Normativity and Naturalism 3110327368, 9783110327366

Citation preview

PRACTICAL PHILOSOPHY Herausgegeben von / Edited by Heinrich Ganthaler • Neil Roughley Peter Schaber • Herlinde Pauer-Studer Band 5 / Volume 5

Peter Schaber (Ed.)

Normativity and Naturalism

ontos verlag Frankfurt

.

Lancaster

Bibliographic information published by Die Deutsche Bibliothek Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliographie; detailed bibliographic data is available in the Internet at http://dnb.ddb.de

North and South America by Transaction Books Rutgers University Piscataway, NJ 08854-8042 [email protected]

United Kingdom, Ire Iceland, Turkey, Malta, Portugal by Gazelle Books Services Limited White Cross Mills Hightown LANCASTER, LA1 4XS [email protected]



2004 ontos verlag P.O. Box 15 41, D-63133 Heusenstamm www.ontosverlag.com ISBN 3-937202-41-2

2004

No part of this book may be reproduced, stored in retrieval systems 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 of the purchaser of the work Printed on acid-free paper ISO-Norm 970-6 Printed in Germany.

PRACTICAL PHILOSOPHY Edited by Heinrich Ganthaler • Neil Roughley Herlinde Pauer-Studer • Peter Schaber

The aim of the series is to publish high-quality work that deals with questions in practical philosophy from a broadly analytic perspective. These include questions in meta-ethics, normative ethics and "applied" ethics, as well as in political philosophy, philosophy of law and the philosophy of action. Through the publication of work in both German and English the series aims to facilitate discussion between English- and German-speaking practical philosophers.

In september 2003 several philosophers met at the ethics centre of the University of Zurich to discuss the nature of normativity. This collection contains the papers presented at the conference "Naturalism and Normativity". Representing a wide range of metaethical views, the authors develop different accounts of normativity and its relation to the natural world. I would like to thank the Arnold Corti-Stamm-Stiftung for its financial support.

Zurich, July 2004

Peter Schaber

Contents

DAVID COPP Moral Naturalism and Three Grades of Normativity

7

NEIL ROUGHLEY Naturalism and Expressivism On the “Natural” Stuff of Moral Normativity and Problems with its “Naturalisation”

47

NORBERT ANWANDER Normative Facts: Metaphysical not Conceptual

87

PETER SCHABER Good and Right as Non-Natural Properties

105

THOMAS SCHMIDT Moral Values and the Fabric of the World. A Reconsideration of Mackie’s Arguments against Moral Realism

121

THEO VAN WILLIGENBURG Conceptual Analysis, Normativity and the Empirical

135

ULVI DOGUOGLU Naturalism and Rule-Following Practices: Finding Fault with Kripke’s Notion of Objectivity

151

List of Contributors

175

Name Index

176

Subject Index

178

David Copp Moral Naturalism and Three Grades of Normativity Is moral naturalism ruled out by the fact that morality is normative? I want to consider this question in a systematic way, explaining the central thesis of moral naturalism as I understand it, and then clarifying the idea of normativity. The chief point that I want to make is that the issue raised by this question is much more complex than it might seem to be, mainly because of complexity in what philosophers have in mind when they speak of morality as “normative”. There are several dimensions to the complexity, but the dimension I will stress is that philosophers disagree about what we might call the “stringency” of moral normativity. I will distinguish three “grades of normativity”. I will investigate, for each of these grades, the plausibility of the idea that morality has that grade of normativity, and the viability of naturalistic accounts of it. My conclusion will be tentative, but it will be optimistic for moral naturalism. Moral naturalism holds that in thinking of things as morally right or wrong, good or bad, we ascribe moral properties to these things – properties such as moral rightness and wrongness, goodness and evil. It holds that there are such properties, and it adds that these properties are ordinary garden-variety natural properties – properties that have the same basic metaphysical and epistemological status as the properties a tree can have of being deciduous, and the property a piece of paper can have of being an Australian twenty dollar bill. I will have more to say about this in what follows. The question is whether moral naturalism can accommodate the normativity of morality. This is basically the question whether it can account for the fact that morality is, in a characteristic way, action-guiding and choice-guiding. Moral thought and discourse concern how we are to act, what we are to choose, and how we are to live; they involve us in evaluating, prescribing, and recommending. I use the term “normative” to speak of this phenomenon.1 1

I want to avoid the term “prescriptive” because it has been given a technical meaning by Hare (1952).

8

David Copp

Unfortunately, the precise nature of the phenomenon, and how we might account for it, are both unclear and contested. Anti-naturalists tend to view naturalist proposals as inadequate, mainly, I think, because they view them as too weak to account for the importance of morality, especially its importance to rational decision-making and action. I therefore attempt in this paper to distinguish different proposals about what the normativity of morality might consist in. These are the three “grades of normativity” that I mentioned, but one might think that the grades can be subdivided, or one might think there are additional grades. There is also room to disagree about how best to understand each of the grades that I propose. Most important, there is no agreement that morality has all three grades of normativity. Hence, to show that moral naturalism cannot accommodate the fact that morality is normative, one would have to specify which grade of normativity is at issue, to argue that morality does have that grade of normativity, and then to argue that moral naturalism cannot account for this. I begin the paper with an attempt to motivate the view that moral naturalism cannot account for normativity. We obviously cannot make genuine progress, however, unless we have a clear grasp of what is meant both by “moral naturalism” and by “normativity”, so this is where I then turn. In the second section of the paper, I provide a brief explication of moral naturalism. In the third section, I provide an initial, pre-theoretical explication of the three grades of normativity, which I call generic, motivational, and authoritative normativity. In the rest of the paper, I discuss the two central questions about each of the grades: Does morality have that grade of normativity? And can naturalism account for it? I think that morality does possess generic normativity, which is the weakest grade of normativity. But I argue that it is not plausible that morality has either motivational or authoritative normativity. If I am correct, then even if for some reason there cannot be naturalistic accounts of motivational or authoritative normativity, this would not argue against naturalism. I claim, nevertheless, that it is possible to provide naturalistic accounts of all three grades of normativity. In short, my conclusion will be optimistic for moral naturalism. At the very least, I aim to show the difficulties standing in the way of an argument from normativity against naturalism.

Moral Naturalism and Three Grades of Normativity

9

1. The Problem Why might one think that naturalism cannot account for normativity? Let me ignore, for a moment, the complexity introduced by the idea that there are different grades of normativity. Since naturalism holds that the moral properties are natural properties, the place to begin is with arguments claiming to show that no natural property can be normative. An anti-naturalist might invoke a version of G. E. Moore’s open question argument.2 It appears that whatever natural property we might consider, it is an open question whether the fact that something instantiates the property deserves to be given any special significance in our decisions and choices. Normative properties, however, it may seem, are such that there is no question but that things that instantiate them deserve to be given special significance in decision-making. For example, the wrongness of a contemplated action must be taken into account in deciding what to do. It seems to follow that no natural property is normative. Alternatively, an anti-naturalist might appeal to J. L. Mackie’s argument from queerness. Mackie argued that a property with “to-be-doneness” built into it would be “queer”, and that such a property would be “utterly different from anything else in the universe” (Mackie 1977, chapter 1). One might take this to mean that moral properties must be non-natural, if there are any such properties. In response, a naturalist might object to the accounts of normativity that are implicit in these arguments. She might claim that there actually is an open question whether to pay attention to moral considerations, and so, if moral properties are normative, the key premise of the Moorean argument must be rejected. Or she might claim that there is nothing queer about normativity, once it is properly understood. She might exhibit a naturalistic analysis of normativity and then argue that normativity so understood is compatible with naturalism. But there are other strategies the naturalist might consider as well, strategies that attempt to account for normativity without providing an analysis, or that attempt to accommodate the normativity of morality without supposing that moral properties are normative. 2

Moore intended the open question argument to rule out the thesis that the property of goodness is “complex” (Moore 1993, section 13, pp. 66-68).

10

David Copp

A naturalist might contend, for example, that normativity is best seen as consisting in a relation between moral thoughts and the nature of rational agency, viz., that rational people are motivated appropriately by their moral beliefs. Or she might contend that it is the moral concepts that are normative, not the properties they represent. Or she might suggest that normativity is a property of moral speech acts, not of moral properties. These strategies may seem unsatisfactory, for it may seem that in making a moral judgment we are responding to something normative that is independent of us, something that is the object of our judgment. It might seem that we cannot adequately account for this without supposing the moral properties themselves to be normative. It is nevertheless important to recognize the availability of these other strategies. Expressivist opponents of moral naturalism adopt strategies of these kinds (see, for instance, Gibbard 1990). The naturalist should not be in any greater difficulty in taking up these strategies than her opponents. I will return to this issue. These different strategies for defending naturalism lie along one dimension of the complexity that an anti-naturalist argument must navigate. Along this dimension are different views about what exactly it is that is characterized by normativity.3 In what follows, to simplify matters, I will tend to blur distinctions along this dimension by speaking of the normativity of “morality”, or of “moral claims”, except when I need to bring the discussion into sharper focus. Along a second dimension of complexity are different strategies a naturalist could adopt in order to show that normativity is a natural phenomenon. Assume that normativity is a second-order property of the moral properties. There are reductive and non-reductive strategies. One familiar reductive approach analyzes normativity as a complex motivational property. Another identifies normativity with a complex norm-relative or standard-relative property.4 In general terms, a reductive strategy specifies a property, using a 3

In thinking about this dimension of complexity, I was helped by Jon Tresan. My own “standard-based” approach is an example of the latter (see Copp 1995, chapter 1). Michael Smith’s proposal is an example of the former kind of approach. Smith denies that his account is reductive, but he also denies that a reductive analysis is needed in order to vindicate naturalism (see Smith 1994, pp. 161-164, pp. 184-186).

4

Moral Naturalism and Three Grades of Normativity

11

term or description that would antecedently be agreed to pick out a natural property – perhaps a complex property the constituents of which are natural properties – and it proposes that the property of being normative is identical to that property.5 A non-reductive strategy argues that the property of being normative is a natural property, but attempts to do this without proposing a reductive identity statement.6 In order to show that moral naturalism cannot accommodate the fact, if it is a fact, that moral properties are normative, one would have to show that no naturalistic strategy of any of these kinds could be successful. In what follows, instead of attempting to answer anti-naturalist arguments directly, I will consider each of the three grades of normativity, and ask how a naturalist might attempt to account for the idea that morality has that grade of normativity. If I am correct, it will become clear that the anti-naturalist arguments are unsuccessful.

2. Moral Naturalism As I use the term, a moral naturalist is a moral realist. She agrees with other moral realists – those we call “non-naturalists” – in accepting a number of doctrines. First, there are moral properties and relations, such as the properties of rightness, wrongness, goodness, virtuousness, and the like.7 Second, although moral properties differ from non-moral properties in that, after all, they are moral properties and the others are not, they have the same basic metaphysical status that any property has – whatever that status is. Third, one of the chief semantic roles of the moral predicates is to express such properties. For instance, in calling an action “right”, one thing we are doing is ascribing rightness to the action. Fourth, a moral assertion expresses a moral belief that represents the world as being a certain way, and what is thereby asserted is true if and only if the expressed belief is true. This would 5

For an account of philosophical analysis as consisting in the analysis of complex properties, see King (1998). Other recent accounts of analysis are found in Jackson (1998), Smith (1994) and Gibbard (1990). 6 For a discussion of non-reductive naturalism, see Sturgeon (forthcoming). 7 In what follows, for the most part, I include relations as a kind of “property”.

12

David Copp

involve, at least in the case of a basic moral belief, the instantiation of a moral property.8 And finally, fifth, some moral properties are actually instantiated. For instance, some actions are actually right and some are actually wrong. The moral naturalist differs from other moral realists in holding that the moral properties are “natural” properties. In brief, then, moral naturalism is the view that there are moral properties and that these properties are natural properties. For our purposes, the most important question raised by the naturalist’s doctrine is, What is meant in this context by a “natural” property? There is room for disagreement about this, and I do not need to insist on any particular account. In introducing the idea, I suggested that natural properties can be taken to be ordinary “garden-variety” properties with the same basic metaphysical and epistemological status as the property a tree can have of being deciduous or the property of being an Australian twenty dollar bill. This intuitive understanding of the idea of a natural property might be adequate for my purposes in this paper. I cannot go into detail here, but I have proposed elsewhere that we should construe the moral naturalist as holding that moral properties are empirical properties.9 The idea of the empirical is itself problematic, but the proposal at least has the virtue of explaining naturalism in terms of a concept that is not special to moral theory. Any adequate epistemology needs to provide an account of the empirical, so my proposal can piggyback on a concept that is widely used and understood, even if it is contestable. To be more specific about my proposal, the basic idea is that, if a property is natural, then empirical evidence is relevant to the justification of our beliefs about its instantiation. Elaborating further, I would propose that a property is natural if and only if – leaving aside analytic truths, if there are any – any proposition 8

A “basic” moral proposition is such that, for some moral property M, it entails that something instantiates M. An example is the proposition that capital punishment is wrong. Among non-basic moral propositions are propositions such as that nothing is morally wrong and that either abortion is wrong or 2 + 2 = 4. Call a belief a “basic” moral belief just in case the content of the belief is a basic moral proposition. 9 I have developed this idea in more detail in Copp (2003). For criticism of this conception of a natural property, and discussion of alternatives, see Sturgeon (forthcoming).

Moral Naturalism and Three Grades of Normativity

13

about the instantiation of the property that can be known, can be known only empirically, or by means of empirical observation and standard modes of inductive inference.10 This proposal is in line with Moore’s basic ideas about moral naturalism. Moore took naturalism to be the view that the property goodness is a natural property (Moore 1993, section 26, p. 93), but he got into difficulty in attempting to explain the idea of a natural property.11 My proposal is similar to his first brief characterization of naturalism, in section 25 of Principia Ethica, where he wrote that, according to naturalism, “Ethics is an empirical or positive science: its conclusions could all be established by means of empirical observation and induction.” (Moore 1993, s. 25, p. 91). It clearly would be controversial to hold that ethics is a science. Instead, what the naturalist might say, following Moore’s suggestion, is that ethics is empirical in that any substantive ethical knowledge is empirical, or based in “empirical observation and induction”. My proposal is in line with this idea. Given this construal, moral naturalism is not entailed by metaphysical naturalism, for a metaphysical naturalist could deny that there are any moral properties at all, thereby rejecting moral naturalism. Nor is moral naturalism a kind of physicalism. It is left open whether there are natural properties that are non-physical. For the most part, I will simply assume the realist position that there are moral properties. The central issue I want to address is whether the normativity of morality somehow rules out the idea that these properties are natural.

10

Even a naturalist may want to concede that some ethical knowledge is not empirical. For example, it seems plausible that the concept of murder is the concept of a wrongful killing, and if this is so, then we can know a priori that murder is wrong. It is no part of ethical naturalism as such to deny that there are conceptual truths in ethics and that our knowledge of any such truths is not empirical. 11 He later acknowledged that his attempts to explain the idea of a natural property in Principia were “hopelessly confused” (in “Preface to the Second Edition”, in Moore 1993, p. 13). He admitted that one of his proposals was “utterly silly and preposterous.” He added that in Principia he had not given “any tenable explanation” of what he meant by saying that goodness is not a natural property (see Moore 1968, pp. 581-582).

14

David Copp

3. Three Grades of Normativity The basic idea idea, as I said, is that morality concerns how to act, what to choose, and how to live. It prescribes, recommends, and evaluates actions and choices. This, however, is a characteristic that morality shares with etiquette, prudence, epistemology, aesthetics, and judgments of individual selfgrounded rationality. In judging that it is rude to carry on a loud conversation during a concert, we evaluate such conversations. Apparently, then, there is a kind of generic normativity shared by these various kinds of thought and discourse. But one might think that moral thought and discourse are normative in a further respect. For one might think that when a rational person accepts a moral claim, she takes it into account in deciding how to act or choose. This is a kind of motivational normativity that distinguishes morality from etiquette and aesthetics, since rationality does not ensure that one will take into account considerations of etiquette or aesthetics. But one might think that moral thought and discourse are normative in still a further respect. For one might think that there are reasons to do what is morally required even if we are unaware of them. And one might think that a rational person who understood this would not take there to be a genuine question about whether to be moral, or at least about whether to give moral considerations due weight in her deliberation. Morality has authoritative normativity, we might say. When philosophers allude to the “normativity” of moral thought and discourse, they could have any of these ideas in mind. One might wonder how these so-called grades of normativity are related to the traditional Kantian distinction between hypothetical and categorical imperatives. The answer is that there are different conceptions of the categorical. Each of the grades of normativity could be understood as a proposal about what the categorical nature of moral requirements comes to. On one view, generic normativity of a certain kind is sufficient for a requirement to be categorical. On another view, moral requirements are genuinely categorical only if they have authoritative normativity.12 12

Philippa Foot pointed out that, on one account of the categorical, the requirements of etiquette qualify as categorical, but that this account would be viewed as too weak by those who think that morality is distinctive in being a source of categorical requirements. Foot denied that moral considerations “necessarily give reasons for acting to any man,” and in so doing she took herself to be denying that morality is a source of categorical reasons (see Foot 1978, p. 161).

Moral Naturalism and Three Grades of Normativity

15

Some philosophers might be tempted to explicate normativity by invoking the idea of a reason. One problem for this approach is that claims about what we have reason to do are themselves normative, so such an approach does not go very deep. Another problem is that there seem to be reasons of different kinds. There appear to be moral reasons, reasons of etiquette, epistemic reasons, and so on. Moreover, these different kinds of reasons appear to have different grades of normativity. Reasons of etiquette might be less stringently normative than moral reasons, and moral reasons might be less stringently normative than rationally compelling or strongly authoritative reasons – unless, of course, they are themselves strongly authoritative. Because of this, before we could explicate normativity by invoking the idea of a reason, we would need to ask what kind of reason is at issue and what is its grade of normativity. This is not a way to avoid issues about the grades of normativity. To a moral realist, it may seem natural to explain the normativity of morality by supposing that moral properties have the second-order property of being normative, and, if there are different grades of normativity, by supposing that there are second-order properties corresponding to the different grades. As I mentioned before, however, a realist does not need to ascribe normativity to the moral properties; there are alternatives. Moreover, at this stage of the discussion, I want to provide an intuitive account of the different grades of normativity without presupposing moral realism. Hence, as I said, I will present the grades of normativity as putative properties of “morality”, or of “moral claims”. (1) Generic Normativity. To begin, then, compare the claims that lying and deception are wrong with the claim that lying and deception are widespread.13 There is a characteristic difference between these claims. The non-moral claim is simply descriptive of aspects of human behavior, but the moral claim is not merely descriptive. The claim that lying is morally wrong is an evaluation of lying and deception. It is a proscription of lying. Moreover, it is evaluative or 13

In this and the next two paragraphs I draw on Copp (1999).

16

David Copp

proscriptive in virtue of what it says. That is, it has these properties in virtue of its semantics, meaning, or content; it is essentially recommendatory or proscriptive. To generalize, it appears that moral claims have a characteristic essential semantic connection to decisions about action or choice, a connection that can be referred to by describing them as “normative.”14 Call this property of moral claims, “generic normativity.”15 It is the weakest grade of normativity. I say that moral claims are not merely “descriptive”, but to a moral realist, basic moral claims express propositions that attribute moral properties to things, and in this sense they do describe those things. To a realist, for instance, we ascribe wrongness to torture in saying or thinking that torture is wrong. In addition, however, such claims evaluate, recommend, or prescribe; they have (at least) generic normativity. The problem for the realist, then – whether or not she is a naturalist – is to provide a philosophical account of the generic normativity of moral claims on the assumption that such claims express propositions and ascribe properties. Various non-moral claims are also normative in the generic sense, including claims about what we have prudential reason to do or epistemic reason to believe, and claims about what we are required by etiquette to do. There are, in addition, obvious differences among moral claims. An adequate theory of generic normativity must account for these other kinds of generically normative claim as well as moral claims, and do so without obscuring the differences among them. (2) Motivational Normativity. One might think that moral thought and discourse are normative in a deeper way than, say, thought and discourse about the requirements of etiquette. A person who thinks she is required in etiquette to do something cannot be convicted of any kind of irrationality even if she views etiquette as beneath contempt and even if she pays it no attention in deciding how to act. Morality seems to be different from this. It has seemed to many philosophers that if a 14

This is not meant as a definition. It is at best a characterization of generic normativity. 15 The term, “generic normativity”, was suggested by Gene Witmer.

Moral Naturalism and Three Grades of Normativity

17

rational person thinks she is morally required to do something, she will be motivated to do it. More generally, and perhaps more cautiously, a person who is practically rational takes her moral beliefs appropriately into account in deciding how to act or choose, and if such a person believes she is morally required to do something all things considered, she is motivated do it.16 We can call this alleged property of moral claims, “motivational normativity”. The thesis that moral claims have “motivational normativity” is a close relative of a stronger view, a view I will call “motivational internalism”, according to which, practical rationality aside, it is a necessary truth that anyone who thinks she is morally required to do something will be motivated to do it.17 There are familiar objections to motivational internalism, however. It appears that people who are depressed might lack motivation to do what they think they morally ought to do. And it appears that people with unusual second-order beliefs about morality might also fail to be motivated to do what they think they ought to do. For example, in Graham Greene’s The Heart of the Matter, Scobie does something he believes to be forbidden in order to secure his own damnation, believing that he is thereby expressing love of God. Perhaps Scobie still has some motivation not to do the forbidden thing, but it is not clear that this is so, and it is doubtful that it must be so. For a different example, suppose someone thinks that God’s commands determine what we are morally required to do, but suppose she is in rebellion against God. She could conceivably lack any motivation to do what she thinks morality requires. There is an additional issue, which is more important for our purposes. It is not clear that motivational internalism is a plausible account of a kind of normativity. Consider, for example, that people are motivated to eat things that they take to be sweet. Of course, this is not a necessary truth, and motivational internalism is supposed to be necessarily true. But we can fix this asymmetry by stipulating that a person counts as believing something is tweet if and only if she believes it is sweet and in addition she is motivated 16

For such a view, see Korsgaard (1986); also Smith (1994), p. 62. The term “internalism” is used for many different positions. The doctrine in question here is the one that Stephen Darwall has called “judgment internalism” (see Darwall 1983, pp. 54-55). 17

18

David Copp

to some degree to eat it. It is a necessary truth that a person who thinks that something is tweet is motivated to eat it, but I doubt that anyone would think on this basis that the claim that something is tweet is normative. There is a motivational condition on the sincere use of “tweet” that there is not on the sincere use of “sweet”, but this does not mean the terms express different properties or that otherwise identical sentences in which the terms are used express different propositions. The proposition expressed by an assertion that sugar is “tweet” is the same proposition as is expressed by the assertion that sugar is “sweet”. This proposition obviously is not normative, and to me it seems implausible that our assertion should count as normative simply because of the special motivational sincerity condition on the use of the word “tweet”. Since what holds in this imaginary case is parallel to what holds in the case of moral assertion, if motivational internalism is correct, I do not think that motivational internalism is plausibly viewed as an account of a kind of normativity. Troubled by the artificiality of the example, one might reply that there is no property of tweetness. There is only sweetness and the contingent fact that we are motivated to eat sweet things. But there is the property an action can have of being morally required, and this property is such that, necessarily, a person who believes he is morally required to do something is motivated to do it. This reply raises difficult issues. Perhaps the property expressed by the term “morally required” is such that it is merely contingent that people are motivated to do what they take to have that property except that they do not count as conceiving of an action as morally required unless they are motivated accordingly. That is, perhaps the parallel between the property expressed by “tweet” and that expressed by “morally required” is quite close, even though, contingently, the term in common use that expresses the property expressed by “tweet” is “sweet”, which does not implicate motivation, while the term “morally required” is in common use and, at least according to motivational internalists, it does implicate motivation.18

18

The suggestion in this paragraph is similar to a view developed by Jon Tresan. Tresan distinguishes between de re internalism and de dicto internalism. In this terminology, the view suggested in this paragraph is de dicto internalist (see Tresan forthcoming).

Moral Naturalism and Three Grades of Normativity

19

The “tweet” example shows two things, I believe. First, it shows that motivational internalism is compatible with naturalism. And second, it shows that motivational internalism is not a plausible account of a kind of normativity. Motivational normativity is relevantly different, however, for it is a thesis about the impact of moral beliefs on the motivation of practically rational persons. Hence, it seems to capture a normative aspect of moral belief that goes beyond mere generic normativity.19 (3) Authoritative Normativity. It might seem that the ideas of generic normativity and motivational normativity are too weak to capture everything we have in mind in thinking of moral claims as normative. For example, even if moral claims are generically and motivationally normative, such that a rational agent would be motivated by her moral beliefs, it might be that she could experience this motivation as alien and unwelcome in something like the way that hunger can be experienced as unwelcome when one is attempting to concentrate on an important matter. This would be compatible with the generic and motivational normativity of moral claims. In short, for a variety of reasons, it might seem that something beyond generic and motivational normativity must be cited in order to explain the normativity of morality. What might we have left out? A variety of ideas could be proposed at this point, and it is somewhat artificial to talk as if there is one thing at issue, “authoritative normativity”. All that we have is the idea that there is a gap here, a place where something 19

Suppose we stipulate that a person counts as believing something is schpreet if and only if she believes it is sweet and, in addition, unless she is not practically rational, she is motivated to some degree to eat it. Is it implausible in this case that the claim that something is schpreet is normative in the way that, I said in the text, it is implausible that the claim that something is tweet is normative? If so, then my argument perhaps goes too far. But it seems to me that, given my stipulations, the claim that something is schpreet would be normative. For suppose I am competent with the use of “schpreet” and I assert that sugar is schpreet. In this case, it seems, I have expressed the belief that sugar is sweet as well as the belief that I am motivated to eat sugar to the extent that I am practically rational. It is plausible, then, that what I have asserted is normative since it concerns what I will do if I am practically rational.

20

David Copp

seems to be missing. This could be a mistake. It could be that morality has no further feature that can fill the gap we seem to detect. The thought, however, is that morality has an objective authority over us, an authority that is not merely a matter of generic normativity, and an authority that is not reducible to the responses or motivations a rational person would have in virtue of her moral beliefs. The thought is that the reason that a rational person would take her moral beliefs appropriately into account in deciding how to live is that her moral beliefs concern something with objective authority over her. What could this come to? There seem to be two chief ideas in the literature. First, objective authority is meant to explain why morality has motivational normativity. It is meant to explain why rational people pay attention to the demands of morality. They pay attention, and they are motivated appropriately, because, to the extent at least that they are fully rational and thinking clearly, they understand that morality is a source of authoritative reasons, reasons of a kind that any rational person would take into account, if she were aware of them, just in virtue of being rational. As Philippa Foot says, the idea is that moral considerations “necessarily give reasons for acting to any man”, or, as Stephen Darwall says, they are a source of reasons “with genuine deliberative weight”.20 Call this the “authoritative reasons proposal”. Second, objective authority is meant to answer the traditional challenge to morality posed by the question, Why be moral? It is taken to provide the definitive answer to the challenge.21 Any putative requirement or ideal can 20

Foot rejects the idea that moral considerations “necessarily give reasons for acting to any man,” and in so doing, she in effect denies that morality has authoritative normativity (see Foot 1978, p. 161). She thinks that if moral considerations do not “give reasons for acting to any man,” then, even if we say they give us “moral reasons”, these would not be robust enough to be genuine (see Foot 1978, p. 168, note 8). Referring to Foot’s discussion, Darwall suggests that the idea is that moral considerations give us reasons “with genuine deliberative weight.” He agrees with her that so-called reasons of etiquette “are not guaranteed to be reasons unqualifiedly.” The same would be true of moral reasons, he suggests, if morality lacks authoritative normativity (see Darwall 1997, p. 306). 21 Darwall suggests that the issue whether morality is normative is posed by the ‘Why be Moral?’ question (see Darwall 2001).

Moral Naturalism and Three Grades of Normativity

21

be questioned. We can ask, Why pay attention to etiquette? Perhaps we can even ask, Why be rational? But if a requirement or ideal has authoritative normativity, then, the idea is, a rational person who was thinking clearly would understand that the relevant why-question is closed. That is, the question could not arise as a genuine practical question for any rational person who was thinking clearly about, and who understood the nature of, a requirement or ideal that had authoritative normativity. Call this the “closed question proposal”. These are different proposals, but they are compatible with each other. According to the authoritative reasons proposal, a person who was rational and thinking clearly would understand that morality is a source of authoritative reasons, reasons of a kind that any rational person would take into account if she were aware of them, just in virtue of being rational. Yet understanding that morality is a source of authoritative reasons might not close the Why be moral? question for her, for she might think there are non-moral sources of authoritative reasons, and that such reasons could override the reasons deriving from morality. Hence, it seems, the closed question proposal is stronger than the reasons proposal. On the closed question proposal, a rational person who is thinking clearly and who understands the nature of morality thereby understands something about morality such that, for her, the Why be moral? question cannot arise as a live question. It might seem that what she must understand is that the authoritative reasons that derive from morality are not overridden by any other reasons.22 Consider the myth of Gyges, as told by Plato in the Republic (359d - 360b). Gyges is a shepherd who discovers a ring that makes him magically invisible when he turns it on his finger. Using the ring, Gyges seduces the queen, kills the king, and takes over the kingdom. It seems fair to say that this course of behavior was morally unacceptable. Now suppose that Gyges becomes persuaded of the wrongness of his plan in advance of the seduction. He might ask himself, “Why should I be moral?”, especially given that doing his duty would mean giving up a life of power and love. Merely responding, “Because 22

But since there may be other ways of understanding the proposal, I do not want to build the thesis that moral reasons are overriding into the closed question proposal itself.

