless there is a full moon. If there is a full moon, he will act for the reasons he is aware of, and his reasons will figure in explaining his actions. Otherwise he will not. But he has the capacity to act for reasons. It therefore is compatible with the Humean theory of action to allow that a person might be able to do something for a reason even if he has no relevant desire and even if full information and sound reasoning from his antecedent motivations will not lead him to be motivated to act. I agree with Williams that if a person has an authoritative reason to do something, she can act for that reason. But it does not follow that all authoritative reasons are internal, not even if we assume a Humean theory of action. And it remains an open question whether moral reasons are internal, for it remains an open question whether

152

D. Copp

there is “sound reasoning” that would lead a (rational and) fully informed person to be motivated to do what she has moral reason to do. To summarize, we have been considering Joyce’s argument for the error theory. I conclude that we can reject the third, negative premise of the argument. Moreover, if we accept the first two premises, we ought to reject the third premise, for the three premises are not jointly plausible. I turn now to what I take to be the most interesting premise of Joyce’s argument, the authoritative reasons proposal.

3 The Authoritative Reasons Proposal According to this proposal, there are moral obligations only if there are moral reasons and only if these reasons are authoritative, or irrational to ignore. I shall take it as given that there are moral obligations only if there are moral reasons. The crucial issue for my purposes is whether the reasons entailed by moral obligations are authoritative in the sense given by the genuine normativity thesis. The proposition I want to assess, then, is the claim that if a person is fully rational, and if she believes that she is morally obligated or has a moral reason to do something, she will take this obligation or this reason appropriately into account in deciding what to do. In what follows, I will write as if this proposition by itself is the authoritative reasons proposal.16 On the authoritative reasons proposal, a rational person who believes she has a moral obligation or a moral reason will take this “appropriately into account” in decision-making. This does not mean that a rational person who is deciding what to do must necessarily pay explicit attention to her beliefs about relevant moral obligations and reasons. Nor does it mean that such a person must necessarily go through a process of deliberation before making a decision. It might be “second nature” to a virtuous person that she does not lie, for instance, since she believes dishonesty is wrong. She might “automatically” be truthful without having to give truthfulness any thought, and this might be “appropriate.” There could even be cases in which there are good reasons not to consider reasons for a decision since doing so might waste time, risk error, or risk unexpected side effects.17 In cases of these kinds, it would not be “appropriate” to take one’s beliefs about reasons into account in decision-making. The idea is simply that rational persons are appropriately responsive to their beliefs about their moral reasons and obligations. If a rational person believes she has a moral reason to do something, or believes she is morally obligated to do it, she is thereby motivated to do it, and she is thereby disposed such that, if she were to deliberate about what to do, she would take this reason or obligation to count in

16 I

thank Al Casullo for helpful discussion of the complexities. thank Jon Tresan for raising this worry in personal correspondence and John Skorupski and Justin D’Arms for raising it at the Dubrovnik conference.

17 I

Normativity, Deliberation, and Queerness

153

favor of doing it.18 According to the proposal, it would be irrational to ignore one’s beliefs about one’s obligations or moral reasons in deliberation about what to do, but this does not mean it would necessarily be irrational to fail to deliberate. It is worth mentioning that the authoritative reasons proposal does not entail that rational agents necessarily do what they believe they are morally obligated to do. Nor does it entail that rational persons necessarily take moral reasons to be conclusive.19 It remains open whether a rational person might decide in a given case that the applicable moral reasons are overridden. As I have said, I do not accept the authoritative reasons proposal. I deny that merely being rational is enough to ensure that a person would not ignore moral reasons (Copp 2007c). Nevertheless, I need briefly to discuss some arguments that have been given for the proposal.

3.1 The Argument from Inescapability Joyce’s central argument begins from the idea that we take moral reasons to be “inescapable” (Joyce 2001, p. 102). If moral reasons are inescapable, he suggests, they must bind us robustly (Joyce 2006, p. 193). He then argues that only practical rationality has “the kind of immunity from legitimate questioning that is attributed to morality” (Joyce 2001, pp. 102, 104). Moral reasons are relevantly inescapable only if it would be irrational to ignore them (Joyce 2001, p. 104). One might object that practical rationality and morality are simply different normative systems that are not interestingly different in status (Joyce 2001, p. 50). It would be irrational to ignore reasons of practical rationality and it would be immoral to ignore moral reasons. Moral reasons are considerations that are morally inescapable, and practical reasons are rationally inescapable. I think that something like this is correct (Copp 2007b, 2007c). Joyce contends, however, that there is an asymmetry between morality and practical reason. He thinks it would be incoherent to question practical reason but not to question morality (Joyce 2001, p. 50). The crucial point, Joyce thinks, is that to ask for a reason of any kind “is to imply the acceptance of practical rationality.” As he says, “asking any practical question [ – any question about what to do or what to care about – ] implies an acceptance of practical rationality” (Joyce 2001, p. 50). Moral reasons can be questioned, since one can ask for a practical reason to care about what one has moral reason to do. But Joyce thinks it would be incoherent to question whether one’s practical reasons should be decisive in one’s decision-making since to question this is to ask for a reason while at the same time “implying that 18 Michael

Smith’s “practicality requirement” says that an agent who believes that she is morally required to do something is motivated to do it unless she is practically irrational (Smith 1994, p. 61). The authoritative reasons proposal adds something about how such an agent will deliberate unless she is irrational. 19 Even if moral reasons are authoritative, there could be other kinds of authoritative reasons and moral reasons might not be normatively more important than they are. See Copp (2007e, 2007b).

154

D. Copp

no reason will be adequate” (Joyce 2001, p. 50). “Practical rationality is not available for legitimate questioning,” and this is a sense in which the reasons of practical rationality are “inescapable” (Joyce 2001, p. 51). To show that moral reasons are inescapable, then, one must show that moral reasons are reasons that it would be irrational to ignore. This argument strikes me as implausible. Even if an instrumental theory of practical reason is true, it is easy to imagine someone doubting it. She might ask why she ought to take the means to satisfying a desire, value, or interest. Perhaps, in asking this question, she would commit herself to there being some truths about reasons, but she would not seem to commit herself to there being reasons it would be irrational to ignore. She might agree that there are moral reasons, for example, but deny that it would be irrational to ignore one’s moral reasons. She might doubt whether she has any moral reason to take the means to satisfying her desires, values, or interests, given that these might be immoral. A problem with Joyce’s argument is that, I think, it confuses two senses of “practical” reason. There is the “broad” idea of a practical reason as a reason of any kind for action. In this sense, moral reasons and reasons of etiquette are “practical” reasons. In this sense, it is not necessarily irrational to ignore a practical reason. But there is also the “narrow” idea of a practical reason as a particular kind of reason for action, a reason that it would be irrational to ignore. It appears that a person can question the force of practical reasons in the narrow sense just as she can question the force of moral reasons. She can ask why she should avoid irrationality. She might thereby be asking for a reason of some kind within the broad class of reasons, to avoid irrationality, but she need not be asking for a reason that it would be irrational to ignore. She might be seeking a moral reason to avoid irrationality. Most fundamentally, I think, one can question whether to do what practical reason requires without thereby asking for a reason of any kind. Reasonable people can disagree about what to do even if they agree about all the reasons bearing on a decision.20 In face of such reasonable disagreement, one can ask oneself what to do. Suppose that Alice realizes she is morally required to tell the truth in a situation where doing so would be enormously embarrassing, and suppose she values her public face much more than she values telling the truth. In this situation, at the point of making a decision, Alice might ask herself whether to be guided by what she takes to be her practical reasons or whether instead to be guided by her moral duty. Her question might be the question of what to do that anyone can ask when all her reasons are on the table. When we ask this question, we are not asking for more reasons, for, even given additional reasons, we would still need to decide what to do in light of the reasons. It seems to me, then, that it is not incoherent to question or doubt the force of reasons of practical rationality any more than it is incoherent to question or doubt the force of moral reasons.

20 See

John Rawls’ doctrine of the “burdens of judgment” (1993, pp. 55–57).

Normativity, Deliberation, and Queerness

155

3.2 The Argument from the Idea of a Deliberative Consideration Steven Darwall says we take moral reasons to “have genuine deliberative weight” (Darwall 1997, p. 306). Joyce claims it is a conceptual truth that moral obligations have “inescapable practical authority” and he suggests that if this is so, a moral reason must be a “genuine deliberative consideration” (Joyce 2006, p. 194). That is, Joyce suggests, moral reasons must be such that an agent who realizes she has moral reason to do something, all things considered, and who reasons correctly, will decide to do that thing (Joyce 2006, pp. 194–195). But a failure to reason correctly would be a failure of rationality. So moral reasons must be such that a fully rational agent who believes she has moral reason to do something will take that reason appropriately into account in deliberation. We may agree that a reason is not a “genuine deliberative consideration” for a person if the person would be making no mistake in decision-making if she ignored it. But we need to think carefully about the idea of a mistake in deliberation. What is meant in this context by “correct reasoning”? If we understand “correct reasoning” to rule out any mistake in deliberation, including any failure to take any kind of reason into account in the right kind of way, then it is trivial that a person who believed she had a moral reason to do something would take the reason appropriately into account in decision-making, provided she reasoned correctly. This would be uninteresting. A more interesting approach would take “correct reasoning” more substantively, perhaps as reasoning that tended to maximize the truth and coherence of the reasoner’s overall system of belief and desire (Smith 1994, pp. 156–161). If we have in mind this idea of correct reasoning, however, it seems that a person who sees she has a moral reason to do something would not necessarily be making a mistake in reasoning if she were to ignore the reason in deciding what to do. That is, she would not necessarily be failing to maximize the truth and coherence of her overall view. Suppose Brenda has moral reason not to tickle Carl and suppose she realizes this. Suppose she engages in reasoning directed at maximizing the overall truth and coherence of her system of beliefs and desires. It does not follow that she necessarily would take her moral reason into account in deciding what to do. Perhaps she does not care about moral considerations. This would be a moral failing, but not obviously a failure of rationality.21

3.3 The Argument from the Concept of a Rational Person A third argument for the authoritative reasons proposal rests on the concept of a rational person, on the idea that a rational person is responsive to all the reasons she knows of that bear on her decision. This idea might seem to be an obvious truism

21 I

have discussed Smith’s arguments more fully in Copp (1997).

156

D. Copp

or platitude.22 If it is correct, it follows that all reasons are authoritative. Hence, if there are moral reasons, they must be authoritative. This argument has immediate intuitive appeal, but it neglects the apparent fact that there are reasons whose force and significance is compelling only to those with a special sensitivity. It is not necessarily irrational to ignore such reasons in deciding what to do. Aesthetic reasons illustrate the point. A person with no musical sensitivity might agree that there were good aesthetic reasons for Beethoven to take a variation in a certain direction at a certain point in a composition rather than, say, to insert a round of “Row, Row, Row your Boat” at that point. She might agree that there were reasons for Beethoven, near the end of the finale of the opus 95 string quartet, to change abruptly from a minor key to a major key and to speed up the tempo. But she might reject the significance of these reasons. A composer obviously is not guaranteed to have musical sensitivity just in virtue of being rational, and so, without irrationality, she might ignore reasons of the kind that moved Beethoven. I think, then, that aesthetic reasons are not authoritative. Rationality must be supplemented with aesthetic appreciation in order for an agent to recognize and be moved by the aesthetic significance of aesthetic reasons. It will not help to stipulate that an agent is not “fully rational” unless she is maximally informed (Smith 1994, pp. 156–161). For a composer who is maximally informed is not guaranteed to have a sensitivity that ensures she will appreciate musical reasons and take them into account in her compositions. It is arguable, moreover, that moral reasons are in company with aesthetic reasons. A morally good person would be sensitive to the significance of moral reasons. But even if I understand that the fact that I promised to do something is a moral reason to do it, I might ignore this reason in deciding what to do. It seems to me that a failure of this kind would not be sufficient to indict me as less than fully rational. At the very least, it seems, we need a better argument to support the authoritative reasons proposal.

3.4 The Argument from Blame This argument is due to Darwall, although it turns on an idea of Bernard Williams’ (Darwall 2006a, pp. 290–292; see also Skorupski 2007). First, an agent is morally required all things considered to do something only if she would be blameworthy if she were to fail to do it, unless she had an excuse. But, second, a person is not blameworthy for doing something unless she had a good reason not to do it (Williams 1995, pp. 40–44). Darwall claims beyond this that it would be “incoherent” to blame “while allowing that the wrong action. . . was nonetheless the sensible thing to do, all things considered.” For if someone could show that she had a “good

22 This

idea is a close relative of Christine Korsgaard’s “internalism requirement” (Korsgaard 1986, p. 11), which Smith describes as a “platitude” (Smith 1994, pp. 150–151). See Note 18 above.

Normativity, Deliberation, and Queerness

157

and sufficient reason” to do what she did, she would thereby have shown that she is not blameworthy (Darwall 2006a, p. 292). It follows, then, that an agent is morally required all things considered to do something only if she has reason to do it, and only if it is not the case that she has better reason not to do it. The fundamental idea is that a person is not blameworthy if she had “good and sufficient” reasons to do what he did. I agree that, in blaming someone, we imply that the person had sufficient moral reasons not to have done what she did, no matter what non-moral reasons she had for doing it. But I believe we do not commit ourselves to denying that she had “good” non-moral reasons that were sufficient on balance to make her instrumentally rational to do what she did. Consider a gangster who seeks to maximize his power and enjoyment. We view him as blameworthy for living a life of crime, but we need not deny that he has good and sufficient nonmoral reasons to live this way, reasons that are sufficient on balance to make him instrumentally rational to live this way. We blame him because we think he could give up his life of crime and because, in not doing so, he acts wrongly, contrary to all relevant moral reasons, lacking an excuse.23 In any event, the argument from blame does not support the authoritative reasons proposal. It shows, at most, that a person with a moral obligation has a reason to act accordingly and no better reason to act otherwise. But despite this, for all the argument shows, the reason to act accordingly need not be a reason that it would be irrational to ignore. The argument is compatible with the view that the reasons entailed by moral requirements are moral reasons but not reasons of practical rationality that it would be irrational to ignore. Accordingly, the argument leaves us with no support for the thesis that rational persons necessarily would take their beliefs about moral reasons appropriately into account in deciding what to do.

3.5 The Argument from “Ought Implies Can” The final argument is Kantian in flavor. Michael Smith and Steven Darwall have given it a contemporary formulation (Smith 1994, p. 85; Darwall 2006a, pp. 289– 290).24 The first premise is that moral obligations are incumbent on all rational agents. Anyone who is rational in the sense of being capable of deliberation leading to action and capable of acting for reasons is subject to the requirements of morality (Darwall 2006a, p. 289). The second premise is the maxim that “ought implies can” (hereafter, simply “the Maxim”). An agent who, all things considered, is morally obligated to do something must be able to do it (Darwall 2006a, p. 289). But then,

23 Darwall’s

response to a similar objection strikes me as unconvincing (Darwall 2006a, p. 292, n. 27). 24 I present the argument without relying on Darwall’s idea that morality is essentially “secondpersonal” in that it involves “second-personal address” (Darwall 2006a, 2006b). I do not believe that this idea adds anything to the argument, and investigating it would take me far afield.

158

D. Copp

third, anyone who is subject to a moral obligation must be able to fulfill the obligation simply in virtue of exercising her rational capacities, irrespective of contingent features of her circumstances or of her desires or values and the like (Darwall 2006a, p. 290). We can now draw the desired conclusion. For if an agent can comply with the demands of morality simply by exercising her capacity for rational deliberation and action, it must be that the reasons to act in accord with these demands would motivate her to act accordingly insofar as she were thinking rationally. Otherwise, to explain why a rational agent complies with the demands of morality we would have to cite something over and above her mere rationality. Since this is ruled out, we can conclude that a rational person would necessarily take her beliefs about moral reasons appropriately into account in deciding what to do. This is the authoritative reasons proposal. Moral requirements are rational requirements, for “the only thing we can legitimately expect of rational agents as such is that they will do what they are rationally required to do” (Smith 1994, p. 85). Reduced to its essentials, the argument is that anyone who is subject to a moral obligation must be able to fulfill the obligation “simply by exercising the [rational] capacities by virtue of which we are subject to them” (Darwall 2006a, p. 290), and so she would be irrational to ignore her moral obligations in deciding what to do. This seems clearly to be a non sequitur, and I hesitate to attribute it to Darwall or Smith. Unfortunately, this step is needed in order for the argument to conclude with the authoritative reasons proposal. To see the problem, suppose the first two premises of the argument are correct – we are subject to moral requirements simply in virtue of having the rational capacities, and “ought implies can.” What follows is perhaps that if we have the relevant rational capacities, we are subject to moral requirements, and so we have the ability (in the sense relevant to the Maxim) to fulfill these requirements. Perhaps it follows in turn that if we have the rational capacities, we can take our beliefs about moral requirements and reasons properly into account in deciding what to do. But it does not follow that we would be irrational not to do so. Perhaps I am subject to expectations about swimming given that I am able to swim, but it does not follow that if I ignore these expectations, I am swimming poorly. The argument seems to confuse what we can do, given the rational capacities without which we would be non-rational, with what we must do in order to avoid a failure in rational performance that would count as an irrationality. A person who has the rational capacities is perhaps able to that extent to fulfill her moral requirements, but this does not mean that she would be irrational to ignore her moral requirements in deciding what to do. In any event, it would be implausible to think that one must be able to fulfill one’s moral obligations merely by exercising the rational capacities. Otherwise, given the Maxim, it would be implausibly easy to defeat the claim that one is morally obligated to do something. Sometimes we would have to be different psychologically in some important way in order to do what we ought to do. If I am unloving, careless, and lazy, I will not take care of my children. I could not cease to be unloving, careless, and lazy by a simple exercise of my rational capacities. But I certainly cannot invoke the Maxim to argue on this basis that I have no duty to feed my children. Or

Normativity, Deliberation, and Queerness

159

consider a gangster who could not give up his life of crime without changing psychologically in ways that he could not bring about simply by exercising his rational capacities. He certainly cannot invoke the Maxim to argue on this basis that he has no duty to give up his life of crime. So it appears that, in the sense relevant to the Maxim, one can be able to fulfill a moral requirement without being able to do it simply by exercising one’s rational capacities. No doubt there are additional arguments for the authoritative reasons proposal. Clearly, however, there is no chance of discussing all the arguments that might be proposed. Nor am I able, in this one paper, to do justice to the five arguments I have considered. The proposal might also be put forward on programmatic grounds, rather than on the basis of a particular argument, as one element in a theory aimed at reducing morality to rationality. Obviously a discussion of such a program is well beyond the scope of this paper. I propose, therefore, to grant for the sake of argument that the proposal is true. What follows?

4 Realism, Naturalism, and the Authoritative Reasons Proposal Mackie argues for the moral error theory on the ground that normative facts would be metaphysically queer. Joyce argues for the error theory on the ground that, first, if there are any moral obligations, there are moral reasons that are both non-instrumental and authoritative, but second, only instrumental reasons are authoritative. I responded to Mackie, essentially, that we haven’t been shown there is something metaphysically queer about normative facts. And I responded to Joyce that the authoritative reasons proposal is implausible when combined with an instrumental theory of practical reason. I believe as well that there is not a good reason for accepting the authoritative reasons proposal. For the sake of argument, however, I am assuming the authoritative reasons proposal is true. In this section of the paper I shall argue that the authoritative reasons proposal can coherently be combined with moral naturalism. If I am correct about this, I claim it shows three things. First, against Mackie, there is a robust conception of normativity such that normative facts are not metaphysically problematic. Second, against Joyce, moral naturalism can make sense of there being authoritative reasons, given a suitable conception of rationality. The main obstacle to thinking of moral reasons as authoritative is the instrumental theory of practical reason. Third, a moral naturalist is not forced by her naturalism to accept an instrumental theory of practical reason. I shall use the society-centered theory to illustrate the implications of the authoritative reasons proposal for naturalistic moral realism. What I say about this theory could be said, with appropriate adjustments, about any version of naturalistic moral realism. Remember that my goal is merely to show that an argument against moral naturalism does not work. It is beyond the scope of this paper to provide an argument in favor of moral naturalism. According to the society-centered theory, very roughly, the moral facts are facts about the content of the optimal system of moral rules – the system the currency of

160

D. Copp

which in society would maximize expected overall human welfare. There is a moral obligation for me to do something just in case the optimal moral system calls for me to do it. Likewise, there is a moral reason for me to do something just in case the optimal moral system calls for me to do it. The fact that there is a moral reason to do something would be the fact that the optimal system calls for this to be done. This is a necessary truth, although obviously not a conceptual truth. The society-centered theory and the authoritative reason proposal might seem to be in fundamental opposition. On the authoritative reasons proposal, a moral reason is such that rational agents would take it appropriately into account in deciding what to do, but on the society-centered theory, our moral reasons depend, not on how rational agents would decide, but on the content of the optimal moral code. One might think it obvious that these views cannot coherently be combined. But this is a mistake. According to the authoritative reasons proposal, rational agents take their beliefs about their moral reasons or obligations appropriately into account in deciding what to do. That is, an agent counts as rational only if, were there to be a moral reason for her to do something, and were she to believe that this is so, she would take this reason appropriately into account in deciding what to do. The society-centered theory says nothing that conflicts with this. It says simply that facts about our moral reasons are facts about the choices that are called for by the optimal moral system. That is, in taking facts about moral reasons into account in decision-making, rational agents would be taking into account facts about the choices that are called for by the optimal moral system. Even if the society-centered theory and the authoritative reasons proposal can coherently be combined, their conjunction might seem to entail the obviously false proposition that rational agents necessarily take their beliefs about the requirements of the optimal moral code into account in decision-making. But this is not entailed. It obviously is not a necessary truth that rational agents believe that the societycentered theory is true. So even if the theory is true, and even given the authoritative reasons proposal, it does not follow that a rational agent who realized that a given choice is called for by the optimal moral code would take this appropriately into account in decision-making. She might not realize that facts about the choices that are called for by the optimal moral system are facts about moral reasons. She might believe that a given choice is called for by the optimal moral system but deny that there is a moral reason in favor of that choice. According to the society-centered theory, this would indicate a failure of understanding rather than a failure of rationality. The authoritative reasons proposal implies only that rational agents take into account in decision-making beliefs that are de dicto about reasons or obligations. It implies that rational agents take their beliefs that they have moral reasons or obligations appropriately into account in decision-making. But even if the societycentered theory is true, it does not follow that rational agents take into account in decision-making their beliefs about the requirements of the optimal moral system. To be entirely clear about this, we would need to deal with vexed issues about the nature of belief, the nature of propositions, the nature of facts, and the nature of properties. I cannot address these issues here, but let me try to make it clear what

Normativity, Deliberation, and Queerness

161

I am assuming. I assume, as I said, that facts about moral reasons and obligations are facts about the choices that are called for by the optimal moral system. And I assume that the property an action can have, of being called for by the moral reasons or of being obligatory, is the same property as the property of being called for by the optimal moral code. Yet I assume that beliefs about moral reasons or obligations are distinct from corresponding beliefs about the choices that are called for by the optimal moral system. These beliefs have the same truth conditions, but they are distinct.25 These assumptions commit me to familiar views about awareness, knowledge, and understanding. Someone can believe truly that p without believing that q even if the belief that p and the belief that q are made true by the same fact. One of these beliefs might represent this fact more fully or in a more articulated way than the other. Someone might believe that a glass contains water without believing that it contains H2 O even if these beliefs have the same truth conditions. This has practical significance. A fully rational agent who is thirsty, who knows that water is thirst quenching, and who knows that the glass in her hand contains H2 O, might fail to drink from the glass even though she would drink if she realized that H2 O is water. Her failure to drink clearly would not count against the claim that to be H2 O is to be water. Obviously, a rational agent might not believe the society-centered theory, so even if the theory is true, she might not realize that facts about her moral reasons or obligations are facts about the content of the optimal moral code. A rational agent might know that the optimal moral code calls for her to φ. But she might not realize that the fact that the optimal moral code calls for her to φ is the fact that there is a moral reason for her to φ. In deliberating about what to do, she might fail to take this fact appropriately into account. On the society-centered theory, her failure would be due to a failure to understand the key point that, necessarily, if the optimal moral code calls for her to φ then she has a reason to φ. Her failure to take this fact to count in favor of the choice does not count against the theory, nor does it count against the authoritative reasons proposal, nor does it mean that she must be irrational. If moral reasons are authoritative, then it is a necessary truth that an agent is rational only if, were she to believe she has a moral reason, she would take this belief appropriately into account in decision-making. This is compatible with the society-centered theory. It therefore is compatible with moral naturalism and, of course, with moral realism. It has often been argued that moral realism in general and moral naturalism in particular cannot account for the normativity of morality. Darwall appears to think more specifically that the realist cannot make sense of the idea that moral reasons are authoritative. The realist must think that there is a “normative world” that guides the moral agent’s desires and choices insofar as these desires and choices are morally

25 The

underlying metaphysics can be developed in different ways. One might invoke a theory of complex propositions, for example, as is developed in King (2007).

162

D. Copp

appropriate. But the realist sees this normative world of moral reasons as independent of the nature of rational deliberation, so the realist cannot explain why rational agents would take moral reasons appropriately into account in deliberation. To be practically rational, Darwall proposes, however, is not to have a kind of normative detection device. It is to have a set of deliberative capacities. To make sense of why it is that rational agents would take moral reasons appropriately into account in deliberation, just in virtue of being rational, Darwall thinks we must suppose that “what makes something a normative reason is that it would. . . motivate in rational deliberation” (Darwall 2006a, p. 304, my emphasis). On the society-centered view, it is indeed correct that the content of the “world of moral reasons” is determined by the content of the optimal moral code, not by the nature of rational deliberation. However, on the authoritative reasons proposal, it is not a mystery why rational agents take into account their beliefs about moral reasons in decision-making. For on the authoritative reasons proposal, one does not count as rational if one is disposed to ignore moral reasons or moral obligations in deciding what to do. If the authoritative reasons proposal is a necessary truth, there is no mystery why rational persons respond appropriately to moral reasons. According to Smith, if there are any moral reasons, then, roughly, the deliberations of fully rational and maximally informed agents would converge on a set of morally appropriate desires and intentions, given full information (Smith 1994, pp. 166–174; Smith 2004, pp. 93, 263). If the authoritative reasons proposal is a necessary truth, convergence is guaranteed by the concept of a rational and fully informed agent, assuming that there are facts about moral reasons. Maximally informed agents would reach the same beliefs about what there is moral reason to do, assuming again that there are facts about moral reasons. Moreover, if the authoritative reasons proposal is a necessary truth, it is constitutive of rational agency to take one’s moral reasons appropriately into account in decision-making. Hence convergence in the moral deliberation of rational agents would be explained by convergence in their beliefs about moral reasons (Copp 2006). If the authoritative reasons proposal is a necessary truth, and if we assume the truth of moral realism, it follows that any fully rational agent who is also fully informed takes moral reasons appropriately into account in decision-making. If I am correct, the authoritative reasons proposal does not make trouble for moral realism, whether naturalistic or non-naturalistic. Indeed, if I am correct, the combination of the authoritative reasons proposal with the society-centered theory enables us to see how a moral realist can make sense of the idea that moral reasons have authoritative normativity. To say that a moral reason is authoritative is simply to imply, roughly, that a rational agent would take this reason appropriately into account in decision-making. There is nothing problematic about being this kind of agent – to be an agent of this kind is simply to be such that, if one believes one has a moral reason or obligation to do something, one takes this into account in decisionmaking in the appropriate way. There is nothing metaphysically queer about being such an agent nor, on the society-centered theory, is there anything metaphysically queer about moral reasons.

Normativity, Deliberation, and Queerness

163

One might object that my argument needs to be backed up by an account of how beliefs about any natural facts, other than facts about our desires, values, or interests, could possibly be irrational to ignore.26 The answer to this objection is simply that the theory of practical rationality is what explains why it is irrational to ignore certain beliefs. The authoritative reasons proposal implies a constraint on the theory of practical rationality. It implies that a person does not qualify as fully rational if she ignores her beliefs about relevant moral facts in deciding what to do. This says nothing about the nature of the moral facts. And of course I am here simply assuming the authoritative reasons proposal is true.

5 Conclusion – Genuine Normativity In previous work, I proposed a neo-Humean, instrumental conception of practical rationality and of the reasons that practically rational persons are disposed to take into account in decision-making, just in virtue of being rational (Copp 2007c). I contended that these are reasons associated with the pursuit of whatever it is that one values. I called these reasons “self-grounded.” I proposed that rationality consists basically in the efficient pursuit of what one values. On this view, self-grounded reasons are authoritative. My account of self-grounded reason is in tension with the authoritative reasons proposal. On the one hand, if the authoritative reasons proposal is true, and if there are in fact moral reasons, then my instrumental account of rationality is incomplete at best. Perhaps a person would be irrational to ignore her self-grounded reasons, but on the authoritative reasons proposal, a person would also be irrational to ignore her moral reasons. On the other hand, on my instrumental account of rationality, it would be surprising and non-trivial if moral reasons are authoritative. It is not plausible that there must be self-grounded reasons to act in accord with one’s moral obligations or one’s moral reasons. So on my account, the authoritative reasons proposal is doubtful at best. To resolve the tension, I am inclined to reject the authoritative reasons proposal. There do not seem to be good arguments in its favor. The proposal is motivated in part by the idea that moral reasons are “genuinely normative,” and in part by the genuine normativity thesis, the idea that any genuinely normative reason is one that it would be irrational to ignore. I deny the genuine normativity thesis, however. I deny it because, together with an instrumental theory of practical reason, such as my theory of self-grounded rationality, it implies that it is doubtful that moral reasons are “genuinely normative.” I think that moral reasons are genuinely normative.

26 One might think more specifically that it needs to be backed up by a showing that facts about the

optimal moral system could possibly be identical to facts that it is irrational to ignore. But recall that the society-centered theory is only being used as an example in my argument. I thank Richard Joyce for pressing me to address an objection of this kind.

164

D. Copp

Our study of Joyce’s argument revealed that, if we think moral reasons are noninstrumental, we must choose between the authoritative reasons proposal and an instrumental conception of practical reason, such as my account of self-grounded reason. Moreover, if we think that there are moral reasons and that they are “genuinely” normative, we must choose between the genuine normativity thesis and an instrumental account of practical reason. These choices will reflect how we conceive of practical rationality, and one might reasonably worry that the underlying issue is at least partly verbal. As Smith has said, “the term ‘rationality’ is almost entirely a philosopher’s term of art” (1997, p. 91; see Joyce 2001, p. 74, n. 26). Philosophers have used the term “rationality” to speak of various standards of evaluation, ranging from an instrumental standard to the morally-laden standard implicit in the authoritative reasons proposal. We obviously can evaluate the performance of a person relative to any of these standards. The question of which of these standards is definitive of full “practical rationality” seems not to be substantive.27 People who are instrumentally rational by the standard I have proposed have a certain psychological profile just in virtue of being this way. They are disposed to deliberate in certain ways. They are disposed to live in accord with their values, for instance. People who are morally good presumably have a somewhat different psychological profile. They find moral considerations especially salient, are motivated to do their duty, and so on. It is unclear that one of these psychological profiles better deserves to be considered definitive of full “practical rationality” than the other. Moreover, it is unclear that one of these profiles better deserves to be taken as definitive of “genuine” normativity than the other. It is unclear why we should think that the question of which of these profiles is definitive of genuine normativity can be settled by a decision as to which of them is the profile of the fully practically rational person. Let us say that reasons that a morally good person would take appropriately into account in decision-making, just in virtue of being morally good, are “morallyauthoritative.” Let us say that reasons that an instrumentally rational person would take appropriately into account in decision-making, just in virtue of being instrumentally rational, are “instrumentally-authoritative.” Then, on the view I would like to defend, self-grounded reasons are instrumentally-authoritative and moral reasons are morally-authoritative. I see no basis for thinking that only one or the other of these kinds of reasons qualifies as “genuinely” normative. Moreover, I see no basis for the thought that the normativity of moral reasons is problematic unless it would be irrational to ignore them in decision-making. This thought accords excessive respect to “rationality.” Acknowledgments Versions of this chapter were presented to the philosophy department at the University of Nebraska, Lincoln, and to the 2008 Dubrovnik Conference in Moral Philosophy. I

27 There

are substantive questions in the vicinity. One such question is whether being instrumentally rational in the sense given, say, by the self-grounded standard I have proposed, entails being morally motivated. I think it plainly does not, but that is another issue.

Normativity, Deliberation, and Queerness

165

am grateful to the audiences on both occasions and also to Richard Joyce, Simon Kirchin, David Sobel, and Jon Tresan for helpful and challenging comments.

References Brink, D. 1984. Moral realism and the skeptical arguments from disagreement and queerness. Australasian Journal of Philosophy 62: 111–125. Copp, D. 1995. Morality, normativity, and society. New York: Oxford University Press. Copp, D. 1997. Belief, reason, and motivation: Michael Smith’s The moral problem. Ethics 108: 33–54. Copp, D. 2006. Review of Michael Smith, Ethics and the a priori: Selected essays on moral psychology and meta-ethics. Mind 115: 476–481. Copp, D. 2007a. Morality in a natural world. Cambridge: Cambridge University Press. Copp, D. 2007b. Moral naturalism and three grades of normativity. In his Morality in a natural world, 249–283. Cambridge: Cambridge University Press. Copp, D. 2007c. The normativity of self-grounded reason. In his Morality in a natural world, 309–353. Cambridge: Cambridge University Press. Copp, D. 2007d. Realist-expressivism: A neglected option for moral realism. In his Morality in a natural world, 153–202. Cambridge: Cambridge University Press. Copp, D. 2007e. The ring of Gyges: Overridingness and the Unity of Reason. In his Morality in a natural world, 284–308. Cambridge: Cambridge University Press. Darwall, S. 1997. Reasons, motives, and the demands of morality: An introduction. In Moral discourse and practice, eds. S. Darwall, A. Gibbard, and P. Railton, 305–312. New York: Oxford University Press. Darwall, S. 2006a. Morality and practical reason: A Kantian approach. In The Oxford handbook of ethical theory, ed. D. Copp, 282–320. New York: Oxford University Press. Darwall, S. 2006b. The second person standpoint: Morality, respect, and accountability. Cambridge, MA: Harvard University Press. Joyce, R. 2001. The myth of morality. Cambridge: Cambridge University Press. Joyce, R. 2006. The evolution of morality. Cambridge, MA: MIT Press. King, J. 2007. The nature and structure of content. Oxford: Oxford University Press. Korsgaard, C. 1986. Skepticism about practical reason. Journal of Philosophy 83: 5–25. Korsgaard, C. 1996. The sources of normativity. Cambridge: Cambridge University Press Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin. Rawls, J. 1993. Political liberalism. New York: Columbia University Press. Skorupski, J. 2007. Internal reasons and the scope of blame. In Bernard Williams, ed. A. Thomas, 73–103. Cambridge: University Press. Smith, M. 1994. The moral problem. Oxford: Blackwell. Smith, M. 1997. In defense of The moral problem. Ethics 108: 84–119. Smith, M. 2004. Ethics and the a priori: Selected essays on moral psychology and meta-ethics. Cambridge: Cambridge University Press. Williams, B. 1981. Internal and external reasons. In his Moral luck, 101–113. Cambridge: Cambridge University Press. Williams, B. 1995. Internal reasons and the obscurity of blame. In his Making sense of humanity, 35–45. Cambridge: Cambridge University Press.

A Tension in the Moral Error Theory Simon Kirchin

There exists a tension at work in the moral error theory that has received little if any attention. In order to make their position convincing moral error theorists must alight on a particular idea and argue that it is a crucial commitment of everyday moral thought and language. Further, in order to be sure that they can convict a commitment of error, error theorists might need to specify that commitment so as to rule out alternative defensible understandings. The danger is that the more one specifies the formulation of a commitment, the more it is likely that the commitment is less than crucial to many people’s everyday moral thought and language. (This need not hold for other kinds of error theory; I will explain why suspecting that this danger exists in the moral realm is justified.) Hence, it is questionable whether error theorists can plausibly convict everyday moral thought and language overall of error. It is the nature of my criticism that things are left open, to be closed only if I consider every candidate crucial commitment. I do not exclude the possibility of a commitment that is universally (or near-universally) accepted, is crucial to everyday moral thought and language, and is false. My aim is not to point out a hitherto unnoticed and fundamental contradiction within the moral error theory that shows it to be fatally flawed. Instead, I expose a worrying tension and advertise my hunch that no error theorist has yet resolved it. Despite my qualms, some error theorists might choose to live with the worry. In Section 1 I describe and discuss the tension generally, and in doing so articulate a view of our everyday moral thought and language. In Section 2 I reflect on work by John Mackie and Richard Joyce who concentrate, in different ways, on the thought that categorical imperatives, conceived in a certain fashion, are a crucial commitment of everyday moral thought and language. In Section 3, by way of conclusion, I briefly consider again the rejoinder just mentioned, that I leave things open.

S. Kirchin (B) Department of Philosophy, University of Kent, Canterbury, Kent, UK

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_10, 

167

168

S. Kirchin

1 The Error Theory and Everyday Moral Thought and Language 1.1 The Error Theory A moral error theory (hereafter just “an error theory”) consists of two main elements: (i) a characterization of everyday moral thought and language that picks out some commitment1 that is embodied by such thought and language (or that is based on such thought and language, or is directly implied by it, or similar), and which, in turn, is claimed to be crucial (or central, or key, or fundamental, or similar); and (ii) some argument (or arguments) that purports to show that this crucial commitment is false (or fundamentally misguided, or completely unjustified, or wildly crazy, or similar) such that the whole of our everyday moral thought and language is flawed, not just the commitment in question. In short, our everyday moral thought and language is in error since one of its key commitments is wrong. Furthermore, error theorists might add that the rotten nature of moral thought and language should lead us to consider strongly the possibility of refraining from using such thought and language, at least sincerely. The seeds of the tension are found in the very aspects that define the error theory. The commitment that error theorists focus on has to be both essential to moral thinking (being moral thinking) and erroneous. The essential nature of the commitment means that the error of the individual commitment transfers to all everyday moral thought and language. But finding something that is both fundamental to the whole and wrong-headed is a tall order in the moral case. As I show in Section 1, there are examples of commitments in other sorts of thought and language that satisfy these desiderata. What makes the moral stand out from them is the fact that moral thought and language is messy, disorganized, cluttered, and lacks a unique and specific identifying point. There might be nothing that is fundamental in the necessary way, or if there is it is so general that one has to specify it so as to argue that it is erroneous and in doing so one invites other plausible specifications that are equally legitimate. In the next subsection I sketch why one might believe the moral is messy in this way. Before that, I reflect a little more on the structure of error theory, mainly to show why the messiness of the moral is so significant. It is important – indeed essential – that error theorists describe everyday moral thought and language correctly. They are claiming that everyday moral thought and language itself is questionable, and so their description of it has to be accurate. In this way the error theory differs from its familiar rivals. Realists do not have to argue explicitly from the (supposed) realist-like sympathies in everyday moral thought and language to their view. The same goes for expressivists, relativists, and the rest. Of course, realists would look odd if they did not use everyday moral thought and

1 I use “commitment” throughout as a neutral catch-all. “Assumption,” “claim,” “idea,” “intuition,”

and the like work equally well.

A Tension in the Moral Error Theory

169

language in support. But it is not definitive of realism and these other positions that they do so. Another way in which error theorists are distinguished from most of their rivals concentrates our attention on the move from one commitment to the whole. Whilst most other positions take a positive stance towards everyday moral thought and language and aim to vindicate it overall, the error theorist views it negatively. Realists can highlight and pick apart certain commitments that cut against their realism. But they do so only to defend their position and moral thought and language overall. They do not have to think that the existence of (as they see it) problematic commitments tells against everyday moral thought and language as a whole. Likewise for expressivists. Indeed, some relativists might aim to vindicate everyday moral thought and language if they think there are far stronger and more relativistic commitments in play than non-relativists often presume.2 But the error theory is different. A typical error theorist focuses on one or a few key commitments, argues that these are misguided in some way, and thus seeks to cast doubt not only on those commitments but the whole enterprise. This aspect of error theory is a contingent matter. It is a contingent fact, albeit a widespread one, that everyday users of moral thought and language cannot typically be characterized as having (inchoate) error theoretic commitments. Instead, they think that what they are doing, broadly, is expressing moral ideas about the world that are cogent and have application. If “pretheoretical” error theoretic ideas were common, then the picture would be more complex: error theorists would be positively vindicating a commonly held belief that everyday moral thought and language is bogus. But such an attitude is rare, hence the error theory’s negative air. This radical rejection – of the whole based on one commitment – is worth emphasizing. Consider this from Joyce: [T]he error theorist must establish that moral discourse is centrally committed to some thesis X. The phrase “centrally committed” is supposed to indicate that to deny X would be to cease to participate competently in that discourse. (Joyce 2007, §4.)

For Joyce, it is not just that error theorists have to “rethink greatly” everyday moral thought and language. The commitment isolated by error theorists and shown to be wrong is so important that if one does not hold to it one is no longer thinking morally. Throughout his 2001 book Joyce refers to such crucial commitments as conceptually “non-negotiable.” I venture that Mackie thought the same in his Ethics and other works, and that all other prominent error theorists think similarly. Assuming that everyday moral thought and language incorporates many sorts of moral commitment, it seems probable that all standard metaethical positions are revisionary in a certain fashion. For example, realists will downplay those commitments and phenomena that can be used to support (traditional) expressivism, whilst

2 This

is not true of all relativists, hence my claim in the previous paragraph that error theorists differ from most, not all, of their rivals.

170

S. Kirchin

their opponents must explain away any realist-like commitments. But the error theory represents a radical form of revision. Error theorists try to dispense with the whole of moral thought, not just the individual commitments they think bogus. So, the structure of error theory sets it apart and creates distinct challenges. Any evidence showing our moral thought and language to involve a range and variety of commitments threatens to undermine the idea that any one of them is essential to moral thought and language. And, if one commitment does stand out as essential, it might be so general that a number of legitimate readings of it are possible, some of which render it perfectly unproblematic. This remains an abstract suspicion for the moment. I now turn to flesh it out.

1.2 Everyday Moral Thought and Language For reasons of brevity, I have used “commitment” as a catch-all phrase, but we should seek a little more precision of the commitments that error theorists focus on. Below I sketch three (overlapping) examples to illustrate the tension facing error theorists. In brief: (a) Error theorists often focus on the supposed commitment that moral discourse is assertoric and fact-stating in nature. (b) They often focus on the supposed commitment that moral demands are of a distinctive type. (c) They often focus on the supposed general metaethical commitments that can be read into everyday moral thought and language. In each case, however, different views of the phenomena in question can be provided, casting doubt on whether error theorists have highlighted an essential commitment of everyday moral thought and language. For each case – in the final paragraph of each discussion – I also briefly discuss a generally worded and relevant commitment that might be seen as essential, but which is far from clearly false. (a) Think about the grammatical structure of everyday moral thought and language. Error theorists are typically committed to the idea that any moral claim has to be expressed or be easily expressible in a truth-apt form, so that it can then be shown to be false. There might be quite a few straightforwardly assertoric moral utterances. Those that are not, for example some instances of “Don’t do that!”, can easily be rejigged, in this case to convey the utterer’s belief that the possible course of action is impermissible. But the error theorist is now in a process of extrapolation. This becomes apparent in the case that many error theorists discuss, namely the concept of moral value or reason that is supposedly at the heart of everyday moral thought. Concepts themselves cannot strictly be true or false, and it is unlikely that typical users of moral language ever explicitly say, “My concept of a moral value is [such-and-such].” So, as I say, error theorists will have to extrapolate from everyday moral thought and language. What is key for some error theorists, given other things they say, is that everyday moral thought and language be committed to stating facts about the world. There have to be enough expressions in moral language that function descriptively and

A Tension in the Moral Error Theory

171

enough of the rest (if there are any) that, even if not literally descriptive, can be tweaked so that they function as descriptions, where such tweaking does not distort speakers’ intentions and potential hearers’ inferences. The functioning of moral thought and language is the main battleground between error theorists and expressivists. Joyce, for example, has strongly criticized expressivists for failing to do justice to the fact that much of everyday moral thought and language seems to function descriptively, or can be read in this way perfectly justifiably. (Joyce 2001, pp. 9–16.) To respond, expressivists have to explain why everyday folk use language that typically functions descriptively without really meaning to describe anything. Just as error theorists might extrapolate from everyday language in one direction, expressivists might extrapolate in another. They often argue that what appears fact-stating language is really a disguised form of attitude expression (or similar). Often the distinction between “surface” and “depth” grammar is invoked. Typically expressivists wish us to recall how important the motivational aspect of our moral lives is. They argue that this aspect is difficult to capture with language that is assumed to function solely as descriptive. They argue, further, that if one loses the motivational aspect of the moral, then one might lose the moral altogether. The debate between expressivists and error theorists runs and runs.3 What I draw out is this. Everyday moral thought and language can be and is expressed in many different forms. Error theorists might well emphasize the large number of descriptively-expressed claims, and might well tweak other utterances so that they function to state facts. But, there are many common examples where people express moral ideas in non-descriptive ways, and where changes to descriptive-functioning language are strained. One-word exclamations, approving “hmmmm”s, various types of demand, mild praisings, and many other examples are often used to express noncognitive states, such as expressions of motivational commitment (be they deep-seated or whimsical), and imperatives of different sorts. It is dubious to assume that all moral utterances can be interpreted as surrogates for fact-stating claims since in many cases something will be lost in translation. What of the general claim? Perhaps we can state that, despite this variety, moral thought and language embodies some very general and essential commitments on this matter. For example, it seems an essential commitment of everyday moral thought that people are attempting to articulate their moral view

3 Some further moves. Expressivists can attack by asking why so many people for so long could be

so wrong about a key aspect of their lives, as error theorists claim. Perhaps we should revise the error theory’s starting assumption. (See Blackburn 1993, pp. 149–150; Hare 1999, Section 1.1.) Error theorists can counter by saying that the same can be said about theistic religion, which many question. They might also claim, controversially, that whether or not people mean to, if they use language that functions descriptively “on the surface,” then they are describing.

172

S. Kirchin

of the world when thinking and communicating with others. That seems something that expressivists, realists, and error theorists can agree on. But that is not something that is clearly strange or bogus. (b) Having sketched an idea about the form and function of our moral thought and language, what of the content and the ideas expressed? Consider the following idea that many error theorists cause trouble for: “Imperatives and demands that have a claim on all are essential to the moral life being as it is.” This assumes on behalf of everyday moral thought and language a particular view about demands and, as part of this, it assumes that everyday moral thought and language is pretty clear on what is and what is not part of the moral life. But, this can be questioned. We can begin by noting that the moral life encompasses many different sorts of idea. For example, we are familiar with the coarse-grained distinction between deontic and aretaic ideas, which in turn hides many shades of gray. Even with only this broad distinction, the case of moral demands is thorny. Whilst such things supposedly fit nicely on the deontic side of things, their place in everyday aretaic thought is harder to discern conclusively. Imagine someone judging someone else to be kind. How should we understand this? Is such a judge to be always or typically understood as directly demanding of all similarly placed others that they follow this person’s lead? Or is the presence of such demands always partly determined by context, and often not present? Is the expression of this idea, indeed all moral ideas, always for the benefit of others, or often for the benefit of oneself alone, to help create a social world (as it is sometimes grandly put), or just to let off steam? Indeed, these questions are affected by our ability to demarcate the moral first of all. Kindness, bravery, and wickedness are supposedly straightforward moral concepts. But talented novelists, for example, often describe characters and situations in terms not overtly moral and still convey ethical ideas. Protagonists can be described as “selfconfident” and “cocksure,” or “diffident” and “timid,” say. In some contexts, these can be clear moral characterizations. This matters, for if self-confidence, say, does turn out to be a moral concept of a sort, then we need to consider our previous questions. When we say of someone that he should be more selfconfident, should that typically be understood as implying a general moral claim that everyone similarly placed should be so? Probably not. Indeed, sometimes there might be no demand placed at all on the individual. We might have merely a piece of throwaway personal advice.4

4 For

reasons of space my discussion is a sketch. See Crary (2007) for detailed discussion of what I have in mind and Kirchin (2008) for my review. The example of self-confidence comes from Gilbert Ryle’s discussion of Jane Austen, via Crary. We could add many to such a list: ratty, selfindulgent, gutsy, obsequious, and so on. The idea that everyday moral thought lacks a narrowly distinct point and function is also a recurring theme in Thomas (2006).

A Tension in the Moral Error Theory

173

What of a candidate general claim? Something such as “It is important to many aspects of the moral life that there be demands that often apply to others” might well be true. But, as we had previously, this does not seem to be something dubious. (c) Sometimes error theorists focus on a broad sort of commitment. Instead of thinking narrowly about demands, they might claim that everyday moral thought and language is committed overwhelmingly to realism. On this basis they might make trouble for defenders of morality. Error theorists are not alone in focusing on theoretical commitments. Many metaethicists draw out ideas from everyday moral thinking that bear close resemblance to familiar theoretical positions. Once we get past the seeming oxymoron, we can readily agree that everyday thinkers have “pretheoretical metaethical” ideas as to what they are doing when they think and speak morally. Such ideas might be more or less vaguely expressed, more or less sophisticated, and more or less consistent with everyday thinkers’ other commitments, but they exist nonetheless. Error theorists might make trouble for the realism they find in our everyday moral thought and language. But, whilst it is hard to deny that our everyday moral thought and language has realistic strands, it is unclear how widespread and fundamental it is, particularly when one factors in the point – hitherto in the background – that people hold different views. Some people might be hardened and clear-sighted realists about all issues. (Some religious fundamentalists come readily to mind here.) Others might be far more relativistic about many issues, whether or not they realize it. Furthermore, I suspect that most people are mixtures across a range of issues. Perhaps some people believe that whether or not you agree, and no matter what your culture, murder (or this sort of murder) is wrong. But such people might also be ready to acknowledge the legitimacy of views different from their own regarding certain “lower-level” types of behavior, such as how children treat adults and talk with them in their everyday affairs, and how people treat each other in their workplace and in romantic relationships. Such people are ready to adopt a “live-and-let-live attitude” on these and other matters. Even if we accept that much of moral thought and language is realist in character, I find it hard to believe that it is overwhelmingly so. If it were, then it might well be difficult for expressivistic anti-realism and relativism to get any purchase. Indeed, even if we accept that much moral thought and language is pretheoretically realist in nature, it immediately invites the question “What do you mean by ‘realist’?”5 We, as professional philosophers, are familiar with many different conceptions of what it is for something to be real, and these conceptions will typically have their anchors in everyday thought and language. Many people might think that the laws and moral norms surrounding torture, for example, are objective and based on something “real.” But, once one focuses

5 The

classic response to Mackie along these lines is found in McDowell (1985).