22

David Copp

it’s your duty,” would not cut any ice since he already knows where his duty lies and he is troubled by the thought that he must follow through. Consider, then, whether invoking any of the three grades of normativity would help Gyges with his question. Merely to invoke generic normativity would not address the question. Gyges presumably is aware that the wrongness of his plan has just the kind of relevance to his decision about what to do that makes it generically normative. This still leaves him with the question why he should avoid wrongdoing. And it would not help to invoke motivational normativity. As a rational person, Gyges could find himself motivated to do what he understands to be his duty, but he might still wonder why he should do his duty. Yet if morality has objective authority, then, the idea is, if Gyges understood the nature of moral considerations and if he were rational and thinking clearly, he would realize that the wrongness of his plan gives him an authoritative reason not to carry through with it. Furthermore, if he were rational and thinking clearly and if he understood the nature of moral considerations, his question about whether to be moral would not seem to be a live question for him. One might think that, even given all of this, Gyges could still raise the question, “What are these authoritative reasons to me?” It does need to be said that the authoritative normativity of morality cannot silence every kind of doubt. Theoretical doubts and puzzles might still arise, including doubts about authoritative reasons. Yet the idea is that a rational person who understands the concept of objective authority and believes that moral requirements and ideals have objective authority cannot think there is a serious practical question about whether to be moral. Such a person might raise various theoretical questions, but she would not be uncertain or indecisive about whether to be moral, not if she were fully rational and thinking clearly. This completes my explication of the three grades of normativity. Of course, I distinguished two proposals about authoritative normativity, and there may be other ideas as well. There is, therefore, some reason to think it would be useful to distinguish more than three grades of normativity. But this is how I will organize the discussion.

Moral Naturalism and Three Grades of Normativity

23

4. The Problem Revisited If we think that morality does have these three grades of normativity, the next question to ask would be, Can a moral naturalist explain or account for this? But a naturalist might want to deny that one or more of the grades of normativity is genuinely characteristic of morality. For each of the three grades of normativity, then, there is a pair of questions. How plausible is it that morality is normative in this way? And, can a naturalist account for this kind of normativity? In order to argue against naturalism on the basis of the normativity of morality, therefore, one would have to defend two key premises. The first premise would be to the effect that morality is normative on some given understanding of what this amounts to. And the second premise would be to the effect that moral naturalism cannot accommodate the fact that morality is normative on this understanding. Either of these premises can be controversial. If we let “X-normativity” refer to one of the three grades of normativity, or else to some other property that is a candidate for a kind of normativity, the naturalist might object that morality does not have the property of being X-normative. Alternatively, she might argue that naturalism can account for X-normativity. Her opponent may face some unexpected complications. First, in order to undermine moral naturalism without undermining moral realism, the anti-naturalist must argue that the difficulty in accounting for X-normativity can be traced specifically to the naturalist’s idea that the moral properties are natural. Otherwise, it might appear that what the argument really does is to undermine the realist thesis that there are moral properties. Moreover, second, a successful anti-naturalist argument must rule out the strategies I mentioned before that seek to explain normativity without treating it as a property of the moral properties. Otherwise a naturalist might be able to explain X-normativity using one of these strategies. From this point on, I will discuss the three grades of normativity, considering, first, the plausibility of the idea that morality has that grade of normativity, and second, the plausibility of the idea that moral naturalism is somehow incompatible with the hypothesis that morality has that grade of normativity. But a brief digression is in order.

24

David Copp

5. Expressivism and Realist-Expressivism It may seem that I have missed a well-known and obvious account of normativity as well as an obvious argument against moral realism. One of the realist doctrines is the view, often called “descriptivism”, that a primary semantic role of the moral predicates is to ascribe moral properties. A number of thinkers, from Charles Stevenson and R. M. Hare through Simon Blackburn and Allan Gibbard, have shared the view that “descriptivist” accounts of moral thought and discourse cannot account for normativity (Stevenson 1937; Hare 1952; Blackburn 1993; Gibbard 1990). Their idea is that normative thoughts involve conative attitudes of one kind or another and that descriptivist views cannot satisfactorily account for this. Gibbard claims, for instance, that “the special element that makes normative thought and language normative” is that “it involves a kind of endorsement – an endorsement that any descriptivistic analysis treats inadequately.” (Gibbard 1990, p. 33). If successful, this objection would cut equally against all versions of moral realism, whether naturalist or non-naturalist. But the objection is not successful. Consider, for example, that we would be expressing an attitude of contempt in saying that cursing is typical among “rednecks”. Even so, it is clear that one of the central semantic roles of the term “redneck” is to ascribe the property of, roughly, being “a member”, perhaps an uneducated or “uncouth” member, “of the white rural laboring class.”23 The example illustrates how it could be the case both that we use moral terms, such as “wrong” and “good”, to express attitudes such as disapproval and approval, and that a central semantic role of such terms is to ascribe moral properties. The expressive speech act potential of the terms does not undercut the realist thesis that the terms are used to ascribe moral properties. Expressivists are therefore mistaken to think that moral naturalism cannot account for the fact – if it is a fact – that moral thought and discourse involve having or expressing conative attitudes of one kind or another. The naturalist can say that moral discourse and thought involve a kind of endorsement and 23

Sometimes it appears that the extension of the term is restricted to white people living in rural areas of the southern United States. I quote from the entry for “redneck” in The American Heritage Dictionary of the English Language (third edition), Boston, 1992.

Moral Naturalism and Three Grades of Normativity

25

that they also ascribe moral properties. I have explained elsewhere in some detail how such an account might go. I have called this kind of view “realistexpressivism”. It is realist – it supposes there are moral properties – and it is expressivist – it supposes that moral thought and discourse involve “a kind of endorsement” (see Copp 2001). The example of the term “redneck” suggests, however, the implausibility of the view that, in Gibbard’s words, “the special element that makes normative thought and language normative” is that “it involves a kind of endorsement”. “Redneck” is not a normative term and thought about rednecks is not normative. So it seems to me that even if moral thought and language involve endorsement and similar conative attitudes, it is not plausible that this is sufficient to account for their normativity. It seems that the subject matter of moral thought and discourse is normative. Arguably this means that the normativity of moral thought and discourse is due to a feature of the moral properties that we attribute to things in having moral thoughts rather than being due merely to the expressive force of moral predicates.24 Even if realist-expressivism is part of a fully adequate naturalism, it is not sufficient to tell the full story.25 This brings us to the central issue, which is whether natural properties can be normative. This is the territory on which the central dispute between ethical naturalism and non-naturalism is to be found. There are the three grades of normativity to consider.

6. Generic Normativity and Motivational Normativity It is surely correct that moral thought and discourse are characterized by generic normativity. The idea here is that, for example, the claim that lying is wrong is evaluative or proscriptive in virtue of what it says or in virtue of its content. It is that, as I said before, moral claims have a characteristic essential semantic connection to decisions about action or choice. If we compare the claim that lying is wrong with the claim that lying is widespread, 24

Jonathan Dancy makes a similar claim (see Dancy forthcoming, section 6). In Copp (2001), I argue that realist-expressivism can explain why motivational internalism can appear to be true even though it is not true. 25

26

David Copp

a naturalist will say that different properties are at issue, the property of wrongness and the property of being widespread, respectively. From the perspective of any moral realist, this will seem to be the key difference between the contents of the claims, leaving aside the issue of expressive content. This means that if we are to explain generic normativity by reference to the content of moral claims, it seems we must explain it by reference to a characteristic normativity of moral properties. One might think that this is enough to settle the matter against naturalism since one might have the picture of properties as just being there in things, inertly, with no intrinsic bearing on action or choice. But this would be a mistake. Of course, it would cut against any kind of moral realism, whether naturalist or non-naturalist. And it would be question begging, for moral realism is precisely the view that there are such properties as wrongness, and no property would count as wrongness if it did not have a bearing of the relevant kind on action. This is just to repeat the insight that wrongness and other moral properties must have generic normativity. It is true, however, that a moral realist cannot get away with saying merely this. We are owed an explanation of what the property of generic normativity could amount to.26 I have proposed such an explanation in other work, where I have developed a naturalistic account of generic normativity called the “standard-based” account (see Copp 1995, especially chapter 1 and chapter 11, pp. 223 - 231).27 The idea is that a normative property relates an action or object of choice to a norm or standard with a relevant status. For example, I argued, the property actions can have of being rude, or being ruled out by (the local) etiquette, is a property that relates actions to norms or standards that are shared in the familiar way in the local culture. And I argued that moral wrongness is, roughly, the property of being ruled out by a system of standards the currency of which in the relevant society would contribute better than that of any other such system to enabling the society to meet its needs. When generalized in an appropriate way, this theory provides an account of the 26

I made this complaint – although not in these terms – in two earlier papers on moral naturalism: Copp (1990); Copp (1991). 27 For a brief exposition, see Copp (1997b). For a reply to some objections, see Copp (1998).

Moral Naturalism and Three Grades of Normativity

27

other moral properties, such as the property of being morally required, the property of being supported by a moral reason, and so on. This approach could be disputed, but it is a naturalistic account. Motivational normativity is more controversial. I have argued in other work that it is possible for a rational agent to believe that she is morally required to do something yet not to be motivated to do it. Indeed some counter-examples to motivational internalism seem also to be counter-examples to motivational normativity. Depression can remove a person’s motivation to do what she believes she is morally required to do, and it seems to me that a depressed person is not necessarily irrational in any ordinary sense. A depressed person might think it is more important for the moment to think about her lost loves than to be fastidious about what she is morally required to do, and so she might have no motivation to do something that she believes she ought to do. This might not be any indication of irrationality.28 I therefore find it doubtful that morality is characterized by motivational normativity.29 When a person fails to do what she believes she is morally required to do, having lacked any motivation to do it, she is presumably blameworthy. Of course we think that a person can be excused in some cases for failing to do her duty, but we do not think that a person can be excused for lacking any motivation at all to do what she believes she is morally required to do. This said, however, we would not likely think that such a person must be irrational, as would be implied by the thesis of motivational normativity. Intuitively, her lack of moral motivation reflects on her moral character, not on her rationality. That is, the thesis of motivational normativity seems counter-intuitive, and not at all a matter of ordinary thinking about morality. All of this being said, however, it is important to point out that moral naturalism can accommodate the idea that morality has motivational normativity. For, in the first place, it is plausible that a rational person is disposed to be moved by reasons. More precisely, it is plausible that a rational person is so disposed that, if she took herself to have a reason to do something, 28

Here I follow arguments in Copp (1997a). 29 It is possible that realist-expressivism can help to explain why motivational normativity might seem to be true in something like the way it helps to explain why motivational internalism can appear to be true.

28

David Copp

she would be motivated to some degree to do it, at least under certain circumstances. Perhaps she would not be so motivated if she thought there were stronger reasons against doing the thing than in favor, but in the absence of an awareness of any countervailing reasons, it is plausible that she would be motivated to do the thing. A naturalist can hold that such a disposition is partly constitutive of rational agency. In the second place, it is plausible that moral considerations are a source of reasons. Indeed, on my account of generic normativity, if a person is morally required to do something, then there is a moral reason for her to do it. A naturalist might then hold that a rational person is so disposed that, if she takes herself to be morally required to do something, and if she understands what this involves and is thinking clearly, then she would be motivated to some degree to do the thing, at least under certain circumstances. This position is compatible with moral naturalism, but I think it is mistaken. I agree, of course, that it is constitutive of rational agency to be so disposed that, if one takes oneself to have an authoritative reason to do something, one is motivated to some degree to do it, at least under certain circumstances. This follows from the concept of an authoritative reason, since authoritative reasons are, by stipulation, reasons of a kind that a rational person would be responsive to, if she were aware of them, just in virtue of being rational. Nothing here is incompatible with moral naturalism. But I think it is an oversimplification to suppose that rational persons are disposed to be moved by reasons of every kind. There are reasons of different kinds, and not all of them are authoritative. It is not a failure of rationality to lack any tendency to be moved by reasons of etiquette. More to the point, I think that moral reasons are not in general reasons that a rational person would be responsive to, if she were aware of them, just in virtue of being rational. That is, I believe that moral reasons are not authoritative reasons. Philosophers who hold that morality has motivational normativity presumably would disagree. They might say that moral reasons are not genuine reasons if they are not among the reasons that a rational person would be responsive to, if she were aware of them, just in virtue of being rational. This strikes me as an eccentric use of the word “genuine”, but the substantive issue is whether moral reasons are genuine in this sense. My claim here is simply that there is logical space to accept that there are moral

Moral Naturalism and Three Grades of Normativity

29

reasons while rejecting motivational normativity (compare Foot 1978, esp. pp. 160-162 and p. 168, note 8). It appears, then, that if normativity makes difficulty for moral naturalism, the difficulty lies with the third grade of normativity, authoritative normativity. To be sure, critics will argue that the naturalistic accounts of generic and motivational normativity that I have sketched are unsuccessful. But it would be more difficult to show that no naturalistic account could be successful. In addition, critics would have to defeat skeptical arguments, such as the argument I offered above, against the thesis that morality is characterized by motivational normativity. The better strategy for critics of naturalism, it seems to me, would be to argue that generic and motivational normativity are insufficient to account for the special authority that morality has, and then to argue that no naturalistic theory could explain authoritative normativity.

7. Authoritative Normativity There is a widely shared intuition that generic and motivational normativity are insufficient to explain the normativity of morality. To be sure, if I am correct, morality does not actually have motivational normativity. Yet it might seem that the examples I have pointed to, in which a rational person fails to be motivated appropriately by her moral beliefs, are compatible with an underlying conceptual link between moral belief and motivation, a link that depends on and is explained by the authoritative normativity of morality. There are two proposals about this. According to the first, moral considerations give us authoritative reasons, reasons of a kind that any rational person would take into account, if she were aware of them, just in virtue of being rational. More generally, requirements of some kind K have authoritative normativity just in case, necessarily, there is an authoritative reason to act as one is K-required to act – a reason that any rational person would take into account, if she were aware of it, just in virtue of being rational. According to the second proposal, a rational person who was thinking clearly would understand that the “Why be moral?” question is closed. That is, the question could not arise as a live question for her, a question that indicated any indecision about whether to be moral. More generally, requirements of kind K

30

David Copp

have authoritative normativity just in case, necessarily, no fully rational person who is thinking clearly can take there to be any serious practical question of whether to act in accord with K requirements. My discussion of authoritative normativity will be complex. First, I will raise the skeptical worry that, at least as they stand, neither the authoritative reasons proposal nor the closed question proposal is plausible as an account of a heightened grade of normativity. Basically the same worry also applies to motivational normativity. Second, I will point out that there are naturalistic strategies for explaining authoritative normativity, at least on the authoritative reasons proposal. Third, I will consider, very briefly, a family of “Kantian” accounts of authoritative normativity that turn on a strategy of argument I call the “self-conception strategy”. I will use Christine Korsgaard’s views in order to illustrate how a naturalist might try to turn a self-conception theory to his advantage. Fourth, and finally, I will contend that it is not plausible that morality meets the conditions proposed by either the authoritative reasons proposal or the closed question proposal. It appears that morality does not have authoritative normativity. I begin with the skeptical worry. (1) Doubts about the Significance of Authoritative and Motivational Normativity The idea of rational agency is crucial to the accounts I have given of both motivational and authoritative normativity. Motivational normativity is explained in terms of what would motivate a rational person. Authoritative normativity, on the authoritative reasons proposal, is explained in terms of reasons that a rational person would take into account in her deliberation; on the closed question proposal, it is explained in terms of a key question that no rational person would take to be an open practical question. The idea of rational agency that is invoked in these accounts is the idea of an agent who complies with norms of practical rationality, such as the norm that requires one to take the means to one’s ends. That is, motivational and authoritative normativity relate anything that has them to the norms of rationality. Hence, if morality has motivational normativity, or if it has authoritative normativity, its having this property consists in the obtaining of the relevant specified relation between it and the norms of rationality. This

Moral Naturalism and Three Grades of Normativity

31

means that unless norms of rationality have special normative significance, neither motivational normativity nor authoritative normativity can plausibly be viewed as a special or heightened grade of normativity. Let me explain more fully. Begin with the authoritative reasons proposal, which explains authoritative normativity by reference to so-called authoritative reasons. On the face of it, such reasons are just a kind of reason alongside other kinds of reason. They are reasons that any rational person would take into account in deliberation, if she were aware of them, just in virtue of being rational. Moral reasons are reasons that any moral person would take into account, if she were aware of them, just in virtue of being moral. It would be interesting and perhaps important if moral reasons were also authoritative reasons in the specified sense. But in the absence of a showing that the norms of rationality have a special kind of normative significance, to establish that moral reasons are also authoritative would not be to show that they have a special or heightened grade of normativity. One might protest that authoritative reasons are not just another kind of reason. They are reasons that have “genuine deliberative weight”, which so-called moral reasons and other reasons-of-a-kind, such as reasons of etiquette, need not have (see Darwall 1997, p. 306). It is unclear how to understand this protest, however, except as pointing out that, while reasons of etiquette are not reasons that any rational person would take into account, if she were aware of them, just in virtue of being rational, authoritative reasons are, by stipulation, reasons that any rational person would take into account, if she were aware of them, just in virtue of being rational. We might insist on reserving the term “reason” for authoritative reasons, but this would not accomplish anything substantive. Note that, if there are any requirements of rationality, they give rise to authoritative reasons.30 For if an action is rationally required, then any person who is aware that it is rationally required would take this fact into 30

I take it as given that requirements of rationality would have generic normativity and motivational normativity. They would have motivational normativity, for it is presumably a conceptual truth that any rational person who believes he is rationally required to do something will be motivated to some degree to do it.

32

David Copp

account in her practical deliberation to the extent that she is rational, just in virtue of being rational. This conceptual point implies that there is authoritative reason to comply with the requirements of rationality. There is a corresponding point one could make about moral requirements, for moral requirements are such that, necessarily, any person who is aware that an action is morally required would take this fact into account in her practical deliberation to the extent that she is moral, just in virtue of being moral. This point about moral requirements does not show that they have any special authority, and the similar point about requirements of rationality does not show that they have any special authority. The point I am making might be obscured if we already believe, perhaps on independent grounds, that the requirements of rationality have special normative significance. But in the absence of an argument that norms of rationality have special normative significance, it seems that the fact that the requirements of rationality give rise to authoritative reasons does not show that they have any special status. Similarly, then, the fact, if it is a fact, that moral requirements give rise to authoritative reasons cannot show that they have any special status. In the absence of an argument that norms of rationality have special normative significance, it would not show morality to enjoy a special degree or kind of normativity. It therefore seems that, in order to support the authoritative reasons proposal as a conception of a special or a heightened degree of normativity, we need a showing that rationality has special normative significance. Turn now to the closed question proposal, which explains authoritative normativity by reference a question that no fully rational person who is thinking clearly can take to be a serious practical question. It might seem that no rational person who was thinking clearly could take there to be a serious practical question whether to act in accord with the requirements of rationality. Yet again, it might also seem, by parity of reasoning, that no fully moral person could take there to be a genuine question whether to do her moral duty. But this would not show that moral requirements have any special status, and the corresponding point about requirements of rationality surely could not show by itself that they have a special status. If not, then it seems that even if no fully rational person who was thinking clearly could take there to be a serious practical question whether to act in accord with

Moral Naturalism and Three Grades of Normativity

33

moral requirements, this by itself could not show that moral requirements have a special status. It therefore seems to me, again, that in order to support the closed question proposal as a conception of a special or a heightened degree of normativity, we need a showing that rationality has special normative significance. It should now be clear that a parallel argument can be constructed with respect to motivational normativity. That is, in order to support the idea that motivational normativity is a special or a heightened degree of normativity, we need a showing that rationality has special normative significance. In short, the three proposed grades of normativity merely relate morality or moral properties to requirements or norms of rationality, and so, unless the latter have special normative significance, a showing that morality stands in the specified relation to them cannot amount to a showing that morality has a heightened degree or grade of normative stringency. It may be, of course, that requirements of rationality do have special normative significance. Elsewhere, I have proposed an account of what I call “self-grounded” rationality. In so doing, I argued that the requirements of self-grounded rationality, and the associated reasons, have a (defeasible) centrality in the deliberation of autonomous agents that other kinds of requirements and reasons need not have (Copp 2005). It is not clear that this means they have a special normative significance. But if it does, this isn’t a problem for naturalism, for, although I cannot go into the details here, it can be explained in a way that is compatible with naturalism.31 I have also argued elsewhere that requirements of self-grounded rationality do not have a special status at the level of generic normativity (Copp 1997c). It now appears that any difference at the levels of motivational normativity or authoritative normativity between their status and the status of requirements of morality may be of no normative significance, since motivational and authoritative normativity are defined in terms of rational agency, which in turn is understood in terms of norms of rationality. We cannot use such differences to support the view that requirements of rationality have special normative significance since these grades of normativity are defined in terms

31

The argument in Copp (2005) is entirely compatible with naturalism. Another naturalistic account is found in Hubin (2001).

34

David Copp

of rational agency, which presupposes the idea that norms of rationality have special significance. That is, the idea that motivational normativity or authoritative normativity is a heightened grade of normativity presupposes that norms of rationality have special normative significance. (2) Naturalism and Authoritative Normativity Let me set these worries aside, however. I now want to point out that there are naturalistic strategies for explaining authoritative normativity, at least on the authoritative reasons proposal, and perhaps also, as we will see in section (3), on the closed question proposal. My account of “self-grounded” rationality is fully naturalistic. I cannot go into detail here (see Copp 2005), but the central idea, very roughly, is that rationality consists in efficiently serving one’s values. I propose that rational action is governed by a standard I call the standard of self-grounded reason, which calls on us to serve our values. I argue that this standard is relevantly authoritative in virtue of the fact that conformity with it furthers or instantiates one’s self-government. It is in virtue of this fact, that claims about the requirements of self-grounded reason have generic normativity. For present purposes, however, the important point is that if rationality consists in efficiently serving one’s values, then self-grounded reasons are the reasons that a rational person would take into account in deliberation if she were aware of them, just in virtue of being rational. If I am correct, it follows that selfgrounded reasons are authoritative, on the above definition of authoritative reasons. This means, if I am correct, that naturalism can make room for authoritative reasons. The interesting question that remains, of course, is whether moral reasons are authoritative. I would argue that a rational agent need not have any moral values, and that if she does not, she may not have self-grounded reason to do what she has moral reason to do. That is, a rational person would not necessarily take moral reasons into account in deliberation if she were aware of them, just in virtue of being rational. Moral reasons are not authoritative reasons. Michael Smith disagrees. He has developed a naturalistic theory according to which, in my terminology, moral reasons are authoritative. According to Smith, roughly, the moral rightness of an action in conditions C is the feature we would want acts to have in C if we were fully rational. He claims that

Moral Naturalism and Three Grades of Normativity

35

this feature is a natural feature of acts, at least in worlds like ours (see Smith 1994, esp. chapters 5 and 6).32 The argument for this depends on two theses. First, the rightness of an action is a matter of its being called for by reasons, and second, there is a reason for an agent to do something just in case, roughly, if she were fully rational, she would want herself (as she actually is) to do it. Smith provides a broadly naturalistic account of full rationality in terms of the coherence of a rational person’s beliefs and desires. Given all of this, he argues, to believe that one is morally required to do something is, roughly, to believe that, if one’s beliefs and desires were fully coherent, one would want oneself (as one actually is) to do it. Moreover, if a fully rational person believes such a thing regarding an action she could perform in a situation, she must be motivated to perform that action. Moral reasons therefore qualify as authoritative on Smith’s account, for they are reasons that would motivate a fully rational person just in virtue of her rationality. I have argued elsewhere that Smith’s argument is unsuccessful (see Copp, 1997a), but the important point is that the problem with Smith’s account is not that it is naturalistic. It fails for other reasons. (3) Self-Conception Strategies A family of “Kantian views” agree with Smith that moral reasons are authoritative, and employ a strategy of argument that I call the “self-conception strategy” to show that they are authoritative. On the self-conception strategy, there is a way of conceiving of oneself such that a rational person who is thinking clearly must conceive of herself this way, but if she does not treat moral reasons as authoritative, she cannot coherently conceive of herself in this way. It might be said, for example, that a person who does not treat moral reasons as authoritative cannot see herself as autonomous; or that she cannot see herself or value herself as a rationally reflective agent, acting for reasons; or that she is committed to a practical solipsism; or that she cannot coherently expect other people to respond to the reasons she addresses to them; or the like.33 On the self-conception strategy, rational people who are thinking clearly and coherently respond appropriately to moral reasons. 32

On the identification of rightness with a natural feature of acts, see Smith (1994), pp. 184-186. On the definition of naturalism, see Smith (1994), p. 203, note 1 to chapter 2. 33 I sketch these ideas with apologies, respectively, to the following philosophers: Kant (1981); Korsgaard (1996); Nagel (1970); Darwall (forthcoming).

36

David Copp

Given the complexity and variety of self-conception views, there is little hope of a generic assessment of them. They need to be evaluated one at a time. They make two central claims, each of which needs defense. First, they specify some way of understanding ourselves such that, they claim, rational people who understand morality and are thinking clearly will not be able to understand themselves in this way if they are unmoved by moral considerations. Second, they claim that the specified way of understanding ourselves is non-optional, such that no rational person who was thinking clearly could fail to view herself in this way. It is only if we cannot fail to view ourselves in the relevant way, provided that we are rational and thinking clearly, that it follows from the first claim of a self-conception view that we cannot fail to be moved by moral considerations provided that we are rational and thinking clearly. For my purposes, the key point is that a self-conception view may be compatible with moral naturalism. To illustrate this, I want to look briefly at Christine Korsgaard’s position. The central point is that one cannot use a self-conception view to argue against moral naturalism without showing that the view is incompatible with naturalism. Korsgaard says that the central question that needs to be answered is, “[W]hat justifies the claims that morality makes on us?” She calls this the “normative question” (Korsgaard 1996, pp. 9-10). She suggests that whatever property an action might have, a person can intelligibly ask why that should matter to her, or why she should take that to be a reason for her to act. So it appears we cannot answer the normative question by citing a property of an action, and she argues that an action’s being morally required is not a matter of its having a certain kind of property (Korsgaard 1996, pp. 33-35).34

34

She therefore rejects the position she calls “substantive realism”. The substantive realist postulates the existence of “intrinsically normative entities” such that correct answers to moral questions are correct “because there are moral facts or truths [involving those entities], which those questions ask about.” (Korsgaard 1996, pp. 34-35; emphasis in original), Although Korsgaard rejects “substantive realism”, she accepts “procedural realism”, which she defines as the view “that there are right and wrong ways to answer [moral questions]” (Korsgaard 1996, p. 35). She adds, “Substantive realism is a version of procedural realism, of course” (Korsgaard 1996, p. 37, fn. 58).

Moral Naturalism and Three Grades of Normativity

37

Instead, she holds, an action’s being morally required is a matter – very roughly – of the action’s being required by our “identity” as rationally reflective agents – where a person’s identity is the set of properties she values in herself (Korsgaard 1996, p. 101). Reflective agency, she thinks is a property that we cannot but value having, if we are to value anything. Accordingly, she thinks, the requirements of our identity as rationally reflective agents are not open to the normative question. Korsgaard’s normative question is very much like the traditional “Why be moral?” question, and indeed, in some passages she seems to be asking this question (Korsgaard 1996, p. 33). It appears that, on Korsgaard’s account, the “Why be moral?” question could not arise as a genuine practical question for a fully rational person who was thinking clearly and who understood what it is for an action to be morally required – provided she understood in addition that she must value her nature as a rationally reflective agent if she is to value anything. For such a person, Korsgaard thinks, the normative question, “What justifies the demands that morality makes on us?” does not raise a genuine practical concern. And if not, then to the extent that she is fully rational and thinking clearly, the “Why be moral?” question also does not raise a genuine practical concern. I am not persuaded. I think that a rational person who understood Korsgaard’s account might still ask why she should be moral, and this question might, for her, be a practical question, a question that indicates an indecisiveness about whether to be moral. Gyges values the power and love he will achieve if he carries out his plot, and this means that, if he is fully rational and thinking clearly, and if Korsgaard’s theory is correct, he must value his reflective agency. Moreover, if Korsgaard’s analysis is correct, and if Gyges understands morality, he must then understand that carrying out his plot would conflict with his valuing his reflective agency. Yet he also values power and love, and he understands that carrying out his plot would help him to achieve a life of power and love. So he might ask, “Why should I not carry out my plot? Why should I be moral?” For all that Korsgaard has shown, it seems to me, his asking these questions would not indicate either that he is not fully rational or that he does not understand morality. Nor need it indicate that he does not value his reflective agency. A person who values A can rationally wonder whether to do something that he understands would

38

David Copp

conflict with achieving or sustaining A when doing it would allow him to achieve something else he values. To wonder about this is compatible with his continuing to value A.35 The more important point, however, is that a naturalist can embrace a reconstruction of Korsgaard’s position. Suppose that Korsgaard is correct that if an action is required of a person then the person’s not performing it would conflict with her valuing herself as a rationally reflective agent. Suppose she is also correct that a person cannot but value her rational reflective agency unless she ceases to value anything, and thereby ceases to be able to see herself as an agent. All of this specifies a complex property that, if Korsgaard is correct, is possessed by actions that are morally required – the property an action has if and only if, roughly, (K1) the agent’s failure to perform it would conflict with her valuing herself as a rationally reflective agent and (K2) the agent must value rational reflective agency if she is to value anything and to see herself as an agent. Call this the “Korsgaard property”. I say that it could be viewed as a natural property. How could the Korsgaard property be a natural property? Earlier I suggested that a property is natural just in case, roughly – leaving aside analytic truths – any proposition about the instantiation of the property that can be known, can be known only empirically. Consider, then, propositions to the effect that an action has the Korsgaard property. Any such proposition is equivalent to the conjunction of a “K1 proposition”, a proposition to the effect that the action in question has property K1, with a “K2 proposition”, a proposition to the effect that the action has property K2. Therefore, we can divide our problem in two. Begin with K2. A naturalist who is tempted by 35

My argument here depends on an interpretation of the idea that actions may be required by our identity as rationally reflective agents. I assume this means that actions may be required by our valuing rational reflective agency. And I propose that an action is required by our valuing something just in case a rational and factually informed person who valued that thing and who was thinking clearly could not fail to perform the action provided that nothing else he valued was at stake. One might respond that the final clause of my proposal should be deleted. But if I am correct, deleting this clause would mean in effect that Gyges is not required by his valuing rational reflective agency to abandon his plot. For he faces a conflict of values. He can choose to pursue power and love while continuing to value rational reflective agency (see Copp 1999).