174

S. Kirchin

on this supposed consensus, it is apparent that whilst some everyday folk form views that are best interpreted along mind-independent lines, others are best interpreted as favoring a form of mind-dependence. Indeed, after investigation the supposed realism of those in a third group might turn out to be an inchoate form of constructivism. So the claim that everyday moral thinking is overwhelmingly realist might be false, and even if much of moral thinking is realist, this claim will have to be narrowed. Some views of moral values and reasons being real will render such values and reasons suspect. But not all will. Hence, we should not be so troubled by the error theoretic worry that everyday moral thought and language is dubious because it is realist. Of course, a general commitment such as “Everyday moral thought and language aspires to be about our moral lives and the world we live in” might be true. But, again, this seems something that an expressivist or a relativist could agree with. These three sketches introduce us to the thought that the moral life can be and is presented by metaethicists in different ways. This striking fact is worth emphasizing. Not only do metaethicists present different views of the moral life, there is a lack of agreement about what are the central and important commitments of our moral life. (Metaethics is not unique in modern philosophy in this regard.) Note that I am not claiming that we should be highly suspicious of the error theory, or indeed all metaethical views, because the different descriptions provided of the moral life show that no one metaethical view is correct. Rather, I think we should exercise a little skepticism before accepting descriptive claims about, for example, the overwhelmingly realist character of everyday moral thought and language. (See also Loeb, this volume.) Such descriptions might be skewed towards the theoreticians’ argumentative aims. We should also exercise similar skepticism about the philosophical claims made about such commitments. Very often metaethicists assert that the commitments they have focused on are absolutely central; other times they argue that whilst a commitment might initially seem key, it is in fact not so. Such claims about the status of commitments often have merit. But, again, it is striking that metaethicists arrive at different views about what is and what is not central to our everyday moral thought and language. Note also that although I am concentrating on everyday moral thought and language, I am not claiming that as metaethicists we should have concern only for what the folk think. Technical arguments, sophisticated ideas, and things that come only with philosophical expertise have to play their part. What I do think – but which I do not argue for in detail here – is that metaethical positions and arguments make sense only against a background where everyday people are thinking and acting in ways that connect with the theoretical positions in play. This section has sketched only three examples of the source of the tension in the error theory. To prove my claim that the moral life involves many different sorts of commitment, and from this show that the error theory is itself in trouble, I would need to provide some large sociological-cum-philosophical account. However, although error theorists might draw attention to this absence in my

A Tension in the Moral Error Theory

175

account, they can make good on their position only if they alight on a central commitment that satisfies the desiderata I mentioned at the start. I reckon that my examples are sufficient to make us doubt that it will be an easy task to find such a commitment. It is now time to see how the tension plays out in two specific error theoretic arguments, where we will find distinct echoes of my three examples.

2 Mackie and Joyce 2.1 Mackie Mackie famously raises two arguments against the idea of “objective prescriptivity,” an idea which he thinks is crucial in everyday moral thought and language. Most commentators concentrate on his “Argument from Queerness,” and I follow suit. The central charge is that objective prescriptions are metaphysically queer entities that do not sit well in our natural world. There is an accompanying worry that we would need a strange faculty to perceive them. It is clear from the few pages that constitute Ethics Chapter 1 Section 9 that by saying of something – in this case demands, reasons, and the like – that it is “objective” Mackie means to say that it is independent of people’s responses, desires, and commitments. In contrast, he has no problem with the notion of demands and values that arise through institutions created and sustained by humans. Characterizing further what this means is tricky. There are two main possibilities. Does he think that the only legitimate demands for an agent are those that are generated by certain institutions that the agent herself recognizes as being authoritative and valuable (or, where one could reasonably make a link between the agent’s present commitments and the institution)? Mackie might often be thought to be denying the legitimacy of demands that are not linked in the relevant way to the relevant agent. This first option places him squarely in the same territory as Bernard Williams’ denial of external reasons (Williams 1981; see Phillips, this volume). But a reading of Chapter 1 Section 7, for example, suggests a second option: that he is more liberal, and is willing to countenance legitimate demands existing for an agent that have little or nothing to do with her S, to use Williams’ language.6 (See also the Introduction to this volume.) On this reading, Mackie is doubtful only of demands having authority external to any human-created institution, but is willing for other humans’ wishes and commitments to ground the demands placed on individuals. The idea of “prescriptivity” is also hard to untangle. The term is open to interpretation and Mackie himself muddies the waters in his presentation. In brief, we face two main choices. Assuming an extreme conception of objectivity for now,

6 “S” stands for the agent’s subjective motivational set, which “can contain such things as dispositions of evaluation, patterns of emotional reaction, personal loyalties, and various projects, as they may be abstractly called, embodying commitments of the agent” (Williams 1981, p. 105).

176

S. Kirchin

he may be entertaining the suggestion that if φ-ing is deemed morally right, then (i) there is a reason for us to φ, and (ii) this reason (or its ground) exists independently of what humans think or could think about φ-ing in this case.7 Alternatively, he might be imagining a moral property or reason that, in addition to (i) and (ii), will (iii) itself cause motivation in humans to φ. That is, we are not simply presented with a (response-independent) reason to φ, but the action, or property, or something, will cause suitable motivation in us towards φ-ing. (There is a further question. We might specify (iii) to hold that “the judgment that such a reason obtains will cause motivation in humans to φ.” For there is the further possibility of (iv): that a reason or property, even when an agent is unaware of its existence, can cause said motivation. This last option is extreme, and no commentator attributes it to Mackie to my knowledge. See Sinnott-Armstrong, this volume, Section 4.) Both interpretations of prescriptivity have been voiced.8 If we ignore the debate about whether to include (iii), and if we focus on Mackie’s worry with legitimate demands that are “institution-transcendent,” it is clear that the challenge is strong. The notion of an objective prescription is strange. How could there exist demands without a demander? How can the idea of a reason be something that exists response-independently? How can we account for the existence of values (which might generate or ground reasons) whilst ignoring humans, the valuing creatures?9 It is as if we are countenancing an Alice in Wonderland world where actions have (figurative) labels that state “Do me!” or “Refrain from doing me!” – where the creator of such labels is no human. One could possibly deflect the challenge by postulating the existence of a god or a race of value-creating Martians, but only at the expense of raising more problems. And all of this seems very queer because such things would be utterly unlike anything else we assume exists. Many things stand

7 Presenting the idea in terms of the “present case” is enough for our purposes. If inclined, we could

add complexity and talk about φ-ing and reasons to φ “generally.” 8 Garner (1990) and Joyce (2001) favor the first, and my presentation is influenced by them, whilst

Brink (1984) favors the second. Other writers that have assumed the second interpretation include Crary (2007, pp. 12–13), and Thomas (2006, pp. 14–18). A little reflection shows a third option. The introduction of clause (iii) suggests that motivation will be caused once an agent recognizes the reason. We might weaken this and say only that recognition is able to cause motivation, and cause it directly, if other factors are present. This weaker view is compatible with Garner’s and Joyce’s views; indeed, Joyce (2001, Chapter 5) suggests it. That said, for both it is the responseindependence of demands that is key when thinking about “queerness.” Philip Stratton-Lake’s discussion (2002, pp. 11–12) is arguably compatible with all three views. 9 I switch between “reasons,” “demands,” “oughts,” and “values,” as Mackie and commentators do. Such switches are often less than innocent, but here there is no danger. Mackie’s overall thought applies not just to claims involving “ought” and reasons but also to the action-directing nature of values such as cruelty and kindness, two examples he often uses. So, claiming that some action is kind can often imply that one has some moral reason to do it, and claiming that someone is kind can imply some reason to emulate them. Mackie argues that this reason cannot be of the categorical kind. It is the last claim that is key, so any differences between reasons and values can be ignored whilst we fry other fish.

A Tension in the Moral Error Theory

177

out as being different, of course. Quarks are strange. But objective prescriptions are different in a different and dubious sort of way. There are responses to this. One might question whether Mackie has scientistic prejudices, or question whether objective prescriptions are so strangely different. I am unconcerned here with pursuing such responses, although I think many are good and that Mackie’s argument is flawed in this respect.10 What I highlight is that Mackie, and any other error theorist, should decide which conception of objective prescriptivity he wishes to discuss. This exemplifies the tension I have discussed.11 The two pairs of interpretations of the key terms generate four understandings of “objective prescriptivity.” It is fairly easy to argue that demands that both motivate people and whose authority is grounded in something non-human are queer things. But to what extent is such an idea part of everyday moral thought and language? More specifically, even if it is part of some people’s vague thoughts, at some time, is it widespread, and is it crucial in the way error theorists need? Could moral thought and language continue in some recognizable form if we eschewed it? I suspect that this conception is not as prevalent as Mackie supposes, and that moral thought and language could survive its absence. A different reading of “objective prescriptivity” offers a different scenario. Perhaps the demands do not motivate at all, even amongst those that recognize them. Assuming this motivational-externalist moral realism is just what David Brink does (Brink 1984). Richard Garner chides him for ignoring – as he sees it – the main queerness, namely the response-independence of the demands. But the power of that counter fades if one admits a different reading of objectivity, one that allows for the legitimacy of demands by tying them to human-created institutions generally, rather than the commitments of particular agents for whom the demand supposedly arises: the grounding of the demands is independent of the individual, not humans as such. To be sure, every so often some people might say things that commit them to supposedly queer response-independent reasons and values. But everyday moral thought and language might be able to cope without this claim. It probably could not cope without the weaker version of objectivity, or something like it. But, then, this version does not seem such a conceptually queer idea.12

10 Shepski (2008) offers a good summary and exploration of such responses, as well as an excellent

dissection of the argument. with which conception of objective prescriptvity Mackie preferred; I strongly suspect he was not clear on the distinctions. 12 A contrast. One could endorse Williams’ attack on external reasons. We might question whether we can talk of there being legitimate reasons to act that are not reasonably linked to an agent’s S, even reasons which many others endorse. But note that Williams was not an error theorist. He thought that he was attacking one strain in everyday (and philosophical) moral thought, but once this was jettisoned moral thinking and activity could still survive. 11 I am unconcerned

178

S. Kirchin

2.2 Joyce Joyce advocates a type of moral skepticism motivated by many of the same considerations that move Mackie. Here I summarize and endorse an argument against him from Russ Shafer-Landau, then add my comments. Joyce, throughout The Myth of Morality, denies that there is anything with categorical reason-giving power. The following quotations summarize his view, the second coming when he considers his argument complete. When I said earlier that morality is not merely a set of Dos and Don’ts that we are willing to back up with force, I meant that morality is something more than a set of weak categorical imperatives [i.e., imperatives that say we ought to do something, where we have accepted the authority of the institution that generates them]. Reflect on how differently we treat “Celadus [an imagined gladiator] ought not to throw sand” and “Gyges ought not to kill people.” We’re content to admit that the former case is just a matter of there being a set of rules which someone from the outside is imposing on the gladiator, and these rules can be overridden by the gladiator’s personal desires and interests. They need not bind him; they need not be his rules; they do not present the thing to do; he may legitimately ignore them. But imagine saying something analogous of Plato’s shepherd: “Of course by killing an innocent person the shepherd is breaking a rule of morality, and so according-to-therules-of-morality he ought not do it; nevertheless, if he stands to gain something important by killing, then that’s what he ought (all things considered) to do.” That’s not how we think of morality. Someone who reasoned in such a way might be accused of fundamentally misunderstanding what we mean by “morally ought.” (Joyce 2001, pp. 36–37; italics original) In short, when we say that a person morally ought to act in a certain manner, we imply something about what she would have reason to do regardless of her desires and interests, regardless of whether she cares about her victim, and regardless of whether she can be sure of avoiding penalties. And yet after careful investigation we have found no defensible grounds for thinking that such reasons exist. Few people in the actual world may be so heartless or so impregnable to recrimination, but that is beside the point. Moral judgments are untrue not just because they sometimes ascribe reasons for (say) honesty to people who have no such reasons. They are untrue even when they ascribe reasons for honesty to people who do have reasons for being honest, in that they imply that those reasons would remain in place across counterfactual situations when in fact they would not. The distinctive authoritativeness which characterizes our moral discourse turns out to be well-entrenched bluff. (Joyce 2001, p. 134; italics original)

Shafer-Landau worries whether it is a conceptually non-negotiable commitment of morality that morality is or supplies a raft of categorical reasons to act. He notes, first, that certain relativists and subjectivists, amongst others, claim to be interpreting moral thought and language in a certain way free from the authoritative categorical imperatives to which Joyce and Mackie take exception, but that they claim they are still discussing moral thought and language (Shafer-Landau 2005, pp. 109–110). The error theorist must maintain that such relativists and subjectivists have changed the topic that is being disputed. But the point is that it is not clearly and unarguably true that these theorists have done so. Shafer-Landau then discusses less extreme sorts of theorist as a way of challenging whether morality is categorically reason-giving. He focuses on Joyce’s

A Tension in the Moral Error Theory

179

assumption that “it is necessary and a priori that, for any agent x, if x ought to φ, then x has a reason to φ.” Joyce labels this Mackie’s platitude (Joyce 2001, p. 38). This platitude is used in support of Joyce’s crucial claim that if moral duties are categorically applicable, then they supply categorical reasons for obedience. (He, of course, denies their existence.) Shafer-Landau, however, contends that this is no platitude at all; in fact, it is contentious. Suppose Joyce is right in regarding [the claim that a person can be morally obligated to do something even if doing it fails to serve any of her interests, etc.] . . . as conceptually nonnegotiable. Thus people see morality as imposing duties whose fulfillment can, in cases, only frustrate desire and stifle self-interest. Those who [then] query whether there is reason to undertake such sacrifice don’t seem to be engaged in verbal trickery. Knowing that doing your duty will bring you only misery, you reasonably ask: why do it? It’s possible, as Joyce alleges, that there’s no good answer to this question. But surely it is a sensible one to ask. (Shafer-Landau, 2005, pp. 110–111)

Shafer-Landau immediately notes that whilst Philippa Foot agreed that morality purports to be categorically applicable and argued that it was not, she was no error theorist simply because she denied Mackie’s platitude (Foot 1972). Shafer-Landau claims that Joyce is committed to thinking that Foot has therefore changed the subject and is no longer talking about morality, but this, thinks Shafer-Landau, is surely a “verdict of last resort.” He then lists others who, in his view, have seriously considered whether we necessarily have reason to do our moral duty: Mill, Sidgwick, Gregory Kavka, David Brink, Peter Railton, and David Copp. He also claims that one way of viewing much Western moral theorizing is as a series of attempts to convince the amoralist that she is wrong, the amoralist being someone who acknowledges that she has a categorically applicable moral duty, but who sees no reason to comply. He ends by noting one way out of the predicament for Joyce. Perhaps we should restrict Mackie’s platitude only to all-things-considered “ought”s. It is only these “ought”s that would entail reasons. But, thinks Shafer-Landau, we might well not agree that moral “ought”s are all-things-considered “ought”s. There are some who think that moral “ought”s can be outweighed by other sorts of “ought.” All of this is excellent evidence, thinks Shafer-Landau, to hold that Mackie’s platitude is no platitude at all. Hence, we can question whether the claim that moral duties supply categorical reasons for obedience is a conceptually non-negotiable commitment of morality. I share Shafer-Landau’s views, although I acknowledge that moves might be made on Joyce’s behalf. Error theorists might seek to water down the non-negotiable commitment: perhaps moral duties do not have overall, unquestionable authority, but have some authority nevertheless. Yet we are then into murky waters. “Moral demands have some authority over the individual beyond what the individual thinks, although at times not indefeasible authority,” looks like something that could be part of some people’s everyday conception of morality. But it might not entail anything conceptually queer. And, anyway, we might also say that this would certainly not be part of some people’s moral thinking. Ironically, some might hold to the more extreme sort of authority that Joyce and Mackie find questionable.

180

S. Kirchin

And this is my overall point. We are in the same territory as my discussion of Mackie, and my discussion in Section 1. Error theorists have to narrow their conception of what is crucial to morality being what it is. We have witnessed two kinds of retort to the error theoretic claim that moral thought and language is to be eschewed because it is based on a narrowly conceived and supposedly rotten idea: “Yes, morality is based on that idea, but the idea is not rotten” and “Who says that morality is like that, or that this admittedly crazy idea is so crucial anyway?” My thoughts can be compared with examples often given by Joyce amongst others: cases such as phlogiston discourse and witch discourse (Joyce 2001 pp. 1–5, 156–157). We can perfectly well understand claims such as “That stuff is phlogiston,” or “She is mixing a potion containing phlogiston,” or “Witches exist,” or “The witch wearing gray stands out from the others wearing black.” The first pair are false, to all right-thinking people, since there is no such thing as phlogiston. (Or, at least, we have no reason to believe these claims.) And, as Joyce says: [T]he discovery that we had been wrong in thinking that there is stuff stored in combustible bodies and released during burning was sufficient for us to decide that there is no phlogiston at all. When Lavoisier gave us the concept oxygen, it wasn’t available for Stahl to say “Well, this stuff that Lavoisier is calling ‘oxygen’ just is what I’ve been calling ‘phlogiston’ all along – I was just mistaken about its being stored and released during combustion.” (Joyce 2001, p. 4.)

What is key to the rejection here, as depicted by Joyce, is that there is a clearly defined commitment or commitments that are clearly seen as crucial to the concept. (He points out (Joyce 2001, p. 28), that Lavoisier did worry whether his rivals had made the concept so nebulous that in the end rejection might be difficult because it was unclear what one was rejecting. However, after reflection one can see that there are some core parts of the concept, and from this one can show it to be bogus.) Joyce thinks the same is true of the witch case, and that our third and fourth claims are also false. We simply have no reason to believe that people can fly on broomsticks or literally perform magic. If such things cannot be done then there are no witches, or at least none of the people that Joyce has in mind exist. We can imagine someone responding that there are in fact quite a few “white witches”, and various other kinds of wise woman, healer, shaman and the like, who practice something they call magic. Such a person might on these grounds claim that the third and fourth claims are true. But this suggestion changes the game perhaps, as Joyce imagines Stahl doing. Or better, I think, we might say that there is more than one meaning of “witch,” and these meanings and associated concepts are distinct. The fact that one concept is shown to be bogus need not threaten the legitimacy of every form of witch discourse. The comparison with the story I have been telling with moral thought and language is revealing. In the phlogiston case we go past much bluster and conflicting claims about phlogiston, alight on and specify some crucial claims, and show the concept is bogus, thus forming an error theory of phlogiston discourse. We do the same in the witch case, possibly distinguishing one sort of witch from another sort, with some forms of witch discourse shown to be in error. The moral case presents a different challenge. If my hunch is right, even after hard reflection there is no clearly defined commitment that is clearly seen as crucial to

A Tension in the Moral Error Theory

181

everyday moral thought and language such that if it could be shown to be false then the whole show collapses. It is not even as if we have a number of different moral discourses (bar idiosyncratic cases associated with just a few people) and that on a sympathetic reading Mackie and Joyce have shown these discourses to be illegitimate. Rather, the ideas in the moral realm tumble across one another, and this makes the error theorists’ job a lot harder than in the other two cases. As I have said, the tension is not insurmountable. In the phlogiston and witch cases it can be so easily avoided that it goes unnoticed: the commitment that is specified and then shown to be false can and does remain as something core to everyone’s thinking and talking about phlogiston and witches. In fact, far from there being something small that goes unnoticed, there is no tension, really. On the other hand, the vague and messy nature of the moral case makes the task of finding the lynchpin commitment far more challenging.

3 A Quick Objection and a Conclusion I have focused on Mackie and Joyce, and before that on the related realist nature of demands amongst other things. Perhaps this is unfair. There are surely many other commitments that error theorists could pick. So, how much credence does my worry warrant? I raised this worry in the Introduction and admitted that I leave things open. Let me draw out a few things, however. First, the commitments that I have focused on are decidedly prevalent in key error theoretical writings. Secondly, I allow that any particular error theorist might discuss at length a whole raft of dodgy commitments, although curiously all such theorists (to my knowledge) tend to focus on just one. Even if they were to focus on a few, then they would need to make sure that the tension does not arise for each commitment singularly.13 Lastly, I maintain that it would have to be a special sort of commitment to satisfy the desiderata mentioned. By way of illustration, here is a quick account of how things might go for a different error theory. An error theorist might say, “It seems that everyday moral thought and language assumes that we are all free beings, otherwise one cannot make much sense of responsibility, for example, and hence of praise and blame. But there are good arguments against conceiving of ourselves as free. Therefore, everyday moral thought and language is in error.” The tension repeats. For, as we all know, one can mean many different things by “free being” and “determined,” and different conceptions admit of and imply different conceptions of “responsible moral agency,” “blameworthy action,” and the like. Focusing on the right conception or conceptions of freedom that are all shown to be rotten and that also knock out any notion 13 That

is, unless they detected interesting conjunctive, disjunctive and conditional relations between them. Perhaps a little doubt spread across a number of key commitments might add up to a large doubt about them all, which in turn might put the whole of moral thought and language in doubt. This is an obvious possibility. I note merely that it would be an extra task to argue that such structured commitments were part of our everyday moral thought and language.

182

S. Kirchin

of moral responsibility such that everyday moral thought and language collapses might be hard. Again, as above, I do not claim that it cannot be done. But there is a large hurdle to be jumped. And this hurdle is created in part by the structure of error theory itself, or so I have argued. The link between individual everyday moral commitments and the legitimacy of the whole of moral thought and language is crucial to the error theory. But this link is also its Achilles’ heel. The tension I have discussed lurks in the background of every version of the error theory that has been articulated in the literature. Furthermore, if one suspects that everyday moral thought and language is so messy, and one suspects that it is controversial as to what exactly lies at its core, then one should conclude that it is very likely that no one commitment could possibly be both unarguably central to moral thought being as it is and unarguably wrong-headed. Acknowledgments Thanks to the philosophy departments at Hertfordshire and Sheffield, and to Tim Chappell, Richard Joyce, Jimmy Lenman, Russ Shafer-Landau, Bob Stern, Alan Thomas and an anonymous referee for comments on presentations and previous drafts.

References Blackburn, S. 1993. Errors and the phenomenology of value. In his Essays in quasi-realism, 149–165. New York: Oxford University Press. Brink, D. 1984. Moral realism and the skeptical arguments from disagreement and queerness. Australasian Journal of Philosophy 62: 111–125. Crary, A. 2007. Beyond moral judgment. Cambridge, MA: Harvard University Press. Hare, R. M. 1999. Objective prescriptions. In his Objective prescriptions and other essays, 1–18. Oxford: Oxford University Press. Foot, P. 1972. Morality as a system of hypothetical imperatives. Philosophical Review 81: 305–316. Garner, R. 1990. On the genuine queerness of moral properties and facts. Australasian Journal of Philosophy 68: 137–146. Joyce, R. 2001. The myth of morality. Cambridge: Cambridge University Press. Joyce, R. 2007. Moral anti-realism. Stanford Encyclopedia of Philosophy. http://plato.stanford.edu/archives/fall2008/entries/moral-anti-realism/. Accessed 17 October 2008. Kirchin, S. 2008. Review of Beyond Moral Judgment by Alice Crary. Notre Dame Philosophical Reviews. http://ndpr.nd.edu. Posted January 10 2008. Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin. McDowell, J. 1985. Values and secondary qualities. In Morality and objectivity, ed. T. Honderich, 110–129. London: Routledge and Kegan Paul. Shafer-Landau, R. 2005. Error theory and the possibility of normative ethics. Philosophical Issues, “Normativity” 15: 105–120. Shepski, L. 2008. The vanishing argument from queerness. Australasian Journal of Philosophy 86: 371–387. Stratton-Lake, P. 2002. Introduction. In Ethical intuitionism: Re-evaluations, ed. P. Stratton-Lake, 1–28. Oxford: Oxford University Press. Thomas, A. 2006. Value and context. Oxford: Oxford University Press. Williams, B. 1981. Internal and external reasons. In his Moral luck, 101–113. Cambridge: Cambridge University Press.

Business as Usual? The Error Theory, Internalism, and the Function of Morality Caroline West

1 Introduction Many take internalism to be an analytic truth about ethics. They find compelling the thought that there is a necessary connection between moral judgments and motivation – between, for instance, judging that some action is right and feeling some normative pressure toward being motivated to do it. Someone who judges that performing some action would be right but who feels not even the slightest pull toward doing it seems to be rationally or conceptually defective in some way. But internalism is in tension with the conjunction of three other theses that many find plausible: (1) cognitivism, the view that moral claims express beliefs that can be true or false; (2) the Humean theory of motivation, the thesis that beliefs and desires are distinct existences, and that the motivational states are desires; and (3) the Humean theory of normative reason, which holds that reason alone mandates no revision to a subject’s existing desire set, except instrumentally. If (1) and (2) are both true, then someone could judge that it would be right for her to perform a certain action, yet have no corresponding desire, and so no motivation, to do it. One could reconcile this with internalism by positing a necessary connection between forming the judgment that some action has the property of being right and it being rational to acquire some desire to do it: that is to say, one could reject (3). This is the solution that Michael Smith favors in The Moral Problem (1994). According to Smith, moral judgments express beliefs about what we would desire that we do in a particular situation if we were in objectively more ideal conditions of reflection – if, say, we were “cool, calm, and collected,” properly factually informed, had spent more time considering the matter, and so forth. The desires that we would have at the end of such an objectively ideal process of reflection are, according to Smith, normatively superior to those that we had at the start, so we have reason now (in our actual, less than ideal, state) to desire to do what our ideal (or more fully rational) self would advise us to do in our actual situation.