Moral Naturalism and Three Grades of Normativity

39

Korsgaard’s picture presumably needs to argue that it is a conceptual truth that any agent must value rational reflective agency if she is to value anything at all and to see herself as an agent. But if this is a conceptual truth, then it is also a conceptual truth that any action has the K2 property. (Similarly, if it is a conceptual truth that murder is wrong, then it is a conceptual truth that any action has the property of being such that murder is wrong.) But if it is a conceptual truth that any action has the K2 property, then we can set aside the relevant K2 propositions as analytic. Turn now to K1. A naturalist who is tempted by Korsgaard’s picture presumably must argue that, for some types of action, it is a conceptual matter, and for other types of action, it is an empirical matter, that actions of that type have the K1 property – the property of being such that an agent’s failure to perform an action of that type would conflict with her valuing herself as a rationally reflective agent. Perhaps, for example, it is a conceptual matter that actions that are fair have the K1 property, and perhaps it is an empirical matter that donating money to charity has the K1 property. If the naturalist can make a case for claims of this kind about the K1 and K2 properties, then she can make a case that Korsgaardian rightness is a natural property. In short, if Korsgaard is correct, then an action is right just in case the relevant K1 and K2 propositions are true. If the naturalist can argue that such propositions are either empirical or conceptual, then she can accept Korsgaard’s position. (4) That Morality Lacks Authoritative Normativity As I have suggested, I do not believe that moral reasons are authoritative in a way that puts an end to all doubts about whether to be moral. I said before that it might seem that a rational person could not question whether to live in accord with the requirements of rationality. But this is not the case, and it will be instructive to see why. Consider again the myth of Gyges. Let us suppose that Gyges thinks that in view of his values he ought rationally to carry on with his plan to take over the kingdom but that he ought morally to give up his plan. We imagined him as asking, “Why should I be moral?”, and this seemed an intelligible question for him to ask since doing his duty would mean giving up a life of power and love. But we can also imagine him asking, “Why should I be rational?”, and this also seems an intelligible question for him to ask since pursuing his plan would involve him in murder, and

40

David Copp

so it would mean violating his moral duty. It would be a mistake to suppose that, if he is rational and thinking clearly, he will settle on carrying out his plan, given that he understands that he rationally ought to pursue his plan. He might well understand that his reasons to pursue his plan have “genuine deliberative weight”, such that he will take them into account provided that he is rational. But his question is whether to continue to be rational. It seems to me that he could decide against this. And if Gyges can take there to be a genuine practical question whether to be rational, compatibly with his actually being rational and thinking clearly, he can surely take there to be a genuine practical question whether to be moral, compatibly with his being rational and thinking clearly.36 I think in fact that a naturalist can reject both the authoritative reasons proposal and the closed question proposal, for I think it is a mistake to suppose that morality has authoritative normativity on either of these proposals.37 In my view, as I suggested earlier, moral reasons are not reasons of a kind that any rational person would take into account in her deliberation, if she were aware of them, just in virtue of being rational. I have argued this in other places (see Copp 1997c; Copp 2005). Moreover, there is no definitive answer to the “Why be moral?” question of the kind that believers in authoritative normativity have in mind. Of course there are answers that 36

To take an extreme case, suppose that the evil genius threatens to destroy the earth unless you take a pill that will make you no longer rational. Even though you are rational, you might decide to take the pill. Thomas Schelling has said that, in face of a threat, it is not “invariably an advantage to be rational.” (Schelling 1960, pp. 18-19). We can imagine as well a case in which the evil genius threatens to destroy the earth unless you take a pill that will make you no longer be moral. Even if you are moral and rational, you might take the pill. 37 Indeed, it may seem that I have already argued that it is a mistake in Copp (1997c). But that paper addresses the idea that moral requirements are “overriding”, and argues, in effect, that no account of overridingness can be given solely on the basis of an account of generic normativity. The issue is different from the issue whether moral requirements have authoritative normativity. My arguments in Copp (1997c) seem to leave open the possibility that moral reasons are also reasons that any rational person would take into account, if she were aware of them, just in virtue of being rational. Even if moral reasons are authoritative in this sense, however, it does not follow that they are overriding. For one thing, there might be other kinds of authoritative reasons and moral reasons might not be normatively more important than they are.

Moral Naturalism and Three Grades of Normativity

41

can be given to various individuals in their specific concrete situations. And there is a very general answer that can be given. For in my view, what justifies the demands of morality is that they are the demands imposed on us by a system of standards the currency of which in the relevant society would contribute better than that of any other such system to enabling the society to meet its needs. But this proposal does not yield a definitive answer to the “Why be moral?” question, for one can respond to my account by asking, “What is all of this to me?” There is no way to put all such doubts to rest in rational reflective persons. In my view, there is no answer to the “Why be moral?” question that lays it to rest by showing that it does not raise a serious practical question that could indicate an indecisiveness about morality in a rational person who had a clear understanding of morality. If this is correct, morality does not have authoritative normativity, not, at least, on either of the conceptions of this that we have considered. I do not think that a failure to be motivated by moral belief is a sign of irrationality. It is a sign of immorality, but this is no surprise. Nor do I think that a fully rational person who had a clear understanding of the nature of morality could not be indecisive about whether to be moral. Such a person could be indecisive about whether to continue to be rational, so it would be astonishing if she could not be indecisive about morality. Perhaps a fully moral person would not be indecisive about whether to be moral, but why should we expect rationality to secure moral motivation in a clear thinking and fully informed person? Those who think that morality has authoritative normativity therefore face a challenge. They need to provide an adequate conception of authoritative normativity and an argument to show that morality possesses authoritative normativity. And, if they think this counts against naturalism, they need to provide an argument that it does.

8. Conclusion I have tried to point to some of the complexities that would have to be addressed before we could conclude that moral naturalism is ruled out by the fact that morality is normative. I distinguished three grades of normativity: generic, motivational, and authoritative normativity. I argued that moral naturalism has no obvious difficulty explaining generic and motiva-

42

David Copp

tional normativity. There are naturalistic theories in the literature, including a >theory I have proposed, that purport to explain what is involved in these grades of normativity. Authoritative normativity is another matter. I discussed two proposals about what authoritative normativity consists in, and argued that morality lacks authoritative normativity on both proposals. I argued as well that, on either proposal, to support the idea that authoritative normativity is a special or a heightened degree of normativity, we need a showing that rationality has special normative significance. I argued that there are nevertheless naturalistic accounts of authoritative normativity on both proposals. I myself suspect that morality does not possess authoritative normativity. If I am correct about this, then it is no problem for naturalism if, despite what I have suggested, authoritative normativity cannot be understood naturalistically. Let me conclude by saying that moral non-naturalism faces the challenge of explaining the normativity of morality just as much as does moral naturalism. If normativity needs to be explained, it is not explained by giving up on naturalistic ways of explaining it. Anti-reductionist forms of non-naturalism that view moral properties as sui generis face an especially difficult problem, for they appear simply to postulate normativity. It is unclear how they could explain it. If one of our goals is to explain what normativity consists in, then, as I have suggested, naturalistic accounts can be given of all three grades of normativity. These accounts seek to explain the normativity of moral properties by providing an analysis, first, that explains what normativity consists in, and, second, that exhibits normativity as a second-order natural property of the moral properties. Moral naturalism appears to be alive and in good health.38 38

I presented an early version of these ideas in September, 2003 to a conference on Naturalism and Normativity at the Ethik-Zentrum of the University of Zurich; in October, 2003 to the department of philosophy at Ohio University; and in March, 2004 to the St Louis Philosophy of Social Science Roundtable. I benefitted greatly from the helpful discussion on these occasions. I am also grateful to Philip Clark, Sarah Buss and Evan Tiffany for very helpful correspondence by e-mail that helped me to clarify my thinking about authoritative normativity. I am especially grateful to my colleagues in the Florida ethics reading group who discussed this paper: Peter Barry, John Biro, Marina Oshana, Crystal Thorpe, Jon Tresan, Anton Tupa, and Gene Witmer.

Moral Naturalism and Three Grades of Normativity

43

References Blackburn, S (1993): Essays in Quasi-Realism, New York. Copp D. (1990): “Explanation and Justification in Ethics”, in: Ethics, 100, pp. 237-258. – (1991): “Moral Realism: Facts and Norms”, in: Ethics, 101, pp. 610-624. – (1995): Morality, Normativity, and Society, New York. – (1997a): “Belief, Reason, and Motivation: Michael Smith’s, The Moral Problem”, in: Ethics, 108, pp. 33-54. – (1997b): “Does Moral Theory Need the Concept of Society?”, in: Analyse et Kritik, 19, pp. 189-212. – (1997c): “The Ring of Gyges: Overridingness and the Unity of Reason”, in: Social Philosophy and Policy, Volume 14, pp. 86-106. – (1998): “Morality and Society – The True and the Nasty: Reply to Leist”, in: Analyse et Kritik, 20, pp. 30-45. – (1999): “Korsgaard on Rationality, Identity, and the Grounds of Normativity”, in: J. Nida-Rümelin (ed.): Rationality, Realism, Revision, Berlin, pp. 572-581. – (2001): “Realist-Expressivism: A Neglected Option for Moral Realism”, in: Social Philosophy and Policy, 18, pp. 1-43. – (2003): “Why Naturalism?”, in: Ethical Theory and Moral Practice, 6, pp. 179-200. – (2005): “The Normativity of Self-Grounded Reason”, in: Social Philosophy and Policy, 22. Dancy, J. (forthcoming): “Non-Naturalism”, in: D. Copp (ed.): The Oxford Handbook of Ethical Theory, New York.

44

David Copp

Darwall, S. (1983): Impartial Reason, Ithaca. – (1997): ”Reasons, Motives, and the Demands of Morality: An Introduction”, in: S. Darwall, A. Gibbard/P. Railton (eds.): Moral Discourse and Practice , New York. – (2001): , in E. Craig (ed.): Routledge Encyclopedia of Philosophy. London, accessed 2004/06/10, from http://www.rep.routledge.com/article/L135. – (forthcoming): “Morality and Practical Reason: A Kantian Approach”, in: D. Copp (ed.): The Oxford Handbook of Ethical Theory. Foot, Ph. (1978): “Morality as a System of Hypothetical Imperatives”, in: Virtues and Vices, Berkeley. Gibbard, A. (1990): Wise Choices, Apt Feelings: A Theory of Normative Judgment, Cambridge (MA). Hare, R. M. (1952): The Language of Morals, Oxford. Hubin, D. C. (2001): “The Groundless Normativity of Instrumental Rationality”, in: The Journal of Philosophy, 98, pp. 445-468. Jackson, F. (1998): From Metaphysics to Ethics: A Defence of Conceptual Analysis, Oxford. Kant, I. (1981 [1785]): Groundwork of the Metaphysics of Morals, trans. by J. W. Ellington, Indianapolis. King, J. C. (1998): “What is a Philosophical Analysis?”, in: Philosophical Studies, 90, pp. 155-179. Korsgaard C. (1996): The Sources of Normativity, ed. by O. O’Neill, Cambridge. – (1986): “Skepticism about Practical Reason”, in: Journal of Philosophy, 83, pp. 5-25. Mackie, J. L. (1977): Ethics: Inventing Right and Wrong, Harmondsworth.

Moral Naturalism and Three Grades of Normativity

45

Moore, G. E. (1968 [1942]): “Reply to My Critics”, in: P. A. Schilpp (ed.): The Philosophy of G. E. Moore, third edition, La Salle. – (1993 [1903]): Principia Ethica, ed. by Thomas Baldwin, Cambridge. Nagel, T. (1970): The Possibility of Altruism, Oxford. Schelling, T. C. (1960): The Strategy of Conflict , Oxford. Smith, M. A. (1994): The Moral Problem, Oxford. Stevenson, C. (1937): “The Emotive Meaning of Ethical Terms”, in: Mind, 46, pp. 14-37. Sturgeon, N. (forthcoming): “Ethical Naturalism”, in: D. Copp (ed.): The Oxford Handbook of Ethical Theory, New York. Tresan, J. (forthcoming): “De Dicto Internalist Cognitivism”, in: Noûs.

Neil Roughley Naturalism and Expressivism On the “Natural” Stuff of Moral Normativity and Problems with its “Naturalisation”

Introduction 1. The term “naturalism” is frequently employed these days as a cipher for respectable philosophical methodology. Authors who do so rely on the assumption of an intimate connection between “naturalism” and the natural sciences. We are all impressed by way scientific thought has enabled the production of technologies that have transformed the face of the earth. It is eminently plausible that this has only been possible because science uncovers the structure and workings of reality. Science may thus seem to have slipped into the role traditionally fulfilled by philosophical metaphysics: the structure of “being” appears, at least up to a point, to be an empirically discoverable matter. There are, however, various points at which doubts arise as to how much of the human world is knowable in epistemic terms modelled on those of the natural sciences. One such point is reached where we step into the ethical sphere. 2. Historically, attempts to understand morality have given rise to a number of strange constructions that certainly don’t permit reconstruction along natural scientific lines. These constructions are, however, not just the products of minds that had not yet grasped the full import of empirical science. They are, rather, attempts to capture constitutive features of moral experience. In what follows, I shall be arguing that the relevant features are best captured

48

Neil Roughley

by a version of the expressivist analysis of moral judgement. My central claim will be that it is expressivism that allows a reconciliation between the claims of science and the essentially practical nature of morality. Moreover, only when we get clear on this can we say in what sense we should, and in what sense we should not, be “naturalists”. 3. These claims need some preparation. I begin by distinguishing various senses of the term naturalism. This is important, as the term can be used to designate anything from a broad commitment to keep the “supernatural” out of philosophy to a methodologically highly specific conception of how that has to be done (section 1). There follows a critical discussion of one such conception in its application to the concept of moral normativity: the idea that “the right” could be explicated in functional terms (section 2). I then develop my own reconstructive suggestion, doing so in three steps. The key to understanding moral normativity, I argue, is an understanding of the notion of a standard. I approach this topic via a brief examination of social norms, as it is here that standards are most palpable (section 3). This leads to a discussion of the basic forms of attitudinising at work in the constitution of social norms, what I call optative attitudes (section 4). On this basis, I finally sketch a conception of the attitudinal constitution of moral normativity, which gains its particular profile through the comparison with social norms: it shares with them certain important features, but also differs from them in decisive ways (section 5). 4. An analysis of the optative constitution of moral normativity clarifies why, although the dimension of the human world thus generated is constructed from nothing but natural materials, it remains resistant to analysis in terms of the properties accessible to the natural sciences. Put succinctly, the reason is that the morally normative is a feature internal to optative attitudinising, in the sense that stepping out of the optative mode renders it inaccessible.

Naturalism and Expressivism

49

1. Ethics and the Ways of Being Naturalist 1. “Naturalisation” is a rapidly expanding philosophical industry. These days a great deal of philosophical energy is invested in projects of “naturalising” our understanding of topics that have often seemed resistant to “naturalist” construals. The development of the cognitive sciences has both made the mind the most intensely hunted game in the philosophical jungle and raised the general confidence that a full-scale naturalisation of all interesting components of the human life form will be possible. 2. Hot on the heels of the mind comes morality. Certainly, reflections on ethics have over the centuries given rise to forms of talk that can only be made sense of if “supernatural” entities or capacities are postulated alongside the components of the empirical world. The Platonic “form of the good” exists in another ontological dimension. Actions that realise the Kantian “moral law” come about as a result of a remarkable “non-empirical causality”. Finally, at the inception of what we know as “meta-ethics”, G.E. Moore insisted that first “the good” and later “the right” are in some significant sense “non-natural properties”.1 3. If we want to understand the ethical sphere, we have good reason to be unhappy about these sorts of pronouncements. Where we are after clarity, they offer us mystery. In effect, they are telling us that there is a point beyond which we are unable to understand what is going on and that further striving after clarity is futile. It ought to be obvious that philosophers should be very reluctant to make such claims, as they amount to an admission of defeat in our analytic endeavours. Nevertheless, it would be equally unphilosophical to insist that there must for a priori reasons be a perfect fit between the world and our cognitive capacities, which guarantees our ability to give a “full account” – whatever that may be – of everything we set out to comprehend.

1

For Moore’s (incomplete) thoughts on the “non-natural” character of “right”, see Moore (1993), pp. 4f.

50

Neil Roughley

4. Although they drew very different conclusions, Plato, Kant and Moore were all convinced that, one way or another, we come up against significant cognitive barriers in the ethical sphere. This is in itself not uninteresting, as it reflects what I suspect is the general cognitive condition among personson-the-street with respect to the morally normative. Firstly, “the folk” will manifest considerable perplexity if asked to spell out what they are talking about when they employ these concepts.2 Secondly, further prodding is likely to elicit an eclectic mix of cultural conceptions with a tendency to the transcendent. There is, we the folk tend to think, something supra-positive about the moral “ought”; we feel that its demands are somehow time- and placeless. A plausible metaethics ought to be able to make clear what it is about the ethical sphere that generates “intuitions” of this kind. Certain features of morality, in particular moral normativity, do appear to be candidates for the status of irreducible phenomena. If someone is to succeed in “naturalising” moral normativity then, they will in the process need to explain its “non-naturalistic” appearance. Here, as elsewhere, the maxim holds that if we cannot show how our analyses account for the everyday phenomena, then they forfeit the right to count as analyses of those phenomena. 5. Before turning to approaches to the ethical, we need to spend a moment or two considering what is precisely to be understood by the term “naturalism”. For certain philosophers, being “naturalist” seems to be a necessary if not sufficient condition of the respectability of one’s philosophical methodology. And indeed, there are ways of using the term that make this fairly obviously true. Two such uses are the following:

2

This contrasts starkly and importantly with the epistemic sphere, where just about everybody will serve up some version of the correspondence “theory” of truth without too much prompting. Because the idea of correspondence is the core of what people mean by “truth”, talk of a “correspondence theory” is a misnomer. In this respect, meta-ethical attempts to reconstruct how we think morally when we are on the street are engaged in a different enterprise. The eclectic mixture of cultural conceptions that structure and clutter our ethical thought can only be made sense of by severe pruning. There is a basic asymmetry here that tells against a systematic parallelism of “the true” and “the right”, as postulated for instance in Simon Blackburn’s “quasi-realism”.

Naturalism and Expressivism

51

(N1) A philosophical approach is naturalist iff its procedures are consistent with the assumption that its subject matter has come into being a result of evolutionary processes. (N2) A philosophical approach is naturalist iff its procedures are compatible with those of the natural sciences. 6. The popularity of “naturalism” rests squarely on the unprecedented success of the natural sciences since the enlightenment. No one doing serious philosophy can deny that in the last two centuries methods have been developed that have enabled massive leaps forward in the understanding of the how the world hangs together and may be preparing even greater leaps for the not-too-distant future. Philosophy clearly has to ensure its compatibility with these advances and ought moreover to be aware of the results of empirical research that are relevant to its topics. This means not only that the philosophy of mind should, as is widely acknowledged, be in dialogue with the cognitive sciences, but also that action theory needs to keep abreast of the findings of motivational and social psychology.3 Philosophical attempts to say what actions are, how they are to be explained and justified, and whether and in what sense they can be free need to ensure they are not claiming anything disproven by empirical research. They also need to be aware of the phenomena thus discovered that their own constructions have to be able to account for. 7. Furthermore, philosophical understanding needs to recognise that its subject matter – whether belief, art or virtue – has come into being as the result of a process of becoming that is neither given direction by, nor interfered with by an kind of super-agent. All the agents there are are themselves the result of such processes of becoming, processes that are contingent and which, at least at some level, involve genetic “selection” as a result of survival in a particular environment. How much of contemporary evolutionary theory is

3

For an attempt to bring the relevant empirical findings to bear on a philosophical analysis of wanting and intending, cf. Roughley (forthcoming a), chapters 3, 5, 6 and 7.

52

Neil Roughley

true is unimportant here. What is decisive is simply the fact that we, along with all our distinguishing features, are the products of contingent processes of becoming that could in principle be reconstructed in causal terms. In this respect, we humans and our world are like everything else. This is a fortiori true of the mind, beauty and “the good”. 8. Short of inserting religious premises into philosophical argument, I have difficulties seeing how anyone could fail to be “naturalist” in either of these two senses. Simon Blackburn’s dictum that “to be a naturalist is to see human beings as frail complexes of perishable tissue, and so a part of the natural order” (Blackburn 1998, p. 48), is basically a concrete illustration of N1. Something close to N2 is expressed by Arthur Danto’s definition of naturalism, according to which “whatsoever exists or happens is natural in the sense of being susceptible to explanation through methods which, although paradigmatically exemplified in the natural sciences, are continuous from domain to domain of objects and events” (Danto 1967, p. 448). Note, however, that Danto talks not of compatibility, but of continuity with the methods of the natural sciences. This may well be a stronger condition, depending on the respects in which natural scientific methods are taken to be “paradigmatic”.4 Certainly, what is often meant in contemporary parlance is significantly stronger. 9. A position whose conception of the relation between philosophy and science would be radically understated as one of “continuity” is that influentially advanced by Quine in his article “Epistemology Naturalized”. Quine proposes that epistemology be understood as “a chapter of psychology and hence of natural science” (Quine 1969, p. 825). This suggests a third conception: (N3) A philosophical approach is naturalist iff its aims and procedures are aims and procedures of the natural sciences.

4

For a brief discussion of the notion of “continuity” at issue, cf. Papineau (1993), pp. 2ff. 5 My emphasis.

Naturalism and Expressivism

53

This raises the question of what precisely the aims and procedures of the natural sciences are. If the former are cast in terms of the discovery of laws, then obviously N3 massively restricts what can count as naturalist philosophy. However, the basic problem here goes deeper. If philosophy is a branch of natural science, then philosophy would appear to have no special brief for investigating the conceptual and metaphysical foundations of science and its practice. For this reason Quine refused the Logical Empiricists the title of “naturalists”: because they aimed to rationally reconstruct science from the materials of sense data and logic, they were integrating science within philosophy, rather than philosophy within science (Quine 1969, pp. 74ff.). According to Quine, it is only the latter perspective that is genuinely “naturalistic”. 10. Such a conception rejects what many philosophers, myself included, see as an essential feature of their metier: that when we set out to provide philosophical answers, everything is potentially up for grabs. That includes the question of how we should understand the natural scientific enterprise itself – a commonplace in the debates between realism and instrumentalism. When we turn to the ethical sphere, there can be no presumption that being “naturalist” in the sense of N3 is an obvious virtue. Indeed, I shall be arguing that any conception that tries to work within such a framework is necessarily unable to make sense of the key notion of moral normativity. 11. To complicate matters further, what is meant by “naturalism” has, in ethical contexts, historically been given a very specific, semantic twist. For G.E. Moore, naturalism was a doctrine concerning the definability of “good”. Leaving aside various problems of Moore philology which need not concern us here, we can usefully distil a conception of ethical naturalism from his writings as the claim that ethical concepts are definable entirely by means of predicates that pick out natural properties. “Natural” properties, in turn, are defined as “[properties] with which it is the business of the natural sciences or psychology to deal” (Moore 1993, p. 13). This gives us: (N4) Ethical naturalism is the claim that ethical concepts can be defined entirely in terms of properties that are the subject matter of the natural sciences or psychology.

54

Neil Roughley

One may note the disjunction between natural science and psychology here, which contrasts with Quine’s subsumption of the latter under the former. In view of the controversies within psychology as to the extent to which the discipline is a natural science – the behaviourism accepted by Quine has few remaining adherents – this is no doubt wise.6 However, if the non-subsumption of psychology under natural science is accepted and if psychological properties are among the main candidates for the reductive role, one ends up wondering why the doctrine should go under the name of “naturalism”. 12. This problem is sidestepped in the conception of ethical naturalism developed by Richard Hare. Hare drops all reference to the natural sciences, instead simply assuming that all genuine properties are “natural” in the broad sense of being part of the causal, empirically accessible order of things. He defines ethical naturalism semantically as the claim that evaluative – and normative7 – concepts are definable entirely in terms of non-evaluative concepts (Hare 1952, p. 82; Hare 1963, p. 16). Again, riding slightly rough-shod over Hare’s precise understanding of the relationship between ontology and philosophy of language,8 we can render the conception of naturalism at work here as follows: (N5) Ethical naturalism is the doctrine that evaluative – and normative – concepts pick out only properties belonging to the causal, empirically accessible order of things and nothing more. 13. There are, then, various ways of being a naturalist that don’t commit one to naturalism in other senses. In the absence of special pleading, we can surely assume that all our philosophical interlocutors are nowadays 6

The alternative, conjunctive formulations employed by Moore (1903, p. 92; 1993, p. 15) are also naturally taken to imply a non-subsumptive relation. The formulation chosen by Michael Smith, “the subject matter of a natural or social science” (Smith 1994, p. 17 and p. 35), where the latter is taken to include psychology, also raises the question of the “scientific” character of the social sciences. 7 For Hare, normative concepts are a sub-class of evaluative concepts. Cf. Hare (1963), p. 27, note. 8 Hare develops his considered opinion on these matters in Hare (1985).

Naturalism and Expressivism

55

adherents of N1. N2 and N3 require assumptions about the natural sciences, assumptions that in the former case are completely unproblematic, but in the latter case are certainly not. N4 and N5 combine the specific focus on the ethical sphere with the idea that the forms of “consistency” or “compatibility” referred to in N1 and N2 are to be established by means of semantic reduction. Obviously, nothing in N1 or N2 stipulates that this need be so. 14. On the other hand, if someone rejects the form of naturalism codified in N5 (and a fortiori of that of N4), but claims to be a naturalist in senses N1 and N2, then she has got some explaining to do. The rest of this essay is an attempt to do some of that explaining. Getting clear on the shape of such a “non-naturalist naturalism” is, moreover, not only vital for an adequate meta-ethics. It also provides some indications of how we should understand the philosophical project in general, in particular what the naturalism of our first two conceptions does not commit us to.

2. The Puzzle of Moral Normativity and its Inexplicability in “Functional” Terms 1. Normativity is arguably the feature of morality that poses the greatest challenge to its understanding as a feature of the world that has developed through evolutionary processes and is studied by the natural sciences. That difficulty was raised with particular poignancy by the Logical Empiricists, who first attempted to distil a theory of meaning out of the scientific world view, before showing that the semantics they came up with disqualifies evaluative and normative9 utterances from counting as meaningful. These days, virtually nobody thinks that the truth conditions necessary for explicating large portions of meaningful language use have to be explicable in terms of observables. And virtually nobody thinks that morally normative claims – that someone morally “should”, “ought to”, “must” (not) do 9

The emotivists were not interested in differentiating between the evaluative and the normative. Ayer does remark that, whereas ‘good’ “suggests” what an addressee is to do, ‘ought’ is more of the nature of a “command”. But this is no more than a passing remark (Ayer 1936, p. 111).