C. West (B) Department of Philosophy, University of Sydney, Sydney, NSW, Australia

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_11, 

183

184

C. West

But if the Humean theory of normative reason (3) is true, there is no such thing as rational desire change in Smith’s sense. The recognition that under different circumstances we would desire differently might cause a change in our existing desire set, but it gives us no reason to acquire a new basic desire (or to remove an existing one), except instrumentally (i.e., insofar as this would contribute in some way to the satisfaction of our present desires). So if (1), (2) and (3) are together true, then internalism is false – indeed, necessarily false. There could be no necessary rational connection between moral judgments and motivation, after all. Reasoning roughly along these lines has led a number of philosophers, including famously J.L. Mackie (1977) and more recently Richard Joyce (2001), to endorse an error theory of morality. If internalism is a conceptual truth about morality, and internalism is actually false, then morality is nowhere instantiated.1 There are some who think that this argument, while perhaps of theoretical interest to metaphysicians or philosophers of language, has no practical significance for ethics. They suppose that we could accept this argument for the error theory, yet go on using moral concepts in the just the ways in which we always have. Indeed, Mackie himself seemed to take this “business as usual” view of the matter. Having argued in Chapter 1 of Ethics: Inventing Right and Wrong that we should cease to believe that there are “objectively prescriptive” features of the world, Mackie spends the greater part of the rest of the book seeking first-order principles that can perform what he takes to be the primary function that morality presently plays for us: namely, that of crowd control; or, as Mackie more genially puts it, of “counteracting limited sympathies” (1977, p. 107). First-order and second-order moral questions, he says, “are not merely distinct, but completely independent” (1977, p. 16). It may sometimes be true that a metaphysical view is a mere theoretical epiphenomenon. But my main aim here is to show that this is not the case with respect to the moral error theory. Business as usual is not a possible option for an error theorist persuaded by the kinds of considerations that move Mackie.2 The ways in which we use moral concepts – to persuade others to alter their attitudes, to deter narrowly self-interested behavior, to overcome weakness of will, amongst others – depend on it being widely believed that there is a necessary rational connection between moral judgments and motivation. Or so I will argue. If so, we could not come to believe that there are no “objectively prescriptive” features of the world yet continue to use

1 Mackie

and Joyce tend to speak of “objective prescriptivity” rather than “internalism” as the problematic feature of morality. But on a common interpretation of the argument from queerness, a crucial part of what makes objectively prescriptive properties objectionably “queer” is that they are supposed to be internalist: they are objective features of a situation that are supposed to be intrinsically motivating or action-guiding, independently of an agent’s antecedent desires or purposes. 2 There might be arguments for a moral error theory that do not involve or require the rejection of internalism, holding that it is some other essential feature of morality that is uninstantiated. My argument here is directed only at those arguments for the error theory that turn, in one way or another, on the falsity of internalism.

Business as Usual? The Error Theory, Internalism, and the Function of Morality

185

moral concepts in the ways that we presently do, as Mackie seems to suppose. On the contrary, the moral error theory appears to become a view with enormous ethical consequences that requires serious rethinking and revision of our first-order moral practices. This result is exactly that advocated (though for quite different reasons) by a different brand of error theorist, the moral eliminativist or abolitionist.3 According to the moral abolitionist, if the error theory is true, we ought to abolish moral concepts from our ordinary, everyday thinking and practice altogether. We should cease to think and talk of things as being morally right or wrong, morally good or bad, morally forbidden or obligatory; and instead think and talk only in terms of nonmoral beliefs and preferences. For instance, rather than thinking and saying that sexual discrimination is morally wrong or unjust, we might think and say that we strongly dislike it, that we want others to dislike it too, and perhaps also that we are prepared to punish those who engage in it, along with those who fail to punish those who engage in it. Moral abolitionism is certainly a consistent option, given the argument to be advanced here. The problem is that it is utterly unrealistic. As a matter of practical, biological, and psychological fact, it is extraordinarily unlikely that we would be able to banish moral thoughts and urges from our psyches, even if we wanted to and tried very hard. We know from evolutionary biology and psychology that we have evolved over millennia to exhibit moral behavior; and, while it is in the end an empirical matter whether or not we could succeed in dispensing with the “moral overlay,” all the evidence we have to hand suggests it extremely unlikely.4 Things now look bad, if my argument here is right and one thinks (as many of us do) that internalism is false. If business as usual is incoherent, and the abolitionist alternative is practically impossible, what (if any) options remain? In conclusion, I discuss four remaining possible options, without attempting to resolve the substantive further issue of which of the views that are left standing we ought to accept.

2 Why Internalism? There are many different varieties of internalism.5 But probably the most popular version of internalism, and the one I will focus on here, holds that it is a conceptual 3 The

reason moral abolitionists usually offer in favor of abolitionism over business as usual is not that business as usual is not a coherent option for a Mackie-style error theorist, but rather that the world would be on balance a more salubrious place if we gave up on the whole business of morality. For defenses of moral abolitionism see Hinckfuss (1987); Garner (1994) and this volume. 4 Here I obviously do not mean to claim that we cannot abandon or radically revise particular firstorder moral viewpoints, such as the idea that slavery is justified. Clearly we can and (thankfully) do. 5 For an influential discussion of varieties of internalism see Brink (1989, pp. 37–41). See also Sinnott-Armstrong, this volume.

186

C. West

truth about morality that an agent who judges that she morally ought to φ will, insofar as she is rational, be prima facie motivated to φ.6 Why does internalism of this kind seem so intuitively compelling to many? For some, the motivation for internalism has its source in rationalism. For others, internalism is motivated by the thought that only if there is an internal connection between the actions that are judged right, on the one hand, and there being reason to be motivated to perform those actions upon judging them to be right, on the other, does the moral realist have an answer to Moore’s open question argument: that is, an answer to the question of why people should be motivated to do the things that they call “right.” I will offer a different (though not incompatible) explanation. What seems to me to be the most compelling thought in favor of internalism stems from the pivotal role it plays in underpinning the practices that sustain our use of moral concepts, combined with the thought that a necessary part of what makes a normative system count as a morality is that it is used in the ways in which we use morality in individual and social deliberation. To illustrate, imagine a community of people who use the word “right” to refer to the same things that we call “right,” and who use the word “wrong” to refer to the same actions that we call “wrong,” yet for whom judgments of right and wrong have none of the practical significance that they do for us. Thus, they do not chastise others for acting in ways that they call “wrong”; they do not think that anyone deserves to be in any way punished for acting in ways that they call “wrong”; they do not attempt to persuade others to refrain from acting in ways that they call “wrong”; they feel not the least bit uneasy for being motivated to do things that they judge to be “wrong”; they feel no guilt or remorse for having done things that they happily say were “wrong,” and so forth. (Similarly, let’s suppose, they feel no inclination whatsoever to encourage or persuade others to do the things that they call “right,” or to praise or think well of them when they do, or to feel themselves under even the slightest normative pressure to be motivated to do the things they call “right,” and so forth.) It is very tempting to say that the word “wrong” in their mouths does not express our concept of wrong, precisely because their use of the word “wrong” has none of the connections with motivation, blame, punishment, persuasion, guilt and so forth that judgments of wrong do for us.7 (Ditto for their use of the word “right.”) If we imagine ourselves in the role of anthropologists charged with the task of interpreting and translating their behavior, there would be a strong case for saying that, while this community uses words such as “right” and “wrong,” those words do not express our concepts of right and wrong because their judgments do not have for them the motivational, emotional and behavioral consequences that those judgments have for

6 This

is the version of internalism that David Brink calls “weak hybrid internalism about motivation” (1989, pp. 37–41). 7 It is, I grant, somewhat mysterious why they would even have the word “wrong,” given that it seems to pick out nothing that is of any practical interest or significance to them. Perhaps it serves some recherché theoretical purpose.

Business as Usual? The Error Theory, Internalism, and the Function of Morality

187

us.8 They do not use moral concepts in the ways in which we do; they do not grasp what moral concepts are for. The belief that moral concepts have necessary connections with motivation, at least in rational agents, is vital to moral concepts being used in the ways in which we use moral concepts, and so is a necessary part of our idea of morality. Or so I will argue in what follows. This, of course, raises the large and complicated question of exactly what are the uses to which we put moral concepts. It is obviously a task beyond the scope of a single paper to attempt to undertake an exhaustive catalog of the ways in which we use moral concepts. It is also a question that cannot be answered from the armchair alone – in the end it is an empirical matter. But moral philosophers, in their role as armchair sociologists, have suggested a number of different ways in which we might use moral concepts. (And armchair speculation about the role(s) of morality is the best we have to go on at this point, since, to my knowledge, detailed and systematic empirical investigation of the ways in which morality is used has yet to be undertaken.) Here I will discuss five related roles morality has often been thought to play: (1) counteracting narrowly self-interested inclinations and behavior; (2) keeping the powerless in their place; (3) keeping the powerful in line; (4) overcoming weakness of will; and (5) attempting to persuade others to alter their psychological attitudes and, through this, their behavior. I will examine how each of these uses depends crucially on widespread belief in internalism.

2.1 Counteracting Limited Sympathies (or “Crowd Control”) There is a longstanding tradition of thought that takes the primary function of morality to be that of counteracting limited sympathies so as to enable human beings to live together in relative peace and security, and to prosper from the fruits of social cooperation. Plato’s Protagoras (Plato 2006), Thomas Hobbes (1651/1968), David Hume (1739–1740 book 3, part 2/1978), G. J. Warnock (1971), and Mackie (1977) are among those who have expressed this general view about the function of morality. On this approach, morality is an adaptation (cultural or biological – or, plausibly, both) designed to solve a problem generated by two fundamental, albeit contingent, features of the human predicament: limited sympathies and scarce resources. Human beings are, if not actively malevolent, then at least “almost always concerned more with their own selfish ends than with helping one another” (Mackie 1977, p. 108). Unconstrained, the combination of selfish motives and scarce resources would generate competition among people whose chief motivation would be to secure for themselves as great a share of the available pool of resources as possible, leading to conflict, mutual distrust, and insecurity – a Hobbesian “state of nature,” in which

8 See

also Joyce (2001, pp. 26–27).

188

C. West

all would be at war with all, and human life would be “solitary, poor, nasty, brutish, and short” (Hobbes 1651 part 1, Chapter 13/1968, p. 186). No one has an interest in living in conditions of perpetual conflict and insecurity; and everyone stands to benefit from the goods and opportunities that a system of cooperation makes possible. So each person has a purely selfish reason to agree to accept certain constraints on the direct pursuit of his or her own interests, provided that others do the same. Agreeing to accept these constraints would be mutually advantageous. As Hobbes sees it, the constraints take the form of certain conditional rules (or, as Hobbes calls them, “laws of nature”), such as: “I won’t kill you when it would be to my advantage to do so, provided you don’t kill me when it would be to yours”; “I won’t steal from you when you have something that I want or need, provided you don’t steal from me”; “I will keep the agreements I make with you, provided you abide by agreements you make with me”; and so forth. Every selfinterested agent has a reason to agree to abide by rules of this kind, since his or her various needs and wants are more likely to be satisfied in a situation in which everyone abides by these rules than in a situation where no one does. But there is a well known problem with the Hobbesian view, as is illustrated by the Prisoner’s Dilemma. While everyone has a self-interested reason to agree to abide by the rules, provided others do the same, any particular individual would do best (in narrowly self-interested terms) if others faithfully abide by the rules while he himself defects whenever it would be to his advantage. There is therefore an omnipresent temptation for selfishly-inclined people to “free ride” on the benefits that general compliance with the rules makes possible; enjoying the benefits provided by the general system of cooperation without paying the price. But freeriding threatens to undermine the system of social cooperation, for no self-interested person has a reason to adhere to the rules that make cooperative living possible if they cannot rely on others to do so as well. So in addition to the existence of rules enjoining respect for life and property, repaying good turns and keeping agreements, and so forth, there is needed some mechanism for ensuring that self-interested individuals will comply with the rules – even (or especially) when it would be to their advantage not to comply. The solution favored by Hobbes is to bring in the troops. Thus Hobbes views the establishment of a sovereign political authority, equipped with the formal coercive powers of the nation state, as the solution to the problem of how to keep narrowly self-interested people in line. Individuals then have two reasons for complying with the rules: the expectation of benefits that mutual compliance makes possible and fear of punishment. But, as Mackie says, reliance on purely formal coercive measures of this kind would leave the system of cooperation on which collective peace and prosperity depends as precarious as “a house of cards” (1977, p. 113). People whose motives remained purely selfish would perennially be poised to infringe on the interests of others for personal gain whenever the opportunity presented itself: whenever the benefits of (say) breaking an agreement were sufficiently great and the chance of detection and consequent punishment by agents of the state sufficiently slim – that is to say, they would break the rules when no state authority was looking.

Business as Usual? The Error Theory, Internalism, and the Function of Morality

189

A more robust mechanism for getting people to adhere to the rules is required. The necessary bulwark is to be found in morality: in the array of additional informal external controls that the idea of morality makes possible; and in moral sentiments that act as an internal psychic check on people’s selfish inclinations, creating in them a motivation to behave morally even in the absence of the threat of external discipline. The “annexing of virtue” (in Hume’s phrase) to motives and behavior (such as the keeping of agreements) that promote mutual trust, and of vice to those that threaten these (such as reneging on agreements for personal gain), makes possible the introduction of a powerful new set of informal external mechanisms for encouraging the former and deterring the latter. Actions that are conceived as immoral (as opposed to, say, simply imprudent or illegal) can be punished not only by means of formal legal sanctions, but also by an array of informal social penalties, ranging from public expressions of hostility and contempt to ostracism and even death. The idea of morality makes it possible for people who deviate from the moral code – who act to further their own interests in ways that infringe the interests of others – to be viewed by other members of the community as proper targets for criticism, censure and, in extreme cases, ostracism or death; while those who act morally, subordinating pursuit of their own narrow interests to that of the common good, are viewed as deserving of praise and the good opinion of others. The prospect of enjoying the good opinion of others, and of avoiding their hostility or contempt, is a powerful motivator, more powerful in many cases than is the threat of formal legal punishment or “respect for the rules.” These informal external controls operate in tandem with internal psychic structures, which they sustain and reinforce: moral sentiments that create internalized feelings of obligation, together with a moral conscience, which attaches negative emotions (such as shame and guilt) to the performance of actions the agent believes to be wrong, and positive emotions (such as pride) to those thought to be right. Shame and guilt are the internal analogs of external punishment, attaching even to undetected offences. An agent who is tempted to act in a narrowly selfish way that she knows to be wrong may thus be deterred from doing so by the knowledge that, even if she may not lose the good opinion of others, she will no longer be able to think well of herself. In this way, a well-functioning moral conscience renders punishment for acting wrongly, and reward for acting well, inescapable. Such tendencies as self-interested agents might have to seek their own good at the cost of infringing the interests of others thus come to be regulated both internally and externally by morality. Let us suppose that this picture is right: at least one of the roles that morality plays is to get each person to sometimes resist acting in his or her narrow self-interest – for instance, to keep agreements made even when it would be to his or her material advantage to renege – in order that we may live together in relative peace and safety and prosper from the fruits of social cooperation. What enables morality to function as an external check on narrowly selfinterested behavior and motives in this way is that each person knows that other members of the community will regard them as culpable for acting in ways that

190

C. West

violate established moral norms. It is the threat of punishment – of incurring the resentment or wrath of others – that enables morality to function to keep people in line. But we take ourselves to be justified in holding others culpable for acting immorally because we believe that there is a normative requirement for people to act morally (and therefore to be motivated to act morally), and not just insofar as it suits them. If we each came to believe that others would regard us as being in no way defective for failing to be so motivated, then the mere fact that others might call what we do “wrong” would provide little disincentive against doing it. Consider now the function that morality is supposed to play in regulating conduct internally. Here the idea is that someone who represents to himself some course of action as wrong will be, as a result, less inclined to do it – or, at least, will feel himself under considerable psychological pressure to be motivated not to do it. Moral sentiments function within people’s psychologies as an internalized authority figure (part policeman, part life-coach), exerting psychological pressure on individuals to be motivated to perform right actions and not to perform wrong ones, even when there is no chance of detection. It is easy to see how moral concepts could play this kind of regulatory role within the psyche of agents if we suppose that people tacitly accept a kind of internalism. If part of what is involved in judging that an action is wrong is that you take yourself to have a reason to be motivated not to do it, then it is easy to see why you would feel yourself under pressure to be motivated not to do it, and why you would feel uneasy (defective, ashamed, or guilty) should you remain unmoved. You would have failed to be motivated to do what you yourself feel you ought to be motivated to do in virtue of judging the action wrong. On the other hand, if the connection between moral judgment and motivation is contingent, then, while someone who happens to desire to do the moral thing may experience the discomfort or tension of conflicting desires, it is hard to explain why she should feel culpable or uneasy for failing to be motivated to do what she calls “right.”

2.2 Oppression There are other views about the role of morality that share the idea that morality functions as a constraint on the pursuit of self-interest, but that differ as to whose interests it promotes (or which crowd it functions to control) and also over whether its net effect is beneficial for all. In The Communist Manifesto, Marx famously lists morality along with law and religion as “bourgeois prejudices, behind which lurk in ambush just as many bourgeois interests” (1848/1977, p. 230). Marx takes morality to be part of the ideological superstructure of society that prevents people from seeing things as they really are. The legal freedom of the worker to sell his labor to whomever he likes on whatever terms he likes conceals the fact that he is really no more able to avoid exploitation by capitalists than the feudal serf is free to avoid working on the land of his lord. Class morality adds an extra layer of false consciousness, leading the worker to believe that the capitalist has a moral right to the proceeds of his investment. (See Singer 1980.) Class morality thus functions to protect and promote the

Business as Usual? The Error Theory, Internalism, and the Function of Morality

191

interests of those who own the means of production, at the expense of the workers from whose forced labor they profit. But it hypocritically cloaks its sectional nature in universal garb, pretending to serve the interests of all while in reality serving only those of a property-owning elite. This is not, of course, a very appealing picture of the function of morality. It makes morality itself appear positively immoral, as a vehicle for oppression and exploitation. Still, it is an empirical question whether morality functions in this way; and while we might be skeptical that morality is in fact always or primarily used to deceive and oppress a certain segment of the population, it is not implausible that morality is at least sometimes used for this purpose. (The Marxist thesis might thus be seen as a piece of common sense writ large.9 ) Here again it is hard to see how morality could be used to oppress in this way if moral concepts were not widely supposed to have a necessary rational connection with motivation. For morality to function to keep the working classes docile and compliant, workers must both accept the idea that the capitalist has a moral right to his or her profits and, crucially, be moved by it. If the workers could happily acknowledge the capitalist’s right to the proceeds of his or her investment, yet see themselves as under no (even prima facie) obligation to be motivated to respect that right, then it is unclear how morality could function to keep the workers docile and compliant in their own subjugation. If moral concepts were not intrinsically motivating, at least in rational agents, there would be nothing intrinsically oppressive about morality. The roots of oppression would lie elsewhere: not in the content of morality, but in the workers’ contingent and exogenous desire to be good moral agents. But then this desire could be dispensed with, whilst leaving the content of morality entirely untouched. That is not what Marx had in mind. He thought that class morality should be replaced by non-sectarian communist morality precisely because he assumed that moral judgments were inseparably connected to motivation and action. Suppose instead, as others have suggested, that morality serves the interests of the masses by keeping the talented in line. Nietzsche (1887/1994) is widely supposed to have held that morality – specifically, Christian or “slave” morality – was a device employed by the mediocre masses to valorize their own weaknesses, functioning to prevent a talented elite from realizing their potential greatness (i.e., from becoming Ubermensche).10 This too supposes that there is a widespread belief in the rational connection between moral beliefs and motivation. Without such a belief, the elite might well acknowledge the “moral” claims of the weak, but then, without experiencing even the slightest pressure to act in accordance with those claims, could act in any way they chose. Morality could not function to keep the great men down unless they are necessarily moved by it, at least to some extent.

9 Thanks 10 See

to Richard Joyce for pointing this out. Pigden, this volume, for an analysis of Nietzsche’s metaethical views.

192

C. West

2.3 Emancipation Others have suggested that the idea of morality plays a quite different, emancipatory role. Unlike those who think that morality – or, at least, certain first-order moral views – should be dispensed with because they are necessarily oppressive, some contemporary feminists think that morality (even if founded in error) should be retained.11 The view has two related components. First, the idea of morality functions to prevent the powerful from abusing their position of power in ways that, without the checks imposed by morality, they are likely to be inclined to do. Ideas such as that women have a moral right to be treated with equal concern and respect, that current unequal social arrangements are unjust, that sexual discrimination is wrong, that men ought not be differentially advantaged, and so on, function as a check on the behavior of the powerful, protecting the comparatively powerless from suffering further at their hands. Second, the idea of morality provides members of presently oppressed social groups with a powerful shared framework within which to mount the case for social reform. (I take up the role of persuasion in the section after next.) Here, too, belief in internalism plays a central role. The powerful, let’s suppose, have a vested interest in preserving and perhaps extending the socio-economic advantages they currently enjoy. What’s more, being powerful, they are well positioned to do so. Morality is supposed to operate as a check on abuses of power. If the powerful recognize that the powerless have, for instance, a right to a reasonable standard of living and other basic entitlements, then, as a practical matter, they might endorse a 40-hour working week and even support (however grudgingly) increasing the minimum wage. But this requires that the powerful both recognize that they have certain moral obligations towards more vulnerable members of society and, crucially, that they are moved at least to some extent to comply with those obligations. In the absence of morality, agents may simply have reason to do whatever will contribute to the satisfaction of their existing desires; and if those desires are to hang on to their present privileged position, then we cannot say that they are in any way irrational or defective for doing so.

2.4 Overcoming Weakness of Will Joyce (2001) suggests that another role that morality plays is to overcome motivational infirmities, such as weakness of will. Someone who represents a course of action to herself as morally obligatory – as something that she is categorically or unconditionally required to do – may be better able to withstand the temptation of doing otherwise than will someone who simply thinks that performing the action

11 For

discussion, see Lovibond (1992).

Business as Usual? The Error Theory, Internalism, and the Function of Morality

193

will contribute to the satisfaction of the desires she happens to have. But, of course, thinking of some course of action as something that “simply must be done, regardless” will not help to motivate, unless this judgment is supposed to have a necessary connection with motivation.

2.5 Persuasion We also use moral discourse to attempt to persuade others to alter their psychological attitudes and behavior. According to Michael McDermott (1978), this is what morality is all about: “how to preach” – where “preaching” is a matter of trying to persuade your society to accept principles you favor as general rules of conduct from which deviation will be punished (if not by law, then by socially sanctioned criticism and disapproval). We seek to persuade others to change their moral opinions by offering reasons and arguments in the expectation that, by getting others to change their moral opinions, we will thereby bring about a corresponding change in their attitudes and so their behavior. Our aim in seeking to persuade someone that eating meat is wrong, for instance, is not to effect an epiphenomenal change in her belief system, but to get her to stop doing it, and perhaps to get her to criticize others for eating meat.12 There would be no point attempting to persuade someone to change her moral opinion unless you believe that a change in moral opinion will produce in her a corresponding change in psychological attitudes and behavioral dispositions – and not just contingently. If you knew that your interlocutors were amoralists – people who felt themselves under no (even prima facie) obligation to be motivated to do something that they recognize to be right (or to refrain from doing something that they judge to be wrong) – attempting to alter their moral opinions would be pointless. Of course, there is no guarantee that attempts at persuasion will succeed. But what gives moral persuasion its point is the thought that if you succeed in persuading others to change their moral beliefs then necessarily they will feel themselves under normative pressure to change their attitudes and so actually to change them, insofar as they are rational. The externalist, of course, has an explanation of how moral argument can succeed in changing other’s attitudes. According to externalists, moral persuasion has a point not because there is any necessary conceptual connection between judging something to have the property of being right and feeling oneself under any normative pressure to approve of it, but rather because good agents have an independent standing de dicto desire to do what they judge to be right. So, for instance, if you persuade me that someone in my situation has a moral obligation to give money to charity I will, if I am a good and rational agent, come to be in favor of people in my situation giving money to charity because I want to do what is right and you have

12 This

is also broadly the view of Charles Stevenson (1944).

194

C. West

persuaded me that giving money to charity is right. (See Smith 1994 for comparison of internalist and externalist explanations.) I do not find this externalist explanation very compelling. But note that in any case it is not available to the “business as usual” error theorist – at least, not if the error theorist actually succeeds in converting people to the truth of the error theory. For the externalist account of persuasion works only if our interlocutors are realists: that is, only if they believe that there are things that are right. But those who believe the error theory believe that nothing is right (or wrong). They believe that there is nothing that falls under the extension of moral concepts. So the error theorist cannot avail herself of externalism to explain how moral persuasion could retain its point if the error theory is, or comes to be, actually believed.

3 No “Business as Usual” In the previous sections, I have discussed how a number of ways in which we use morality seem to depend on its being widely believed that there is a rational requirement for people to be motivated by their judgments that actions are morally right and wrong. If a necessary part of what it takes for a normative framework to constitute a morality is that it can be used in these sorts of ways, then widespread belief in internalism is a necessary condition for morality. But suppose that, as many of us believe, internalism is actually false: there are no properties such that the judgment that they obtain rationally requires people to be motivated in a particular way, independent of their existing desires or purposes. What should we say? One thing we clearly cannot say, if I am right, is that we could all come to believe that internalism is false but nonetheless carry on using moral concepts just as before. For, as we have seen, the practices that sustain our use of moral concepts require that internalism is widely believed to be true. If, having been persuaded by the arguments against internalism, we all came to believe that there was no necessary rational connection between moral judgment, then we could not use morality in all of the ways that we presently do. The “business as usual” error theory of Mackie (1977) is therefore not a possible option. Nor can we recommend that moral practice be abolished; for that, as I argued earlier, is almost certainly psychologically impossible. If neither business as usual nor moral abolitionism is an option, then what alternatives remain?

4 Four Options Left Standing I will conclude by sketching what seem to me to be the four remaining options. The first two options are ways in which the “business as usual” approach might be salvaged or resurrected. The third and fourth options are somewhat revisionary, but probably not so revolutionary as to be psychologically impossible.

Business as Usual? The Error Theory, Internalism, and the Function of Morality

195

4.1 Belief in Theory vs. Belief in Practice A first option might be to distinguish between belief in theory and belief in practice.13 It might be that a belief can be domain- or context-specific, operative in one domain of a person’s life or behavior, but not in another. If so, a person might come to believe something in theory, yet that belief might not inform her practice. For instance, someone might believe theoretically that there is a God, and that God forbids some behavior X (pre-marital sex, say), yet nonetheless continue to engage in X. Such a person would have a divided mind; believing that p in one context, and that not-p in another. Perhaps phobias are like this. Someone who has agoraphobia might say, if asked, that she does not believe that there is anything especially dangerous about open public places. She might even be willing to stake a lot of money on her being right about this. Yet other aspects of her behavior (never leaving the house, reacting in a terrified or horrified way to the suggestion that she might or should leave the house, and so forth) might indicate that – in another, practical context – she does believe that open public spaces harbor dangers. If it is plausible that beliefs can be domain-specific in this way, then it would be possible for someone to come to believe in theory that internalism is false, yet to believe in practice that it is not. She might continue behaving exactly as before in all practical, first-order ethical matters, whilst believing in theory that there are no necessary rational connections between moral judgments and motivation (and being prepared to bet a lot of money on the truth of this claim).