56

Neil Roughley

something; that some action is morally right, wrong, forbidden, obligatory – are meaningless. Rather, it seems to many as if the loosening of the truthconditional constraints on meaningful language use has opened up the space in which to understand the meaning of normative utterances: if within such a less tightly-knit semantics there is room for meaningful claims about positrons and the universe, about hopes and fears, then why should not the same be possible for the right and the wrong? 2. Norms, so it may seem, are real in a sense that can only be understood by placing them squarely in the middle of empirical reality. Indeed, it ought to be obvious that everyday human life is permeated by a large variety of norms and that such “permeation” involves norms playing demonstrable causal roles. The “existence” of norms – however we are precisely to make sense of it – leads to all sorts of actions that would not occur without them. If causality is to be analysed at least in part in counterfactual terms, then the “rightness” of certain kinds of actions is plausibly assigned some causal responsibility for the fact that people perform actions of that type, in as far as they would not act thus were there to be no such “normative fact”. 3. However, it ought to be equally obvious that norms, moral or other, are not definable in terms of their causal roles. It is an essential characteristic of norms that their existence does not necessitate behaviour that realises their contents. Indeed, there appears to be nothing self-contradictory about the claim that there “are” norms which have never had any influence on human action. If the causal role norms tend to play is indicative of their reality in some sense or another, the significant possibility of their complete causal inefficacy marks what may appear to be their special ontological status. 4. One can call this the puzzle of normativity. If one lets oneself be hypnotised by it, norms may seem to be entities with a strange ontological dual status – both empirical or “natural”, in as far as they are part of the causal network of the real world, and somehow non-empirical or non-”natural”, in that they also appear to stand outside that network. I think we should understand Kant as articulating this puzzle. He located moral norms outside the empirical reality framed by the constraints of time and space “in a completely

Naturalism and Expressivism

57

different … order from the order of nature” (Kant 1787, p. 373)10. Although they are supposed to be immune to influence from empirical causation, they are somehow nevertheless able to – “non-empirically” – influence the behaviour of agents in the empirical world. Thus understood, Kant’s position clearly does not count as naturalist in the sense of N1 – and is for this reason not a serious option. Nevertheless, Kant is helpful in providing us with a phenomenologically accurate picture of the puzzle itself. 5. The fact that moral oughts need not be causes makes a type of post-verificationist solution that has become popular for theoretical entities and mental states a non-starter here. This is the idea that the relevant kind of entity can be defined in terms of a characteristic causal role: by postulating certain items as intermediaries between typically observable causes and effects. The idea gains its plausibility from the fact that, just because we are – as yet, or perhaps in principle – unable to devise a means of observing goings on of a certain scale or kind, that does not mean that there is nothing there. If, after the occurrence of events of type e1…en, events of type f1…fn tend to occur, although direct causation of the latter by the former can be excluded, and if the missing links can be formally represented by intervening variables, then it is a plausible assumption that there are real, yet unobservable entities fulfilling the functions thus represented. In this way the introduction of theoretical terms can be made sense of within a truth-conditional semantics. Clearly, this is not going to work with whatever it is we mean by what morally ought to be done or what is morally right. These ideas are quite simply not matters of fulfilling causal roles. 6. Nevertheless, there are two further features of definitional “functionalism” that have made it seem attractive in metaethics, features that had first proven particularly fruitful where the functionalist apparatus was applied in the philosophy of mind. The first is that functionalism provides a way of making sense of the idea of different “levels” of reality, whilst keeping the levels thus distinguished firmly anchored in some basic stratum: where the vocabulary

10

My translation.

58

Neil Roughley

and theoretical apparatus used to describe things in some particular way can be defined in terms of causal roles, the question is open as to what otherwise identifiable entity realises that role. A description of some item in functional vocabulary can be added to a non-functional description without giving cause to believe that the item has acquired properties that somehow go beyond what is natural in senses corresponding to N1, N2 or even N4. A significant number of philosophers have come to believe that mental states are simply functional states, which can thus be shown to be unproblematically identical to physical states, once they are described in the vocabulary of the physical sciences.11 Irrespective of whether the idea really works for mental states,12 the distinctive normative and evaluative vocabulary of morals seems to be an analogous addition to all the terminologies used by the sciences. Moreover, the systematic dependence of ethical judgements on justification by reference to descriptive facts seems to parallel the systematic dependence of the mental on the physical. We have what in both cases may be termed a supervenience relation13. 7. The second feature of the functionalist analysis of mind that has appeared attractive in meta-ethics is the definitional holism of the mental. According to its self-understanding, functionalism differs from its behaviourist predecessor in insisting that mental terms are not definable merely in terms of input-output relations.14 Rather, a theory of how inputs come to produce outputs in a law-like manner has to reckon with complex causal interrelations between the items represented by the intervening variables, relations that are in turn not reducible to their relations to behavioural input and output. 11

There is actually disagreement among functionalists as to whether the doctrine supports or undermines the identity theory. Lewis and Armstrong have argued for the former, Putnam and Fodor for the latter claim (cf. Block 1980, pp. 177ff.). 12 In 4.10-12 below, I provide indications of why the “role” that would have to be played by neurophysiological items cannot in the case of our practical attitudes be circumscribed in purely causal terms. The arguments are developed in detail in Roughley (forthcoming a) chapter 3. 13 Cf. Hare 1952, pp. 80f., p. 131; Blackburn 1973, pp. 114ff.; Smith 1994, pp. 21ff.; Jackson 1998, pp. 118ff. 14 As a matter of historical fact, this is incorrect with respect to behaviourism. See Roughley (forthcoming a) chapter 3.

Naturalism and Expressivism

59

Definition of the theoretical constructs in the black box of the mind thus requires reference to other such constructs. Again, there looks to be a promising analogy with the ethical case. There is a fairly widespread consensus that the project of defining individual ethical concepts by adducing necessary and sufficient conditions whose description contains no reference to ethical terms is a lost cause. So perhaps something analogous to a functionalist philosophy of the mental is possible, if rules for entry into, and exit from the ethical can be specified and combined with interdefinition of the ethical concepts themselves. 8. What has been called “moral functionalism” suggests an analysis of the ethical, including the normative, which adopts these latter two features of the functionalist analysis whilst making clear that the relevant “functional” relations cannot be causal in character (Jackson/Pettit 1995, p. 25; Jackson 1998, p. 131; cf. Smith 1983, pp. 44ff.). This idea raises an obvious question: if the notion of “function” is not to be cashed in in causal terms, then what is it a placeholder for? Surprisingly, the “moral functionalists” have virtually nothing to say in answer to this question. The reply that the relevant kind of relation is a matter of “which properties typically go together” (Jackson 1981, p. 131) can hardly count as a clarification. As it turns out, where the so-called “output” clauses concern what “accompanies” what, that connection is at least in some cases a causal matter – as in the “platitude” that certain actions tend to be performed by agents with certain moral attitudes. However, what is decisive for the question of moral normativity is the relation specified by the “input clauses”. It is after all the question of the transition from the descriptive to the ethical, including the normative, that has been primarily at stake here since Hume. The answer tendered by Frank Jackson is disappointing, to say the least: the input clauses, he says, “tell us what kind of situations described in descriptive (sic), non-moral terms warrant kinds of description in ethical terms” (Jackson 1981, p. 130)15. Certainly, we get from “is” to “ought” by means of “is” descriptions warranting “ought” claims: Jackson has here faithfully reproduced a “platitude” of everyday thinking about these things. The holistic underwriting of interdefinition

15

My emphasis.

60

Neil Roughley

allows us to partly define both “warranting” as what takes us from “is” to “ought” and “the right” as wherever we are taken from “is” claims by the “warranting” relation. Add to this the idea that this network of terms refers to a layer of properties (globally and necessarily) supervening on the nonethical and it follows that there will be non-ethical properties that constitute the reality picked out by our ethical vocabulary (Jackson 1981, p. 123), just as there are presumably neurological properties that constitute the reality picked out by our mental vocabulary. 9. “Moral functionalism” is a strange animal, not least because the conception of definition in terms of causal roles from which it takes its name is simply dropped. What remain are a holism of the concepts up for explication and the supervenience premise. But once we drop the idea that it is causal relations that permit interdefinition of the concepts, the notion of “function” at work seems to be no more than that of the discursive contiguity of the relevant terms. But the fact that we tend to use these in a systematic manner relative to the descriptions in virtue of which they are applied, or at least to believe that we ought to, is no guarantee that they are systematically used to ascribe properties. Where “supervenience” is taken to refer to a relation between properties, the articulation of this feature of our moral discourse in terms of supervenience can obscure this point.16 Further, where “function” means “causal role”, we are clearly talking description; where, however, “function” means no more than “characteristic terminological environment in discourse of the relevant kind”, this entails nothing about the real relations picked out by the co-occurring terms. 10. If we want to understand moral normativity, we need answers to the two central questions raised by Hume in this connection: what constitutes the ought relation and what constitutes the relation of warranting that takes us there from descriptive reasoning (Hume 1739-40, III, i, 1). To say that 16

Note, however, that Hare characterises “good”, “right” and “ought” as supervenient “epithets” (Hare 1952, p. 131, pp. 153ff.; my emphasis). Blackburn uses both the property idiom (Blackburn 1973, pp. 114ff.) and expresses the point in terms of supervening “claims” (Blackburn 1985, p. 137).

Naturalism and Expressivism

61

they hang together in everyday discourse is hardly to provide answers. Only where we have an answer to how they hang together will we know what sort of properties we can reasonably expect to correlate with them on other “levels” of reality. Above all, only then will we be able to judge if the systematic relationship between “ought” judgements and the descriptions that mark their grounds results from their picking out supervenient properties relative to the properties ascribed in the descriptions. Should it turn out that “the right” is not a property at all, then the only correlated descriptions available may turn out to be ascriptions of neurophysiological properties to the person making the normative judgement. 11. This happens to be what I think is the case. “Moral functionalism” is, it seems to me, on the wrong track from the start because it begins with the assumption that “rightness” is a theoretical term. In what follows, I hope to show why this assumption is mistaken.17 I hope to show that the notion of “rightness” we operate with in moral contexts is best analysed not by listing the moral platitudes in which it occurs and in which neighbouring terms occur. Rather, we can shed considerably more light on the matter if we first analyse certain non-moral normative judgements, before going on to inquire what they have in common with, and in what ways they differ from judgements of rightness in moral contexts.

3. Social Standards 1. The central factor common to all judgements of “rightness” is the notion of conformity with a standard.18 This is the case whether we are dealing with standards for the production of norm-sized copy paper, standards for the qualification for a sports competition, rules for behaviour at the table or with moral norms. If a piece of behaviour is located within the relevant field, then 17

There is a related misconstrual at work in the functionalist philosophy of mind. Cf. above note 12. 18 This point has been emphasized in their different ways by Hare (1952), pp. 111ff. and Copp (1995), pp. 19ff.

62

Neil Roughley

it is open to judgement as to whether it satisfies the relevant standards or not. Where it does so, we can say that the agent behaves “correctly”, “rightly” or as she “ought” to have. 2. If this much is accepted as a starting point, the key question to be answered by an analysis of moral normativity concerns the metaphysical status of standards. As far as I can see, there are basically two possibilities. I shall label these the “symmetrical” and the “asymmetrical model”. In the symmetrical model, standards are primitive entities with a status in the practical sphere analogous to that of propositions in the theoretical sphere (Castañeda 1975, pp. 32ff., pp. 169ff., pp. 280ff.; Copp 2001, p. 21). The great advantage of such a model is that it allows us to make (fairly) easy sense of what appear to be truth claims about the moral rightness or wrongness of some action. It is part of our everyday moral practice for people to claim that the way someone acted was right, and for others to contest that claim. The symmetrical model reconstructs such disputes in terms of claims and counterclaims as to whether the action realises a legitimate standard. This is seen as parallel to non-normative factual disputes, in which the assertion that some particular instantiates a property is a claim for the truth of the relevant proposition. Such a conception is to be contrasted with the asymmetrical model, in which standards are conceived as entities consisting of propositions plus something else. On such a construal, there is no parallel between standards and propositions, because the former incorporate the latter (whereas the converse is not true). 3. The nature of propositions is a notoriously controversial matter. For the purposes of this paper, I am simply going to assume that they exist. However, although their controversial nature should not lead us to reject their existence, I do think it should cause us to be highly reluctant to postulate further entities with the same kind of status. This is a first reason to be sceptical about the symmetrical model. As it stands, it does not carry much weight, being merely a consideration of parsimony that will have to be weighed against other considerations. It is however supported by a far weightier reason.

Naturalism and Expressivism

63

4. This grounds in the fact that, unlike propositions, standards appear to be entities that can have a highly familiar place in everyday life. Think of the various non-moral norms and rules mentioned above (in 3.1). These are standards that not only have a clear place in human social life, but whose sociological objectivity is uncontroversially the result of a genesis through social processes: the norm specifying the dimensions of A4 paper and the offside rule in football exist because social agents have created them. 5. This genetic fact provides the key to grasping what it is these standards consist in. In many cases, of course, there are written regulations, but the pieces of paper are not the standards themselves. The core of such social norms lies in the attitudes taken on by the relevant agents. Let me illustrate: a group of people stranded on a desert island agree to regulate certain everyday processes by means of a mechanism of turn-taking, to take it in turns, say, to hunt the evening meal. Such an agreement institutes a rule within the group. If it turns out that certain members are reluctant to keep the agreement, the group might decide to reinforce it by means of penalties to be imposed under circumstances of its contravention. Depending on the characters of the persons involved, the breaking of the such rules may also trigger feelings of anger, indignation and shame. However, it is perfectly conceivable that some or all of the agents are extremely easy-going, and therefore respond to the contravention of the rules by simply completing the task themselves. Similarly, there is no a priori reason why the rule might not do its coordinating job without requiring any kind of punitive reinforcement. 6. What then is the “mode of existence” of the turn-taking rule? The obvious answer is: the rule is constituted by the coordinated attitudinising of the group members. The founding agreement could have all sorts of complications built into it concerning what is to happen should one of the parties opt out, and other eventualities. But let’s imagine the agents have not had time to consider any such possibilities. All they need is an attitude with a content of the form “Under circumstances c, action a is performed by the person who last performed it n days ago”. Obviously, the attitude with that content cannot be just any kind of attitude. More precisely, a doxastic attitude is not going to do the trick. Believing that the specified agents do in fact

64

Neil Roughley

perform the action in the specified sequence is not going to help establishing the sequence. Moreover, short of collective hypnosis, the belief is not going to arise if the sequence has not already been established, either in practice or at least by the agreement itself. That agreement has to involve the agents taking on an attitude by means of which some such content (“p”) is framed in the attitudinal mode expressible by the utterance “Let it be the case that p”. 7. Coordinated attitudes of this kind are the core of social standards, norms and rules. This point is of course no more than the starting point of their analysis. Social standards vary according to many additional parameters. These concern matters such as the means by which attitudinal coordination is established; the distribution of the relevant attitudes among all or only a few; their distribution relative to the distribution of the addressee role; the question as to how much detail is attitudinally encapsulated and how much content can, or needs to be assigned to some other storage medium; and the question of whether, and if so how, the standards are to be enforced. What I am claiming is that these are all elaborations of the core datum, which is an essentially attitudinal matter. Note that, if the desert island community decides from the start to fit its norms out with sanctions, it can only do this by taking on a second-order attitude of the same sort. This has the basic form: “Let it be the case that: if (we attitudinise: let it be the case that: if c then a) & (c but ¬a) then the agent responsible for ¬a be subject to sanction s”.19 8. In what follows I shall, following Kenny and Carnap (Kenny 1963, p. 221; Carnap 1963, p. 1001; cf. Seebass 1993, p. 71), refer to the “let-it-bethe-case” component as the “optative mode” and to those mental states that are at least in part constituted by optative framing of a representational content as optative attitudes. Social standards are the products of beings capable of taking on, and coordinating optative attitudes. If this is correct, then we have very good reason not to elevate the concept of a standard to the same kind of metaphysical status occupied by that of propositions. Indeed, the asymmetry between the two notions is now obvious: talk of social standards 19

In other words, the installation of sanctions is itself a matter of establishing second-order norms that specify what is to be done when first-order norms are contravened.

Naturalism and Expressivism

65

requires talk of propositions, these being the characteristic content of the attitudes that constitute those standards. In contrast, talk of propositions is independent of talk of social standards. Finally, if social standards are indeed artefacts generated by coordinated optative attitudinising, that ought to make us suspicious of the idea that there are action-guiding standards of one special type, namely of the moral type, that are not the product of some such genesis. 9. With the exception of Hobbesian contractarians and other moral relativists,20 people tend to shy away from taking this suspicion too seriously, because they believe it commits them to seeing moral norms as social standards. And, whereas no-one is tempted by the idea that the laws of some particular land or the rules of football are timeless and independent of what agents at a particular time think, the idea of “supra-positive validity” seems essential to at least some, certainly to the most important, moral claims. I shall be arguing that people are right to shy away from the conclusion that moral norms are social norms. Moral norms are nevertheless like social norms in being constituted by optative attitudinising. The decisive question is precisely how.

4. Attitudinally Internal Standards 1. Before advancing my proposal for an understanding of moral normativity, I need to say something more about the nature of optative attitudinising. Am I postulating some strange kind of essentially normative faculty possessed exclusively by humans? And if that is supposed to be the solution, then, someone might object,21 the solution is going to fare no better than that of the intuitionists, who also want to fit us out with some special faculty in order to make sense of our peculiar experience of the right, the good and the beautiful. Now, there is a sense in which our optative capacity is indeed 20

Authors who accept the normatively relativistic consequences of the meta-ethical claim that moral norms are social norms are Harman (1975) and Stemmer (2000). 21 Christoph Fehige has repeatedly done so.

66

Neil Roughley

somewhat puzzling. However, this is far removed from the mysteries of intuitionism. For one thing, the optative conception will not lead us to postulate non-empirical properties to make sense of morality. For another, optative attitudinising is not restricted to the moral sphere, nor indeed, to the normative in general. On the contrary, as I see it, optative attitudinising is something we are more or less continually involved in. 2. Mentally representing some propositional content in the optative mode is representing it in such a way that its non-realisation will, from the point of view of the attitude’s bearer, brand what is real as in this respect dissatisfactory. This contrasts with the assertoric mode, which frames the content represented in such a way that its non-realisation will brand the representational content as dissatisfactory. There is indeed something puzzling here, and it seems to me to be extremely important to get clear on what that is. It is this: the two basic modes of propositional attitudinising cannot be analysed without making – implicit or explicit – reference to the idea of a standard. Moreover, this point is true of both kinds of attitude. Various discussions of assertoric attitudinising bring this out in different ways, for instance Bernard Williams’ remark that beliefs “aim at truth” (Williams 1970, p. 136) or Donald Davidson’s claim that being a believer essentially involves having the concept of a mistake (Davidson 1982, p. 104). Belief, at least in the full-blooded sense, is not merely a matter of having certain propositions pass before one’s mind’s eye. Rather, it involves their affirmation as meeting the standard that is constitutively brought to bear in believing, namely that of representing the way things are. 3. If this is right, then the idea of a standard is internal to our most basic propositional attitudinising. Where we are dealing with beliefs, the attitudinally internal standard is fixed. A believer takes it as given that it’s not being met means that there is something “wrong” with the relevant attitude’s content. There is an important sense, then, in which merely believing involves taking ones mental representations to be subject to a standard. Once one sees this, then it becomes easier to see that optative attitudinising also involves a normative component. Where someone takes on an attitude of the type expressible by the utterance “Let p be the case”, she is operating

Naturalism and Expressivism

67

with a standard in the converse way to that involved in belief. Representing a content in the optative mode is precisely setting that content as a standard against which to measure reality. 4. Such standards are merely subjective. As such, they operate well below the level of moral, or indeed any other form of transsubjective normativity. Indeed, they may be of so little importance, even to their bearer, that she abandons them almost immediately after their adoption. Optative attitudes of such short-lived nature and such minimal integration in a person’s life, for instance that there should be parsley on the moon, are often called “whims”. Other optatives, such as biologically generated basic wants, individual longterm goals and spontaneous, but hedonically charged desires tend to have a much stronger hold on their bearers. Note, though, that the differences between these cases all derive from additional components beyond the common optative core. The normativity involved in optative attitudinising itself is not responsible, for instance, if a person feels “bound” to realise an attitude’s content. All the optative component specifies is that, where its content does not represent the way things really are, there is, from the perspective of the attitude’s bearer, something “wrong” with the world with respect to that content. The necessarily “problematic” nature of unrealised optatives is the root of the Buddhist counsel to minimise one’s desires. It is also the key to the marketing strategies of a consumer society whose dynamics depend on the “demand” created through the generation of new wants. 5. The strategy that I think bears the greatest promise for the attempt to reconstruct norms, whether moral or social, involves getting clear on this subjectively normative dimension that necessarily arises for all human attitudinisers. The strategy is promising because the evidence for this component is independent of the analysis of norms. But, one might wonder, how “naturalist” can such a strategy claim to be? To begin with, if the foregoing is correct, optative attitudinising is something we all do empirically: the optative component is simply the feature that is common to our desires, longings, intentions and hopes and that distinguishes them all from our beliefs. Moreover, there is no reason to assume that this feature of these attitudes need have a history of becoming that is not covered by evolutionary

68

Neil Roughley

processes. As far as the ontogenesis of mind is concerned, it seems likely that full-blown attitudinising develops out of rudimentary capacities of neonates that involve no such normative dimension and that it only does so as a result of their interaction with full-blown attitudinisers. Furthermore, there is evidence that some of our evolutionarily nearest relatives have the concept of a mistaken belief,22 which would mean that they, like us, would also merit the title of standard-bearers. 6. There thus need be no conflict between the optative conception and N1. “Naturalist” worries are more likely to concern the relationship of such a conception to natural scientific procedures. In particular, they are likely to focus on the idea of taking certain criteria as a standard, whether for the content of one’s attitudes or for the way the world is. The worry is that this idea only has a place as a feature of a subjective perspective: it is only from the perspective of an attitudiniser that the non-realisation of the content of an optative attitude is necessarily problematic. Put another way: the attitude expressed by the utterance “Let it be the case that p” needs an author or at least a bearer, whose mental stance on the matter of p is thus articulated. But natural science sees its task as one of decomposing the subject in order to discover the laws, or law-like regularities, which constitute its functioning. There can, so it seems, be no place for subjective perspectives within a scientific view of the world. This is, of course, the primary reason why there have been so many attempts to show that, in what is taken to be the most important sense of “exist”, qualia don’t. 7. We should not forget that a great deal of empirical work in motivational and developmental psychology involves investigating the effects of subjects’ optative stands. For instance, there is an entire research programme that focuses on the effects of persons forming so-called “implementation intentions”, that is, intentions to perform specific actions under specified conditions. Forming such concrete conditional intentions apparently increases 22

Cf. Savage-Rumbaugh (1986), p. 193; Savage-Rumbaugh/Lewin (1994), p. 82. As I am using the expression here, “having the concept of x” does not entail being able to formulate explicit semantic rules for the use of “x” or even being able to use the term “x” or an equivalent.

Naturalism and Expressivism

69

the probability of the agent performing the relevant action, decreases reaction times within certain experimental designs and, where the intentions are furnished with appropriate content, leads to the suppression of epistemic stereotyping effects as well as of emotional reactions. The evidence is that getting people to commit themselves by taking a specific kind of optative stand on some future practical matter leads to effects measurable in milliseconds or in terms of EEG frequencies.23 8. The mental stand taken in optative attitudinising, for instance in deciding, is thus an operationalisable datum for empirical research within experimental designs geared to produce quantifiable results. This is an indication that there need be no incompatibility between the optative conception and that of the sciences. One can thus consistently be a naturalist of type N2 and accept the optative conception. However, if it is stipulated that “compatibility” with science is only given where we have definability in terms of the concepts employed in science, that is, if one takes the step from N2 to N4, then the optative analysis clearly does not qualify. Although motivational psychologists can, indeed must, work with concepts grounded in subjective perspectives – there is no other way to get an agent to form an intention with a specific content – neither the notion of a subjective perspective nor a fortiori that of its linguistic expression are likely to be definable in anything like the terms of physics or biology. The important philosophical question is whether there is any good reason to reject as illicit an analysis that presupposes a subjective perspective. An affirmative answer must depend on being able to say everything we need to say about the relevant field of investigation without making that presupposition. It ought to be clear that this is not only not the case in ethics; it is out of the question in any sphere of our practical lives. 9. It may well be the case that, whenever someone adopts an attitude best expressible in optative terms, one of a complicated disjunction of interacting neural networks is firing. There are well-known reasons why some people would like to say that this complex disjunction of conjoined event patterns would be what optative attitudinising “really is”, presumably at least so long 23

The relevant experiments are summarised in Achtziger/Gollwitzer (forthcoming).

70

Neil Roughley

as the rest of the neural and other biological stuff is there within which the relevant networks have to be embedded for the sequence to generate a subjective state thus expressible. Of course, we have no idea at the moment what this all might be, but that is not the point. What is essential is that the criterion for whether these event patterns and their surrounding states are the right ones remains the expressibility in optative terms of the subjective state thus generated. 10. It reflects a lack of genuine attention to the “folk’s” use of concepts such as “wanting” to suggest that the criterion of optative expression is dispensable in favour of the scientifically respectable idea of causal role. Our main use of these notions simply does not involve the application of an explanatory “theory”. Of course, as Hume pointed out (Hume 1739-40, II, iii, 1), it is important that we can make successful assumptions about what people are going to do on the basis of what we know about their motives. But that doesn’t make prediction and explanation of behaviour the primary function of our psychological idioms (pace Lewis 1972, pp. 256f.). It’s certainly not what we are doing when we express our own “desires”. Rather, we need the language of concerns, wants and aims first and foremost in order to facilitate and represent forms of practical deliberation. We generally think about our concerns, clarify and mull over what is bothering us and relate these things to our overarching aims not in order to discover what we are going to do or explain why we have done what we have done. Reflective episodes in which we take on such a distanced theoretical perspective towards ourselves are the exception and require specific explanations. What we are normally doing in such episodes of reflection in terms of our optative attitudes is wondering what to do. As the folk all know, this is not at all the same as wondering what you are going to do. 11. The kind of thought in which agents engage in order to answer the first-person practical question “What shall I do?” constitutively requires an optative answer.24 For the deliberator, no amount of scientific or other descriptive information can substitute for the mental step expressed by 24

On the notion of “the practical question”, see Hare (1963), pp. 54ff.; Tugendhat (1979), pp. 182ff.; Williams (1985), pp. 18f. Strictly speaking, there are also derivative second- and third-person practical questions. I can ask what you or she “should” do – where “should”

Naturalism and Expressivism

71

the words “Let me a”. This point would be valid even if neuroscience had managed to develop an exact model of what is going on neurally during our optative moments. It is only because we are bearers of the subjective states expressible in optative terms that there are such things as practical deliberation and decision. It would be very surprising if there were not to be neurological and other explanatory stories to be told about how these subjective states come into being. The essential point for our purposes though is that, where agents are fitted out with them, they create for their bearers a relation to the world that only makes sense for a mind entertaining the same sort of modal relation to the world. 12. To summarise: Taking on the practical perspective on one’s immediate or distant future involves taking an attitudinal stand of the form “Let me a”, or at least working towards such a stand in order to answer the practical question as to what to do. Optative attitudinising opens up a dimension of human existence that is only accessible to subjects fitted out with this attitudinal capacity. When we are making decisions or asking for advice, we are thinking in ways that have an optative component which, at least for our purposes as agents, is obviously irreducible. If it should turn out that it is possible to map optative occurrences exactly onto neurological events from the “absolute” perspective of science,25 that would be irrelevant from the agentially internal perspective of the optative attitudiniser. 13. Let us now return to the conception of social norms proposed in the previous section, which I claimed are constituted by the coordinated optative attitudinising of relevant participants in the social practice in question. Two points now need to be made about social norms, before I can sketch a proposal as to how we should understand moral normativity. These two points concern firstly, a specification of the internal structure of those optative attitudes that can count as norm-constitutive and secondly, the application of the distinction between optatively internal and external perspectives to the case of social norms. is the indirect rendering of the optative mode – and appropriately propose an optative answer in the form of advice. Like decisions, pieces of advice are neither proffered for explanatory purposes nor typically employed in order to cause behaviour. 25 On “the absolute perspective” of the natural sciences, cf. Williams (1985) p. 111, pp. 139f.

72

Neil Roughley

14. On the first point: I have claimed that optative attitudinising involves the setting of subjective standards for the realisation of states of affairs. Here, talk of “standards” means no more than that the realisation of some representational content is taken as “correct” and its non-realisation as marking something “wrong” with the world.26 Because the notion of “correctness” at work here is purely subjective, what I have been loosely terming the “coordination” of the optative attitudes of members of a social group is insufficient to constitute social norms. Norm-constitutive optatives need to take on an impersonal character. This is done by excising all proper names and definite descriptions, employing only structures that incorporate the universal quantifier and variables. This gives us attitudes of the form: “Let it be the case that: all agents of type g perform (or refrain from performing) actions of type a under conditions of type c”. Coordinated optatives with this structure constitute behavioural “standards” in the everyday sense of the term, according to which the criteria of correctness are not only accepted by more than one agent, but also avoid any specifications in terms of particular individuals. 15. The second point to be made before we move on to morality concerns the importance of the distinction between the practical perspective of the optative attitudiniser and the external theoretical perspective. This distinction is vitally important for an understanding of the functioning of social norms. Above all, it allows us to see what I shall argue is crucial difference, that between being a norm and being normative. 16. Someone standing outside a society (an ethnographer) or some particular practice within that society (a sports sociologist) can gather data that enables her to say what coordinated optative attitudes characterise the participants. This enables her to describe the norms or standards in question in sentences such as “In society s it is legally forbidden to have sex before marriage” or “In football, it is forbidden for everyone but the goalkeeper to handle the ball within the field of play”. These are claims that can be true or false according 26

We acknowledge the role of such subjective standards in everyday English in turns of phrase such as “That’s all right (for me)” and “Things went wrong (for her)”.

Naturalism and Expressivism

73

as to whether these proscriptions are indeed laid down by the relevant human powers-that-be. Thus, their truth-value can alter with changes in the relevant attitudes. 17. The ethnographical or sociological stance is merely a systematisation of a perspective that participants in any practice can adopt by “taking a step back from” that practice. The mental move thus designated is the relinquishing, or at least ignoring, of one’s optative relationship to the contents of the relevant standards. Instead, one focuses epistemically on what normconstitutive optatives are played host to by relevant agents. The role of this attitudinal shift is decisive for our topic. In as far someone has such a purely epistemic relationship to existing social norms, these norms are grasped only as social facts devoid of normativity. In other words, the normativity of social norms for an agent arises only in so far as she is optatively engaged with them. That some content is of normative import means that it is framed in the “to-be-the-case” mode. But no matter how many other people take on attitudes according to which an agent “should” a, as long as we only register the fact of the others’ coordinated attitudinising, the relevant norm remains devoid of normativity. The normativity of a norm is dependent on its content recurring in the content of a concurrent optative stand. 18. It is because people generally concur optatively with the attitudinal stands constitutive of the norms of their practical environments that there is no “puzzle of social normativity” analogous to the puzzle of moral normativity enshrined in the Kantian picture (2.4). The causal efficacy of social norms results from the fact that people tend to “want” to realise the contents they prescribe. This is because a central part of socialisation into a human culture or sub-culture is what is often called the “internalisation” of social norms, that is, the adoption of optative attitudes concurrent with those of our educators. Such processes prepare the ground so that the impersonal optatives of social standards, rules, laws and conventions can be taken by social agents as supporting their answers to their own practical questions. It is because we know this – and not only because we know that people try to avoid sanctions – that we can frequently predict people’s behaviour on the basis of our knowledge of social norms.