4.2 Non-revisionary Moral Realism A second option would be to adopt a kind of realism such as that recently suggested by David Braddon-Mitchell (2006), where widespread false belief in something like internalism becomes a partial truth-maker for moral claims. According to Braddon-Mitchell, the fact that it is believed that certain properties have a necessary connection to motivation in rational agents plays a crucial subject-setting role in ethics. It is what enables us to locate the moral properties in the natural world: Roughly, the moral properties are those natural properties that (1) we would ideally desire (in a thin, non-rationalistic sense of “ideally”) and (2) are widely believed to give an agent reason for being motivated in a particular way when he judges that they obtain. The existence of morality depends on enough people believing (perhaps falsely) that there are certain properties whose apprehension rationally requires motivation. On this view, it is business as usual, even if internalism is in fact untrue, just so long as enough people believe (falsely) in the truth of internalism.

13 Thanks

to David Braddon-Mitchell for pointing out this option.

196

C. West

4.3 Revolutionary Noncognitivist Fictionalism A third option is a form of antirealism that Joyce (2005) calls “revolutionary noncognitivist fictionalism.” On this approach, realists are correct to think that when we currently make moral claims we are at least trying to say something true, but wrong to think that we ever in fact succeed. The error theory is correct, so the moral claims we presently make (e.g., “Murder is wrong”) are in fact untrue. Nonetheless, we can – and (pragmatically) should – carry on using moral concepts and vocabulary, but we should cease to believe it and cease to assert it. We ought to give up our existing realist conception of morality in favor of a noncognitivist conception according to which we “accept” or make-believe (rather than believe) that there are objectively prescriptive features of the world (e.g., actions that “simply must be done, regardless”) and fictively-utter (rather than assert) that things are morally thus and so. It is an open question how much of our existing practice would continue unchanged if we shift from belief to make-belief and from assertion to fictiveutterance, but it is at least logically possible that business might proceed (more or less) as usual employing realist morality as a (useful) fiction.

4.4 Revisionary Moral Realism Finally, it might be that there is a subset of the roles that morality currently plays that could be played even if internalism came widely to be believed false. If the Humean theory of normative reason is true, there is no rational requirement for us to be motivated to do what we would desire in objectively more ideal conditions. But there is a weaker thesis that might be true and that might enable morality to play at least some of the roles that it currently plays. The weaker thesis says that we are rationally required in our current state to desire to do what we would desire to do under subjectively more ideal conditions of reflection – that is, under those conditions that we ourselves actually approve of or regard as more ideal. There may be things (probably many things) that we currently desire non-derivatively but that we would no longer desire under conditions which we ourselves regard as more ideal – say, after acquiring more information, more careful reflection, lengthy discussion with our friends, imaginative acquaintance with what it would be like to have (or lose) those desires or to act on them, and so forth. Suppose that many of us came to believe this weaker thesis instead of the stronger internalist thesis. Then moral concepts could arguably continue to be used for some, but not all, of the purposes for which they are presently used. For instance, belief in the weaker thesis would probably still be of use in overcoming weakness of will. Weakness of will is the failure to be sufficiently motivated to do what we take ourselves to have most reason to do: it is failure to be motivated in light of the ends we do, or think we would, endorse. You do not need to believe the stronger thesis to feel yourself under a rational obligation to act in ways that further those ends that, on reflection, you would endorse. The weaker thesis explains why we are rationally bound to be motivated to do what contributes to the satisfaction

Business as Usual? The Error Theory, Internalism, and the Function of Morality

197

of the ends that we would endorse under conditions that we regard as improved. So it tells you that you ought act to achieve those ends that you would ideally desire, regardless of urges or desires you might have to the contrary. If your ideal self would desire that you stay home to finish a manuscript by the promised date rather than going out to party with friends, then the weaker thesis says that you are rationally obliged to do so, whether you want to go out partying or not. Staying home to finish the manuscript is something that simply must be done, given the very ends that you yourself endorse. The weaker thesis could also explain why moral argument has a point – at least in debates with the like-minded. It could be that in attempting to persuade our friends to give money to famine relief we are attempting to show them that in conditions that they themselves would regard as more ideal they would desire that they give money to famine relief. So moral argument could retain its ad hominem force if the weaker thesis came generally to be believed – its point being to get others (and perhaps ourselves) into more ideal conditions, or to reveal to people the respects in which their current moral views are not aligned with their own fundamental values (what they would desire (subjectively) ideally) and should be revised. But if the weaker thesis came widely to be believed, moral concepts could not be used in all the ways we presently use them. Engaging in moral persuasion and argument with people who do not share your fundamental values would come to appear pointless. People with different fundamental values differ as to the conditions that they take to be more ideal; and, if the weaker thesis is true, neither we nor they need be guilty of making any rational mistake. So, if we believe the weaker thesis (and it informs our practice), attempting to alter their attitudes by rational persuasion would be pointless; engaging in moral argument with those who do not share our fundamental values would be mere browbeating.

5 Conclusion This is far from a comprehensive discussion of the uses to which we could (and couldn’t) continue to put moral concepts if the weaker (but possibly true) thesis came widely to be believed. My aim is simply to sketch the outlines of the option. On this fourth option, neither false belief (option 2) nor make-believe (option 3) is required to sustain moral practice; and business can proceed, though not quite as usual. The question is whether belief in something like the weaker thesis would enable moral concepts to play enough of the roles that moral concepts currently play for us to deserve to count as a morality. The moral error theorist claims that it does not; it would be a mere “schmorality” (Joyce 2007). The revisionary realist insists that it would; or near enough, so as to be, at worst, only a limited change of subject. The choice between option 3 (revolutionary noncognitivist fictionalism) and option 4 (revisionary realism) turns on whether the felicity conditions for utterances in the revised discourse are truth conditions or not. How would we tell? In my view, it depends on how much we would hanker after the old way of thinking if we came explicitly to believe the strong internalist thesis to be false. But it is

198

C. West

also worth noting that a slightly more complicated combination of these options is a possibility. Let me illustrate these points with a familiar sort of analogy. Suppose we start with a conception of solidity according to which solid things are impenetrable matter. We then discover that the things we’ve been calling “solid” are in fact reasonably penetrable. But suppose that we are compelled to go on thinking and talking of things as solid. Perhaps initially, with the old concept lingering fresh in our minds, we might be inclined to think that being reasonably penetrable is simply not a good enough deserver for the label “solid.” Since we can’t dispense with the notion, we might think that although we can go on talking of things as solid, we’re not really saying anything true. But then, after a generation or two of engaging in the practice, the memory of the old concept might fade, and we might come literally to believe what we now think and say. So it could well turn out to be with morality. Revolutionary noncognitivist fictionalism might be the correct analysis of the first phase of a conceptual revolution that culminates in the installment of a new form of realism. Acknowledgments Special thanks are due to Robert Bezimienny, David Braddon-Mitchell, and the editors of this volume for very helpful comments and suggestions.

References Braddon-Mitchell, D. 2006. Believing falsely makes it so. Mind 115: 833–866. Brink, D. 1989. Moral realism and the foundations of ethics. Cambridge: Cambridge University Press. Garner, R. 1994. Beyond morality. Philadelphia: Temple University Press. Hinckfuss, I. 1987. The moral society: Its structure and effects. Canberra: Australian National University. Hobbes, T. 1651. Leviathan. 1968. London: Penguin. Hume, D. 1739–1740. A treatise of human nature, ed. L. A. Selby-Bigge and P. H. Nidditch, 2nd ed. 1978. Oxford: Clarendon Press. Joyce, R. 2001. The myth of morality. Cambridge: Cambridge University Press Joyce, R. 2005. Moral fictionalism. In Fictionalism in metaphysics, ed. M. Kalderon, 287–313. Oxford: Oxford University Press. Joyce, R. 2007. Morality, schmorality. In Morality and self-interest, ed. P. Bloomfield, 51–75. Oxford: Oxford University Press. Lovibond, S. 1992. Feminism and pragmatism: A reply to Richard Rorty. New Left Review 1: 56–74. Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin. Marx, K. 1848. Karl Marx: Selected writings, ed. D. McLellan. 1977. Oxford: Oxford University Press. McDermott, M. 1978. How to preach. Canadian Journal of Philosophy 8: 633–652. Nietzsche, F. 1887. The genealogy of morals, Trans. C. Diethe; ed. K. Ansell-Pearson. 1994. Cambridge: Cambridge University Press. Plato. 2006. Protagoras and Meno. Trans. A. Beresford. London: Penguin Classics. Singer, P. 1980. Marx. Oxford: Oxford University Press. Smith, M. 1994. The moral problem. Oxford: Blackwell. Stevenson, C. L. 1944. Ethics and language. New Haven: Yale University Press. Warnock, G. J. 1971. The object of morality. London: Methuen.

The Fictionalist’s Attitude Problem Graham Oddie and Daniel Demetriou

On the face of it, moral claims characteristically involve reference to various moral properties and relations: not only such thin moral properties and relations as permissibility and obligatoriness, goodness and badness, betterness and worseness, virtuousness and viciousness, but also thick properties such as being compassionate, cruel, selfish, kind, greedy, generous, honest, or wicked. To accept a moral claim is to believe a proposition ascribing such properties, thick or thin, to various entities – persons, acts, states of affairs, dispositions, and so on. And to utter a moral claim is to express one’s belief in the associated proposition. But thin moral properties have seemed problematic to many, for both epistemic and metaphysical reasons, and their thick offspring, to the extent that they implicate the thin properties, inherit those problems. If moral properties and propositions are problematic, how can one rescue moral talk without being burdened with a problematic ontology? For much of the twentieth century the favored solution was a bold and striking denial of the accuracy of the fundamental appearances. If there are no genuine moral properties, thick or thin, the meaningfulness of moral discourse does not involve reference to such properties. To make a moral claim is not to express one’s belief in a moral proposition, but to express some other attitude – approval, desire, or some other noncognitive attitude. Finally, the meaningfulness of a moral claim is somehow to be cashed out in terms of those attitudes. That is to say, noncognitivism became the ruling paradigm. John Mackie broke out of the noncognitivist paradigm (Mackie 1977). According to Mackie, the realists are right that moral claims should be taken at their face value: moral predicates really do refer, or attempt to refer, to thick and thin moral properties; moral claims involve truth-evaluable propositions about such properties; and moral utterances express beliefs in moral propositions. Mackie thought the noncognitivists were just wrong about all that. What the noncognitivists got right, according to Mackie, was the extraordinary implausibility of the metaphysics which moral talk presupposes, and hence of the truth of moral claims taken at face value. So, Mackie G. Oddie (B) Department of Philosophy, University of Colorado, Boulder, CO, USA

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_12, 

199

200

G. Oddie and D. Demetriou

put forward his now famous error theory of moral discourse: moral talk is cognitively meaningful, representational talk all right, but it is at worst wildly false, and at best wildly unbelievable. In his seminal book on the subject, Ethics: Inventing Right and Wrong, after arguing for this error theory Mackie went on to formulate, and apparently endorse, various substantive moral claims. This seems odd. If you think moral claims are just plain false, and radically misleading, because the basic properties and relations they presuppose do not exist, then presumably you have good reason not to believe, utter, or endorse them. Even if you might have a reason to appear to endorse them at times – say, while playing backgammon – you shouldn’t be overtly endorsing them in the context of a philosophical treatise in which you are arguing for their radical defectiveness on the score of truth. Recently Richard Joyce has suggested that Mackie can be rescued from this awkward predicament by an explicitly fictionalist account of moral practice (Joyce 2001). The fictionalist about a certain discourse acknowledges that, although the characteristic claims of the discourse have propositional content, those claims are false (or likely to be false, or truth-valueless, or rationally unjustifiable), but thinks there is nevertheless a good reason to go on using the discourse, not only making the characteristic utterances, but in some sense accepting them. Morality is, strictly speaking, an elaborate fiction, but it is a valuable fiction, in the sense that it is worth our while going about our business as though it were true. Well, perhaps not valuable, at least not in any realist sense. But useful, or helpful, or something like that. So, an error theorist about a discourse who thinks that there is reason to continue engaging in the discourse should go fictionalist. Notice that this fictionalist response to Mackie’s error theory requires a doubleattitude approach to the acceptance of propositions. On the one hand, there is the kind of acceptance of a proposition that is simply belief in that very proposition. There are all manner of propositions that we have reason to believe. If Mackie is right then all the substantive moral claims – that torture is wrong, that abortion is permissible, that cruelty is bad, that courage is a virtue, and so on – are either false or at least unbelievable. We have no good reason to believe them. But, despite this, we should carry on somehow as though we accept them. There are two possibilities if acceptance is somehow a matter of belief. One is that in accepting a moral proposition one straightforwardly believes some other related proposition, perhaps the proposition that in the moral fiction, the moral proposition in question is true. So to accept that lying is wrong is to believe that in, or according to, the moral fiction, it is wrong to lie. The other possibility is that in accepting a moral proposition one adopts a belief-like attitude to the proposition, perhaps the attitude of makebelief. To accept that lying is wrong, for example, is not to believe it is wrong, but to make-believe that it is wrong. It has to be admitted that make-belief sounds a bit feeble for the kind of attitude which morality, on the surface, demands. Make-belief is, after all, a highly overridable attitude. Whenever we make-believe something we are primed to abandon the attitude if reality intrudes in a rude or demanding way, and it is entirely appropriate

The Fictionalist’s Attitude Problem

201

to be so primed. Suppose make-belief is the appropriate attitude of an audience to the propositions explicitly or implicitly true in a work of fiction – like a play one is attending. Suppose it is true in the play that two people are chatting comfortably on a couch in their home, and that their home is not on fire. As we get into the play, we make-believe that that is true. Now, if smoke starts seeping onto the stage from backstage but it is clearly true, in the play, that there is no smoke in the room, we tend in such circumstances to abandon the make-belief (that there is no smoke in the room) and go with the belief (that there is smoke), and it is entirely reasonable to do so. When push comes to practical shove, make-belief will rightly give way to genuine belief. Suppose I am a moral error theorist, but also a fictionalist. So I make-believe that it is wrong to lie, and generally act as though that were true, conducting myself accordingly, avoiding both large and small lies. By analogy with the fire in the theatre, however, when the consequences of acting as though I really believe that lying is wrong compete with rude self-interest, then why wouldn’t I have more reason to be guided by my actual beliefs about lying (i.e., that it is not wrong to lie)? The other possibility is for the moral fictionalist to equate acceptance of a moral proposition with belief in some related proposition – an unexceptional non-moral proposition associated with that moral proposition. If morality is an elaborate fiction, then at least it is a fiction according to which, or in which, various moral propositions are true. Suppose it is true, in the moral fiction, that torture is morally wrong. To accept that torture is morally wrong is thus to believe the proposition that according to the moral fiction, torture is morally wrong. Again, however, this seems too feeble for genuine full-blooded moral acceptance. An error theorist who is also a moral nihilist might well believe that according to the moral fiction lying is morally wrong but he does not thereby accept that lying is morally wrong. He doesn’t accept that anything at all is morally wrong. Double-attitude fictionalism need not hold that the appropriate attitude to moral propositions is either belief or belief-like. Mark Kalderon, in his recent defense of moral fictionalism, explicitly argues for a double-attitude account (Kalderon 2005). He argues both that the acceptance of a moral proposition is quite different from belief, and also that fictionalism need not be grounded in error theory. Briefly, Kalderon thinks that the realists and error theorists are right about the semantics of moral talk, while the noncognitivists are right about acceptance and utterance of moral claims. Kalderon argues that two different issues have traditionally been systematically conflated within the debate about cognitivism versus noncognitivism. One issue is this: 1. Do the target sentences of a certain discourse have genuine propositional content; are they truth-apt, do they have truth conditions? The other is this: 2. Is acceptance of a sentence in that particular discourse wholly a matter of belief?

202

G. Oddie and D. Demetriou

Traditionally, cognitivism has been thought of as the affirmation of both of these, and noncognitivism the denial of both. Assuming that the object of a belief has to be a proposition, if the answer to the second question is YES then the answer to the first question should also be YES. If the acceptance of a moral claim, for example, is a matter of belief, then moral sentences must have truth-apt propositional content. To believe P is to believe that P is true. But the two questions can be decoupled, because answering YES to the first question is quite compatible with answering either YES or NO to the second. You might hold that moral claims, say, have propositional content, but go on to deny that acceptance of a moral claim is a matter of belief. There may be contexts in which acceptance of an utterance is not a matter of belief. And that is precisely what Kalderon does say. A positive answer to the first question Kalderon calls factualism, a negative answer nonfactualism. This is not perhaps the best terminology – it sounds as though a factualist about morality, say, accepts that there are some moral facts, but this is clearly not Kalderon’s intention. It might be better to call a positive answer to the first question propositionalism, and a negative answer non-propositionalism. In any case, Kalderon calls it factualism, and asks us to reserve the term cognitivism for the positive answer to the second question, and noncognitivism for the negative answer. So a nonfactualist is ipso facto a noncognitivist, but a factualist need not be a cognitivist. The possibility of accepting a proposition without believing it provides logical space for noncognitivist factualism – and that, Kalderon contends, is what fictionalism (or at least the best version of fictionalism) amounts to. A nonfactualist is under some pressure to explain what the content of a moral utterance amounts to if it does not pick out a moral proposition. Some version or other of expressivism has been the standard response. That is to say: “the content of a moral sentence consists wholly or partly in the noncognitive attitudes conveyed by its utterance” (Kalderon 2005, p. 53). One of the standard objections to expressivism is, of course, the Frege-Geach problem. Suppose you think that a moral or value claim does not come along with a truth-evaluable proposition. When you utter “Torture is morally wrong” you are not attributing a moral property – the property of moral wrongness – to an action type, torturing. Rather you are expressing, say, your disapproval of torturing. The trouble with this is that that sentence can be embedded in other claims such as: It is simply false that torture is morally wrong. If it is morally wrong to torture, then condoning others torturing is also morally wrong. If torture is morally wrong, then it is morally wrong to condone the torture of suspected terrorists.

Now, in giving utterance to these claims you evidently do not express your disapproval of torture, since each of them is compatible with the rejection of the proposition that torture is morally wrong. Furthermore, moral argumentation makes extensive use of such embedded uses of the original moral sentence. For example: 1. Torture is morally wrong. 2. Waterboarding is a form of torture.

The Fictionalist’s Attitude Problem

203

3. If some action is morally wrong, then it is morally wrong to condone someone’s acting in that way. 4. Cheney condoned the waterboarding of suspected terrorists. Therefore: 5. Cheney did something morally wrong.

Or consider this argument: 1. If torture is morally wrong, then it is morally wrong to torture a suspected terrorist who may have information about the imminent explosion of a dirty bomb. 2. It is morally permissible to torture a suspected terrorist who may have information about the imminent explosion of a dirty bomb. Therefore: 3. It is false that torture is morally wrong.

Unless “Torture is morally wrong” has a truth-evaluable propositional content, it is hard to see how these arguments can be classified as valid. Validity is a matter of the guaranteed transfer of truth and falsity. Validity guarantees a downward transfer of truth; truth flows from the premises of a valid argument down to the conclusion. For the same reason validity guarantees an upward transfer of falsity. Falsity seeps (rather than flows) up from the conclusion to the premises – not all the premises of a valid argument with a false conclusion can be true. As Kalderon makes clear, the apparent validity of these arguments presents the noncognitivist with a dilemma (Kalderon 2005, pp. 58–59). Either: The meaning of an utterance of “Torture is morally wrong” is not the same in both embedded and unembedded contexts (in which case we have a fallacy of equivocation in all arguments involving a moral sentence that is both embedded and unembedded).

Or: The meaning of the utterance “Torture is morally wrong” is the same in both embedded and unembedded contexts (in which case the meaning is not given by any expressivist semantics).

This dilemma captures a version of the famous Frege-Geach problem – we will call it the Frege-Geach truth-transfer problem. As Kalderon rightly points out, it is a problem for nonfactualism, and specifically for expressivism, but not for noncognitivism as he characterizes it. Provided one rejects nonfactualism and unabashedly embraces propositional content for moral sentences, then even if one is a noncognitivist about acceptance, one can happily deny the first horn and embrace the second. The good arguments for traditional noncognitivism are arguments to the effect that moral acceptance is not wholly or even partly a matter of belief in a moral

204

G. Oddie and D. Demetriou

proposition. (As the internalists have maintained, it is plausible that there is at least something more than belief involved in accepting a moral judgment.) They are not arguments that moral acceptance cannot be some attitude (other than belief) towards a moral proposition. Kalderon argues that if acceptance of a moral proposition is not belief in that proposition (or in some other related proposition) but rather some noncognitive attitude to that proposition, then we can combine the good features of noncognitivism while avoiding the bad features of a problematic nonfactualist semantics. The Frege-Geach truth-transfer problem, in particular, is not a problem for the noncognitivist factualist. Kalderon has an interesting if somewhat controversial view of what acceptance of a moral proposition amounts to: to accept a moral proposition is to decide that things are going to appear to one, affectively, in a phenomenologically vivid way – presumably just the way they would appear if that moral proposition were true and you were the sort of being who experienced moral states of affairs appropriately (see, for example, Kalderon 2005, p. 147). Whatever acceptance of a moral proposition is, the important thing for a noncognitivist like Kalderon is that such acceptance be fundamentally different from belief, which is the attitude of acceptance appropriate to a non-moral proposition. So, following Kalderon, let acceptance be, in effect, a determinable, which embraces quite distinct determinate realizing attitudes. Different kinds, or determinates, of acceptance are appropriate to different kinds of propositions. In the case of a non-moral proposition acceptance is the attitude of belief, and in the case of a moral proposition it is some other attitude, involving an affective or conative component – let’s label it endorsement. (Analogously, rejection of a proposition is a determinable of which disbelief is the cognitive determinate. We can label the noncognitive determinate of rejection appropriate to a moral proposition repudiation.) Thus we have: The two-attitude account of acceptance: Endorsement/repudiation and belief/disbelief are distinct determinates of the determinable acceptance/rejection. If P is a non-moral proposition, to accept P is to believe P (and to reject P is to disbelieve P); if P is a moral proposition, to accept P is to endorse P (and to reject P is to repudiate P).

Belief and endorsement are thus two fundamentally different attitudes – one cognitive, the other noncognitive – appropriate to these two quite different types of propositions. Consider the simplest possible argument that involves both moral and non-moral propositions. Consider the following trivial inference: C. Cheney condoned torture. T. Condoning torture is morally wrong. Therefore: W.

In condoning torture, Cheney did something morally wrong.

The Fictionalist’s Attitude Problem

205

The argument is surely valid. C is (let’s say) a non-moral proposition. Perhaps it is shorthand for some purely non-moral report such as “Cheney said it was a ‘nobrainer’ to torture a suspected terrorist by waterboarding.” T is clearly a moral proposition, attributing the property of moral wrongness to the condoning of torture. The conclusion, W, we may suppose, is simply tantamount to the conjunction of the premises. (If you are skeptical of that then you can substitute C&T for the conclusion as stated.) If we embrace nonfactualism then we have no problem with the validity – the guarantee of a downward transfer of truth – of this argument. But what exactly is the point of offering or uttering a valid argument? Typically the point of arguing is to get someone (perhaps oneself) who accepts the premises to accept the conclusion (or alternatively, to get someone who rejects the conclusion to reject the conjunction of the premises). C, the first premise of our argument is, we stipulated, a purely descriptive, non-moral proposition. Assume that you believe C; you accept that Cheney condoned torture. Suppose you also endorse the moral proposition T; you accept that condoning torture is morally wrong. So, you accept both the premises, and the conclusion follows from the premises because it is simply logically equivalent to their conjunction. Since the argument is valid, the conclusion is a demonstrated consequence of the premises, so. . . surely you are rationally obliged to accept the conclusion. Why? Because acceptance, like truth, should transfer from premises to conclusion of a demonstrably valid argument. In the case where we have non-moral propositions, transfer of acceptance is unproblematic, because presumably we have a principle of transfer for belief. As a first stab at this, let’s take a formulation that is often cited in elementary logic courses. Naive transfer principle for belief/disbelief: A rational person who believes (i.e., accepts) some propositions is rationally obliged to believe (accept) any proposition which is a demonstrated logical consequence of those propositions. A rational person who disbelieves (rejects) some proposition is rationally obliged to disbelieve (reject) the conjunction of any propositions of which it is a demonstrated logical consequence.

Gilbert Harman (2002) has criticized principles like this one for failing to take into account the distinction between principles of logic, which describe logical relations between propositions, and the principles of reasoning, which tell us how to proceed epistemically. Harman argues that the fact that some of your beliefs can be shown to entail some proposition does not mean that you are irrational to disbelieve that proposition (Harman 2002, pp. 173–174). Of course, you might be just as rational to disbelieve the conjunction of the premises instead, and the transfer principle articulated above embraces this in its second clause. However, even if you continue to believe the premises, and acknowledge the entailment, you may still not be obliged to adopt the conclusion as an additional belief. It may be just too costly, in some sense, to add that proposition to your stock of beliefs. For example, each of us has limited epistemic storage resources, and one may not be obliged to clutter

206

G. Oddie and D. Demetriou

up one’s belief-box with the trivial or uninteresting consequences of other elements in the belief-box. Or, if you find the conclusion unacceptable, you may find it too difficult either to dislodge the antecedent beliefs or to accept their implication. You might have to forget about it in the meantime while you pursue some other task. Our little argument above illustrates Harman’s point. Clearly this argument would never be used for the purposes of exerting rational pressure on someone to accept the conclusion on the basis of their acceptance of the premises. Since the conclusion of this argument is the conjunction of the two premises, it is just too obvious. But we have a reason for keeping our example simple, and the simplicity of this particular argument does not undermine the observation that one of the main points of classifying arguments as valid or invalid is to enable us to reason soundly, and one of the main points of reasoning, at least in the theoretical realm, is to improve or maintain the rationality of what one accepts. Valid arguments can exert rational pressure on us as regards which propositions to accept and which to reject, and they do so by channeling reasoned acceptance down from the premises of a valid argument to their conclusion, or by channeling reasoned rejection from the conclusion of a valid argument up to the premise set. Another kind of objection to naive belief-transfer trades on the familiar lottery paradox. There are a million tickets in the lottery, and as it happens you believe each of the following propositions: Ticket #1 will not win the lottery. Ticket #2 will not win the lottery. ... Ticket #1,000,000 will not win the lottery.