74

Neil Roughley

5. Moral Standards 1. Social norms, I have been suggesting, are constituted by specific constellations of coordinated optative attitudinising. This suggestion provides a clear sense in which social standards are real: “cognitivism” about social norms is hardly likely to be a disputed position. However, if what I have just been claiming is correct, the normativity of social norms is another matter. If I merely recognise that some such optative constellation is a contingent, or even essential feature of some practice or of a certain society, that by no means commits me to judgements to the effect that agents ought to act in accordance with those standards. This holds even if I happen to be a member of the relevant society and I am making judgements about what is normatively binding for myself. That commitment only ensues where I am optatively aligned with the relevant optative constellation. 2. There are interesting and complicated questions about when precisely that optative alignment can be said to be given relative to particular kinds of social practice. A comparatively simple case is provided by a participant in some clearly circumscribed sport. Someone who opts to play a game of football thereby accepts the constitutive rules of the game. You cannot play the game whilst rejecting the rule that handling the ball constitutes an infringement. There is a decisive sense in which your running around the pitch under these circumstances would not even amount to playing football. You enter into a normative relationship to the football rules by taking the optative stand you need to take in order to enter into the practice in the first place. 3. Things are not so simple for the citizen of a country with a certain legal system. Accepting the idea of a legal system does not commit anyone to acceptance of all the laws that particular countries happen to have developed. Nevertheless, there is also a significant sense in which opting to live within the constraints of a specific legal system involves, at least pro tanto, buying into a positive optative relationship to the laws of the land.27 Importantly, 27

This is obviously highly schematic. Certainly one would need to differentiate here. One might wonder to what extent acceptance of the penal law involves commitments to the statutes of civil law. There also is an important question as to when precisely someone is to be seen as having taken this step. If I decide to take on the citizenship of some

Naturalism and Expressivism

75

this is a matter of their acceptance as legal standards and does not entail anything about a person’s moral perspective on those standards. Where, for instance, individuals or ideological groups reject particular laws of their country, such rejections are frequently moral rejections. This is then a matter of the relativisation of one set of standards in the light of another set. This, however, need by no means involve the dissolution of the normativity of the relativised standard. There is a difference between protesting against a particular law, for instance by demonstratively breaking it on some occasion, and living systematically in such a way that you simply ignore the collective demand that constitutes it. Only in the latter case has a person retreated from the optative relationship that endows the law with normativity for her. 4. Imagine that there is some quasi-religious practice that involves the contravention of a law of some country and that there are sub-cultural groups which nevertheless continue to practice it. A, who is considering joining one of these groups, has asked B for advice. One thing B might do is point out to A that taking this step would involve him subjecting himself to normative demands that conflict with those of the country’s legal system. Independently of any moral considerations, he would be creating a problem for himself by taking an optative step that has local normative consequences that conflict with the normative demands of the more extensive legal system he has also bought into optatively. 5. The question of moral normativity is raised in particularly stark terms where someone relativises legal norms by appeal to moral standards. It is particularly in such cases that the puzzle of moral normativity arises. Moral norms are at least sometimes causally efficacious. Yet this seems not to depend on them being constituted by coordinated optative attitudinising. Indeed, we tend to think that what is morally right is independent of what any individuals think, however many of them there may be. Plato’s and Kant’s parallel universes and Moore’s sphere of non-natural properties acquire a certain appeal precisely because the only plausible reconstruction of social country other than that of my native land, then in doing so I accept the legal system of that country. This is a very good reason not to seek the citizenship of countries with laws one rejects on moral grounds. Clearly though, the standard situation of someone born into a particular legal system cannot be analysed in exactly the same way.

76

Neil Roughley

norms seems obviously inapplicable to the moral sphere. On the other hand, the fundamental mystery on offer here is massively unattractive, not least so because of its incompatibility with even the broadest forms of naturalism encoded in N1 and N2. This is the dilemma faced by any attempt to provide an explanation of moral normativity and thus to dissolve the puzzle it poses. 6. This seems to me to be the point at which the appeal of meta-ethical expressivism becomes manifest. What I call optative expressivism28 accepts the following three points: firstly, the claim that certain kinds of behaviour in far-off places ought morally not to take place is a paradigmatic case of moral judgement. Secondly, for that judgement to be coherent, there need be no structures of coordinated optative attitudinising in place that involve both the author of the judgement and its addressees. Thirdly, the only analytic option by means of which we can make sense of the idea of practical standards sees them as generated through optative attitudinising. 7. It follows from the second point that, in paradigmatic cases of the kind referred to in the first, the cognitive dimension of judgements about social rights and wrongs is not available. There is nothing in the world about which I am making an assertion when I morally condemn certain actions in Africa or the Middle East. The moral standard contravened by the relevant forms of behaviour is – unfortunately – not “out there” at all. Rather, moral standards are simply set by the optative attitudinising of the persons who make moral judgements. Strictly speaking, there is no such thing as a moral norm. There is only the moral normativity that emanates from the stands of the individual bearers of moral standards. 8. Obviously, moral normativity is not simply generated by means of any kind of optative stand. Everyday desires, longings and intentions need have nothing whatsoever to do with morality. Rather, moral standards appear to be characterised by a form of “objectivity” that contrasts with the subjectivity of the standards set by mere wants. Moreover, the transcendence of the subjectivity of their bearers actually seems to go beyond the

28

The following brief sketch is based on Roughley (forthcoming b).

Naturalism and Expressivism

77

intersubjectivity of social norms. Thus, although the internal features that distinguish the norm-constitutive optatives in the case of social norms, namely their impersonal formulation (4.14), are also necessary here, they are obviously not in themselves sufficient to make the relevant difference. An adequate explication of moral normativity should make it clear how the claim of the relevant standards on agents exceeds the mere intersubjectivity of social norms. And surely, it might plausibly be argued, this is exactly the opposite movement to their localisation in the individual optatives of the bearers of moral judgements. 9. An explication of moral normativity has indeed to clarify the sense in which the claims being made when we take moral stands transcend our own subjective perspectives. This can be labelled “the objectivist challenge”. Moreover, it also has to clarify a further feature of moral normativity that is naturally seen as a correlate of the belief that moral norms are more objective than social norms. This is the idea that they are also in some sense more “real” than the norms of some social constellation. What is morally wrong is, in some important way that is hard to pin down, unavoidably so. We have a strong sense that moral “oughts” are forced on us by features of the way things are – in a way that may feel similar to the unavoidability of snow in the Swiss Alps. Call this “the realist challenge”. It is essential to the expressivist case to meet both challenges. This is best done by showing that, contrary to appearances, the two challenges are rooted in two different features of the optatives constitutive of moral normativity. That is, the expressivist strategy here is one of divide-and-rule. 10. The objectivist challenge, I want to suggest, is met by the following move: as there is no standard out there in the world against which a moral claim can be measuring actions, the transsubjective feature at work also has to be optative in character. My suggestion is that the relevant feature is a higher-order optative concerning the attitudes of all other persons who consider the matter at hand: making a moral judgement involves taking a stand according to which all those who consider what optative stand to take on someone’s a-ing under certain circumstances should adopt the same stand as oneself. In other words, a normative moral stand involves at least conjoined optative attitudes of the following form:

78

Neil Roughley

(1) Let it be the case that: all agents of type g perform action a, or refrain from a-ing, in circumstances c. & (2) Let it be the case that: all persons who consider whether it should be the case that agents of type g a under circumstances c take the optative stand specified in (1). The “should” in (2) is the indirect rendering of an optative stand. Moral claims thus do have a dimension that can be characterised as transsubjective, whilst not being dependent on the factual intersubjective stands of other attitudinisers. However, we “reach out beyond” our subjectivity in making moral claims not by referring to any particular kinds of objects, or properties of objects, but by making claims on the optatives of other subjects. 11. What I called the realist challenge is, I think, to be met in a different manner. Unlike what “realist” and “objectivist” metaethicists assume, moral normativity is unavoidable in a way that is independent of its transsubjective component. The unavoidability is psychological in character: the optatives constitutive of our morally normative stances are attitudes we cannot help adopting. One way to gloss this is to say that they are attitudes we are disposed to continue to want to uphold. In other words, they are at least the potential, if not the actual objects of further higher-order optative stands of the form: (3) Let it be the case that: I retain the optative stand specified in (1). 12. On my understanding, this makes the objects of morally normative claims propositions that are valued by their bearers (see Roughley (2002), pp. 180ff.).29 Certain authors have argued that there are further dispositions that need to be added here. These tend to include dispositions to react with specific emotions, such as indignation, resentment or guilt when the relevant optatives are flaunted (cf. Gibbard 1990, pp. 41ff.; Copp 1995, pp. 94f.; Blackburn 1998, p. 9). My impression is that such dispositions are only 29

This part of the proposal is consonant with a suggestion by Blackburn (1998), pp. 66ff.

Naturalism and Expressivism

79

typically present where we are dealing with contraventions that are taken to be particularly serious. Here, however, let me simply note that I would be open to such extensions of the strategy sketched here should they turn out to be necessary. 13. Clearly, there are serious objections with which this proposal should be confronted. I attempt to reply to the most important of these elsewhere (see Roughley forthcoming b). However, there is one central objection that I do think needs commenting on here in order to avoid a basic misunderstanding. Someone might accept something like the construction I have sketched, but object that the question of moral normativity is the question of when optative stands of the structure detailed in (1) and (2) and backed by dispositions to optatives along the lines of (3) are justified. What I have called the constitution of normativity would then at most be the constitution of putative normativity. For genuine normative force to be thus generated, we would need to be dealing with warranted optative attitudinising. And if that is the case, then that makes normatively decisive the descriptive properties that justify optative stands. That should, so the objection goes, lead us to conceive normativity as constituted by those descriptive properties that warrant the relevant optative steps. 14. The objector has a point: we should indeed see only those optative stands as generating moral normativity that are justifiably adopted. Moreover, justification here does involve reference to properties that can at least in principle be characterised in purely descriptive terms: such as whether certain actions cause people pain or pleasure, and what their consequences are for the autonomy or self-esteem of those agents they affect. Where the objection goes wrong, however, is in the assumption that characterising an optative stand as justified on the basis of certain descriptions is itself a purely descriptive matter. It is the defining move of expressivism to reject this assumption. 15. Once again the comparison with social norms is instructive. We can describe an action as wrong or a judgement about some action as correct or justified when there is some relevant established social standard, i.e. factual culture of coordinated optative attitudinising. A description in these terms is

80

Neil Roughley

then quite simply a matter of measuring the action or the judgement against that standard. In doing so we are adopting exactly the same kind of procedure as an inspector whose job is to check whether the paper being produced at some factory conforms to A4 norm size, or a meta-inspector checking on whether the inspector has been passing judgements that are guided accurately by that standard. The problem with this model, once again, is that there is in the moral case no ontological niche for a standard against which actions or judgements of actions could be measured. 16. But are we not effectively left with a rejection of the idea of ethical justification, replacing it with whatever may happen to move individuals to take on the relevant kinds of optative attitudes? Surely the contingent facts of what moves individuals in this way do not invest their optative stands with the kind of “authority” we attribute to justified moral judgements.30 That all depends on what kind of authority we think people attribute to justified moral judgements in everyday contexts. That seems to me to be anything but clear.31 The feature that I take to be central is that moral judgements make claims on their addressees, where the type of claim at issue cannot be reconstructed as a demand that one’s addressee simply accept that something is the case. The “authority of truth” – the ultimate warrant for assertoric claims – is quite simply not the kind of authority we are after. Rather, the type of claim in question demands certain things of its addressees. I have suggested that the contents of such demands are, firstly, the performance or non-performance of actions by all agents of a certain type in situations sufficiently similar to that of the addressee and, secondly, the adoption of a stand that itself demands the content of the first demand. 17. It is a further contingent fact about us humans, but decisive for the content of our morality, that we only tend to make such demands on all potential considerers – as opposed to demands on those individuals who happen to be 30

Peter Schaber pressed this point. Unlike some authors, I see very little evidence that there is a philosophically untutored agreement that moral claims describe something in the world. The fact that Mackie’s interpretation of “the apparent authority of ethics” in descriptive terms (Mackie 1977, pp. 32f.) led him to see the ethical as a fictional system tells first and foremost against a descriptivist reading of the kind of “authority” in question. 31

Naturalism and Expressivism

81

standing in our way – where we are ourselves moved by considering certain kinds of properties, in general, properties such as those mentioned above in 5.15. To take some normative moral judgement to be justified is to take a norm-constitutive optative stand on the basis of assertoric attitudes that sway us to take on that stand. And to take as normatively binding a metajudgement that some moral judgement is justified involves adopting exactly the same attitudinal structure. There is nothing more substantial to be had. (What could it be?) 18. In spite of what many seem to feel, seeing this does not force us to conceive morality as a mere subjective construct in the heads of various individuals. On the contrary, the optative structures set out in (1) and (2) provide the basis for the rational discussion of moral claims between attitudinisers prepared to interact optatively. We demand demands – and meta-demands – from each other, supporting our own claims by appeals to descriptions which we hope will sway our interlocutor to adopt congruent stances. Where congruent responses are established, this may increase our confidence in our own optative stands: we may feel more strongly “authorised” to take moral stands. There is, however, no metaphysical safety net we can fall back on to guarantee these stands independent “authority”. There are undoubtedly people who expect morality to provide authority in that sense. But that seems to be primarily a hangover from religious conceptions within which there is an author of the relevant standards whose essence invests him with the right to demand compliance. 19. The human life form, like those of other animals, can only be lived in confrontation with the way things are. There is a great deal of stuff around that we cannot ignore, but moral norms just do not belong to that stuff. Of course, those who want to live an amoral life do well to take note of the kind of behaviour that is typical of those who think morally. Those of us who do so can be led to produce very unpleasant reactions when faced with contraventions of those optative stands constitutive of moral normativity. In interpersonal dealings, these are often forms of emotional behaviour; in the sphere of international relations, these can taken on the form of diplomatic, economic, even military sanctions. The fact that we are often driven

82

Neil Roughley

to adopt such measures is, on the other hand, an indication of the fact that moral normativity is, in itself, often causally impotent. Many people simply do not see the transsubjective optatives of others as providing them with serious candidates for good reasons for action. Nevertheless, normative claims can unfold causal power by the simple and direct means of being raised, furnished with grounds, listened to and then adopted by their addressees. Whether this happens depends on whether the addressees are prepared to see them as answers to their own practical questions. Where this is the case, moral normativity can turn out to be a causal force to be reckoned with.

Conclusion 1. What I called the puzzle of moral normativity, according to which moral norms seem to be located both inside and outside the causal network of empirical reality, is to be explained by moral normativity’s character as an internal feature of our optative attitudinising. In as far as other agents are prepared to see my transsubjective optative stands as potential answers to their own practical questions, they may end up developing congruent opatives and acting accordingly. Where this is not the case, they will carry on regardless. 2. We also have an explanation why, although the stuff of morality is “natural” in the sense of resulting from a causal process of becoming, moral normativity cannot be “naturalised”, where “naturalisation” involves the reduction of claims about rightness to statements about the way things are. Saying that an action is right or wrong obviously does not involve saying anything about the attitudes people have towards such actions. It is, rather, articulating the internal perspective constituted by such a stand. For the same reason, although the empirical “science” of psychology, including neuropsychology, may one day be able to tell us everything we want to know about the mental states of attitudinisers, just as sociology may provide extensive information about the norms they set up, neither have access to the normativity thus generated. It is constitutive of the scientific perspective to approach the world epistemically.

Naturalism and Expressivism

83

Normativity, however, does not only exist because there are agents that can take on an optative relationship to the world. It only exists for such agents and it only exists for them in as far as they are optatively engaged.32

References Achtziger, A./Gollwitzer, P. (forthcoming): „Motivation and Volition in the Course of Action“, in: J. Heckhausen/H. Heckhausen (eds.): Motivation and Action, Cambridge. Ayer, A. J. (1990 [1936]): Language, Truth and Logic, London. Blackburn, S. (1993 [1973]): “Moral Realism”, in: Essays in Quasi-Realism, Oxford, pp. 111-129. – (1985): “Supervenience Revisited”, in: Essays in Quasi-Realism, Oxford, pp. 130-148. – (1998): Ruling Passions. A Theory of Practical Reasoning, Oxford. Block, N. (1980): “Introduction: What is Functionalism?”, in: Readings in Philosophy of Psychology, Cambridge, pp. 171-184. Carnap, R. (1963): “Replies and Systematic Expositions”, in: P. A. Schilpp (ed.): The Philosophy of Rudolf Carnap, La Salle, IL, pp. 859-1013. Castañeda, H. (1975): Thinking and Doing. The Philosophical Foundations of Institutions, Dordrecht. Copp, D. (1995): Morality, Normativity and Society, New York/Oxford. – (2001): “Realist-Expressivism: A Neglected Option for Moral Realism”, in: Social Philosophy and Policy, 18, pp. 1-43.

32

This paper and Roughley (forthcoming b) have both been influenced by the helpful discussion of my talk at the workshop documented by this volume. My thanks to all the participants, as well as to Julius Schaelike for his comments.

84

Neil Roughley

Danto, A. (1967): “Naturalism”, in: P. Edwards (ed.): The Encyclopedia of Philosophy, New York/London, pp. 448-450. Davidson, D. (2001 [1984]): “Rational Animals”, in: Subjective, Intersubjective, Objective, Oxford, pp. 95-105. Gibbard, A. (1990): Wise Choices, Apt Feelings. A Theory of Normative Judgement, Oxford. Hare, R. (1952): The Language of Morals, Oxford. – (1963): Freedom and Reason, Oxford. – (1985): “Ontology in Ethics”, in: T. Honderich (ed.): Morality and Objectivity. A Tribute to J.L. Mackie, London, pp. 39-53. Harman, G. (1975): “Moral Relativism Defended”, in: The Philosophical Review, 84, pp. 3-22. Hume, D. (1978 [1739-40]): A Treatise of Human Nature, Oxford. Jackson, F. (1998): From Metaphysics to Ethics. A Defence of Conceptual Analysis, Oxford. Jackson, F./Pettit, P. (1995): “Moral Functionalism and Moral Motivation”, in: Philosophical Quarterly, 45, pp. 20-40. Kant, I. (1911 [1787]): Kritik der reinen Vernunft, Prussian Academy Edition, vol. III, Berlin. Kenny, A. (1994 [1963]): Action, Emotion and Will, Bristol. Lewis, D. (1972): “Psychophysical and Theoretical Identifications”, in: Australasian Journal of Philosophy, 50, pp. 249-258. Mackie, J.L. (1977): Ethics. Inventing Right and Wrong, London. Moore, G.E. (1993 [1903]): Principia Ethica, Cambridge. – (1993): “Preface to the second edition”, in: Principia ethica, pp. 1-32. Papineau, D. (1993): Philosophical Naturalism, Oxford/Cambridge (MA).

Naturalism and Expressivism

85

Quine, W.V. (1969): “Epistemology Naturalized”, in: Ontological Relativity and Other Essays, New York/London, pp. 69-90. Roughley, N. (2002): “The Uses of Hierarchy: Autonomy and Valuing”, in: Philosophical Explorations, V, pp. 167-185. – (forthcoming a): Wanting and Intending. Elements of a Philosophy of Practical Mind. – (forthcoming b): “Optative Expressivism”. Savage-Rumbaugh, S.(1986): Ape Language, New York. – /R. Lewin: Kanzi. An Ape at the Brink of Human Mind, New York. Seebass, G. (1993): Wollen, Frankfurt am Main. Smith, M. (1994): The Moral Problem, Oxford/Malden (MA). Stemmer, P. (2000): Handeln zugunsten anderer. Eine moralphilosophische Untersuchung, Berlin/New York. Tugendhat, E. (1979): Selbstbewusstsein und Selbstbestimmung. Sprachanalytische Interpretationen, Frankfurt am Main. Williams, B. (1973 [1970]): “Deciding to Believe”, in: Problems of the Self. Philosophical Papers 1956-1972, Cambridge, pp. 136-151. – (1985): Ethics and the Limits of Philosophy, London.

Norbert Anwander Normative Facts: Metaphysical not Conceptual1 Suppose someone tells you that he is able to build an airplane out of cheese. When you ask him how it manages to fly, he looks surprised and claims that flying is a completely different issue. I suspect your initial admiration will give way to puzzlement when he continues to insist that his huge lump of cheese is just a non-flying airplane. Somewhat similar is my reaction when I come across naturalist versions of moral realism. We are assured that, contrary to some rumours, morality is nothing queer for moral properties are just ordinary well-behaved properties. What it is for a person to be generous, for a society to be just, and for an act to be morally right, is just for them to have certain natural properties with which we are perfectly familiar. Just before we get excited or maybe suspicious, we learn that the presence of the relevant natural properties does not mean that there is reason to be generous, that we have reason to aim at a just society, or even that we have reason to do that which is morally right. Questions such as what we have reason to do, or what we ought to do, they tell us, are a completely different issue. Again, most of us are simply puzzled when the naturalist continues to insist that his account is just an externalist moral theory (see Boyd 1988; Brink 1989; Schaber 1997)2. And it is difficult not to feel a bit disappointed since the real challenge, it seems, has always been to account for morality’s normativity.

1

I am grateful to the participants of the conference on Naturalism and Normativity at the University of Zurich, September 2003, for helpful suggestions and challenging questions. Had I done more to accommodate the latter, a rather different paper might have been the result. 2 These authors are externalists about morality in the following sense: They deny that ‘It is morally wrong for S to f’ implies ‘S has reason not to f’. One non-naturalist who holds this position is Shafer-Landau (2003). Advocates of naturalism in epistemology are usually even more explicit about not being in the normative business.

88

Norbert Anwander

That does not mean that we should feel cheated. It is not just their opponents who have pointed out that normativity might be a problem for naturalists. This, for instance, is how James Rachels begins the final section of a piece on ethical naturalism: „The most important objection to ethical naturalism is that it leaves out the normative aspect of ethics” (Rachels 2000, p. 89). And David Copp has almost made a habit of admitting in the first or second paragraph of his numerous articles on ethical naturalism that it has difficulty accounting for the normativity of moral claims.3 In this paper, I am going to argue that on this particular point these naturalists are indeed right: Normativity is beyond their reach. I am suggesting that, for better or worse, if we want to take normativity seriously, we are committed to a view about normative properties and facts according to which they turn out to be nearly as queer as John L. Mackie thought objective values would have to be (see Mackie 1977, pp. 38-42).4 While Mackie was right to claim that his considerations would apply to aesthetic values no less than to moral values (Mackie 1977, p. 15), he was mistaken if he meant to imply that prudential values and epistemic values are less queer. It is indeed often thought that the relevant gap is between epistemic reasons and prudential reasons on the one hand, and moral (and aesthetic) reasons on the other, with the former presumably being less suspicious than the latter. But that does not seem right: Surely, what matters is the divide between the nonnormative and the normative, and what needs to be accounted for are facts about values, reasons, and oughts, regardless of whether they are epistemic, prudential, aesthetic or moral. If naturalists have indeed a problem explaining the normativity of moral judgements, why should they fare any better with regard to the normativity of epistemic judgements and prudential judgements? It might turn out that there are less moral reasons than we thought or 3

Two examples: “I believe that in fact it [sc. synthetic moral naturalism] has difficulty accounting for the normativity of moral claims” (Copp 2000, p. 113). “There are various familiar objections to it, including, most importantly, the objection that naturalism cannot explain the normativity of moral judgement” (Copp 2003, p. 179). 4 The view argued for in this paper is not quite as queer as the one characterized by Mackie because it does not include the claim that knowledge of normative properties and facts necessarily motivates.

Normative Facts: Metaphysical not Conceptual

89

none at all, but that would not be something we could establish by debating naturalism. It would have to be a substantial result of moral inquiry. While it might be true that there is reason to believe that the earth rotates around the sun and reason to care about your own future, but no reason not to torture innocent people for fun, this could not depend on the issue of naturalism. If naturalism is a threat to normativity, it is a global one.5 Actually, I am less interested in naturalism than this might suggest. One of the many confusions that G.E. Moore created was to confuse two quite distinct questions. His central claim that goodness is not a natural property can be read and has been read in either of two ways: i) Non-reductionism: Goodness is not identical with any natural (or supernatural) property (apart, of course, from itself). ii) Non-naturalism: Goodness does not belong to the domain of the natural. These are two separate issues. For instance, you could claim that some fundamental physical property is a natural one but is not identical with any other property, natural or non-natural. Similarly, Moorean goodness could be a natural property without being identical with any other natural property (see Baldwin 1990, pp. 80-100; Sturgeon 2003, pp. 534-538). With regard to Moore’s claim, we can distinguish the following questions: i) Can normative properties be reduced to non-normative properties? ii) If so, do those non-normative properties count as natural? iii) If not, can normative properties nevertheless count as natural? In this paper, I shall concentrate exclusively on the first question. With regard to the second question, this focus is justified because if the answer to the first question is, as I hope to show, No, we can simply forget about the second 5

Jean Hampton similarly argues that “moral objectivism is just an instance of a more general position that might be called normative objectivism” (Hampton 1998, p. 4). – As for moral naturalism, it is interesting to note that those naturalists who start out with realism, but put normativity to the side, tend to be consequentialists, while those naturalists who take normativity seriously tend to be contractarians because they think that the only way morality can be accounted for is via unproblematic prudential reasons. As I indicate in the main text, I find both views deeply mistaken, the former because it leaves out normativity, the second because in terms of metaphysical economy/profligacy all reasons are equal.

90

Norbert Anwander

one. As to the third question, it strikes me as hopelessly ill-defined. The question whether normative properties are natural, it appears, is just as much a question about what counts as natural.6 This becomes obvious when we ask about the significance of the question. The main motivation for enquiring whether the normative is natural, is, I suspect, the assumption that only the natural is real. Depending on our view of the natural, however, this is either just dogmatic or quite vacuous. Alternatively, it might be the question of how inquiry into the normative is to proceed. But again, if for instance, you are prepared to count intuitions as natural, as some naturalists do (see Audi 1997; Copp 2003), it becomes difficult to see what the dispute between the naturalist and the non-naturalist amounts to. Furthermore, reductionism rather than naturalism is what was at stake for Moore. Whatever sense we can make of Principia Ethica’s epigram, “Everything is what it is and not another thing”, the quotation from Bishop Butler is much more plausibly understood as an attack on identifying a normative property with some other property than as an attack on thinking that it belongs to the domain of the natural. The same holds for Moore’s (in)famous Open Question Argument, which in the end may fail to show that goodness is not to be identified with some other natural property (see Baldwin 1990, pp. 86-100), but could not even get started if it was understood as an argument against goodness belonging to the realm of the natural. This paper consists of two parts: In the first and much longer part, I discuss the question of how we should argue for non-reductionism. While my major target is, of course, (naturalist) reductionism, I will also show why the familiar strategy that appeals to the ineliminability of normative concepts is unsatisfactory. It is not just that normative properties and facts can only be captured in normative terms, rather they are metaphysically different – hence the paper’s title. In the second part, I try to articulate this view by looking at several kinds of necessity, of which normativity is one irreducible species.

6

What might be the point of arguing about whether normative properties and facts are natural if there is no agreed account of what it is to be natural? The fact that most authors (seem to feel obliged to) spend quite a lot of space establishing their own account of the natural suggests to me that they are barking up the wrong tree.

Normative Facts: Metaphysical not Conceptual

91

1. Against Reducing the Normative What is the distinction between the normative and that which is not normative? We have a pretty good idea of the contrast: Statements using concepts such as “reason”, “ought”, “justified”, “right” are normative.7 These concepts refer to properties and facts, and we use them with regard to agents and their acts (which covers actions proper, beliefs, and affective states). All other statements are not normative. These include statements about what happens, happened, will happen, or would happen, but also about what some agent does, did, will do, or would do. Since reductionism sometimes gets a bad name, it bears emphasis that there are some perfectly good reasons to look for a reductionist account of the normative. I want to mention but three considerations. The first adverts to a question to which an answer must be given. The second refers to a general consideration why we should always opt for irreducibility only as a last resort. The third refers to something which on some views it would be nice to have. First, there is a consideration specific to the normative, and that is its supervenience on the non-normative. It is almost universally agreed that the normative does not vary independently of the non-normative. While, for instance, the colour of objects can change as their weight stays the same, it is impossible for the way things normatively are to vary while they stay the same non-normatively. This impossibility, that is this dependence of the normative on the non-normative, calls for explanation. Indeed, it is a challenge that any account of normativity must meet. Reductionism can do that rather easily; some even think that a reductionist account follows from supervenience (see Jackson 2003). Second, there is the general consideration to do with what the sociobiologist E.O. Wilson calls consilience, the unification of our knowledge.8 Accounting for one domain in terms of some more basic domain deepens our understanding. An impressive example of such a connection of different 7 8

Terms like „appropriate“ or „merit“ also refer to the normative. For a recent attempt at unification, see Pinker (2002).