It follows from these propositions that no ticket will win the lottery. But you are certainly not rationally obliged to believe that, even if you do believe each and every one of the premises. Finally, it does seems odd to say that someone who happens to believe some outlandish or highly improbable propositions is thereby rationally obliged to believe some even more outlandish and improbable consequence of their conjunction. What this last objection suggests is that what is at issue is not so much the transfer of brute acceptance as the transfer of good reasons for acceptance. If you have good reason to believe some propositions, and some other proposition is a demonstrable logical consequence of them, then you have a good reason to believe that consequence. We can state a more nuanced transfer principle in order to make it compatible with these criticisms. We need to soften the rational pressure involved. Transfer principle for belief/disbelief: A rational person who has a good reason to believe some propositions has a good reason to believe any proposition that is a demonstrated logical consequence of those propositions. A rational person who has a good reason to disbelieve some proposition has a good reason to disbelieve the conjunction of any propositions of which it is a demonstrated logical consequence.

The Fictionalist’s Attitude Problem

207

Does this principle deal with the objections to the naive principle? First, it deals with Harman’s objections. Having a good reason to accept a proposition does not entail that one is rationally obliged to accept it. There might be other reasons in the offing. Second, clearly it blocks the rational transfer of acceptance from crazy beliefs to their consequences; unless you have some good reason to believe the premises, the transfer principle alone does not guarantee that you have a good reason to believe the conclusion. Third, the second half of the principle applies straightforwardly to the lottery paradox even if the first does not. Anyone who rejects the proposition that no ticket will win the lottery has a good reason to reject the conjunction of the premises of the paradox – and that seems right. In any case, assume that this or some suitably refined transfer principle for belief links logic and argument to good reasons for acceptance and rejection of propositions. There must be some such principle and it must work something like this. Now consider the case where we have a bunch of purely moral propositions, and acceptance is a matter of (noncognitive) endorsement rather than belief. For example, suppose you endorse the proposition that torture is wrong, and you also endorse the proposition that condoning someone else’s wrongdoing is itself wrong. Then it seems that you have a good reason to endorse the consequential proposition that condoning someone else’s torturing is wrong. What we apparently require here is a parallel principle of transfer for the noncognitive attitudes of endorsement and repudiation. Transfer principle for endorsement/repudiation: A rational person who has good reason to endorse some propositions has a good reason to endorse any proposition that is a demonstrated logical consequence of those propositions. A rational person who has a good reason to repudiate some proposition has a good reason to repudiate the conjunction of any propositions of which it is a demonstrated logical consequence.

At first blush this transfer principle looks as sound as the parallel principle for belief. But on reflection, it is not at all obvious why a double-attitude fictionalist about morality is entitled to it. In the case of belief, truth-seeking is built in. We aim at having true beliefs, and a necessary condition for the truth of our beliefs is their joint consistency. If our beliefs are inconsistent and we know this then we also know they cannot all be true. So if we come across an apparently unbelievable consequence C of some our beliefs, the goal of truth provides us with a reason either to believe C after all, or else to reject the conjunction of those antecedent beliefs. One cannot maintain consistency in one’s beliefs if one continues to believe the premises but believes the negation of the conclusion. But why, if you are a fictionalist about some domain, should you aim at consistency of what you accept? As a fictionalist about morality, for example, you are not concerned that the moral propositions you endorse be true, so it is not obvious that the consistency of the moral propositions one endorses is required. And if it is not required, then we lack the corresponding argument for endorsement-transfer.

208

G. Oddie and D. Demetriou

One could argue that the propositions we endorse are intended to be action guiding, and a set of endorsements that are logically incompatible might well imply contradictory injunctions, and that would leave us without any guidance how to act. Hence the desirability of consistency follows from the action-guiding nature of our endorsements. But of course, this presupposes that if our endorsements can be shown to entail some moral injunction, then we have a reason to endorse that moral injunction. Absent endorsement-transfer, it is not clear that endorsements that jointly entail some contradiction will actually yield confusing injunctions. In any case, let us grant to the double-attitude fictionalist this endorsementtransfer principle in addition to the belief-transfer principle. Suppose we have these two transfer principles for the two determinates of acceptance. Does that guarantee what we really need: namely, a transfer principle for acceptance itself? This, or something like it, is the principle the double-attitude fictionalist has to end up with: Generalized transfer principle for acceptance/rejection: A rational person who has good reason to accept some propositions has a good reason to accept a demonstrated logical consequence of those propositions. A rational person who has good reason to reject some proposition has a good reason to reject the conjunction of propositions of which it is a demonstrated logical consequence.

For the factualist cognitivist about morality, the belief-transfer principle guarantees a generalized acceptance-transfer principle, whether the propositions involved are moral or non-moral. But for the noncognitivist, it is not clear that the conjunction of the two specific principles guarantees generalized acceptance-transfer. Consider our little argument concerning torture. Does either of our two initial transfer principles apply, given the double-attitude thesis? Belief-transfer applies when the premises are all believed. Endorsement-transfer applies when all the premises are endorsed. The first premise is believed but not endorsed, while the second premise is endorsed but not believed. So neither principle taken by itself can give one a good reason to move from rational acceptance of the premises to rational acceptance of the conclusion. We need some additional transfer principles. Note that this is not the standard Frege-Geach problem, which is a problem for the transfer of truth and falsity. Rather it is a Frege-Geach-like problem – the problem of transfer of rational acceptance and rejection. Why not just postulate a generalized principle of acceptance/rejection-transfer in addition to belief-transfer and endorsement-transfer? Even disregarding the rather ad hoc nature of such a postulation, we still face a problem: what does it take to accept the conclusion of the little argument? The conclusion, since it is tantamount to the conjunction of two perfectly good propositions which are logically compatible, is itself a proposition – but is it either a fit object of belief, or a fit object of endorsement, or a fit object of some other attitude? Remember that belief and endorsement are, according to the noncognitivist, quite distinct attitudes – one of them a cognitive attitude appropriate to a non-moral proposition, the other a noncognitive attitude appropriate to a moral proposition. So the attitude appropriate to acceptance of the conclusion will depend on what kind of proposition the

The Fictionalist’s Attitude Problem

209

conclusion is. Is it a moral proposition, or is it a non-moral proposition, or is it something else? Since it is equivalent to the conjunction of two propositions, one moral the other non-moral, it is not clear what it is. Call any proposition that is equivalent to the conjunction of some moral and non-moral propositions (like W and the logically equivalent proposition C&T) a fusion. What does acceptance of a fusion amount to? There are three possibilities. Acceptance of a fusion is either a matter of belief, or a matter of endorsement, or it is some other attitude. If it is some third attitude, that attitude might be a combination of endorsement and belief (or, more accurately, it may supervene on endorsement and belief), or else it may be a distinct determinate of acceptance that does not supervene on endorsement and belief. Suppose, first, that acceptance of a fusion is a matter of endorsement. Then, anyone who has a good reason to accept C and T would have a good reason to accept W. So (given that acceptance of a fusion = endorsement) such a person would have a good reason to endorse W. But W entails C. So, by the transfer principle for endorsement, upon grasping a very simple demonstration, she would have a good reason to endorse a non-moral proposition (C). But if it is not entirely clear what the noncognitive attitude of endorsement of a moral proposition is, it is even less clear what it would be to take that attitude to a non-moral proposition like “Cheney condoned torture.” Recall that this is a non-moral proposition, a report of Cheney’s saying in an interview that it was a “no brainer” to waterboard a suspected terrorist. Can one really endorse that non-moral proposition (viz. noncognitively) in exactly the way in which one – noncognitively – endorses the proposition that torture is morally wrong? Of course, one can certainly accept (i.e., believe) that Cheney condoned torture without – in the colloquial sense – endorsing Cheney’s condoning of torture. For, in the colloquial sense, to endorse Cheney’s condoning of torture is to deem it morally acceptable, or even morally obligatory. In our technical sense, that would be tantamount to accepting (i.e., endorsing), not the proposition that Cheney condoned torture, but the proposition that it is morally permissible (or obligatory) for Cheney to condone torture. But these are, or course, two logically distinct acceptances. In general, it is hard to see how one could endorse a purely non-moral proposition in exactly the same way that the noncognitivist thinks one endorses a moral proposition, and at the very least the noncognitivist would owe us an account of that. A corollary of this observation is that the noncognitivist is in trouble accepting endorsement-transfer quite independently of the problem of the correct acceptance attitude for fusions. Suppose one endorses the proposition that torture is morally wrong. Then, by endorsement-transfer, one has a good reason to endorse any logically necessary proposition at all, all of which follow from this. And at least some logically necessary propositions (e.g., “It is either raining or not raining”) are non-moral propositions. There is a good reason to accept any logically necessary proposition (they are, after all, necessarily true), but this applies only if acceptance is belief. If endorsement-transfer is in force, and one accepts any moral proposition at all, then one has a good reason to endorse any logically necessary proposition as

210

G. Oddie and D. Demetriou

well. So one would have a good reason to believe any logically necessary proposition, but also, if one endorses any moral proposition at all, one also has a good reason to endorse any logically necessary proposition. This would constitute an odd case where two distinct determinates (belief and endorsement) of the one determinable (acceptance) apply to a whole slew of entities – namely, all the logically necessary propositions. Perhaps the noncognitivist could solve this particular anomaly by restricting the domain of the moral to logically contingent propositions. That is not totally implausible. Even so, we are left with the problem that for any given proposition, one might still bear any one of four pairs of acceptance/rejection attitudes to it. Suppose, then, that acceptance of a fusion is a matter of simple belief. Anyone who accepts (believes) C and accepts (endorses) T would (by the generalized transfer principle) have good reason to accept W, and so would have a good reason to believe W. Then, by the transfer principle for belief alone, upon grasping a very simple demonstration, she would have a good reason to believe T: that torture is morally wrong. But the moral fictionalist clearly does not want one’s acceptance of a moral proposition together with acceptance of some non-moral propositions to endow one with a good reason to believe that same moral proposition. The whole point of noncognitivism is to avoid having to say that our acceptance of moral propositions creates rational pressure to believe those moral propositions. This result can be generalized, for there is nothing special about C, T and W. Given the two particular transfer principles and the generalized transfer principle, if acceptance of a fusion is a matter of belief, then a person will have good reason to believe all the moral propositions she endorses. And if acceptance of a fusion is endorsement, then a person will have a good reason to endorse all the propositions she believes. (Suppose you have a good reason to endorse moral proposition M. Suppose you also have a good reason to believe some non-moral proposition N. Then you have a good reason to accept both N and M, and so by the generalized transfer principle for acceptance, you have a good reason to accept N&M. If acceptance of the fusion N&M is belief, then you have a good reason to believe N&M. By the transfer principle for belief, you have a good reason to believe the moral proposition M. If acceptance of N&M is endorsement, then by the transfer principle for endorsement you have a good reason to endorse the non-moral proposition N.) So a rational being with a minimal grasp of conjunction-elimination and conjunction-introduction would have good reason to believe all and only the propositions she has good reason to endorse. Clearly this is not a result that would please the noncognitivist, or indeed any two-attitude theorist – it obviously contradicts the thesis that there are two quite distinct attitudes appropriate to the two distinct kinds of propositions at issue. To avoid this acceptance-transfer problem, the fictionalist is going to have to extend the notion of acceptance to fusions without making them the direct objects of either belief or endorsement. This is actually not too hard. A fusion P can be divided into two parts. The non-moral content of P, N(P), is the strongest nonmoral proposition following from P. The moral content of P, M(P), is the strongest

The Fictionalist’s Attitude Problem

211

moral proposition following from P. And since P is a fusion, P is equivalent to N(P)&M(P). (That’s what a fusion is defined as.) Now we can characterize an extended notion of acceptance for fusions which is itself a mixture of belief and endorsement. Acceptance of fusions: A person accepts a fusion P if and only if she believes N(P) and endorses M(P).

Notice that we have two limiting cases. Where P is equivalent to its non-moral component N(P) (that is to say, P is wholly non-moral), acceptance of P will simply amount to acceptance of its non-moral component, N(P), and since N(P) is a nonmoral proposition, acceptance of P will be a matter of belief that P. Analogously, in the case where P is equivalent to its moral component, M(P) (that is to say, where P is a wholly moral proposition) acceptance will just be endorsement. On this account, there is no problem with accepting W: that in condoning torture Cheney did something wrong. The non-moral content of W is C. The moral content is T. So you accept W just in case you believe C and endorse T. So far so good. Unfortunately this move is not sufficient to solve the acceptance-transfer problem. Notice that accepting a fusion is neither belief nor endorsement. It is a third attitude, distinct from the other two. Maybe we should call acceptance of a fusion, beldorsement, a third determinate of the determinable of acceptance. The twoattitude account has thus become a three-attitude account. Since beldorsement supervenes on belief and endorsement this is not such a huge cost, if it is any cost at all. More importantly, not all propositions that are neither purely moral nor purely non-moral are fusions. All propositions have moral and non-moral components, but some propositions are not equivalent to the conjunction of their moral and non-moral components.1 These are what we might call organic hybrids – propositions that have moral and non-moral components but are not the conjunction of those two components. Consider the proposition C⇔T: “Cheney condoned torture if and only if condoning torture is morally wrong.” What is the moral content of C⇔T? C⇔T does not entail either C or ~C. So it doesn’t entail this purely non-moral proposition or indeed 1 Note that we are assuming that the tautology is a limiting case of both (purely) moral content and

(purely) non-moral content. Hence, even a purely moral proposition M still has a purely non-moral content – albeit the empty, tautologous moral content T. And so M is also equivalent to the conjunction M&T of its purely moral and purely non-moral contents. The purely moral propositions are those that do not “cut across the diagonal” of the moral/non-moral matrix outlined in the main text below. So if M is purely moral so too is ~M. It also follows that the disjunction of two purely moral propositions is also purely moral. Consequently Mv~M is purely moral. Similarly, if N is a purely non-moral proposition, so too is ~N, and hence so too is Nv~N. The logically true proposition is thus a limiting case in that it falls into both pure categories. (Note that we are assuming a coarse-grained notion of proposition here.) Now, we could just stipulate that any purely moral proposition (or a purely non-moral proposition) has to be non-tautologous. However, that strikes us as somewhat ad hoc and unnatural. In any case, the same points could still be made with a suitably amended definition of a hybrid proposition.

212

G. Oddie and D. Demetriou

any other non-tautologous non-moral proposition. Nor does it entail either T or ~T, or any other non-tautologous purely moral proposition. Both its moral and nonmoral contents are thus tautologous. But C⇔T itself is not a tautology. So C⇔T is not equivalent to the conjunction of its moral and non-moral components. It is not a fusion but a hybrid, and so we clearly cannot cash out acceptance of C⇔T in terms of endorsement of its moral content and belief in its non-moral content. A factualist can think of propositions as picking out classes of worlds, where each world is a combination of a maximal non-moral component n and a maximal moral component m. This suggests a matrix with the columns representing maximally specific non-moral propositions and the rows maximally specific moral propositions. A purely non-moral proposition is any collection (disjunction) of maximal non-moral components – it is a “vertical” proposition in the matrix. A purely moral proposition is any collection, or disjunction, of maximally specific moral components – it is a “horizontal” proposition in the matrix. The two-attitude theorist holds that belief is the determinate of acceptance appropriate to vertical propositions, and endorsement the determinate of acceptance appropriate to horizontal propositions. Fusions are the intersections of the vertical and horizontal propositions: they are rectangles in the matrix, and so we can define acceptance of those in terms of belief in the vertical component and endorsement of the horizontal component. But hybrids cut across the vertical and horizontal lines. They are propositions on the diagonal. Confining the logical space to just one non-moral proposition (C) and one moral proposition (T) apiece, the situation looks like this:

T: Condoning torture is wrong ~T: Condoning torture is not wrong

C: Cheney condoned torture

~C: Cheney did not condone torture

C&T C&~T

~C&T ~C&~T

The moral propositions are horizontal rows (T and ~T); the non-moral propositions are vertical columns (C and ~C). Fusions are the intersections of the moral and non-moral propositions (C&T, ~C&T, C&~T, ~C&~T). Hybrids are everything else – any propositions on the diagonal: like C⇔T, but also C⇒T, T⇒C, ~C⇒T, and C⇒~T.2

2 A referee made the following observation which is well worth clarifying: “Consider ‘Cheney con-

doned torture if and only if condoning torture is wrong.’ Isn’t this simply a fusion of two fusions, namely: ‘If Cheney condoned torture condoning torture is wrong’ and ‘If condoning torture is wrong, Cheney condoned torture’? And isn’t a fusion of two fusions simply a fusion, instead of a hybrid that requires yet another different attitude of enlief?” It is true that the conjunction of two fusions is a fusion, but the two conditionals involved here, C⇒T, T⇒C, are not fusions but hybrids. They both cut across the vertical and horizontal divides. Of course, this does not guarantee that their conjunction is a hybrid. T⇒C is a hybrid, as is ~T⇒C, but their conjunction is equivalent to C, neither a fusion nor a hybrid.

The Fictionalist’s Attitude Problem

213

What, then, is the acceptance attitude appropriate to hybrid propositions? It is neither belief, nor endorsement, nor beldorsement. Let’s call it enlief. Now, the postulation of enlief is more radical than the embrace of beldorsement. The latter supervenes on, and is definable in terms of, belief and endorsement. Not so for enlief. Recall the nature of supervenience. A property Q supervenes on a base of other properties R just in case sameness of distribution of R guarantees sameness of Q. So, is it the case that sameness in beliefs and endorsements guarantees sameness of enliefs? Can two people share the very same beliefs and endorsements but accept different hybrid propositions? Consider Cheney Hater – a person who is agnostic both about the morality of torture and about whether or not Cheney condoned it. He doesn’t know what to think about torture or the condoning of it, and also he doesn’t know what Cheney is reported to have said, but accepts that whatever Cheney does on this issue, he is bound to be wrong, to unerringly choose the morally wrong thing to do. So Cheney Hater accepts the biconditional C⇔T: “Cheney condoned torture if and only if condoning torture is morally wrong.” Now consider Cheney Lover, who as it happens is also agnostic about both the morality of condoning torture and about Cheney’s actual behavior. He hasn’t seen the TV interview and he distrusts the reports in the liberal media. But he accepts that whatever Cheney does, he unerringly chooses the morally right thing to do. So Cheney Lover accepts C⇔~T: “Cheney condoned torture if and only if it is not the case that condoning torture is wrong.” Now C⇔~T is tantamount to ~(C⇔T). Like C⇔T, ~(C⇔T) is a hybrid (the negation of a hybrid is always a hybrid) and one which, like C⇔T, has tautologous moral and non-moral components. Cheney Lover and Cheney Hater share the same beliefs and endorsements (and ipso facto they accept the same fusions) but they differ with respect to the hybrids that they accept. They differ in their enliefs. Enlief then, is a separate category, over and above belief and endorsement. So to fully describe your acceptances you would have to list the non-moral propositions you believe, the moral propositions you endorse (which jointly fix the propositions you beldorse), but as well, and independently, the hybrid propositions you enlieve. But although these are separate attitudes, they have to mesh together appropriately in a rational being. Thus multi-attitude fictionalism is a bit too much to accept. A fictionalist might well be tempted here to think that the source of these problems with fictionalism is the intrusion of the noncognitive element, and be tempted back to some variant of cognitivism.3 After all, if acceptance is a matter of belief, and belief-transfer is unproblematic, then presumably we get acceptance-transfer for free. There are two options for the cognitivist fictionalist. One option is that to accept a proposition from the fictional discourse is to believe in the literal truth of that very proposition. However, since the fictionalist holds that there is no good reason to believe the fictional propositions, the belief-transfer principle has no purchase

3 This was suggested to us by Michael Tooley, and he cited van Fraassen’s constructive empiricism

as a promising model.

214

G. Oddie and D. Demetriou

on arguments involving fictional premises. If there is no good reason to believe fictional propositions, the fact that they demonstrably entail some consequence does not provide one with a good reason to believe that consequence. The much more popular option is that to accept a proposition A about some fictional domain is to believe not A itself, but to believe some other proposition, f(A), which we have called the factual or real content of A. The factual content of a proposition is that proposition which, according to this kind of fictionalist, one has to believe in order to accept A. Typically f(A) is the proposition that in the relevant fiction, A is true. For example, to accept the moral proposition that torture is wrong is to believe that in the moral fiction, torture is wrong. It will be instructive here to briefly consider a fictionalist account completely outside the moral realm – a fictionalist account of theoretical entities. This is a simplified version of Bas van Fraassen’s constructive empiricism, which holds that to accept a theory A is simply to believe that A is observationally adequate – i.e., that all its observational consequences are true (van Fraassen 1980). In this case f(A) is the strongest purely observational consequence of A itself. Does acceptance-transfer fall out of normal belief-transfer? Let P be any purely theoretical claim (like there goes a proton) and O any purely observational claim (like there is such-and-such a streak in the cloud chamber). Then since P is purely theoretical, its observational content f(P) is empty. f(P) is logically true. Consider the conditional claim: P⇒O. This also has no observational content, as evidenced by the fact that its ramsification is a second-order logical truth. So f(P⇒O) is also logically true. Assume, plausibly, that one has good reason to believe any proposition that is demonstrably a logical truth. So one has good reason to believe both f(P) and f(P⇒O). To accept P is to believe f(P). So one has good reason to accept both P and (P⇒O). These entail O by a single application of modus ponens. So if transfer acceptance holds, then one also has good reason to accept O. To accept O is to believe f(O), and since f(O) = O, that means that one has good reason to believe O itself. So, simply in virtue of believing demonstrable logical truths, one would have good reason to believe any observational proposition at all. And that is clearly absurd. Quite generally, acceptance-transfer will be problematic for any cognitive fictionalist whenever there are what we might call f-hybrids: that is, where there are propositions A and B such that the f(A&B) is logically stronger than f(A)&f(B). The virtue of fictionalism over its antirealist rival, expressivism, is that it avoids the Frege-Geach truth-transfer problem. That’s a real advantage, however, only if the fictionalist can sustain some reasonable version of acceptance-transfer. What we have shown is that fictionalism almost invariably requires a double-attitude or multiattitude account of acceptance, and that generates problems for acceptance-transfer. Fictionalists of both the cognitivist and noncognitivist stripes thus have an attitude problem. Acknowledgments The authors would like to thank Eric Chwang, Richard Joyce, Simon Kirchin, Charles Pigden, and Michael Tooley, for their valuable criticisms and comments.

The Fictionalist’s Attitude Problem

215

References Harman, G. 2002. Internal critique: A logic is not a theory of reasoning and a theory of reasoning is not a logic. Studies in Logic and Practical Reasoning 1: 171–186. Joyce, R. 2001. The myth of morality. Cambridge: Cambridge University Press. Kalderon, M. E. 2005. Moral fictionalism. Oxford: Clarendon Press. Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin Books. van Fraassen, B. 1980. The scientific image. Oxford: Oxford University Press.

Abolishing Morality Richard Garner

The call to abolish morality is rarely heard but often rejected, which is not surprising because, on the face of it, abolishing morality seems like a horrible idea, one that would lead to massive suffering and disaster. But would it? Or is it possible that, as a few “abolitionists” have argued, morality is responsible for more suffering than it prevents? Difficult as it was, and remains, to abolish slavery, hunger, and torture, abolishing morality would be more difficult still. We can’t legislate it out of existence, eliminate its cause, or condemn morality as itself immoral. Morality is embedded in our language and in our vision of ourselves and our place in the world in a way that, we must hope, slavery, hunger, and torture are not. There is nothing that morality may not touch, and no side of any dispute that it cannot be called upon to support. But apart from its familiarity and its apparent usefulness, the main reason why most people would reject the call to stop speaking and acting as if there are moral truths about what is good, bad, right and wrong is that they believe that there are such truths. To convince someone to abolish or abandon morality, you must first get them to stop believing in moral properties, facts, and truths. You must get them to become moral antirealists, but, as we shall see, even that is not enough. The moral antirealist has arrived at the idea that moral judgments are not true statements of objective facts; but moral antirealism comes in two traditional flavors – noncognitivism and the error theory. Both begin by denying that that there are such things as moral facts, but the absence of moral facts matters little to the noncognitivist, who thinks that when we make moral judgments we aren’t even trying to state moral facts. Instead, we are expressing our feelings, emotions, attitudes, or stances, and/or attempting to influence the attitudes and behavior of others. A. J. Ayer startled a moralistic world with his claim that our “exhortations to moral virtue are not propositions at all, but ejaculations or commands which are designed to provoke the reader to action of a certain sort” (Ayer 1946, p. 103). Noncognitivists are rarely troubled by our use of moral language to express attitudes and to manipulate others, so they are unlikely to be moral abolitionists. R. Garner (B) Department of Philosophy, Ohio State University, Columbus, OH, USA

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0_13, 

217

218

R. Garner

By contrast, error theorists claim that our “exhortations to moral virtue” either are propositions, or presuppose propositions, about what is objectively good, bad, right, or wrong. Here the absence of moral facts is very important because it is what makes each of these moral judgments false. This second way to develop moral antirealism was first seriously defended thirty years ago by John Mackie in Ethics: Inventing Right and Wrong, the book we honor in this special edition of Ethical Theory and Moral Practice.1 In this chapter I want to explore a dispute among error theorists about how to deal with what all sides admit are false moral judgments. The advice of the moral abolitionist is to give up moral language, and with it the arrogance and interference that a belief in the objectivity of morality often occasions. But the contrary advice of the moral fictionalist is to retain moral language and moral thinking because morality is too useful to be abandoned. My aim here is to clarify the choice that arises for one who has embraced a moral error theory, and to argue that moral abolitionism has much to recommend it. I will assemble some considerations that suggest that the death of moralizing might be good for the individual and for society, and that moral abolitionism may be more useful than moral fictionalism in helping us reach the goals that most compassionate moralists seek.