92

Norbert Anwander

fields of knowledge is our coherent understanding of life on the levels of evolutionary theory, genetics, and biochemistry. To postulate irreducible entities, on the other hand, is in general not much of an answer to our questions, and we should do so only as a last resort. I take this to be the real thrust of Ockham’s razor: The more independent entities there are, the less understanding we have. It may be true that the only way to account for some phenomenon x is by postulating some irreducible property p(x), but we should admit that in terms of explanation nothing has been gained. Third, there is the point that reductionism is the obvious route to naturalism. If normative facts can be reduced to non-normative facts, and these non-normative facts are all natural, then normativity turns out to be well placed in the natural. We need to get a clearer understanding, however, of what a reduction amounts to. By a reduction, I mean an answer to the question: “What is it for x to be N?” The reduction tells us that being F is what it is for x to be N. Being N is said to consist in being F. Thus, the reductionist who is our target holds that normative properties consist in non-normative properties. The term “consisting in” is deliberately ambiguous between the relations of identity and constitution. For many naturalists, the distinction between identity and constitution is crucial because they think that their position is not a reductive one, as long as they merely claim that the normative is constituted by the non-normative, but deny that the properties in question are identical (see Brink 1989; Shafer-Landau 2003). What matters, however, is not whether we want to count constitution as a kind of reduction, but whether it makes any difference with regard to getting us from the non-normative to the normative. It is hard to see, though, how it could. There are two respects in which the relations differ: First, constitution allows for multiple realizability. For instance, sticking your tongue out, at least in specific contexts, constitutes an offence. But since there are other ways of offending people, we should not say that sticking your tongue out is identical with offending people. It is merely that the former is a way of realizing the latter. Second, constitution does not imply necessity. If A and B are identical, they are necessarily identical. Again, in a different world, sticking your tongue out might not be a way of offending people. Neither feature – multiple realizability and non-necessity – seems even remotely relevant for the issue at stake.

Normative Facts: Metaphysical not Conceptual

93

What matters, though, is that our question is not whether normative concepts can be reduced to non-normative concepts. Many naturalists, particularly those who emphasize the distinction between identity and constitution, would readily admit that this cannot be done. The question we are asking concerns metaphysics: Can normative properties such as being a reason, being what ought to be done, being right, be reduced to non-normative properties such that we could, for instance, say that being right is identical with or constituted by being F? The reason for insisting on metaphysics is this: Suppose some normative concept, say “right”, cannot be reduced, for instance, to the non-normative concept “maximizing net pleasure”, but nevertheless the two concepts refer to the same property. We can then ask whether that property is a normative one or a non-normative one. One option that is excluded is that it is both normative and non-normative since the latter just means that it is not normative. Now consider this: While it could correctly be said that rightness just consists in maximizing net pleasure, that maximizing net pleasure just is what it is for some action to be right, nobody would be tempted to say that maximizing net pleasure consists in rightness, that being right just is what it is for some action to maximize net pleasure. This makes it plausible to think that the property both concepts refer to is a non-normative property. Though there are normative concepts, there are no normative properties and facts. It might be suggested that we should just be happy with concepts and get rid of a misguided desire for “real” properties and facts. But that is beside the point. For what worries me is the asymmetry: It is not as if everything was on a par. In whatever way we want to put it, on this account, maximizing net pleasure is metaphysically deeper than rightness. I cannot help thinking that this undermines the reality of normativity. Against the Pattern-Argument The most popular argument against the possibility of reducing the normative to the non-normative is the Pattern-Problem, also known as the Problem of Shapelessness (see Crisp 1996; Dancy 1995; McDowell 1998; Wiggins 1998). In this section, I want to show why we should not appeal to this argument. The normative is irreducible, according to the Pattern-Argument, because the grouping of items effected by some normative concept makes no

94

Norbert Anwander

sense from a non-normative point of view. For instance, from a non-moral point of view there is no pattern recognizable among all the distributions that can be called just. There is no non-normative property that is shared by all and only those distributions that are just; all they have exclusively in common is that they are just. Contrast this with the successful identification of water as H20: There is this single property of being H20 that is shared by all and only those items that are grouped together as water. If someone was given the task of sorting everything in the world into those things that are water and those that are not, he would not have to be familiar with the concept of water. His mere ability to recognize when something is H20 would guarantee his getting the right result, and this not as a matter of luck, but of necessity. No such thing, it is argued, is feasible with regard to the normative: We could not design a machine operating with only non-normative distinctions that reliably picks out all and only those actions that are just. Since we are not merely talking about some specific moral concepts, but about normativity in general, the task may be even a lot more difficult: Is there some non-normative property such that for all and only those actions that ought to be done and all and only those beliefs that ought to be believed, they and only they necessarily have it? The argument says that the answer is no and concludes that normative concepts are indispensable. The obvious reason why we cannot appeal to this argument is, of course, that it establishes at best the irreducibility of normative concepts. But there are a number of more specific objections to be made against this argument. First, the argument as it stands does not show that the task cannot be done, but simply assumes that there is no property such that all right actions have it. That is, it assumes that, among other monistic candidates, Hedonic Utilitarianism, which tells you that the property which all actions that ought to be done share is that they maximize net pleasure, is false. Now, Hedonic Utilitarianism may well be wrong, but it would be nice to have an argument that works even if Hedonic Utilitarianism is right. More generally, I am worried about an argument that makes questions in metaethics depend on our views in normative ethics. That is, I cannot see how the answer to the question of whether normativity is sui generis could be contingent on whether some monistic moral theory is right. Attempts to establish non-reductionism on the grounds of moral pluralism or particularism seem quite misguided.

Normative Facts: Metaphysical not Conceptual

95

Second, the argument fails against constitution views. If the property of being right is claimed to be identical with some non-normative property, it is indeed necessary that whenever an action is right it has that property. If we then find no non-normative property that all and only those actions that are right share, there is indeed a problem. But suppose instead that the reductionist claims that this particular action’s being right just consists in having those non-normative properties. This gives us no reason to conclude that all (or indeed any) other action’s being right consists in having the same nonnormative properties. Thus, this reductionist can happily admit that there is no non-normative property that all right actions share, but hold on to his view that rightness just consists in having certain non-normative properties, although not always the same ones. The Pattern-Problem only arises when we talk about groupings and types, which we are indeed forced to do when identity-claims are being made. But the reductionist who appeals to constitution can safely stick with action-tokens and evade the objection. Third, the argument seems to rely on a phenomenon that is far too widespread to capture anything specific about normativity and support grand claims about its non-reducibility. Consider our concept of “vegetable”. The configuration effected by this concept is not intelligible from the botanist’s point of view or indeed any practice other than the culinary arts.9 Not only do vegetables differ considerably in size, shape, and colour, there is no biological property that all and only vegetables share: Carrots are roots, lettuce is just leaves, tomatoes are actually berries, and some, for instance, peas are, well, fruits! Vegetables are botanically patternless. What should we conclude from the fact that vegetable is not a natural kind? Some people might stop being cognitivists about vegetables: Talk about vegetables, they would say, is not in the business of reporting facts, but we use that v-word for recommending things for use at certain stages in a meal. Others might become nihilists about vegetables: Although we describe things as being vegetables, there is no such property out there. This would be bitterly opposed by a realist, who echoing G.E. Moore, would say: “If I am asked ‘What is a vegetable?’ my answer is that a vegetable is a vegetable, and that is the end of the matter.”10 He would be committed to the view 9 10

I am grateful to Barbara Bleisch, who first drew my attention to this example. See Moore (1993, p. 58) on goodness.

96

Norbert Anwander

that there are vegetable facts which cannot be reduced to any other facts. A fourth party might finally call for relaxation and point out that our vegetable practices are just fine and we need not worry about whether there is anything out there in the world that underpins our vegetable discourse. The morale of ‘Vegetable’ is that if what is operating in the PatternArgument is such a widespread phenomenon that by the same reasoning we might be forced to be Mooreans about vegetables, and indeed lots of other things, that amounts to a reductio. It is yet another reason why non-reductionists had better not rely on the Pattern-Argument. – So let me turn to a more promising line of reasoning. An Alternative Argument We can start by noting that there is something odd about the way the task of reducing morality is often understood. Put crudely, the aim is to find a nonmoral way of dividing up the world that necessarily matches the partitioning that is effected by the moral. This is just what the machine I mentioned earlier was supposed to do: Operating on purely non-moral characteristics, it can reliably distinguish between distributions that are just and those that are not, between the actions that are right and those that are not. We need not go into the technical details of competing accounts of how this is to be done. And we need not ask whether the distinctions on which the machine operates show a pattern that is discernible, if not to the machine, at least to people just like us except for their lack of moral concepts. And let us just assume that it can be done. What would be the point of such a partitioning by non-moral characteristics? The only significance I can see in such an account would be for a person who happens to be interested in dividing up the world along moral lines but unfortunately lacks moral concepts. I admit this is a bit of a caricature, but I think the essential point stands: No account of the authority of morality would have been given. I do not just mean that it would still be an open question whether you ought not to do what happens to be an action that falls under the category of the morally wrong. For the same job could presumably be done for all actions that ought to be done, all beliefs that ought to be believed, and all feelings that ought to be felt. And with regard to those,

Normative Facts: Metaphysical not Conceptual

97

the question clearly would not make sense. But I do not think that we would have captured normative force. To use two ugly words that however get us to the core, we would not have got an account of what reasonhood and oughtness is. All we would have done is to find a non-normative property whose presence reliably tells us that there is also some normative property around. The basic mistake about this partitioning business is that it ignores an important distinction between two ways in which properties can be related. Remember that reductions are offered as answers to questions of the form “What is it for x to be N?” Suppose for some fairy-tale country it is true that what it is for you to be its king is to be its absolute ruler. Since that is meant to capture the essence of kingness, it can be said that there is nothing to your being this country’s king over and above your being its absolute ruler. But when you are challenged why you are its king, you are not going to say, because I am its absolute ruler, but you point out to us what it is that makes you its king, such as for instance being the person with the longest nose (this is after all a fairy-tale country). Having the longest nose is, of course, not what it is to be the king, and it would be quite wrong to say that there is nothing over and above to being the king than to have the longest nose. Another way to see that these are different questions is to notice that changes can be made independently: We could limit your power, or we could establish that the next king shall be the person with the highest IQ score. In the former case, being king would no longer be the same, in the second, being the person with the longest nose would no longer make you king. We might say that in the former case, kingness had changed, while in the second, having the longest nose had lost its king-making force. But suppose instead that there never are any changes. The partition effected by the property of being the king would be exactly the same as the one effected by the property of having the longest nose. Would that make anybody believe that the property of kingness had been successfully reduced to the property of having the longest nose? Clearly not. And this answer stands even if there never could be any changes. Now, substitute rightness for king, and for instance maximizing net pleasure for longest nose. It is just the same structure: Maximizing net pleasure is what, if Hedonistic Utilitarianism is correct, makes an action right.

98

Norbert Anwander

Although the action is right because it maximizes net pleasure, that is not what it is for the action to be right. Furthermore, if Hedonistic Utilitarianism or any other monistic theory is incorrect, there will be several other properties that make actions right. Nevertheless, what it is for an action to be right is always going to be the same thing. And just as before, it would be quite wrong to say that there is nothing over and above to being the right action than to be, for instance, the one that maximizes net pleasure. This is, of course, just another way of saying that the partition story will not work in this case either. It is essential to see that there are three different questions to be asked: i) What is it for some action to be right? ii) What makes an action right? iii) Which actions are right? The most important thing is that i) and ii) concern distinct properties. We can further ask which of these properties are normative and which are nonnormative. iii) is uninteresting, since I take it for granted that the action itself is not normative. There is some controversy over whether ii) – more generally, that which is or provides a reason for some action – is always non-normative. Since I think it is, there is no need for dispute on that score in the present context.11 What is crucial, is whether i) is normative, or since it clearly is, whether the property in question can in turn be reduced to some non-normative property. My answer comes in two parts. The first part is just to mention that it has always been one of the great pleasures of non-reductionists to spot the normative concepts that have been smuggled into proposed reductions (see Dancy 1996, p. 180; McNaughton/Rawling 2003, p. 44). Here is an example from Frank Jackson’s From Metaphysics to Ethics: “We want rightness to be what makes an action right, not in the causal sense, but in the sense of being what ought to be aimed at” (Jackson 1998, p. 141). There are a number of things to say about this sentence: First, we can note that according to 11

My view is that on a non-normative, i.e. externalist, reading of terms such as generous, just, right, it is true that the fact that something is generous etc. is a reason for acting in a certain way. On a normative reading, however, whereby to call something generous etc. implies that it provides a reason, the fact that something is generous etc. provides no reason for acting in a certain way. To think otherwise would amount to double-counting.

Normative Facts: Metaphysical not Conceptual

99

Jackson’s analysis, in the example above, kingness would be the property of having the longest nose, because that is what makes you a king. This is rather odd. Jackson elides the distinction between the status – rightness, kingness etc. – and the conditions for obtaining it – e.g. maximizing net pleasure, having the longest nose etc. We may grant that the latter can be captured in non-normative terms but insist that this is quite irrelevant since the dispute is about the former. Second, it is not true that being the thing that makes an action right is the same as being what ought to be aimed at. At least, it had better not be true, if, for instance, indirect consequentialism is not to be ruled out by definition alone. Furthermore, as the hedonistic paradox teaches us, you ought not to aim at your own happiness because that is most likely to make you unhappy. But it should not follow from this that the fact that some action makes you happy cannot make it right. Third, and this is the point that is crucial in our context, we will want to ask what Jackson is going to do about that “ought” in his analysis. For surely this is a normative concept, if any is. The second part involves looking for a more principled answer. Establishing that whenever it is true that some action is the right one, there are in fact two distinct properties on the scene – the normative property of being right and the non-normative property that makes the action right – is not enough. The naturalist may have been mistaken in taking the property that makes the action right as his target, but this is no proof at all that it is impossible for him to provide a reduction of the property that he should aim at, viz. the property of being right. True, the history of attempts to reduce rightness is a history of failures, but that is no guarantee that the reduction cannot be done. The more principled answer, I think, must be along lines Derek Parfit suggested: It is conceptually impossible that normative facts could be the same as non-normative facts. That is, of course, not the trivial point, which we already touched upon, that it is conceptually excluded that the same fact is both normative and non-normative. That nothing can be both p and not-p is something we knew all along. What is meant is that it is conceptually excluded that normative concepts and non-normative concepts refer to the same facts. Parfit illustrates this claim in the following way: While it was conceptually possible that heat should turn out to be molecular kinetic energy, it could not have turned out to be a shade of blue, or a medieval king

100

Norbert Anwander

(Parfit 1997, p. 122). In advance of trying to reduce heat to a medieval king, we could already know that such a reduction would fail. Similarly, we do not have to await further investigation, but can decide on conceptual grounds that the normative cannot be reduced to the non-normative. At this stage, the reductionist can in turn point out that the history of claims about conceptual impossibilities is hardly one that inspires confidence. For a long time, it seemed plain obvious to many people that the living could never be the same as or be constituted by the-nonliving. Now that the reduction has been carried out, we no longer feel tempted to postulate some irreducible élan vital. What justifies our thinking that the same will not happen to goodness, reasonhood, and oughtness? Something of an answer is to be found, I suggest, if we can make clear, in what way the normative must be seen as sui generis.

2. What Normative Facts Sui Generis Might Be When it is claimed that ethical facts are sui generis, what is often meant is that they are irreducible to non-ethical facts in the same manner as psychological facts cannot be reduced to physiological facts and biological facts cannot be reduced to physical facts. Ethics, on this view, is an autonomous discipline that will not be replaced by, for instance, psychology or economics. However, if all we are concerned about is the status of ethics as an independent discipline, there is no need to be worried by naturalism. On the view that ethical properties are constituted by (other) natural properties, for instance, it is still true that ethics cannot be replaced by the disciplines studying those properties. But that is not how we should think about normativity being sui generis. For the asserted distinctness of the disciplines is not metaphysically deep, but operates merely on the conceptual level. When it comes to metaphysics, all the properties referred to in the higher disciplines are thought to be, if not identical with, at least constituted by the properties referred to in the more basic disciplines. Normativity, I suggest, is in a very different way sui generis, it is, we might want to say, in a sui generis way sui generis. It is a mistake, that is, to think of the normative as yet another layer on

Normative Facts: Metaphysical not Conceptual

101

top of the non-normative. The view according to which the facts ethics and epistemology – understood as normative disciplines that tell us what there is reason to do and what there is reason to believe – deal with are, to put it metaphorically, somehow growing out of or emerging from the facts that are the subject of the other disciplines strikes me as utterly misleading. (This is also why the Pattern-Argument is no good for our purpose: At best, it establishes that ethics is an independent layer – in the same way as biology or psychology, in fact any discipline, are. It does nothing, however, to capture the way normativity is different from all those layers.) Here is a very brief sketch how we should think about normative facts: According to Parfit, “[n]ormative concepts form a fundamental category – like, say, temporal or logical concepts” (Parfit 1997, p. 121). This is indeed the degree of distinctness that we should be looking for, but I want to suggest a slightly more specific analogy. We are familiar with several distinct kinds of necessity, for example logical, metaphysical, and causal necessity. It is logically necessary that if p or q and not p, then q. It is metaphysically necessary that if x is water, then x is H20. And it is a causal necessity that if water is heated to a certain temperature it starts boiling. Most people accept that logical necessity is not reducible to causal necessity. For instance, we would refuse an understanding of the above logical necessity in psychological terms, such as for instance ‘when people believe both that p or q and that not p, this causes them to believe that q’. Neither do modifications such as that this is what would happen under normal or under ideal circumstances have much to recommend them. Normativity, I am saying, is to be understood as a further kind of necessity.12 It is a sui generis kind of necessity in the same way as logical necessity is sui generis with regard to causal necessity. Christine Korsgaard 12

The term ‘normative necessity’ as it is used here may easily be misunderstood: It may be thought to refer to the way normative properties are related to non-normative properties. While it is universally agreed that the normative properties of a situation supervene on its non-normative properties, there is dispute about how this connection is to be explained. (For an interesting attack on ‘neo-naturalist’ reductions of normative necessity to metaphysical necessity, see Fine (2002).) This is not, however, what is at issue here. Jean Hampton usefully distinguishes between “necessity that” and “necessity to” (Hampton 1998, p. 103). The latter refers to a relationship between certain aspects of the

102

Norbert Anwander

can be seen as suggesting a similar view: „What the normativity of reasons and the power of causes seem to have in common is that they are forms of necessitation: a cause makes its effect happen, and so necessitates it (all else equal); a reason for action or belief necessitates that for which it is a reason in another way, namely it necessitates a person to act or believe as it directs (again all else equal)“ (Korsgaard 1996, p. 226). There is a compelling structural analogy between causality and normativity: With causality, we have some event or state cause another event. For example, the fact that it is very hot (or alternatively, the sun’s shining very intensely) causes John’s sweating. Similarly, with normativity, we have some fact reason13 some agent’s acting in a certain way. Thus, the fact that it is very hot reasons John’s going swimming. While the former is a causal fact, the latter is a normative fact. Normative facts are facts about the way some (non-normative) fact normatively necessitates a person to act in a certain manner. Of course, normative necessity does not ensure that you do anything; it is all too easy not to do the things you ought to do. The explanation for this, however, just is that normative necessity is sui generis; it has no impact on the causal order.14 True, this account is dissatisfyingly vague, and it is not unfair to argue that it is merely a restatement of the non-reductionist position. But it is hard to see how much more could be said about something that is supposed to be sui generis. For the reasons mentioned at the beginning of world and agents. For instance, the fact that the building is on fire makes it necessary for me to jump out of the window (even though it is not necessary that I do it because for all sorts of reasons I might actually not do it). It seems to me a major mistake of much recent work in metaethics, above all in the debate between naturalists and their opponents, to have focussed exclusively on necessity that, where the issue of normativity can easily be overlooked. The real dispute, it seems to me, is whether necessity to is reducible or not. – A further clarification is in order: The term ‘necessity’ may suggest that normative relations are always of a strongly requiring kind. This is not so: The fact that eating an ice-cream would cause me pleasure is a reason for having one but it would be odd to say that I am thereby required to eat one. Clearly, I am not necessitated to have an ice-cream (see fn. 13, and Dancy 2004). 13 I am using the verb ‘to reason’ as an umbrella term for all possible ways of (positively) contributing to the normative status of an agent’s action. It is often said that reasons are considerations that speak in favour (of some agent’s action). It is this general notion of ‘speaking in favour’ that is relevant here. 14 See Schaber (this volume) on the explanatory impotence of normative properties.

Normative Facts: Metaphysical not Conceptual

103

part I, non-reductionist accounts are bound to be unilluminating. It might also be said that the non-reductionist makes normativity appear to be a queer thing. In the present context, however, this will not be an objection since the non-naturalist’s point is exactly that normativity cannot be easily absorbed into the world as we know it otherwise. On this, he agrees with Mackie’s views on morality. Unlike Mackie, he thinks that not only morality but normativity altogether does not come for free, which is probably why he is more prepared to pay the metaphysical price.

References Audi, R. (1997): Moral Knowledge and Ethical Character, Oxford. Baldwin, Th. (1990): G. E. Moore, London. Boyd, R. (1988): “How to Be a Moral Realist”, in: G. Sayre-McCord (ed.): Essays on Moral Realism, Ithaca, pp. 181-228. Brink, D. (1989): Moral Realism and the Foundations of Ethics, Cambridge. Copp, D. (2000): “Milk, Honey, and the Good Life on Moral Twin Earth”, in: Synthese, 124, pp. 113-137. – (2003): “Why Naturalism?”, in: Ethical Theory and Moral Practice, 6, pp. 179-200. Crisp, R. (1996): „Naturalism and Non-Naturalism in Ethics“, in: S. Lovibond/S.G. Williams (eds.): Essays for David Wiggins. Identity, Truth, and Value, Oxford, pp. 113-129. Dancy, J. (1995): “In Defense of Thick Concepts”, in: Midwest Studies in Philosophy, 20. Moral Concepts, ed. by P. A. French, Notre Dame, pp. 263-279. – (1996): “Real Values in a Humean Context” (Review of Michael Smith’s The Moral Problem), in: Ratio, IX, pp. 171-183.

104

Norbert Anwander

Dancy, J. (2004): “Enticing Reasons”, in: R. J. Wallace/M. Smith/S. Scheffler/P. Pettit (eds.): Reason and Value. Essays on the Moral Philosophy of Joseph Raz, Oxford, pp. 91-118. Fine, K. (2002): “The Varieties of Necessity”, in: T. Szabo Gendler/ J. Hawthorne (eds.): Conceivability and Possibility, Oxford, pp. 253-281. Hampton, J. (1998): The Authority of Reason, Cambridge. Jackson, F. (1998): From Metaphysics to Ethics. A Defence of Conceptual Analysis, Oxford. – (2003): “Cognitivism, A Priori Deduction, and Moore”, in: Ethics, 113, pp. 557-575. Korsgaard, C. M. (1996): The Sources of Normativity, Cambridge. Mackie, J. L. (1977): Ethics. Inventing Right and Wrong, London. McDowell, J. (1998): Mind, Value, and Reality, Cambridge (MA). McNaughton, D./Rawling, P. (2003): “Naturalism and Normativity”, in: Proceedings of the Aristotelian Society, Supplementary Volume LXXVII, pp. 23-45. Moore, G. E. (1993 [1903]): Principia Ethica, ed. by Th. Baldwin, Cambridge. Parfit, D. (1997): „Reasons and Motivation“, in: Proceedings of the Aristotelian Society, 71, pp. 99-130. Pinker, S. (2002): The Blank Slate, London. Rachels, J. (2000): „Naturalism“, in: H. LaFollette (ed.): The Blackwell Guide to Ethical Theory, Oxford, pp. 74-91. Schaber, P. (1997): Moralischer Realismus, Freiburg. Shafer-Landau, R. (2003): Moral Realism, Oxford. Sturgeon, N. (2003): „Moore on Ethical Naturalism“, in: Ethics, 113, pp. 528-556.

Peter Schaber Good and Right as Non-Natural Properties1 There is widespread agreement that Georg Edward Moore takes ‘good’ to be a non-natural property. It is his well-known “open question argument” that is supposed to show that good cannot be a natural property. “(G)ood does not, by definition, mean anything that is natural; and it is therefore always an open question whether anything that is natural is good” (Moore 1903, p. 44). The idea is this: For any natural property N it can sensibly be asked whether x which is N is at the same time also good. Thus, the term “good” cannot refer to a natural property, otherwise - Moore thinks - the question ‘but is it good?’ would not make sense. Moore’s open question argument did not finish off ethical naturalism. As has been argued by various philosophers, the argument does neither show that good cannot be a natural property nor even that the term “good” cannot have the same meaning as a term which refers to a natural property (see Brink 1989, p. 154; Putnam 1981, p. 207). The former because “good” could refer to the same property as, for instance, the term “pleasant” without having the same meaning. This is, for instance, the case with the two terms “water” and “H2O”. They refer to the same items without having the same meaning. The identity of property does not imply the identity of meaning. Thus the fact that “good” might not have the same meaning as “pleasant” does not mean that they do not refer to the same property. But then Moore’s open question argument also fails with regard to the identity of meaning. The fact that the question ‘X is pleasant, but is it good?’ does make sense does not imply that these two terms differ in meaning. If a person does not know the meaning of “pleasant” or of “good” she might sensibly ask whether x which is 1

Thanks to Norbert Anwander, Susanne Boshammer, David Copp, Neil Roughley, Thomas Schmidt, Tatjana Tarkian, and Theo van Willigenburg for their helpful comments.

106

Peter Schaber

pleasant is at the same time also good. The open question ‘Is it good?’ could be a question of clarification, asked by a person who does not know or is maybe not sure about the meaning of pleasant. Thus, open questions can be meaningful with regard to two terms which have the same meaning. But even though Moore’s open question argument does not succeed I think we should agree with him that good is indeed not a natural, but rather a non-natural property sui generis. We should agree with him, I will argue in this paper, of course for reasons other than Moore himself offered. And we should not just believe that good is a non-natural property. We should have the same belief about right and all other so called thin ethical properties such as bad and wrong. Thin properties are properties referred to by thin evaluative terms. Thin terms inform us about the relevant evaluative properties such as good. Thick terms such as “polite”, “generous”, “courageous” and “just” on the other hand inform us also about non-evaluative properties of the actions they are ascribed to (see also Williams 1985, p. 129). If the queen’s donation to Oxfam was generous, the amount of money exceeded one thousand pound. Thin ethical terms just inform you that the action in question has the relevant evaluative property. And the claim I will argue for is that thin ethical terms refer to non-natural properties.

1. Ethical Naturalism In his “Principia Ethica”, Moore aims at refuting ethical naturalism. He takes ethical naturalism to be the view that the term “good” refers to a natural property, and that good is whatever the natural property the term “good” refers to (see Moore 1903, p. 40). Most ethical naturalist think that this property can of course be referred to by another term such as for instance “pleasant”. Thus, good is what is pleasant. Or good is what is desired. The view that good is a natural property is compatible with the widely shared view that good supervenes on certain natural properties. That is to say, it is compatible with the view that good depends on certain natural properties. An action,

Good and Right as Non-Natural Properties

107

say, could not cease to be good without a change in its natural properties.2 Natural properties supervene on other natural properties. The properties described by chemistry, for instance, supervene on the properties described by physics. Supervenience does not have to be a relation betwen non-natural and natural properties. Thus, ethical naturalists can easily accept the claim that moral properties supervene on natural properties. Two forms of ethical naturalism have to be distinguished: a) reductive naturalism and b) non-reductive naturalism. a) According to reductive naturalists good is identical with a natural property. Good could be, for instance, identical with the pleasant. Being this property it supervenes, of course, on other natural properties. Something is pleasant in virtue of other properties it has. b) Non-reductive naturalists on the other hand hold the view that good is a natural property sui generis, that is, a property not identical with another natural property, even though also supervening on other natural properties. This is the view David Brink, for instance, holds. According to Brink, the property good (or right) is a natural property constituted by natural properties. “Moral facts and properties, so construed, are constituted, composed or realized by organized combinations of natural and social scientific facts and properties ... This naturalist claim should be understood on the model of other common constitution claims: for instance, tables are constituted by certain combinations of microphysical particles, large scale social events such as wars and elections are constituted by enourmously complex combinations of smaller scale social events and processes ...” (Brink 1989, p. 159). Non-naturalists also hold the view that moral properties supervene on natural properties. But unlike a reductive as well as a non-reductive naturalist a non-naturalist thinks that good and right are supervening non-natural properties. 2

Moore himself holds this view in Moore (1942), p. 588: “I should never have thought of suggesting that goodness was nonnatural, unless I had supposed that it was derivative in the sense that, whenever a thing is good … its goodness depends on the presence of certain non-ethical characteristics possessed by the thing in question … if a thing is good … then that this is so follows from the fact that it has certain natural intrinsic properties.”