1 Inventing Morality In discussing Mackie’s error theory, Simon Blackburn observed that since Mackie thought morality was infested with error, perhaps he should have abandoned moral concepts and ways of thought, or at least replaced them “by ones that serve our legitimate needs but avoid the mistake” (Blackburn 1993, p. 149).2 But Mackie was less inclined to abandon or replace moral thought and language than to make use of them. Like Hume, he believed that our habits of selfishness and confined generosity are incapable of generating sufficiently sociable behavior, and he concluded from this that “it is important that there should be a widespread tendency to act on moral grounds” (Mackie 1977, p. 124).3 Important social practices like recognizing property and keeping agreements need to be supplemented by a “moral overlay” that allows us to say that the actions we favor are objectively required by objective circumstances and that some of the things we desire have a value of their own. Unlike many moralists and moral philosophers who take the usefulness of moral beliefs for granted, Mackie directly addressed the question of whether morality does more harm than good. In Hume’s Moral Theory, published three years after Ethics, he defended his belief in the need for morality against those who recommend that

1 [Editors’

note: The present paper was first published in a special issue of Ethical Theory and Moral Practice that marked the 30-year anniversary of the publication of Mackie’s 1977 book.] 2 Blackburn could say this because, as a “quasi-realist,” he held that no mistake is involved in moral judgments. 3 Mackie later adds that the notion of a right is “valuable and indeed vital.” (1977, p. 173).

Abolishing Morality

219

we abandon or replace it. He argued that if we supplement important social practices like recognizing property and keeping agreements with the moral overlay, then people will be more likely to leave our possessions alone and to abide by their words. Mackie put forward a second reason for preserving the moral overlay. He said that when people believe in an objective truth about what is right and wrong it is easier to support and rationalize legal decisions and sanctions. If we could not anchor the use of force in some claim to a legitimate, objective, and moral ground, that use of force and the useful practices it supports could be more easily challenged. (See Mackie 1980, p. 154.) There may be something to this, but these two reasons for adopting the moral overlay are surprisingly weak. Consider the first. We all want others to respect our property, to tell the truth, and to keep their agreements, and it may be that when we insist that some practices are required by morality some people will be encouraged to adhere more closely to them. But it is not true that a practice plus a moral overlay will always be as fully, regularly, or willingly followed as that practice supplemented by some of the many available non-moral devices. We might promote the practice of keeping promises by early and extensive training in empathy, by strict surveillance and strong penalties for promise-breakers, or by massive doses of advertising by celebrities. Since we have access to these and other powerful ways to encourage promise-keeping, the abolitionist can argue that the moral overlay may be set aside in favor of more effective and less peculiar devices, some of which are already operating at full strength. It is hard to know what role “moral considerations” actually play in our choices because our behavior is a function of countless known and unknown factors.4 It is possible that when some of us make decisions, moral considerations are completely idle. But even when moral beliefs do play a role in choices, morality is flexible enough to be available to support any choice anyone is likely to want to make, including the choice of government officials to suppress what they choose to call “immorality.” This throws a different light on the idea that the moral overlay is useful when we want to justify social and legal sanctions, the second of the reasons Mackie gave for preserving our error-infested moral judgments and language. What good is morality if it can so readily be marshaled to defend the sanctions of a tyrant? After offering this brief and unconvincing rationale for exploiting the language of morality, Mackie admitted that “a case can be made out for the view that though what I am calling the basic practice is beneficial, it would be better without the moral overlay” (Mackie 1980, p. 154). The case he presented is impressive, and Ian Hinckfuss has even claimed that Mackie’s remarks here were an expression of “doubts about the usefulness of the moral institution” that he had begun to entertain “shortly before his untimely death” (Hinckfuss 1987, p. 18). In any event, the case Mackie constructed for what we are calling “abolitionism” involves three charges: morality inflames disputes and makes compromise difficult, it preserves

4 See

what Blackburn has to say about “the holism of the mental” (1998).

220

R. Garner

unfair arrangements and facilitates the misuse of power, and it makes global war possible. Morality inflames disputes because moralizing an issue tends to excite and confuse the parties involved. If we hope to resolve conflicts by arriving at a compromise, our task will be easier if moral disagreements are seen as partial conflicts of interest “without the embroidery of rights and moral justification” (Mackie 1980, p. 154). The controversy over abortion would not be nearly as intractable as it has become if the fiction of moral rights had not been appropriated by both sides. If the issue is not moralized, Roe v. Wade looks like a sensible compromise between two extreme positions, but when the right to life is set against the right to choose, neither side can yield without violating morality. A human embryo is what it is, but someone who insists on describing it with morally loaded terms like “person” or “innocent human baby” leaves no room for compromise over issues like abortion or embryonic stem cell research. How can anyone compromise with someone they see as wanting to murder babies? Not only does the moral overlay inflame disputes and make compromise difficult, the lack of an actual truth of the matter opens the game to everyone. Every possible moral value and argument can be met by an equal and opposing value or argument. The moral overlay adds an entire level of controversy to any dispute, and it introduces unanswerable questions that usurp the original question, which is always some practical question about what to do or support. This “moral turn” guarantees that the participants will be distracted from the real issue, and that the disagreement will flounder in rhetoric, confusion, or metaethics. The dangers of the moral overlay are far worse than Mackie thought. The second problem Mackie found with the moral overlay is that the addition of moral overtones to a practice “will tend to stabilize whatever differential advantages the various parties initially have” (Mackie 1980, p. 154). Property laws are made by those with property, and when the laws are given a moral defense, so are the inequalities. Then “the moral overtones of the duty of allegiance, with the associated concepts of loyalty, patriotism, and the like” can be used to condemn reformers as traitors and critics as criminals, and to facilitate the misuse of power (Mackie 1980, p. 155). In fairness to the moralist and the moral fictionalist, it should be admitted, as Mackie did, that moral principles can also be called upon to protect us from the abuse of the powerful. But these principles protect us only if those with the power to abuse us accept the idea of morality and are moved by it. It is likely that they do accept morality and that they are moved by it when it tells them what they want to hear; but it is also likely that they will never accept a moral directive as authentic if it threatens to deprive them of any property or privilege they cherish. Mackie’s third point was that “without morality there might, indeed, be more small-scale fighting; but war as we know it, organized on a national or an international scale, would be impossible” (Mackie 1980, p. 155). Mackie did not explain why he thought that without morality we would not have global or international wars. Perhaps he was thinking that we need the extra fictional power of morality to mobilize against a nation of strangers, but not against our neighbors, who always give us concrete causes for annoyance. He had no way of knowing that the threat

Abolishing Morality

221

from terrorists with weapons of mass destruction would eclipse the threat of global war at the start of the 21st century, nor could he have predicted the damage one individual or a few individuals might bring about in the name of morality. The truth seems to be that morality is invariably called upon to underwrite the actions of both sides of any violent conflict – large, medium, or small – and this does seem to be a reason for thinking that we might be better off without it. As strong as these considerations for eliminating the moral overlay are, they were not strong enough to turn Mackie into an abolitionist. He admitted that morality has “side effects some of which benefit some people at the expense of others, while others do more harm than good to almost everyone,” but he avoided a final summation by asking his imagined abolitionist opponent “What would work in its place?” Answering for that abolitionist, Mackie suggested that the basic practices might be supplemented with the “social psychologist’s techniques of conflict resolution.” Even so, he said, there is a “case for the other side.” The abolitionist’s replacement of the moral overlay with techniques of conflict resolution may not be the best option because it is “easiest to understand the techniques of conflict resolution working within an agreed framework of prima facie – not absolute – rights possessed by all the parties to an issue” (Mackie 1980, p. 155). Mackie may have been inclined to support the institution of morality, but if his final thought on the matter is that that the policy of appealing to moral rights, even prima facie ones, will resolve more conflicts than it nourishes, we may observe that this is not obviously true. We will be better off with prima facie moral rights than with absolute ones, but Mackie still needs to show that it will be easier to resolve practical conflicts when the parties believe in prima facie moral rights than when they believe in no moral rights. Unfortunately, reluctance to compromise can be resolute even when prima facie rights are the only rights mentioned. My prima facie right to peace and quiet is likely to seem stronger to me than your prima facie right to practice your trombone, and the very idea that a right, even a prima facie one, is objective, natural, or God-given, will lead to holding it with that special fondness that makes compromise difficult and violence attractive. A world where the concept of moral rights has been abandoned is a long way from anarchy. Take away moral rights and there remain legal rights, civil rights, conventional rights, and countless entitlements and liberties we regularly give to and demand from one another. You don’t need a right to privacy if others have a habit of respecting your need for privacy, or if the laws have the power to secure it. It is easy to think of ways to encourage people to take the needs and interests of others more seriously. We could work harder at teaching and promoting communication skills, tolerance, and empathy, and we could stop looking at compromise as if it were surrender. Mackie concluded his case for keeping the moral overlay by saying that morality is less likely to have regrettable side effects if it is understood (Mackie 1980, p. 156). But if he was right about what morality is, then when it is understood it will be seen as the fiction, invention, or projection it is, and those who understand this may no longer be interested in doing and avoiding things for moral reasons. So it is possible that, contrary to what Mackie said, to the extent that morality is understood it will

222

R. Garner

either fall out of use, or it will be used with even more cynicism and regrettable side effects than it now is.

2 Assertive Moral Fictionalism In a recent paper, Daniel Nolan, Greg Restall, and Caroline West endorse Mackie’s characterization of our deep moralist tendencies. They agree that most people do believe that there is a realm of objective moral facts with prescriptive and motivational force, and they add that when we make moral judgments we are saying things that are “literally false” (Nolan et al. 2005, p. 307). Nevertheless they claim that there are practical advantages to continuing to use moral language, and they recommend that we reject moral abolitionism, which they call “eliminativism” and characterize as the view that “we ought to abandon false talk of rightness and wrongness, goodness, and badness, duties and obligations, justice and injustice altogether.” Moral fictionalists, they say, believe that “realist moral discourse should be retained, even though it is strictly speaking false, because it is useful” (p. 308). Their claim is not that moral discourse should be employed when and only when it is useful, but rather that moral error theorists should embrace and encourage the practice of moralizing because everyone is better off with that practice in place. The moral abolitionist, more impressed by the harm of not seeing the error as an error, recommends that we abandon the practice, or better, replace it with some motivational aids that allow us to acknowledge and deal with things as they are. Of course, even a dedicated moral abolitionist may find an occasional moralistic utterance overwhelmingly useful, just as someone who has resolved to be truthful might resort to a lie to save a life. However, no moral abolitionist worthy of the name will find himself or herself affirming, in his or her own mind, judgments of moral right and wrong, or assumptions of intrinsic worth. Like Mackie, Nolan, Restall, and West offer several considerations to support their recommendation. They say that moral judgments help us coordinate attitudes and regulate interpersonal conflict, but our discussion of Mackie’s reflections suggests that the opposite may be true. If both sides can call upon morality to ratify their positions, then their attitudes will continue to clash, and the conflict is likely to increase. If we hope to resolve conflicts by arriving at a compromise, our task may actually be easier if moral disagreements are seen as conflicts of interest without, as Mackie said, “the embroidery of rights and moral justification.” By embracing moral fictionalism, error theorists regress to the comfortable familiarity of normative ethics, pretending to seek answers where, as they already know, no answers are available. Nolan, Restall, and West say that “the institution of realist morality has taught us to discuss the resolution of such matters in terms of rights, duties, and obligations, and to abandon these frameworks for discussion would deprive us of many of the procedures and tacit understandings that provide a well established framework within which such discussions take place” (p. 312). Exactly; but one of the advantages of eliminativism is that it frees us from that well-established framework

Abolishing Morality

223

and those tacit understandings that lock us into interminable arguments, ultimate impasses, righteousness, rhetoric, and error. The major argument used by Mackie and by Nolan, Restall, and West is that we should keep the moral overlay because we are better off with it than we are without it. But to this they add that moral concepts “so pervade ordinary thinking and discourse” that we would find it “difficult and inconvenient” to do without moral discourse (p. 311). This, of course, is not a very good reason for holding on to a false belief, or for encouraging others to do so. It was difficult to abandon geocentrism and the belief in the efficacy of human sacrifice, but we are better off for having done so. Actually, how difficult or inconvenient it will be to abandon the moral overlay, or any other set of false beliefs, will depend on how rigid and indoctrinated we are, and on whether we can find other ways to say what has to be said and do what has to be done. It might be difficult for some dedicated moralists to break the habit of moralizing, but not everyone is so inflexible. Contrary to what Nolan, Restall, and West suggest (p. 311), the changes we would have to make if we stopped employing the moral overlay are not nearly as great as those that would be required of us if we decided to abandon the language of folk psychology. Others (who apparently have not tried it) make a point of how difficult it might be to break the habit of moralizing. William Lycan says that “to produce a genuine freedom from moral intuitions, one would need a steady diet of hard drugs, or some other very powerful alienating force” (1985, n. 29). Richard Joyce quotes Peter Singer, who claims that even if we avoid all moral language “we will find it impossible to prevent ourselves inwardly classifying actions as right or wrong.” Joyce observes, and I agree with him, that these remarks are overstated (Joyce 2001, p. 171).5 Nolan, Restall, and West, emphasizing the tenaciousness of the habit of moralizing, mention a kind of “inconsistent nihilism” where error theorists “find themselves unwilling or unable to refrain from making positive moral judgments” (2005, p. 314). There may be some people who are addicted in this way to morality, but almost any habit can be broken. Even if, as Joyce argues, there is a genetic basis to think morally, we have managed to alter and suppress other once useful but currently troublesome behaviors like territoriality and aggression that are also grounded in dispositions based on natural selection.6 Another benefit of moral fictionalism alleged by Nolan, Restall, and West is that it “enables us to avoid raising complicated metaethical issues every time we wish to discuss an applied ethical question” (p. 311). But in reality the opposite seems to be true. Any time we find ourselves involved in some difficult question of applied ethics, questions of metaethics inevitably arise. It is often reasonable to ask if some disagreement may be based on a semantic misunderstanding, and it is customary for both sides of any moral argument to display and demand reasons. So how does

5 Joyce’s

reference is to Peter Singer’s The Expanding Circle (1981, p. ix). (2006). In Chapter 6 of The Myth of Morality (2001), Joyce explains how evolution may have given us a “hardwired predilection to believe that moral obligations exist” (p. 146).

6 Joyce

224

R. Garner

fictionalism bypass metaethics? The best way to avoid metaethical questions, or at least the ones we cannot answer, is once and for all to reject normative ethics – that is, to become eliminativists. No one is likely to offer moral fictionalism as a descriptive account of what people now mean or think when they make moral judgments. For one thing, most people are not moral error theorists. When Nolan, Restall, and West say that the descriptive option “is not likely to be plausible” (p. 322), they are rejecting the obviously false claim that we are all error-theorists intent on preserving the moral overlay. Their form of moral fictionalism can be seen as a recommendation to other error theorists not to abandon the moral overlay. They make this recommendation because they believe that a society that subscribes to moralistic beliefs and practices will do better than one that does not. The moral eliminativist or abolitionist disagrees, and urges those same error theorists to replace the moral overlay with more effective and less duplicitous devices. When Mackie defended his version of moral fictionalism, he was thinking of our actual world and advising the rare error theorist to continue to operate with, and even improve upon, the moral overlay. Nolan, Restall, and West agree that when we make moral judgments we are making assertions that are “literally false,” but it is not clear what sort of world they are talking about. In arguing for the value of their fictionalism, they imagine a case where “all the parties to the debate are moral fictionalists.” Does this mean that we are to try to answer the question of whether moral fictionalism is sufficiently useful by imagining a world where everyone, or nearly everyone, is a moral fictionalist, and then asking if that world generates more satisfactions than one where everyone, or nearly everyone, is a moral abolitionist? It is hard to know even how to begin such a thought experiment, but we can be sure that of those who try it, some will favor the world of fictionalists and others will disagree.

3 Assertive Moral Abolitionism The assertive moral fictionalist is an error theorist who urges us (or at least urges other error theorists) to continue to assert moral judgments, and to speak and argue as moral realists do. Assertive moral abolitionists also construe moral judgments as false assertions, but they urge us to stop making them because they believe that any benefits that come from pretending that moral realism is true are outweighed by the harm that comes from having to promote and defend a series of easily questioned falsehoods. If the question of whether we are better off with or without the institution of morality is an empirical question, it is one to which we are not likely to find a definitive answer. There will be plenty of opinions, but the future is unknowable and our imaginations are both limited and hostage to our desires. That said, it is also true that we can sometimes tell when an idea (revenge) or an institution (slavery) is

Abolishing Morality

225

likely to have bad consequences. After rehearsing their reasons for recommending that we keep the moral overlay, moral fictionalists often acknowledge that morality has, as Mackie conceded, some side-effects that “do more harm than good to almost everyone,” but they do not go into this very deeply (or at all). One error theorist who did accentuate the negative was Ian Hinckfuss (1987), who has offered some impressive considerations in support of the abolitionist view that we would be better off without morality. Hinckfuss claimed that we live in a “moral society,” that is, a society in which almost everyone does accept the literal truth of the moral overlay. He acknowledged that false moral beliefs may have some benefits, but argued that any moral society will come with a number of disagreeable features that are essential to its propagation and preservation, features like “elitism, authoritarianism, guilt complexes, ego competition, economic inequality and war” (Hinckfuss 1987, p. v). Moral societies, he added, are intellectually dishonest, “inefficient in maximizing human happiness, satisfaction, or self-esteem,” and “because of the threat of war with other societies, physically dangerous.” The moral overlay actually hinders the resolution of conflicts and fosters the exploitation of the “poor and the weak by the rich and powerful” (Hinckfuss 1987, pp. 20–21). In a moral society, children will be raised in “an environment of continual moral injunctions,” and they will be conditioned to want to be good, and trained to respect the moral authorities and their values. When they reach this point they are “in a position to be morally propagandized by those whom they regard as their ‘betters,’ that is, those who they feel know more about what is right and what is wrong than they do” (Hinckfuss 1987, p. 23). Ordinary members of a moral society will not be able to find the moral truth (because there is none), but they will have learned to believe in moral truth, and to rely on the members of a moral “elite,” who take themselves to be superior in knowledge, virtue, and worth. When one is an honored member of a moral elite “it is easy to believe that what one wishes for oneself is morally permissible, and how one wants others to behave is morally obligatory” (Hinckfuss 1987, p. 27). Hinckfuss is more emphatic than Mackie about the tendency of those with strong moral beliefs to go to war. After mentioning a dozen or so of the world’s bloodiest conflicts, he asks us to “think of how the situation would have been if, by a miracle, moral thought could have been eradicated from the minds of all the agents involved. . . . There would be no sense of duty, no sense of loyalty, no patriotism, no feeling morally obliged to fight for a cause, no sense that the people one is trying to kill or subjugate are less worthy of survival or freedom than oneself or anyone else” (Hinckfuss 1987, pp. 45–46). Joyce (2001, p. 180) replies to this by pointing out that morality didn’t cause these wars, which is right, but he neglects the fact that morality made it more difficult to resolve the conflicts that led to them, and the fact that once a war starts, morality can be used to justify inflicting any cruelty deemed necessary for victory. Getting rid of morality will not solve all the world’s problems, and no moral abolitionist would claim that it would, but it would allow us to see conflicts of interests

226

R. Garner

for what they are and other people for who they are, and that by itself will undermine demagoguery and fanaticism. If Hinckfuss is even partially right about what it takes to institute and enforce morality, it may be a far more dangerous institution than its fictionalist supporters have allowed. Although he thought that all of the concepts of morality are dangerous, Hinckfuss found the notion of moral desert particularly harmful. We have been conditioned to think that some people are better than others, and that because of this they deserve the advantages they have acquired. These aristocrats and masters have “a vested interest in preserving the system,” and they have the power to do so. If somehow we could manage to “discard” the idea that some people deserve their unfair advantages – or better, if we could abandon the doctrine of moral desert entirely – “the perpetuation mechanism of morality would be lost and morality itself would rapidly become non-existent.” This, Hinckfuss says, “is the bloodless. . . revolution that I, for one, would welcome” (1987, p. 40).

4 Practical Rationality If we do decide to abandon our false moral beliefs and to refrain from using moral categories and language, we seem to be left with “practical rationality.” We must figure out what is in our short-, middle-, and long-term interest, and base our decisions on that. Practical rationality might tell us that widespread use of the moral overlay is in our interest, or it might support abolitionism. No one can say for sure; but no one who is interested in achieving his or her goals is a stranger to practical rationality. Moral abolitionists have no desire to abolish it, and neither do moralists or moral fictionalists. The real question is: “Is practical rationality enough, and if not, what else do we have?” It is possible, and perhaps even likely, that practical rationality is another fiction. We have never been very good at calculating consequences, and usually the proponents of each side of a dispute predict those consequences that bolster their own positions. The same complex of computational problems that keeps utilitarians from ever being sure of their bottom lines also undermines those who dream about calculating costs and benefits. How, for example, do we weigh the claims of our short-, middle-, and long-term goals? Nevertheless, if practical rationality is a fiction, it is less obviously a fiction than is morality. Sometimes when we predict a disaster, disaster comes, and some things that appear to be in our interest really are in our interest. Joyce says that Hume’s answer to the sensible knave is “roughly correct.” It may be in our interest (almost all of the time) to refrain from “defecting,” and even when it appears that we can get away with some breach of trust, we can’t be sure, and the penalty for a mistake is often great. So we usually do have “good instrumental reasons for acting in a cooperative manner.” But, he adds, this answer would only “be good if we lived in a world populated entirely by humans enjoying perfect rationality.” Since we don’t, practical rationality, a concern for our own welfare, will not

Abolishing Morality

227

prevent all defections (Joyce 2001, p. 210). So, Joyce concludes, we need to supplement practical rationality with the fiction of morality, but not by embracing or promoting, or even asserting, false moral beliefs. His moral fictionalist will still be an error theorist, but one who expresses moral judgments without asserting them and who thinks moral thoughts but has no moral beliefs.

5 Non-assertive Moral Fictionalism If we add up the negative features of the moral overlay mentioned by Mackie and Hinckfuss, and then factor in the criticisms of the arguments brought forward in its support, we do have a case against using morality. Since moral judgments, we are now assuming, are false, what we say is sure to conflict with reality at many points, and then we will need to resort to evasion, obfuscation, or sophistry just to maintain our fiction. It is hard to estimate the damage this can do. If we continue to insist on the truth of our fiction, if we defend it as strongly as a convinced moralist would, then we are courting doxastic disaster, Orwellian epistemology, and perhaps a nervous breakdown. (See Joyce 2001, pp. 214–215). If we want the benefits of the moral overlay without the damage spawned by the ill-advised worship of false beliefs, we need to find some way to make moral judgments without making or believing false assertions. This is the aim of “nonassertive moral fictionalism,” a development of the error theory explored by Joyce, who sees this position as one last way an error theorist might be able to reject the moral abolitionist’s disconcerting recommendation. The non-assertive moral fictionalist, says Joyce, wants to continue with the moral overlay because he believes that “morality, though in error, is a justified practice in light of its usefulness” (2001, p. 173). Joyce writes as if when the truth of the error theory dawns on us it will be relatively easy to continue speaking and even thinking as moralists. He speaks of someone who has become an error theorist but who is so used to employing the concepts of moral right and moral wrong that she carries on using them in everyday life and “is happy to do so” (2001, p. 224). But getting clear about metaethics is hard work, and an actual insight into the truth of the error theory is revolutionary and not easily put aside. A clear-headed and convinced error theorist is not likely to fall mindlessly into old habits of moral speech, so continuing with the moral overlay will be a decision, at least at first, to make and defend what one believes to be false assertions about value and obligation. When Fred, newly converted to the error theory, decides to continue using moral language with his moralist friends, he will not be slipping into a kind of non-assertive fictionalism because he will know that his moralist friends will take him to be making full strength moral assertions. Joyce says that Fred, moralizing in the midst of moralists, may fail to assert because “the conventional understanding between speaker and audience has broken down to such a degree that Fred’s speech acts are. . . ‘misfiring’.” Fred, he adds, has “ceased to speak” (2001, p. 173). But unless Fred has completely lost touch with

228

R. Garner

the existing conventions, and with the fact that willingness to participate in those conventions is signaled by the straightforward utterance of a non-bizarre descriptive sentence, he knows that he will be understood as asserting. This means that he can’t even intend not to be asserting. So he is asserting. But now suppose that Fred does somehow forget or repress his understanding of the error theory. Imagine that, as Joyce says, Fred has pushed his former acceptance of the error theory so far to the back of his mind that moral thoughts like “But it would be wrong” seem like beliefs to him. He says that Fred’s “well-rooted” moralistic habits are likely to be brought up short only when he “enters a very critical context of discussion, such as a conversation on metaethics” (2001, pp. 218–219). Joyce may believe that these metaethical excursions would be rare, but I suspect that questions about meaning and justification will turn up as soon as we disagree about anything that really matters to us. Since very many things do matter to us, frequent metaethical digressions will remind us of our belief in the error theory, and that will make it increasingly difficult to stay immersed in our fiction. If these metaethical episodes do not alert Fred to the objective (as opposed to the fictional) situation, if he continues to ignore his former metaethical insights, if he begins to have moral feelings, moral outrage, moral guilt, and moral arguments, then we have every reason to say that he has reverted to his former moral beliefs, and to the error he once identified and abandoned. So how can Fred become a non-assertive moral fictionalist? How can he say things he knows to be false and yet avoid making false assertions? One way to utter a sentence without asserting it is to give some sign that neutralizes the assertive force of one’s utterance. A sarcastic tone, a wink, or extreme exaggeration will withdraw assertive force and set an interpretation problem for the listener. But unless he relies on some such device, Fred cannot unilaterally withdraw assertive force from his utterances. The other way Fred can manage not to assert the sentences he utters, says Joyce, is for him to become a noncognitivist about his own moral discourse. Noncognitivism “might become true if we were to alter our attitude towards moral discourse.” Then we would be able to utter a sentence like “That was morally wrong,” not as an assertion, an expression of a belief, but “as an expression of a thought” (Joyce 2001, p. 201). The content would remain the same – namely, the false claim that an objective wrongness qualifies the act – but we would be uttering the sentence with a non-assertive force (or with no force). This can be done, but not unilaterally. Fred can address moral utterances to others without asserting them if he is participating in a convention that makes this possible. As Joyce notes, “whether one asserts. . . depends upon a framework of linguistic conventions within which one’s utterance occurs” (2001, p. 203). The non-assertive moral fictionalist looks forward to the day when we are all (or almost all) error theorists, devoted to the moral overlay and committed by convention to understand moral utterances not as assertions but as expressions of thoughts, or emotions, or whatever. This might come to pass, but it is not easy to see how we could get there from our present deeply moralistic society, or to see very clearly what “there” might look like. Perhaps we can begin to make sense of a world filled with non-assertive

Abolishing Morality

229

moral fictionalists by considering a group of error theorists for whom the idea of moral fictionalism has become an option. Only in such a group could a convention develop that would allow speakers to unhook their moral utterances from assertive discourse.

6 A Group of Moral Fictionalists Becoming a non-assertive moral fictionalist is a three-step process. The first step is to accept the error theory, the second is to choose to continue with the moral overlay, and the third is to form, or to become part of, a group with a convention that enables one to utter moral judgments without asserting them. A group of refugees from a seminar in metaethics who have decided to embrace the error theory could agree to speak in moral terms with each other, and agree that in doing so they were expressing but not asserting thoughts. But if these same error theorists express moral judgments in books, or in public, or to “outsiders,” those moral judgments will correctly be taken to be assertions of belief. If we are to be non-assertive moral fictionalists we will either have to speak “morality” only with the members of our group, or bring it about that the larger group accepts a convention to the effect that uttering a moral judgment triggers a withdrawal of assertoric force. When Fred, our newly minted error theorist, looks around, he will find that almost everyone (including the moral realists and both the assertive and the non-assertive moral fictionalists) will be deeply involved in the moral overlay, and will be uttering moral judgments, engaging in moral arguments, expressing moral outrage, and making moral demands. If he is surrounded by non-assertive moral fictionalists among whom expressing a moral judgment is understood not to be making an assertion, how will he ever find this out? It will not be apparent from anything anyone says or does. If he asks someone who happens to be “immersed” in the fiction, that person may fail to admit, or even to remember, his or her disbelief, and simply respond to the question as a genuine moral realist would. If indeed it is possible to be so out of touch with what one believes (which I doubt), then that forgetful error theorist might even say “Of course I really believe that it is morally wrong to torture innocents for fun.”

7 Problems for Non-assertive Fictionalism: Beliefs vs. Thoughts and Assertion vs. Expression Non-assertive moral fictionalists are error theorists who set out to express moral judgments without asserting or believing them. In this way they hope to reap the advantages of belief in morality and to avoid the bad effects of asserting and having to defend false beliefs. Joyce says that we can use the moral overlay in our day-today life as much as we please, but as long as we would deny it in our most critical moments, we do not assert the content of our utterance, nor do we believe it – we just think it (Joyce 2001, p. 201). The difference between a thought and a belief, he

230

R. Garner

says, is the disposition to dissent (2001, p. 219). As long as Fred once embraced the error theory (in a critical mode – perhaps in a philosophy class), and as long as he might, under similar or different circumstances, do so again, we don’t have to admit that he did have moral beliefs in the interim. (2001, pp. 190–194). That is an unusual interpretation of belief. The distinction between thoughts and beliefs is real enough, but pushing it to allow the non-assertive moral fictionalist to say that Fred didn’t have moral beliefs even while he was thinking and acting as if he did seems unnecessary. In any case, it is the thinking and the acting that cause the trouble and that lead to the endless arguments and intensified conflict. The bad effects of believing in moral facts are as likely to result from the pretence as from the belief, especially since it is sometimes difficult to distinguish the one from the other. Perhaps we can say this: the deeper we go into our moralistic trances, the more likely we are to experience the full range of the bad effects attributed to moral belief; and the less we immerse ourselves in the moral fiction, the less likely we are to see any of its alleged beneficial effects. What now of those error theorists who continue with the moral overlay but do not lose themselves and revert to moral belief? This group will include assertive moral fictionalists like Mackie, but also non-assertive moral fictionalists who do not intend for their moral utterances to be (or to be taken as) expressions of moral beliefs. We know that if the conventions for doing so are in place we can use moral language without making moral assertions. But why would a group of self-conscious and mutually aware error theorists revisit their moralist past? Assertive moral fictionalists at least believe that others are open to moral persuasion, but if you know that your audience doesn’t believe in an objective morality any more than you do, what is the point? Falling back into the moral framework would require an immense expenditure of time and effort on a project both you and your target audience would know to be doomed. There may be differences between moral beliefs and moral thoughts, but when they go public they look alike and they both need to be defended. If we defend them with vigor, then we will be indistinguishable from the moralists, and that will facilitate the return of those interminable debates and ritualistic arguments. But if we defend our fictive moral utterances and thoughts half-heartedly, we will show others and remind ourselves that we are not really serious about what we are saying. Joyce claims that non-assertive moral fictionalism is useful in overcoming lapses due to weakness of will. He observes that when we are tempted to act in ways that are not in our interest, moral beliefs might help us do what we know we (practically) ought, but, he adds, so can moral thoughts. Just thinking of a fire destroying one’s house and possessions might cause anxiety or fear. Similarly, just repeating or hearing the familiar language of moral obligation might trigger a familiar impulse to obey, and that may be enough to get us to resist some temptation or to eat our vegetables. Apparently the magic works on others in the same way. If I tell you that something is morally wrong, that very thought might elicit in you a mild but negative emotion against the act, even if we are both non-assertive fictionalists sharing an assertion-canceling convention. A “mild” emotional reaction to a moral thought may add some subliminal motivation, but it is unlikely to be strong enough to combat an impulse that has already

Abolishing Morality

231

managed to defy our self-interest. Joyce says that even if the influence of the thought were minimal, “maintaining the moral discourse. . . would at least be a reliable improvement on straight thinking (and straight talk) about one’s long-term preference satisfaction” (Joyce 2001, p. 217). This may be so, but moral discourse may not be an improvement on straight thinking and talk that ranges a little more widely than a single person’s long-term preference satisfaction. Many of us automatically factor in the interests of others we know and cherish, and sometimes we give weight to the interests of complete strangers. A vast number of known and unknown factors resolve into our decisions: thoughts, beliefs, fears, desires, emotions, habits, memories, resolutions, impulses generated by chemicals and hormones, or by what someone has just said to us. Joyce emphasizes “how easily influenced humans are, concerning important decisions, by apparently trifling factors: what story has just been on the radio, whether information is handwritten, what one is wearing” (2001, pp. 217–218). In this nexus of causes, it is unlikely that the echo of emotion from a moral thought will find a way to play much of a role in our actual decisions. Given the effort we must spend to keep the fiction going, and given what it costs us in integrity, it is far from clear that non-assertive moral fictionalism, even if it were achievable, would be a wise policy.

8 A Personal Choice We may never find an indisputable answer to the question of whether the moral overlay helps us more than it harms us. But I hope it is now clear that there are more problems with morality than moralists and moral fictionalists usually admit. Anyone who has come to accept the error theory will have to decide what to say in situations where a moralist would naturally resort to the use of moral language. An isolated error theorist cannot decide unilaterally to stop using moral language assertively, but he or she can decide unilaterally to stop using moral language. Future error theorists may someday blaze a trail to non-assertive moral fictionalism by spreading a convention that cancels the assertive force of a moral judgment. But that day is not likely to come in our lifetimes, so the practical question for an error theorist remains whether to abandon the habitual day-to-day use of the moral overlay. This is a question one can understand, and, unlike the question of whether “we” would be better off if (almost) everyone were an abolitionist, there is a simple and straightforward way to deal with it. We can just cut back on our use of moral language and see how things go. Even error theorists who aren’t ready to give up the moral overlay can try out abolitionism in order to see what happens when we minimize our moralizing. This experiment will not take some arduous act of self-control. All one has to do is to watch oneself, and to anticipate, and then restrain, the impulse to pronounce moral judgments, publicly or privately. For amoralists – that is, error theorists who have already come to believe in the falsity of all moral judgments – cutting back on moral pronouncements will be no more difficult than cutting back on swearing, and not nearly as difficult as getting rid of an accent.

232

R. Garner

Error theorists who can refrain from falling into actual moralism or any kind of moral fictionalism will soon come to understand that we are in no way limited in our ability to express and communicate our attitudes, requirements, and feelings. Instead of telling others about their moral obligations, we can tell them what we want them to do, and then we can explain why. We can express annoyance, anger, and enthusiasm, each of which has an effect on what people do, and none of which requires duplicity or moralizing. The moral abolitionist is equipped, as we all are, with uncountable habits, preferences, policies, aims, and impulses that can step up to take the place usually thought to be occupied by moral beliefs and thoughts. If we complete the transition to these purely non-moral considerations, we will find that there will be less to argue about, and that our conflicts and disagreements with others, at last seen for what they are, can be addressed and resolved. The moral overlay may have been useful in the past, but if someone who sees it for what it is can discard it, that person can be a witness to the fact that life without recourse to moral concepts is preferable to pretending to believe what we know to be false, and healthier than allowing ourselves to forget our hard-won insights about what is really going on when people make moral judgments. Assertive moral fictionalism is dangerous because it undermines our integrity by forcing us to find ways to defend things we know to be false. Non-assertive moral fictionalists hope to avoid this danger, but we have seen that there may be no path from our current moralistic society to global non-assertive moral fictionalism. Since the majority of users of moral language are moral realists, non-assertive moral fictionalism that extends beyond a small group of confused philosophers is not at this time a live option for an error theorist. Error theorists do have a choice, but it is between assertive moral fictionalism and moral abolitionism. I hope that by now it is clear that abolishing morality may not be such a “horrible idea.” It might even be an essential step in achieving many of the goals well-meaning moralists and moral fictionalists have always cherished. The first step on the path to both moral fictionalism and moral abolitionism is the acceptance of the error theory. What I share with Joyce, Hinckfuss, Nolan, Restall, and West, and with other error theorists here unmentioned, is a deep intellectual debt to John Mackie. By bringing his exemplary clarity and common sense to moral philosophy he opened the way to a radical revision of our ideas about morality. His attack on moral objectivity is finally being acknowledged and taken seriously by moral philosophers, and his respect for the power and usefulness of morality is shared by both assertive and non-assertive moral fictionalists. In the work we here celebrate, Ethics: Inventing Right and Wrong, Mackie made his case for the error theory, but he expressed no doubts about the usefulness of the moral overlay. The later reflections that appear in Hume’s Moral Philosophy reveal that he subsequently came to take moral abolitionism quite seriously. Perhaps further reflection would have led him to a position closer to that of Hinckfuss (which is what Hinckfuss thought), or to Joyce’s variety of non-assertive moral fictionalism. It may be useless or even impious to speculate, but I like to think that if he had been given time to carry his reflections further, Mackie may not have drawn back from what seems to the moral abolitionist to be the reasonable outcome of a belief in the error theory: the

Abolishing Morality

233

resolution to purge the error from our words and our thoughts. What serious philosopher can long recommend that we promote a policy of expressing and supporting, for an uncertain future advantage, beliefs, or even thoughts, that we understand to be totally, completely, and unquestionably false?

References Ayer, A. J. 1946. Language, truth and logic. 2nd ed. New York: Dover. Blackburn, S. 1993. Errors and the phenomenology of value. In his Essays in quasi-realism, 149– 165. New York: Oxford University Press. Blackburn, S. 1998. Ruling passions. Oxford: Clarendon Press. Hinckfuss, I. 1987. The moral society: Its structure and effects. Canberra: Australian National University. Joyce, R. 2001. The myth of morality. Cambridge: Cambridge University Press. Joyce, R. 2006. The evolution of morality. Cambridge, MA: MIT Press. Lycan, W. 1985. Moral facts and moral knowledge. Southern Journal of Philosophy 24: 79–94. Mackie, J. L. 1977. Ethics: Inventing right and wrong. London: Penguin. Mackie, J. L. 1980. Hume’s moral theory. London: Routledge and Kegan Paul. Nolan, D., G. Restall, and C. West. 2005. Moral fictionalism versus the rest. Australasian Journal of Philosophy 83: 307–330. Singer, P. 1981. The expanding circle. Oxford: Oxford University Press.

Index

A Abolitionism, moral, xiii, xxii, xxiii, 184-185, 185 n3, 194, 217-233 Adams, Ernest, 75 n10 Amoralism, 26, 179, 193, 231 Anaxarchus, xi Ansell-Pearson, Keith, 17 n1 Anscombe, Elizabeth, 50 Anthropomorphism, 36-37 Aristocles of Messina, xiv Atheism, ix, x, xii, xiii, xv, xvi, 1, 8, 11, 12, 50 n10, 71, 71 n1, 75-76 Audi, Robert, 73 Ayer, A. J., 71 n2, 217 B Bennett, Jonathan, 75 n10 Bentham, Jeremy, 11 Bierce, Ambrose, 3, 14 Blackburn, Simon, xx, 22-25, 27, 32, 33, 73, 75, 103, 108 n18, 119, 171 n3, 218, 218 n2, 219 n4 Blame, argument from, 156-157 Blixen, Karen, 8-9 Bloomfield, Paul, 102 n3, 111 n23 Braddon-Mitchell, David, 195 Brink, David, 45, 57, 76, 95, 102, 102 n3, 104, 104 n11, 105, 107-109, 146, 176 n8, 177, 179, 185 n5, 186 n6 Burgess, John P., xix-xx C Callicles, ix Carroll, Lewis, 11 and Alice in Wonderland 176 Casati, Roberto, 41 Cawkwell, George, 21 Charlotte’s Web, 82-85 Chisholm, Roderick, 112 n25

Christensen, David, 110 n22 Clarke, Samuel, xviii, 56, 57, 61 Cognitivism, moral, xx, 1, 40, 50, 103 n7, 104 n11, 183, 201-214 Conservatism, epistemic, 45, 52, 109-110, 113 Copp, David, xxi-xxii, 141, 142, 142 n3, 143 n6, 145, 145 n12, 146, 153, 153 n19, 155 n21, 162, 163, 179 Craig, Edward, 43 n6 Crary, Alice, 172 n4, 176 n8 Credulity, Principle of, 110-112

D Dancy, Jonathan, 45, 102, 102 n3, 104, 108 n17 Daniels, Norman, 136 Danto, Arthur, 17 n1 Darwall, Stephen, 92, 97, 155, 156-159, 157 n23, 157 n24, 161-162 Davidson, Donald, 123 Demetriou, Daniel, xix, xxiii Disagreement, argument from: See Relativity, argument from Doppelganger problem, 27-29 “Reinforced Doppelganger” problem, 29-32 Dostoyevsky, Fyodor, xiv, xiv n3 Dreier, Jamie, xi, xii, xxi, 72 n4, 102 n3 Dworkin, Ronald, 31-32, 113-114, 113 n26

E Edgington, Dorothy, 75 n10 Eliminativism, moral: See Abolitionism, moral Emotivism, 1, 2, 3. See also Noncognitivism

R. Joyce, S. Kirchin (eds.), A World Without Values, Philosophical Studies Series 114, C Springer Science+Business Media B.V. 2010 DOI 10.1007/978-90-481-3339-0, 

235

236 Eusebius of Caesaria, xiv n2 Expressivism, 73-77, 77 n14, 79, 82, 85, 145, 145 n12, 168, 169, 171, 171 n3, 173-174, 202-203, 214. See also Noncognitivism F Fictionalism, moral, xiii, xxii-xxiii, xxiii n5, 196, 197-198, 200-214, 218, 222-224, 227-233 Field, Hartry, 20, 28 Fine, Cordelia, 67 Flaubert, Gustave, 19 Foot, Philippa, 179 Frege-Geach problem, 202-214 Freud, Sigmund, 7, 43 n7, 76, 76 n11 Fungus (example from Mackie), 36-37, 39-42 G Garner, Richard, xv, xviii, xix, xxiii, 21, 26, 96, 103 n9, 119, 176 n8, 177, 185 n3 Gauthier, David, 124 Geach, Peter, 2, 13 Gert, Bernard, 69 Gibbard, Allan, 73, 74-75, 75 n7, 103 n8 Gladiators (example from Joyce), 149, 178 Greene, Joshua, 103 n9 H Hare, R. M., 3, 13, 18, 77 n14, 77 n15, 78, 103 n8, 128-130, 132, 171 n3 Harman, Gilbert, 4, 5, 13, 108 n18, 108 n19, 136, 205-207 Hinckfuss, Ian, 14, 21, 26, 185 n3, 219, 225-226, 227, 232 Hobbes, Thomas, 187-188 Hookway, Christopher, 105 n13 Horgan, Terry, 73, 74, 74 n6, 76 Huemer, Michael, 45, 45 n8 Hume, David, 3, 51, 187, 189, 218, 226 on projectivism, 35, 37 n2, 42 n5, 43 n6 theory of motivation/action, 59, 151, 183 theory of reasons, 88, 97, 122, 163, 183-184, 196 Hypostatization, 36 I Imperative: categorical, xvii, xix, 2-3, 5-6, 51, 90, 95-97, 166, 171, 172, 176 n9, 178-180, 192 hypothetical, xviii, 25-26, 32, 89-90, 96-97, 171

Index Inference to the best explanation, 107-110, 112, 116 Internalism, xx-xxi, xxii, 82, 85, 88, 92-93, 156 n22 many versions distinguished, 55-70 motivational internalism, 5, 56-60, 64-67, 70, 81, 89, 93, 96, 144, 183-198 reasons internalism, 56-59, 65-67, 70, 81, 175 J Jackson, Frank, 101 n1 Johnson, Samuel, xiv Johnston, Mark, 41 Joyce, Richard, xiii, xx, 21, 38, 39 n4, 41, 42 n5, 45 n8, 76 n13, 87, 95-97, 103 n9, 109 n20, 119, 142, 142 n5, 143, 147-164, 147 n14, 149 n15, 167, 169, 171, 176 n8, 178-181, 184, 184 n1, 192, 196, 197, 200, 223, 223 n5, 223 n6, 225, 226-231, 232 Julian (the Apostate), 1, 14 K Kalderon, Mark, xxiii, 201-214 Kant, Immanuel, 78, 81 n18, 106 n16, 120, 128, 130, 141, 157 Kavka, Gregory, 179 Kennett, Jeanette, 67 King, Jeffrey, 161 n25 Kirchin, Simon, xi, xxi, xxii, 104, 172 n4 Korsgaard, Christine, 90 n2, 93, 96, 102 n6, 122, 143 n6, 149, 156 n22 L Leiter, Brian, 17 n3 Lewis, David, 119, 125 Lillehammer, Hallvard, 103 n9 Loeb, Don, xix, xxi, 45, 106, 106 n15, 108 n19, 174 Lovibond, Sabina, 102 n3, 192 n11 Lycan, William, 223 M Mackie, John L. (quotations), xi, xi-xii, xvi, xvii, xviii, xix, 19, 35, 36, 37, 38, 44, 45, 46, 50, 50 n10, 51, 51-52, 52, 56, 58, 59, 63, 68, 71-72 n3, 77, 78, 80, 80 n17, 81, 89, 95, 98, 103, 120, 144, 144 n10, 145, 184, 187, 188, 218, 218 n3, 219, 220, 221 Marx, Karl, 190-191 McDermott, Michael, 193 McDowell, John, 119, 173 n5

Index McNaughton, David, 102, 102 n3, 104, 105 n14, 108 n17, 111 n23 Mill, J. S., 179 Moore, G. E., 9, 76, 78, 79, 120, 121, 186 Mooreanism, 73, 85, 134, 138 Moorean non-natural qualities, 121, 122-124, 125, 126, 127, 128, 130, 134, 135, 186 Moral experience, argument from, xxi, 101-116 Moral realism: See Realism, moral Motivational internalism: See Internalism N Nagel, Thomas, 1, 13, 102 n3, 102 n4, 108 n17, 113 n26 Naturalism, moral, xxii, 52, 142-143, 143 n6, 159-163 Nichols, Shaun, 105 n13 Nihilism, moral, x, xi, xx, 11, 17-33, 223 Nietzsche, Friedrich, xx, 17, 18, 19-22, 25, 29, 31, 33, 191, 191 n10 Nolan, Daniel, 222-224, 232 Noncognitivism, x, xii, xxiii, xxiii n5, 27, 30, 37-38, 44, 51, 73-74, 103, 103 n7, 103 n8, 104 n11, 145, 196, 197, 199, 201-214, 217, 228 Nonnaturalism, moral, 85, 134, 138, 145 O Objectification, xx, 35-52, 95 Objective prescriptions (and objective prescriptivity), xi, xviii, xxi, xxii, 35, 46-47, 48-52, 56, 80, 175-177, 177 n11, 184 n1 Oddie, Graham, xix, xxiii O’Leary-Hawthorne, John, 75 n9 P Parfit, Derek, 90 n2, 91, 130, 133, 134 Phillips, David, xix, xxi, 149, 175 Phlogiston (example), x, xvi, 180-181 Pigden, Charles R., xi, xii, xix, xxi, 144 n9, 191 n10, Plato, ix, xviii, 56, 61, 76, 78, 81, 81 n18, 120, 178, 187 Pettit, Philip, 41, 92 Prescriptivism, 1, 2, 3, 5, 78. See also Noncognitivism Price, Huw, 75 n9 Price, Richard, 111 n24 Prisoner’s dilemma, 188 Projectivism: See Hume; Objectification Protagoras, 187

237 Pybus, Cassandra, 21 Pyrrho, x Q Quasi-realism, xii, xxi, 23, 73-74, 103, 218 n2 Queerness, argument from, xv, xvii-xix, xxi, xxii, 35, 45-46, 51, 52, 55, 58, 59, 64, 65, 68, 79-85, 90, 95-96, 120, 121, 122-123, 128, 135, 144-147, 159, 175-177, 176 n8, 184 n1 Quine, W. V., 6, 109 n21 R Railton, Peter, 75 n8, 179 Rawls, John, 136, 154 n20 Realism, moral, x, xii, xxi, xxii, 27-28, 30, 40, 41-44, 43 n6, 51, 71-85, 101-109, 108 n17, 142, 142 n3, 143-145, 143 n6, 159, 160, 162, 168, 169, 173, 174, 177, 195, 196-197, 198, 201, 224 Reasons, xiii, xviii-xix, xxi, xxii, 48, 55-70, 119-138, 147, 174, 175-177, 176 n9, 178, 183-184, 221, 226 authority of, 144-164, 178, 179 internal vs. external reasons, xviii-xix, 86-99, 149-152, 175, 177 n12 Reasons internalism: See Internalism Reflective equilibrium, 112-116, 136-137 Reid, Thomas, 111 n23, 111 n24 Reification, 36 Relativity, argument from, xv, xvii, 44, 45, 46, 47, 52, 90, 95 Relativism, moral, 4, 5, 13, 77, 168, 169, 169 n2, 173-174, 178 Restall, Greg, 222-224, 232 Rights, 2, 11, 218 n2, 220, 221, 222 Roskies, Adina, 66 Ruskin, John, 36 Russell, Bertrand, xi, 10, 11, 18, 26 S Sainsbury, R. M., 43 n6 Sayre-McCord, Geoffrey, 27-28, 29, 30, 41 Scanlon, Thomas, 90 n2, 92, 130, 133 Schacht, Richard, 17 n2, 21-22, 33 Schiffer, Stephen, 103 n9 Schroeder, Mark, 74 n5 Shafer-Landau, Russ, 73, 102 n3, 121, 178-179 Shepski, Lee, 177 n10 Sidgwick, Henry, 81 n18, 120, 121, 122, 123, 128, 130, 138, 179 Singer, Peter, 190, 223, 223 n5 Sinnott-Armstrong, Walter, xviii, xx-xxi, 176, 185 n5

238 Skorupski, John, 156 Smith, Michael, xxi-xxii, 19, 67, 69, 75 n8, 90, 92, 96, 101 n1, 102 n3, 108 n17, 119, 122, 124-125, 131, 153 n18, 155, 155 n21, 156, 156 n22, 157-158, 162, 164, 183-184, 194. Stevenson, Charles L., 78, 103 n8, 193 n12 Stratton-Lake, Philip, 176 n8 Strawson, P. F., xii Stroud, Barry, 37 n2, 42 n5 Sturgeon, Nicholas, 108 n19 Supervenience, 74, 75, 80, 213 Swinburne, Richard, 111 n23 T Tappolet, Christine, 41 Thin and thick concepts, 18, 199 Thomas, Alan, 172 n4, 176 n8 Thrasymachus, ix Timmons, Mark, 73, 74, 74 n6, 76, 103 n7, 103 n8, 104 n12, 108 n17 Tolhurst, William, 63

Index V Van Fraassen, Bas, 213 n3, 214 W Warnock, G. J., 187 Warren, James, xi n1 Weakness of will, 184, 187, 192-193, 196, 230 West, Caroline, xi, xv, xxi, xxii, 222-224, 232 White, E. B.: See Charlotte’s Web Wiggins, David, 102 n3 Wilde, Oscar, 6 Williams, Bernard, xviii, xxi, 57-58, 87-99, 122, 149-151, 156, 175, 175 n6, 177 n12 Owen Wingrave (example), 88, 94 subjective motivational set, 88, 91-94, 98, 150-151, 175 n6 Wingrave, Owen (example): See Williams, Bernard Witches (example), xxii, 79, 180-181 Wright, Crispin, xx, 22-24, 25, 27, 32, 33, 41, 73-74, 103 n8, 113