108

Peter Schaber

2. Natural Properties But what are natural and non-natural properties? Let us start with natural properties. In his “Principia Ethica” Moore holds the view that natural properties are those properties which do not depend on the existence of their objects (see Moore 1903, p. 41). A natural property exist by itself without being the property of this or another object. Moore thinks that this does not apply to the non-natural property good. “Can we imagine ‘good’ as existing by itself in time and not merely as a property of some natural object? For myself, I cannot so imagine it, whereas with the greater number of properties of objects - those which I call the natural properties - their existence does seem to me to be independent of the existence of those objects” (Moore 1903, p. 41). In his “Reply to my Critics” he advocates the view that natural properties do not depend on other properties (see Moore 1942, p. 588). I think that both views should be rejected: The problem with the view advocated in his “Principia Ethica” is that it is hard to imagine any property existing by itself in time, independently of the existence of objects. And the problem with the second view is that it would exclude the supervenience of natural on other natural properties. But then the properties described by chemistry which supervene on properties described by physics had to be conceived of as non-natural properties. This is an implausible view. I think that natural properties should rather be taken as those properties that would figure in a perfected version of the natural and social sciences (see also Crisp 1996, p. 117). This is compatible with the view that natural properties are empirical properties: They would figure as properties in the best empirical sciences (see Copp 2003, pp. 185-187). What distinguishes non-natural from natural properties is to my mind the following: Unlike natural properties non-natural properties play no role in the explanation of empirical phenomena, that is, in the explanation of the objects of the empirical sciences. They play no role neither in causal nor in possible non-causal explanations of empirical phenomena.3 And it is not the case that they do not 3

It has to be emphasized that non-natural properties do not just play no role in causal but also in non-causal explanations. For a discussion of the causal criterion of a property’s being natural see Crisp (1996), p. 115.

Good and Right as Non-Natural Properties

109

actually have an explanatory force, but could have one, provided the world was different. This is excluded by their very nature. Non-natural properties are not part of the explanation of how things, states of affairs and events come about. This applies also to the explanations of actions, choices and evaluative beliefs. Thus if ‘good’ and ‘right’ are non-natural properties they have no explanatory force. The truth of ethical non-naturalism depends on the explanatory impotence of the thin ethical properties. I will argue that the properties ‘good’ and ‘right’ do have no explanatory force. They should therefore be taken as non-natural properties, because a property is a natural property if it is part of the best explanations of the best natural and social sciences.

3. Against Reductive Naturalism We should reject ethical naturalism. But first of all, non-reductive naturalism is more plausible than reductive naturalism. There is the problem of reduction: Is good (or right) identical with a single natural property such as pleasant? How could we tell? Let us take Frank Jackson’s methodological proposal “about how to identify the ethical properties: find the properties which are such that, going under their purely descriptive names, they make the clause of mature folk morality come out true ... and then identify each ethical property with the corresponding descriptive property” (Jackson 1998, p. 143). If we proceed this way, could we then find a descriptive property name that would confirm our everyday moral judgments? Could the term “good” in this way be replaced by a purely descriptive term? This is not clear. But then, more importantly could “good” be replaced by one single descriptive term such as pleasant or serving the interests of others or desired by others? This seems to be very unlikely, given the way we use the word “good”. Things are called good for different reasons. Things that are pleasant are good, but then also things that are healthy or life saving or autonomy enhancing or problem solving and so on. Thus, if we follow Jackson’s methodological proposal we will end up with very different descriptive property names,

110

Peter Schaber

and as a consequence with very different natural properties “good” would supposedly refer to. The property would be identical with different properties such as “pleasant”, “healthy”, “life saving”, “autonomy enhancing” etc. Jackson thinks that “good” and “right” might turn out to be disjunctive properties, “possibly infinitely disjunctive descriptive properties” (Jackson 1998, p. 124). If so, the word “good” would function just like the concept “bald” or the concept “tall”. “Bald” refers to many different hair distributions, and there are many ways to be tall. To take good as a disjunctive property has its obvious difficulties. First of all, it is definitely not what we expect naturalism to be.4 One assumes “good” to refer to a property such as “pleasant” or “promoting hedonic value”. But if we follow Jackson’s methodological proposal it is very unlikely that this will be the upshot of our inquiry. But then, more importantly, why do we call all these different natural properties “good”? The different hair distributions we refer to by the word “bald” are not the same, but we call them all “bald” because of their similarity.5 But it is not at all clear whether the properties “good” refers to have any similarity. Thus, if the natural properties “good” refers to are very different it is unclear why they all are supposed to be referred to by the term “good”. If the relevant natural properties are very different something else is meant by “good” than just that the object in question has a certain natural property. Non-reductive naturalism is of course not faced with this problem. A nonnaturalist can claim that the term “good” refers to a natural property sui generis that supervenes on certain natural properties. It is always the same property that is referred to by “good”; a property that is realizable in many different ways, supervening on different natural properties. If so, the non-reductive naturalist is in a better position to account for our use of “good” and the other thin property names. Thus, I think our choice has to be one between non-reductive naturalism and non-naturalism. And with regard to this choice the question of whether thin ethical properties have an explanatory force is crucial. 4

See Shafer-Landau (2003), p. 71: “(I)t is certainly a highly non-standard form of classical naturalism.” 5 See Jackson (1998), p. 124: “You have at some stage to say that to be bald is to be like these exemplars in the ‘bald’ way, hoping that one’s hearers have latched onto the relevant similarity and can go on in the right way.”

Good and Right as Non-Natural Properties

111

4. Good and Right as Secondary-Properties Does good have an explanatory force? We should have a clearer picture of what we mean by saying that x is good, in order to answer this question. I take it that Thomas Scanlon’s buck passing view of goodness is helpful in this context (Scanlon 1998, pp. 95-100). According to Scanlon, to say that x is good means that x has properties that provide us with reasons to respond to x in certain ways. Good is not a property that itself provides us with reasons to respond in a certain way: “Rather, to be good ... is to have other properties that constitute such reasons. Since the claim that some property constitutes a reason is a normative claim, this account also takes goodness and value to be non-natural properties, namely the purely formal, higher-order properties of having some lower-order properties that provide reasons of the relevant kind” (Scanlon 1998, p. 97). Good is related to reason-giving properties, it is not in itself reason-giving. X is good, for instance, because it is pleasant. To say that x is good is not to say that it is pleasant. It is just to say that it has a property that provides us with a reason to choose it or to recommend it to others. The formal property good supervenes on the property of being pleasant: It is good, because it is pleasant. The idea is that good is not itself a reason-giving property. A reason is something that speaks in favour of an action or a response or an attitude etc. I have a reason to choose x, because x is pleasant. The fact that x is good is not a reason for choosing x. It is not the case that I have a reason to choose x, because x is pleasant and then in addition to this, because it is also good. If good was reason-giving we had two reasons in play here. But this is not true. ‘Why should I choose x?’ ‘Because it is pleasant’. This would be a satisfactory answer.6 Nothing had to be added such as: ‘and then it also good’. ‘X is good’ tells you that there are reasons to choose x such as the fact that it is pleasant. 6

See Scanlon (1998), p. 97: “(T)he fact that a resort is pleasant is a reason to visit it or to recommend it to a friend, and the fact that a discovery casts light on the causes of cancer is a reason to applaud it and to support further research of that kind. These natural properties provide a complete explanation of the reasons we have for reacting in these ways that things that are good or valuable. It is not clear what further work could be done by special reason-providing properties of goodness and value, and even less clear how these properties could provide reasons.”

112

Peter Schaber

5. Does Good have an Explanatory Force? Again: Does the second-order property good have an explanatory force? Let us take the action of my choosing x. Does good have an explanatory force with regard to this choice? I think that x is good and also that x is pleasant. And I choose x. One could say I did so, because x was good. That is to say, the goodness of x brought about the belief that choosing is good and this belief then moves me to action. Could this be a satisfactory explanation of my choice? Explaining an action is, I take it, naming the reasons the action has been carried out for. If good has an explanatory force it would be such a motivating reason. But I think that good is not a motivating reason in the same way as it is not a normative reason for doing something. Take the example of Paul’s writing a paper on abortion. ‘Why did he write this paper? His doing so would not be explained by saying that he did it, because it was a good thing to write a paper on this subject. Of course, it might indeed be a good thing to write such a paper. But to say that it is good to do so is just to say that there are reasons to act accordingly. The reasons that speak in favour of doing so differ from its being good. To write a paper on abortion is good due to the reasons that speak in favour of it. It is good to write such a paper, because it is a controversial issue in ethics which needs to be solved. This might be the reason Paul wrote a paper on abortion. To say that it is good to do so is too unspecific with regard to an explanation of the action in question. Why - one can ask - a paper on abortion? Because it is good to deal with this issue. But this holds for many other papers. Why abortion? This question is not answered by saying that it is good to do so. It might be good to do so, but the reasons Paul wrote the paper are different. For instance, he wanted to develop a solution to the problem of abortion people could agree on. Thus good does not figure among the reasons that guide our actions. And good has therefore no explanatory force. It might be the beginning of an explanation of an action, in the sense that there is an explanation of the action, that there are reasons that speak in favour of doing it. Some one might have serious doubts: ‘Why does one write a paper on abortion?’ ‘What does guide a person who does this?’ If one answers, ‘It is good thing to do this’, one says that there are indeed reasons for writing such a paper, reasons that have to be spelled out in order to explain the action in question.

Good and Right as Non-Natural Properties

113

The same I think holds for evaluative beliefs. I think that doing x is good. Again, the property good does not play a role in explaining how I came to have my belief. I do not hold this view because doing x is good, that is to say because I saw that doing x was good. There are reasons why I think that doing x is good: I think so, because doing x is pleasant or autonomy enhancing or very important for my health etc. Such properties provide us with reasons to believe that the things in question are good. They bring about the relevant evaluative beliefs. I think that doing x is good, because I saw that it has good-making properties. And to say that I saw that doing x was good is just to say that I saw that doing x has properties that provide us with reasons to respond to it in a positive way.

6. Under the guise of the good? Berys Gaut does not share the view that good has no explanatory force (see Gaut 2002). He thinks that a full explanation of actions and choices cannot be given without reference to goodness. Take the following explanation: Talking to John is pleasant, and this is why I talk to him. I think that this could be a perfect explanation of the action in question. But Gaut believes that this could not be a satisfactory explanation if it was not clear that things which are pleasant are at the same time good. “It turns out then on closer inspection that one cannot eliminate reference to values in giving a full explanation of one’s reasons for choosing” (Gaut 2002, p. 151). So why did he choose x? One might say: ‘Because he knows that doing x is pleasant’. But then one might ask ‘Is this a reason for choosing x?’ Yes, it is, because things that are pleasant are good. According to Gaut, this reference to good is necessary for a full explanation of an action. His idea seems to be the following: Take the counterfactual test of explanatory relevance (see Sayre-McCord 1986). A property does not play an explanatory role with respect to a certain fact if “we could have explained it just as well even if the assumption had not been invoked in the explanation ...” (Sayre-McCord 1986, p. 272). The question then is: Would the explanandum have occurred even if the putatively explanatory fact had not obtained, other things being equal? Provided that we – as Gaut assumes – are always guided by the good, good seems to meet this counterfactual test

114

Peter Schaber

of explanatory relevance. X would not have been chosen if it was not good. Thus good seems to be explanatorely relevant. But even if we were - as Gaut thinks - always guided by the good, good played no explanatory role. First of all, good has no explanatory relevance when the relevant evaluative belief is wrong and the thing in question is as a matter of fact not good. In these cases only the belief about good would have an explanatory relevance. Good could only play a part in explaining choices in those cases where the belief about goodness is true. But even then I think good does not explain choices. If it is true that x is good, good is just an enabling condition of an agent’s choosing x (see Dancy 2000, p. 127). One has to distinguish between being part of an explanation and just being a necessary condition which is not part of the explanation. X being an enabling condition of Z does not belong to the elements which brought about Z. Z would not have occurred, had X not obtained. But X has not be mentioned to explain why Z came into existence. Take one of Jonathan Dancy’s example: “For instance, that England is not sinking beneath the waves today is a consideration in the absence of which what explains my action would be incapable of doing so. But that does nothing to show that England’s not submerging today is part of the explanation of why I do what I do” (Dancy 2000, p. 127). I think that this also applies to good (and of course also to all other thin ethical properties). If the relevant belief about good is true the choice would not have been made without its object being good. But the goodness of the object chosen is not the reason why it has been chosen, and thus it is not a proper part of the explanation of the choice in question. To say that x is good is not in any way to illuminate us as to why x occurred. Things do not happen, because they are good. Things are chosen because of the reasons that explain why they are good. Take an action you do not understand: Why did she do x? The reasons for doing are not at all obvious. Say one gave the answer ‘Well, doing x is good’. Would this make the action understandable? Would we now know why she did x? Not at all. And referring to the goodness of the action is not even part of a full explanation. It does give you no reason at all. The reasons for her to choose x might be that doing so made her happy or doing x was pleasant etc. These are reasons that make our choices and actions unerstandable. Of course, in some cases it is certainly true that if there were no reasons explaining the goodness of an action, the

Good and Right as Non-Natural Properties

115

action had not been carried out. But this does not mean that good plays any explanatory role. This role is played by the reasons that explain the goodness of the action chosen.

7. Properties, Reasons, and Good How are natural properties, reasons and goodness related to each other? The reasons for choosing x are provided by certain properties x has. It is good to choose x, because x is pleasant. Being pleasant is the reason for choosing x. And it is also the reason that can guide an agent (She chose x, because it was pleasant). ‘X is good’ means x has such reason-providing properties. It does not mean that x is pleasant. Good is a formal property which can supervene on many different natural properties. That x is good is a normative fact. That x is pleasant is not a normative fact. It is rather a normatively relevant fact (it provides us with reasons to respond, see also Parfit 1997, p. 124). ‘S has reason to choose x’ is also a normative fact. Like the property ‘is good’, the property ‘is a reason for S’ has no explanatory force. ‘S chose x, because she had reason to do so’ is not an explanation of the action in question. S chose x due to the reasons she had to do so, but not due to the fact that she had reasons. These reasons are explaining actions, not the the fact that you have a reason. One can compare this with reasons for beliefs. I have reasons to believe that p. These reasons explain my belief. I believe that p, because I saw such and such speaks in favour of holding this belief. I don’t believe that p, because I think I have reasons to do so. The same holds for actions: Actions are explained by reason-giving properties: I did x, because it was a great pleasure to do so.

8. Good as a Property? The fact that good (or any other thin ethical property) does not illuminate us as to why an action has been carried out, could be seen as a reason to conclude that the term “good” does not refer to a property, neither to a nonnatural property nor to a natural one. “Good” one could say has a completely

116

Peter Schaber

different function: It does not refer to anything, it has rather an expressive meaning. It is the expression of an inner state (a positive attitude, an approval). To say ‘x is good’ is to say something like ‘I’m pleased that x has the properties a and b’ or ‘I recommend choosing x due to the properties a and b it has’. But I think that “good” should not be taken this way. I might of course express my approval that x exists by calling x “good”. But I tell you at the same time that x has properties that are worth going for. I myself might not be pleased about x. x might not interest me, but I have to admit that it has properties that speak in favour of choosing x. And I might be right or wrong with regard to this. I’m right if it has such properties and I’m wrong if it has no such properties. Thus, to say that x is good is to say that x has a certain kind of property, namely the second-order property to have reason-giving properties. And this is the case or not the case: S has a reason to choose x or she has not. The talk of a property ‘good’ is related to the possibility of being right and wrong about the existence of reason-providing properties. Thus, the fact that good has no explanatory force is no reason to conclude that “good” does not refer to a property.

9. Supervenience: Two Objections There are two objections to be met with regard to the idea that the thin non-natural properties ‘good’ and ‘right’ supervene on natural properties. According to the first objection (a) supervenient properties have to be natural properties and according to second objection (b) the superveniencerelation between the property good and any natural property is completely mysterious. a) Frank Jackson thinks that ethical properties are as supervenient properties necessarily natural properties: “(F)or any ethical predicate there is a purely descriptive one that is necessarily co-extensive with it. It follows that ethical properties are descriptive properties. For it is a consequence of the way the ethical supervenes on the descriptive that any claim about how things are made in ethical vocabulary makes no distinctions among the possibilities that cannot in principle be made in purely descriptive vocabulary” (Jackson 1998, p. 123).

Good and Right as Non-Natural Properties

117

Jackson’s idea is that ethical properties are natural properties because for any ethical predicate there is a natural predicate which is necessarily coextensive with it. If the ethical property M is instantiated, necessarily also the natural property N. If so, M and N are identical. Identical properties are co-extensive in all possible worlds. And properties that are necessarily co-extensive are identical. The latter seems false. Triangularity and trilaterality would then also be the same property. But this does not seem to be true (see also Shafer-Landau 2003, p. 91). They are not the same property, even though they are necessarily co-extensive. But Jackson holds the view that triangularity and trialaterality are the very same property: “Cases where we think that a triangle is equiangular while failing to think that it is equilateral are ones where we have a separation in modes of representation in thought for what is, all the same, one and the same property in our sense of ‘property’” (Jackson 1998, p. 126). If so, having angles and having sides would be, as Shafer-Landau rightly argues, identical, too. Although they are necessarily co-extensive, they can be distinguished. “[A]ngularity is a funtion of the distance between two intersecting lines. Laterality isn’t that. Thus a difference” (Shafer-Landau 2003, p. 91). This is an example of how two predicates can be necessarily co-extensive without referring to the same property. Thus, ethical properties are not necessarily natural properties. If the natural property N is instantiated, necessarily also the non-natural property M. The fact that ethical properties supervene on natural properties does not imply that they are identical with natural properties. b) Let us turn to the second objection. Simon Blackburn argues that “[s]upervenience ... becomes a mysterious fact” (Blackburn 184, p. 185), if one takes this - as moral realists do - to be a relation between natural and ethical properties. According to Blackburn, the problem consists in explaining “the ban on mixed worlds, and the argument goes that anti-realism does this better than realism” (Blackburn 1984, p. 184). What is meant by “mixed worlds”? Say, M supervenes on N*. But M does not necessarily supervene on N*. It might be the case that X is N* without being at the same time M. Mixed worlds obtain where it is true that M does not necessarily supervene on N*. These mixed worlds seem to be excluded by the supervenience idea. If X is M due to the fact that X is N*, how could it be that X is N*, but at the

118

Peter Schaber

same time not M, other things being equal? There is a ban on mixed worlds, and the problem is, as Blackburn thinks, “that of finding out the authority behind the ban” (Blackburn 1993, p. 135). According to Blackburn, this a particular problem for moral realists who think that all moral properties supervene on natural properties. There is an explanation available with regard to the supervenience of two different natural properties. All competent people will agree, for instance, that chemical structures supervene on physical structures. “The best explanation of why competent people recognize the supervenience of kinds of physical or chemical structure is that we live in a culture in which science has found this out” (Blackburn 1993, p. 144). This sort of empirical explanation is - as Blackburn thinks - not available with regard to the supervenience of moral on natural properties. The supervenience relation is not the result of empirical investigations. So how can we assume that if X is N* (a natural property), then necessarily X is also M (an ethical property)? First of all, I think it is not clear why the same answer should not also be available to a moral realist. Competent people recognize the supervenience of the moral on the natural, because they found out that if X is N*, it is always also M. The moral realist has of course no explanation why this is so, but then I think the same applies to those who believe in the supervenience, for instance, of mental on physical properties. What would explain the fact that there is a necessary connection between these two kinds of properties? But let us assume that there was an explanation of the supervenience of the mental on the physical. Would this create a problem for the moral realist? I think it would not be a problem for the naturalist realist. She could argue that the moral supervenes on the natural, just because the moral and the natural are identical. That is to say, there cannot be a change with regard to the moral properties of X, if there is no change with regard to its natural properties, because they are the same. Say, good was pleasant, X could not cease to be good without ceasing to be pleasant. This answer is not available, one has to admit, to the non-naturalist. But then it would also not be available to the non-reductive naturalist. So if the moral realist has a particular problem with explaining the supervenience of the moral on the natural, it is a particular problem for the non-naturalist as well as for the non-reductive naturalist. If so, Blackburn’s objection is a

Good and Right as Non-Natural Properties

119

challenge not only for the non-naturalist realist. Thus, Blackburn’s problem, provided that there is one, had to be dealt with in a general defense of moral realism. There is no particular problem of the supervenience of non-natural on natural properties.

Conclusion Ethical naturalism which takes ethical properties to be natural properties should be rejected, because thin ethical properties do not play a role in the best explanations of empirical phenomena, including actions, choices and evaluative beliefs. Thin ethical properties (good, right) supervene on natural properties that provide us with reasons to respond to them in certain ways. Something is good or right if it has such reason-giving natural properties. Unlike these reason-giving properties good an right are not part of the natural world.

References Blackburn, S. (1984): Spreading the Word, Oxford. Brink, D. (1989): Moral Realism and the Foundations of Ethics, Cambridge. Copp, D. (2003): “Why Naturalism?”, in: Ethical Theory and Moral Practice, 2, pp. 179-200. Crisp, R. (1996): “Naturalism and Non-Naturalism in Ethics”, in: S. Lovibond/S. G. Williams (eds.): Essays for David Wiggins. Identity, Truth and Value, Aristotelian Society Series, pp. 113-129. Dancy, J. (2000): Practical Reality, Oxford. Gaut, B. (2002): “Justifying Moral Pluralism”, in: Ph. Stratton-Lake (ed.): Ethical Intuitionism. Re-Evaluations, Oxford, pp. 137-160. Jackson, F. (1998): From Metaphysics to Ethics. A Defense of Conceptual Analysis, Oxford.

120

Peter Schaber

Moore, G.E. (1903): Principia Ethica, Cambridge. – (1942): “Reply to my Critics”, in: P.A. Schilpp (ed.): The Philosophy of G.E.Moore, Evanston. Parfit, D. (1997): “Reasons and Motivation”, in: Proceedings of the Aristotelian Society, Supplementary Volume 71, pp. 99-130. Putnam, H. (1981): Reason, Truth, and History, Cambridge. Sayre-McCord, G. (1986): “Moral Theory and Explanatory Impotence”, in: Sayre-McCord (ed.): Essays on Moral Realism, Ithaca, pp. 256-281. Scanlon, Th. (1998): What We Owe to Each Other, Cambridge (MA). Shafer-Landau, R. (2003): Moral Realism. A Defense, Oxford. Williams, B. (1985): Ethics and the Limits of Philosophy, London.

Thomas Schmidt Moral Values and the Fabric of the World. A Reconsideration of Mackie’s Arguments against Moral Realism

1. According to John L. Mackie, values and, a fortiori, moral values “are not part of the fabric of the world” (Mackie 1977, p. 15). This is the metaphysical claim that he has in mind when opening his Ethics. Inventing Right and Wrong with the famous statement that “[t]here are no objective values” (ibid.). Both Mackie’s claim and the reasons he offers in support of it give voice to a profound philosophical uneasiness which many people experience when confronted with the set of views which moral realists subscribe to: there are metaphysically robust moral facts, the point of moral judgements is to describe moral facts correctly, moral judgements are true if, and only if, they succeed in doing so, and knowledge about moral facts is not beyond the scope of our epistemic capacities. Many believe it to be simply obvious that a view of morality committed to or even explicitly stating these claims is erroneous from the very beginning. Those denying moral realism, however, are not always ready to embrace an antirealist position. Rather, quite a number of ethicists turn their back on the issue as to whether a realist or an antirealist interpretation of moral discourse is appropriate. They doubt whether the metaphysical question as to whether there is a matter of fact which moral judgements can be said to be about is the right place to start off from when trying to give an informative account of the philosophical foundations of morality (see, for instance, Scanlon 1998, pp. 2-3). I am convinced that this view of the matter is substantially right. I am not convinced, however, that the philosophical desire giving rise to the realism / antirealism-issue in ethics should be dismissed as being pointless. As I shall suggest towards the end of this paper, the desire to understand what we are doing when judging matters morally and when trying to justify our moral

122

Thomas Schmidt

judgements ought to be taken seriously, and both realism and antirealism in ethics are to be regarded as attempts to deal with this very desire. Providing an alternative way of dealing with it is a task faced by those refusing to take sides in the realism/antirealism-debate in ethics. In this paper, I offer some thoughts on how this task can be framed. I do so by way of an analysis and, in part, a critique of Mackie’s famous arguments against moral realism. Despite the fact that foundational ethical theory has undergone important developments and refinements since Mackie’s book was published, returning to these arguments is worthwhile for several reasons. First, Mackie’s way of arguing against moral realism is philosophically rich. Analysing his influential arguments in detail is, as we shall see, a useful manner of unpacking a number of issues which ought to be kept apart even though it may seem – and, indeed, it did seem to Mackie – that they are fundamentally related to each other. Second, Mackie phrased the more important of his two arguments in a way that makes it relatively easy to circumvent it. His point, as I shall argue, can be carried further than he himself might have intended, which makes his thought, or an extension of it, more subtle than is sometimes believed. Moreover, this result will not only reveal a problem of (certain versions of) moral realism but also shed light on fundamental issues in ethical theory as a whole. 2. Mackie’s first argument against the existence of objective moral values, the argument from relativity, takes up the idea that the diversity in moral outlooks and attitudes makes the assumption of the existence of objective moral values implausible.1 This thought goes back at least to Sextus Empiricus, and today it is often put forward by social scientists and social anthropologists. This is how it is phrased by Mackie: “the actual variations in the moral codes are more readily explained by the hypothesis that they reflect ways of life than by the hypothesis that they express perceptions, most of them seriously inadequate and badly distorted, of objective values” (Mackie 1977, p. 37). 1

In this paper, I shall for the most part follow Mackie in addressing the question as to whether there are objective moral values. However, much of the subsequent analysis can be rephrased in terms of moral facts without substantial loss (which is, of course, not to say that all moral facts, provided that there are any, are facts about moral

Moral Values and the Fabric of the World

123

As an initial reaction to this argument, one might want to attack its premise: quite a number of cases of cross-cultural and/or inter-temporal divergence in moral judgements can be explained with reference to the fact that the very same general moral principles might lead to diverging moral verdicts under different contingent circumstances. This observation, however, does not carry very far: it is quite obvious (and also pointed out by Mackie) that by far not all moral disagreements are ultimately based on different ways of applying the same fundamental moral norms. There are two better reasons for being sceptical about the force of the argument from relativity. First, there are several areas in which disagreement and seemingly irresolvable controversies abound, even though hardly anybody believes that this speaks against the objectivity of the respective discourse. Philosophical disagreement on precisely the issue under discussion provides a nice example. Interestingly, even those who believe that persistent disagreement about anything – a fortiori about the force of the argument under discussion – speaks against the existence of a correct answer, continue to stick to the argument from relativity.2 A second line of thought is this: if the non-existence of objective moral values were the only explanation for moral diversity and disagreement, then the argument from relativity would seem difficult to resist. There are, however, alternative explanations, one of which is available to moral realists. They can point out that disagreement is to be explained by reference to the fact that only one of two people disagreeing on a certain moral issue can be right. (A stronger claim sometimes put forward by moral realists is that the possibility of substantial disagreement presupposes a realist conception of a matter of fact with regard to which one can disagree.) To be sure, moral antirealists will remain unconvinced. However, the consideration just presented shows that rivalling foundational theories of morality can provide different ways of accounting for moral disagreement and that it is difficult to see how criteria for what is to be considered as a better explanation of this phenomenon can be arrived at in a theoretically neutral way. values). Since the notion of a moral fact is central to the current debate on moral realism, I shall sometimes shift to this terminology. 2 The example is borrowed from Putnam (2004), p. 30.

124

Thomas Schmidt

3. The second argument against the existence of objective moral values which Mackie puts forward is, as he himself recognized, more important. In the following, I shall consider the main points of his famous argument from queerness in some detail. Its core is contained in the claim that “[i]f 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” (Mackie 1977, p. 38). Yet, many things are ‘strange’ in one way or another (think of numbers, quarks, and black holes), and it is reasonable to ask in what way objective moral values would be strange and whether there is any straightforward way of arguing from this strangeness against their objectivity. According to Mackie, one way in which objective moral values, if they existed, would be different from anything else in the universe, concerns the epistemological question as to how we would be able to gain knowledge about them. The relevant passages in his book are well-known and have become a locus classicus for a critique of classical intuitionism as maintained by G. E. Moore (1903) and others. Mackie takes it that we would have to posit a faculty of moral intuition – conceived of either as analogous to empirical observation or as some kind of intellectual faculty – in order to make intelligible how moral knowledge is possible. The appeal to “a special sort of intuition”, says Mackie, is a “lame answer” (Mackie 1977, p. 39). There are two obvious reasons for this: first, explaining moral knowledge by invoking something which is likely to sound even more mysterious to moral sceptics than does the assumption of the possibility of moral knowledge itself, does not explain anything. Second, if an appeal to intuition is the end of justificatory chains in moral reasoning, fundamental moral disagreements cannot be solved, not even in principle. Hence, Mackie is right to dismiss classical intuitionism as an unacceptable account of moral epistemology. As opposed to Mackie’s contention, however, those believing in the objectivity of moral values are not compelled to adopt intuitionism of the sort just outlined. Intuitionism is a consequence of moral realism and foundationalism, i. e. the view that there are fundamental moral insights which can be apprehended and justified without reference to other moral beliefs. In ethics, however, a foundationalist epistemology is as unattractive an option as it is elsewhere in philosophy. Most contemporary

Moral Values and the Fabric of the World

125

moral philosophers, moral realists or not, therefore either adopt a coherentist conception of moral justification or else refrain from providing anything like a justificatory model at all. None of these options involves the assumption of basic moral convictions which are fully justified by an appeal to intuition, which appeal Mackie rightly criticizes. Moral realism, therefore, is not threatened by the epistemological part of the argument from queerness. Mackie’s critique of classical intuitionism is easily circumvented by adopting a non-foundationalist view of moral justification, which is a wise step not only for moral realists but for anybody interested in the philosophical foundations of morality. In passing it should be noted that the epistemological part of Mackie’s argument from queerness, if it were to show anything, would not show that objective moral values, if they existed, would be queer. Even if it were necessary to assume a queer faculty in order to explain how moral knowledge is possible, the queerness of this faculty would not imply the queerness of the objects of moral knowledge. If any element of moral realism is threatened by the epistemological part of Mackie’s argument, hence, then moral realism’s epistemological claim is questioned, while the metaphysical view that moral values are part of the fabric of the world is not directly attacked. 4. The metaphysical part of the argument from queerness has more philosophical force. In order to make his case, Mackie draws an analogy between objective values and Plato’s forms, which analogy he takes to “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” (Mackie 1977, p. 40). Note that there are two points in this passage, which we need to discuss separately. First, it is claimed that moral values would have to be such that a person recognizing them would necessarily be motivated to act accordingly. Second, moral values are, according to Mackie, not only motivationally effective but also provide a direction for acting. In order to make the discussion more transparent, it is useful to introduce some terminology. Following several other authors, I shall use the term motivational internalism as referring to the claim that somebody embracing a

126

Thomas Schmidt

moral judgement is necessarily motivated to act accordingly. Motivational internalism needs to be distinguished from normative internalism which says that if a person acknowledges that it would be morally right for her to perform a certain action, then she is committed to accept that there is a reason for so acting. 3 Given this terminology, Mackie’s first point can be rephrased as follows: motivational internalism is true, and motivational internalism requires there to be a special property of moral values which figures in an explanation of how moral motivation arises.4 Let us take up these two claims in reverse order. The second is a non-sequitur: even if it were true that those having moral knowledge are necessarily motivated to act accordingly, it does not follow that one has to assume that moral values have some special property in order to explain how moral motivation comes about. This is so since one can think of alternative possibilities for explaining how this motivation arises. One candidate is the view that moral sensibility necessarily has a motivational dimension, i. e. that it is impossible to have moral insight without being appropriately motivated due to a property of the person having moral insight rather than due to some property of objective values. We can leave it open whether or not this proposal is plausible. For our present purposes it suffices to note that motivational internalism does not straightforwardly imply the queerness of objective moral values in any interesting sense of this term. Not only can the inference from motivational internalism to the queerness of moral values be doubted along the lines just outlined, but motivational internalism itself is also not at all uncontroversial. To begin with, it can hardly be denied that there are situations in which somebody is morally required to 3

Note that normative internalism implies motivational internalism if it were true that somebody acknowledging that he has a reason to act in a certain way is also motivated to act accordingly. Whether this latter claim is true, however, can readily be disputed. – What later has come to be called motivational internalism has been prominently defended by R. M. Hare (see, especially, Hare 1963, Ch. 5); a recent defence of a variant of motivational internalism is provided in Garrard/McNaughton (1998). 4 The arguments to be presented in what follows are substantially unaffected by the fact that there are two slight differences between motivational internalism as introduced above and Mackie’s claim about motivation: (i) Mackie’s motivational requirement is meant to be applied only to those having moral knowledge (rather than just belief); (ii) motivational internalism does not require the presence of an overriding motive.

Moral Values and the Fabric of the World

127

perform a certain action, in which the person is aware of this being so, while refraining from acting as knowingly required. When faced with examples of this sort, those wishing to maintain motivational internalism typically hold that the person either does not seriously believe that she is morally required to act in the way in which she actually does not act; or else she does indeed have the appropriate moral belief and is motivated to act on it, but this motivation is outweighed by a stronger motivation pulling in a different direction. In other words: there is no situation in which a person seriously believes to be morally required to act in a specific way and in which she is not motivated so to act at all. How is this position to be assessed? There surely are situations in which a person does not act according to what appears to be her moral belief simply because she does not seriously entertain that very belief. And there are situations in which somebody’s motivation to act according to his seriously held moral belief is too weak to become effective. The crucial question is whether situations are possible in which somebody seriously entertains a moral belief without being motivated to act accordingly at all. The issue is further complicated by the fact that cases which appear to be examples for situations of this kind can be reinterpreted in one of the two manners just described, since there are no independent criteria for assessing whether a person seriously believes an action to be morally required (and it is difficult, if not equally impossible, to establish independent criteria on whether she is motivated at least to a minimal degree). The truth of motivational internalism depends on the adequacy of so reinterpreting all seemingly recalcitrant cases. Since it is hard to imagine how the required criteria of adequacy can be formulated in a way which is theoretically neutral with regard to the question as to whether motivational internalism is true, the issue cannot be decided by an analysis of example situations. There is, however, a prominent way of arguing in favour of motivational internalism which might seem to start off from theoretically neutral ground. Motivational internalists sometimes point out that within motivational internalism there is no mystery about how moral motivation arises since moral motivation is simply there (provided that the person has an appropriate moral belief). Those rejecting the claim of a necessary connection between moral judgement and moral motivation, however, are supposed to face the problem of explaining how moral motivation can arise at all.

128

Thomas Schmidt

Yet, this way of construing the matter misses an important point. Provided that both motivational internalists and their opponents want to be able to say that someone should act according to his moral insight although he actually does not (since either his motivation is too weak or else there was no appropriate motivation at all), both accept the idea of a normative connection between moral insight and moral motivation. To the extent to which those rejecting motivational internalism have to be able to explain how moral motivation arises when none was present before, motivational internalists are required to explain how an existing motivation can be made stronger than it was before. As regards the problem of explaining how an appropriate moral motivation arises, motivational internalists and their opponents are theoretically on a par. To sum up: there are good reasons for calling both the validity and the internalist premise of the motivational part of Mackie’s argument from queerness into question. Consequently, there is no straightforward way of attacking moral realism on grounds of motivational considerations. 5. Consider again the passage quoted at the outset of the preceding section. Objective moral values, according to Mackie, are queer in that they would not only provide the person with a motive but also with ‘a direction’. Mackie fleshes out this metaphor shortly afterwards by saying that “something’s being good […] tells the person who knows this to pursue it”, an objective good would have “to-be-pursuedness somehow built into it”, and a wrong course of action “would have not-to-be-doneness somehow built into it” (all quotes from Mackie 1977, p. 40). Many believe that Mackie has indeed identified a serious problem of moral realism here. But what exactly is this problem? Mackie’s choice of words such as “direction” and “to-be-pursuedness” is the right place to start. Although people can be indifferent with regard to ordinary facts which they come to know of, there is reason to believe that this is not so in the case of moral facts. The point can be stated as follows: there is a significant difference between, for instance, the judgements that lying is morally wrong, on the one hand, and that lying is widespread, on the other. One way to pinpoint the core of this difference is to maintain that judgements of the first type imply (or even are) judgements about reasons for action; whereas judgements

Moral Values and the Fabric of the World

129

of the second sort do not imply anything of the kind. Insofar as I believe it to be morally right not to lie, I am committed to accept that I have a reason (or that I ought) not to lie. This description of the matter is, I take it, a first step in fleshing out the idea that judgements about moral rightness (and other moral judgements) are normative as opposed to empirical judgements such as the one that lying is widespread. This example also shows that whether a judgement is normative or not depends on the way it is used: one can imagine contexts in which someone, by claiming that lying is widespread, means to say that there is a reason for doing something about this. According to the idea of moral normativity just introduced, to claim that moral judgements are normative is to maintain normative internalism as introduced above. With regard to this preliminary understanding of (moral) normativity, Mackie’s point can be stated as follows: those believing that moral judgements are normative are required to account for the difference between judgements which are normative and those which are not in a theoretically illuminating way. Moral realists subscribing to the view of moral normativity under discussion can try to do so by maintaining that those facts which normative judgements aim at correctly describing have a property (“to-be-pursuedness”, “not-to-be-doneness”) which ordinary facts do not have. Mackie holds that this property is queer. There is, however, a much stronger point against normative moral realism as I shall call the version of moral realism just introduced: positing a property which normative facts share and other facts lack has to accomplish something – and it is hard to see what this could be. To explain the difference in kind between a true normative judgement and a true non-normative judgement by stating that the first judgement is made true by a normative fact, while the second judgement is rendered true by a non-normative fact, is no explanation at all.5 5

Shafer-Landau, in his recent defense of normative moral realism, acknowledges this problem (Shafer-Landau 2003, p. 205), and goes on to argue that there are limits to explaining claims about normativity to normative sceptics (ibid., pp. 209-211). This latter claim is certainly true, but it does not imply that the notion of a normative moral fact can do any illuminating philosophical work.

130

Thomas Schmidt

The proposed understanding of the problem Mackie has tried to capture in saying that moral values are queer is for two reasons an improvement of Mackie’s way of describing the matter. First, the objection against normative moral realism based on the argument from queerness cannot be met by pointing out that all sorts of things can rightly be called ‘queer’. Second, the problem of accounting for moral normativity within a realist framework does not presuppose a specific understanding of moral normativity. In particular, it does not depend on whether one places the idea of a moral ought or rather the notion of a moral reason at the centre of one’s conception of morality.6 6. I have introduced the claim that moral judgements are normative as stating that there is a fundamental difference between moral judgements on the one hand and ordinary descriptive ones on the other; and I have argued that normative moral realism cannot account for this difference in an illuminating way. In order to place the matter in a broader perspective, I shall now briefly indicate why the two most prominent types of naturalist accounts in foundational ethical theory which are discussed in the current literature also have no resources for explaining moral normativity. Naturalist moral realists such as Richard Boyd (1988) and David Brink (1989) reject the claim that there is a fundamental cleavage between moral and empirical judgements. According to naturalist moral realism, there is nothing wrong (in a normative sense) with a person holding a judgement about the moral rightness of some action to be true while at the same time being not motivated to perform the action. This is to say: although such a person might be judged to have an unorthodox psychology, there is, according to naturalist moral realists, no deep difference in kind between this judgement and, for instance, the one that the person has curly hair. Naturalist moral realists, hence, reject the claim that moral judgements are normative in the sense introduced above.

6

Those acquainted with Wittgenstein’s ironic remarks on a “superlative fact” and on a “machine as symbolizing its action” (Philosophical Investigations, §§ 192-194) will recognize the analogy between the problem of normative moral realism pointed out above and Wittgenstein’s critique of an ontological explanation of mathematical necessity.

Moral Values and the Fabric of the World

131

Noncognitivists opt against a descriptivist interpretation of moral discourse and, consequently, reject moral realism. The point of moral judgements, according to noncognitivists, is not to aim at a correct description of a metaphysically robust realm of moral facts. Rather, moral judgements express noncognitive attitudes such as wishes, preferences or other, more complex conative mental states. Noncognitivists subscribe to motivational internalism in holding that moral motivation is conceptually tied to the states which moral judgements express.7 The outlook on moral normativity introduced above rejects naturalist moral realism, since the latter view does not have conceptual space for a distinction between normative judgements which are implying judgements about reasons for action and non-normative judgements which do not. It also is incompatible with noncognitivism since it rejects motivational internalism: moral judgements and moral motivations are not conceptually linked. Rather, there is, as it were, a normative connection between moral judgements and moral motivation, which is a metaphorical way of expressing this: persons seriously believing that they are morally required to do something ought to be accordingly motivated. 7. Neither moral realism (both in its normative and its naturalist version) nor noncognitivism, as I have argued, have resources for a philosophically illuminating account of moral normativity. This result renders some plausibility to a view which I have already mentioned in the first section of this paper: the claim that asking whether moral judgements express noncognitive mental states or else purport to describe some part of objective reality does not offer a helpful perspective on the philosophical foundations of morality. This view of the matter can be supported by another consideration. Moral realism is sometimes put forward as the attempt to secure the objectivity of morality and, to this extent, as a justification of this very objectivity with regard to a point of view located, as it were, outside of morality. An 7

This is true both of the classical theories of Ayer (1936), Ch. 6., Stevenson (1944) and Hare (1952, 1963), as well as of the recent and more sophisticated versions of noncognitivism developed by Gibbard (1990) and Blackburn (1998).

132

Thomas Schmidt

increasing number of philosophers, however, are convinced that justifying morality ‘from the outside’ is neither possible nor necessary in order to secure its objectivity. Moral justification takes place within morality, and hence the project of grounding morality with reference to an external point of view would be pointless.8 But is it pointless? There is an important ambiguity in the word “grounding” here. The above claim, if correct, only implies that one should not attempt to justify morality’s objectivity by appealing to an external point of view. It does not show that there is no point in looking at morality from the outside for reasons other than to give an answer to sceptics about morality’s objectivity. I suggest that the philosophical desire giving rise to the metaphysical question as to whether or not there is a subject matter which moral judgements are about should be interpreted as the desire to understand morality and, more specifically, moral judgements and our ways of justifying them. Is there a way of dealing with this desire without entering metaphysical issues? A prerequisite of doing so would be to flesh out the picture of moral normativity introduced above. One seems to have to engage in a more precise description of the very differences in virtue of which there is reason to introduce the distinction between normative and non-normative judgements in the first place. In what way does our way of dealing with moral judgements differ from our use of ordinary descriptive ones? How are judgements about moral reasons related to judgements about moral requirements and moral obligations? What is the relation between judgements about moral reasons and those about other reasons for action? How is reasonable moral argument to be distinguished from strategic manipulation in disguise? These are, I take it, the most important questions with which one needs to deal

8

Korsgaard (1996), see esp. p. 40, has convincingly argued that normative moral realism presupposes our confidence that we do have moral obligations and, hence, cannot be used to support this confidence. Whereas Korsgaard herself does seem to embark on the project of seeking out justificatory resources for all of morality in an external point of view, others, for instance Dworkin (1996), have argued against the possibility of an ‘archimedean’ justificatory project in ethics.

Moral Values and the Fabric of the World

133

when pursuing a project which addresses the desire to understand morality without entering a debate with the sceptic questioning morality’s objectivity, on the one hand, and without interfering with first-order normative issues ‘within morality’, on the other.9 The suggestion is that the project of describing our ways of dealing with normative moral judgements of different sorts provides what is required by a philosophical theory of understanding morality: a way of understanding what morality is about in, as it were, a nonmetaphysical sense of ‘aboutness’. In pursuing such a project, morality is placed within a larger picture of our activities – not within a framework distinguishing what is part of the fabric of the world and what isn’t, but within our lives.10

References Ayer, A. J. (1936): Language, Truth and Logic, London. Blackburn, S. (1998): Ruling Passions. A Theory of Practical Reasoning, Oxford. Boyd, R. N. (1988): “How to Be a Moral Realist”, in: G. Sayre-McCord (ed.), Essays on Moral Realism, Ithaca/London, pp. 181-228. Brink, D. O. (1989): Moral Realism and the Foundations of Ethics, Cambridge. Dworkin, R. (1996): “Objectivity and Truth: You’d Better Believe It”, Philosophy and Public Affairs 25, pp. 87-139. Garrard, E./McNaughton, D. (1998): “Mapping Moral Motivation”, Ethical Theory and Moral Practice 1, pp. 45-59.

9

A project bearing certain similarities to the one proposed here is outlined in McDowell (1987). 10 Material included in this paper has been presented to conferences at the Universities of Bielefeld and Zurich. I have greatly benefited from the discussions on these occasions and from comments by Holger Baumann on an earlier version of the manuscript.

134

Thomas Schmidt

Gibbard, A. (1990): Wise Choices, Apt Feelings. A Theory of Normative Judgement, Cambridge (MA). Hare, R. M. (1952): The Language of Morals, Oxford. – (1963): Freedom and Reason, Oxford. Korsgaard, C. M. (1996): The Sources of Normativity, Cambridge. Mackie, J. L. (1977): Ethics. Inventing Right and Wrong, Harmondsworth. McDowell, J. (1987): “Projection and Truth in Ethics”, The Lindley Lecture, University of Kansas; reprinted in S. Darwall/A. Gibbard/ P. Railton (eds.) (1997): Moral Discourse and Practice. Some Philosophical Approaches, New York/Oxford, pp. 215-225. Moore, G. E. (1903): Principia Ethica, Cambridge. Putnam, H. (2004): Ethics without Ontology, Cambridge (MA)/London. Scanlon, T. M. (1998): What We Owe to Each Other, Cambridge (MA)/ London. Shafer-Landau, R. (2003): Moral Realism. A Defence, Oxford. Stevenson, C. L. (1944): Ethics and Language, New Haven. Wittgenstein, L. (1958): Philosophical Investigations, transl. by G. E. M. Anscombe, second edition, Oxford.

Theo van Willigenburg Conceptual Analysis, Normativity and the Empirical

Introduction In my view the debate on normativity and naturalism is in general a very confused debate, and this confusion already starts with the way the debaters understand the main terms of their discussion (‘naturalism’ and ‘normativity’). This confusion is nourished by taking the questions on which the debate turns to be primarily of a metaphysical kind. I believe that the interesting debate is really about epistemological questions. Most philosophers, however, take the debate on normativity and naturalism to be a debate about how we may account for normativity in a world that can be completely described in natural, preferably physicalist terms. If physicalism is the position that a complete account of everything contingent in our world can be given in terms of a limited set of physical particulars, properties and relations, then they take the debate to be about how we can account for normativity in these terms. So they take as basic to the discussion questions about the shape of a comprehensive metaphysical account of normativity in terms of more basic, naturalist notions. I do not deny that these are important and interesting questions. My contention is that the metaphysical answers given are often much less interesting and hardly illuminating. Take Frank Jackson’s cognitivist and reductionist account of moral properties and moral content. According to Jackson, normative properties are just descriptive natural properties, which means that there is nothing sui generis about them. A property says Jackson, following Lewis , is ‘the set of all its instances throughout the possible worlds’. So the property of being right, is the infinite disjunction of all the descriptive properties (of acts, presumably) that allows us to call those acts right. We use normative terms, like

136

Theo van Willigenburg

‘just’ or ‘laudable’ or ‘permissible’, to capture certain similarities among those descriptive properties, but these normative terms do not refer to some extra feature or fact over or above the descriptive features, for “we could in principle say it all in descriptive language” (Jackson 1998, p. 125). I am afraid that as such this view is not very interesting or illuminating. It is not very interesting until is becomes clear what is exactly meant with ‘descriptive properties’ in distinction from normative properties (if descriptive means ‘as described by the natural (and social?) sciences’, the question is: described at what point in the future?). And this view is not very illuminating because it seems to introduce a very formal class of properties: the property of being a patternless disjunction of properties. As Sabina Lovibond remarks in an exchange at the 2003 Joint Conference of the Aristotelian Society and the Mind Association: … a ‘property’ taking the form of a possibly infinite, yet patternless, disjunction would not genuinely deserve that name; any determinate character we could picture it as having would be borrowed from the ethical property through which our thought makes contact with it, so that in itself it could be nothing to us but (say) ‘that possible infinite disjunction of descriptive circumstances, whatever it is, that constitutes someone’s being a spoilt brat’. And if, as McDowell argues, there are no conceptual resources at the so-called ‘descriptive’ level that would enable us, on their own, to achieve autonomy in recognizing new instances of spoilt brats, then the ‘descriptive property’ with which the property of being a spoilt brat is to be identified will fall back into the role of a quasi-Kantian noumenon, a mere metaphysical placeholder (Lovibond 2004, p. 194). Jackson identifies normative properties with patternless disjunctions of non-normative ones, but the problematic result is that we have no epistemological access to these natural (non-normative) properties independent of our grasp of the normative concepts that we use to refer to the normative properties that are realised by these patternless non-normative ones. Jackson acknowledges this. We need ethical terms, says Jackson, “to handle” the infinite disjunctions by which ethical properties are in fact describable:

Conceptual Analysis, Normativity and the Empirical

137

…ethical language may be needed in practice to capture the similarities among the various descriptive ways that (…) constitute ethical nature, but ethical properties are, nevertheless, possibly infinitely disjunctive descriptive properties- there is nothing more ‘there’ other than the relevant similarities among those descriptive ways (Jackson 1998, p. 124). But, how interesting and illuminating is the idea that ethical properties are really nothing more than infinitely disjunctive descriptive properties, if we can only have access to these properties via non-descriptive ethical parlance. The interesting question here seems not to be whether there really are such disjunctive properties. The really interesting question is an epistemological one: how can we have access to these properties? Jackson contends that it is, of course, not “by magic” that we gain mastery of ethical terms. But he does not answer the question how we come to identify the non-normative ‘realiser’ properties to which ethical concepts refer and how we can scrutinise our application of ethical terms in view of growing empirical knowledge about the disjunctive descriptive properties that constitute ethical nature. The naturalism that Jackson defends would be illuminating if it were to show possibilities for us to identify non-normative ethical ‘realiser’ properties in such a way that our empirical knowledge of them could yield normative knowledge or could count as counterevidence against our moral beliefs. Theories about the metaphysical shape of normative properties are of interest in so far as they provide for input into the debate about how normative knowledge and empirical knowledge hang together. In my view, the debate on naturalism and normativity should be a debate on the question whether what is normative is dependent on what can be known empirically. Is there a strict is / ought distinction or is there a sense in which what is normative is (partly) dependent on what can be known empirically? Intuitively, it seems that there is some relation between what is normative (what we regard as good, beautiful or admirable etc.) and our physical and psychological ‘make up’ that results in typical human sensibilities and concerns. Take our abilities of aesthetic evaluation. Why would people rather listen to a Bach Trio Sonata or a Lied from Schubert than to some bunch of uncoordinated sounds produced by randomly hitting the manual

138

Theo van Willigenburg

of a piano that is out of tune? Why do we think that a piano is out of tune in the first place? Well, is seems that our love for tonal and rhythmic (instead of atonal and arrhythmic) music has something to do with the structure of the specific auditive sensibilities that we have as humans. Whether noise is counted as music, or even beautiful music, seems to depend on the occurrence of a particular ‘match’ between our auditive sensibilities and certain physical facts about the frequencies and amplitude of sound waves. Some constellations of noises, forms, colours or movements will have aesthetic meaning to us because of the way we are disposed to relate to them. Similarly, we could say that we are disposed to react to acts, states of affairs and persons in a moral way, because of the particular ‘match’ between features of these acts, states of affairs or people and concerns springing from our physical and psychological ‘make up’ as humans. In general, it seems that that there is a relation between what is normative to us and what can be known empirically about human nature. The important question is then whether changes in our human sensibilities that we can describe in an empirical way would result in changes in what is normatively the case (in what is right, wrong, good, bad, beautiful, awful etc.). Say that our auditive sensibilities would change, because of some changes in our biology caused by, say, the global warming process. Say that our sensibilities change so radically that we cease to appreciate tonal music. Bach’s Woltemporierte Klavier comes to sound like horrible noise to us, while we now appreciate the wonderful music that we hear if we listen to the rain falling on the pavement or the wind blowing through the trees. Would that mean that Bach’s Woltemporierte Klavier ceases to be a great work of art? Say that the human sensibilities and concerns that are at the basis of our moral judgements and evaluations would change radically. Would that mean that what is morally right, good, wrong or bad were to change too? This is the question I would like to address in this paper.

Conceptual analysis: semantics and essence I am going to develop an answer to this question by making a comparison to properties of which we have the same intuition as in the case of

Conceptual Analysis, Normativity and the Empirical

139

aesthetic and moral properties, namely that their instantiation somehow is related to a ‘match’ between typically human sensibilities and the world. Colour-properties are paradigmatic for the kind of properties I am thinking of here. The intuition that we have about aesthetic properties, moral properties and ‘secondary order’ properties does not simply boil down to the idea that the instantiation of these properties is constitutively dependent on human opinion. The intuition is rather that these properties instantiate because aspects of the world on the one hand and human sensibilities and concerns on the other hand are in some sense ‘made for each other’. This differs both from the thought that the awareness of these properties is due to some faculty that traces them in reality (a straightforward recognitional view), and from the thought that the awareness of these properties is due to some faculty that projects them in reality (straightforward projectivism) What I will do is try to make explicit the intuitions that we have about the nature of ‘secondary quality’ properties and moral properties, by making explicit what is implicit in our use of the concepts that refer to such properties. As competent users of ‘secondary quality’ concepts and moral concepts we treat at platitudinous certain ideas about the appropriate application of these concepts and thereby also about the nature or essence of the properties they refer to. By making these intuitions or platitudes explicit and trying to systematise them (this is just what conceptual analysis is all about1) we may be moving into the direction of a theory that clarifies what is typical or characteristic of what we treat as platitudinous.2 We may come up, for instance, 1

For a defence of such a understanding of conceptual analysis see Smith (1994), pp. 37-41 and Smith (1998). 2 Conceptual analysis makes explicit and summarizes the network of ‘folk’ intuitions surrounding a concept. These summaries are the start of further theoretical systematizations that teach us something about the nature of a concept and about the essence of the properties the concept refers to. The summaries and systematisations are always critical or normative: conceptual analysis may correct our pre-reflective judgmental and inferential habits connected to a concept, thus improving our competence as owners of the concept and the network that surrounds it. Still, good conceptual analysis should not result in a correction of our ‘folk’ intuitions so as to effect a ‘change of subject’ or ‘shift of category’. Conceptual analysis is a normative and hermeneutical exercise (critical and modest). Conceptual analysis should do justice to the norms that the practice of concept use reveals (it should avoid resorting to ‘error theories’).

140

Theo van Willigenburg

with a response-dependence theory of colours or values that aims to provide for a systematic account of the kind of dependency of the instantiation of these properties on typical human sensibilities.3 It is important to see, however, that the intuitions that we have concerning the dependence of the instantiation of normative properties and colour properties on our human nature do not necessarily enter into the semantics of the concepts that we use to refer to these properties. We have to distinguish the semantic identity of concepts from what is typical of the properties to which they refer. Response-dependence accounts, for instance, tell us that essentially there is some relation between evaluative properties and human reactions, sensibilities and concerns, just as there is an essential relation between the instantiation of colour properties and the workings of specific perceptual mechanisms under suitable conditions. Response-dependence accounts formulate the idea that something is red or admirable or good just in case it is disposed to look red or to elicit feelings of admiration or favourable responses in normal observers in suitable circumstances. As such the response-dependence account is an account about the nature of the property of being red, admirable, beautiful or good. However, the idea of responsedependence not necessarily enters into the semantics of concepts like ‘red’, ‘admirable’, ‘beautiful’, or ‘good’. Essentially redness is the disposition to look red, just as fragility is the disposition to break, but semantically redness is not a disposition like fragility or solubility. Semantically redness is an ostensive property: a property that is directly available to be picked out in perception. Redness is visually present in such a way that the only way to learn the use of the words ‘red’ and ‘redness’ is by being presented with red objects and having the visual redness-experience. So, a dispositional or response-dependence account of the essence of properties need not be part of the semantics of the concepts that we use to refer to those properties. 3

Another example: it may be that our platitudes about the thinking procedures that we use to scrutinize our evaluative and normative judgments can best be summarized by invoking a coherentist epistemological theory, like a reflective equilibrium theory. Invoking such a theory will not only summarize the platitudes about what we mean with ‘normative argumentation’ and ‘justification’, it will also inform us about the nature of the argumentation and justification involved.

Conceptual Analysis, Normativity and the Empirical

141

Still, in many cases there will be some relation between the nature of a concept and the nature of the property the concept refers to. This is for instance clear in the case of evaluative concepts that correlate to bodily sensations and affective states, like ‘nauseating’, ‘disgusting’, ‘tasty’, ‘shymaking’, ‘horrific’ or ‘boring’; and also in case of evaluative concepts that, more generally, correlate to a desire or pro-attitude, like ‘agreeable’, ‘appealing’, ‘charming’ or ‘repellant’. These concepts refer to evocative properties, i.e. the property of evoking in us a feeling of disgust or attraction. But the evocative character of these properties is also directly heard in the semantics of the concepts used to refer to them.4 In the case of evaluative concepts like ‘admirable’, ‘despicable’, or ‘preferable’ things are a bit more complicated. These evaluative concepts have something in common with concepts like ‘disgusting’ and ‘boring’ or, more generally, ‘appealing’ and ‘repellant. The evocative character of the evaluative property of being admirable or despicable is directly heard in the semantics of the concepts used to refer to them. Something is admirable if it evokes admiration. Still, there is an important difference. We say that something is nauseating or disgusting if it evokes actual nausea or disgust in responders in normal circumstances. However the instantiation of the property ‘admirable’ is not actually dependent on our reaction of admiration, because the dependency is a normative dependency: someone is admirable if she deserves to be admired, i.e. if it is appropriate to admire her. We can make a distinction between false admiration and appropriate admiration, but it is difficult to make such a distinction in cases of nausea or disgust. Evaluative properties like ‘admirable’, ‘laudable’ or ‘preferable’ are not simply evocative. They are evocative in a normative way of reading it: they instantiate 4

Mark Johnston therefore speaks about “inherently sensuous values” (Johnston 2001, p. 182). Also in the case of deontic concepts like ‘right’ and ‘wrong’ there is an easily felt relation between the abstractness of their semantics (an abstractness like in ‘flat’ or ‘parallel’), and the formality of the property to which these deontic concepts refer, namely the formal property of having properties that provide for conclusive reasons to perform an act or not to perform it. This felt formality is systematized in so called buck passing accounts of deontic discourse. Such an account reduces deontic talk to reason talk, just like buck passing accounts of thin evaluative discourse (‘good’, ‘bad’) reduces evaluative talk to deontic talk.

142

Theo van Willigenburg

in case the response of admiration, praise or preference is merited (see also McDowell 1985; Wiggins 1987; Smith 1998). We regard something as good, beautiful, admirable or preferable if it merits or makes appropriate such a way of regarding it. Whether a response is appropriate is not just an empirical question, it is always also a normative question. Still, in answering that question empirical data cannot be ignored. Whether a good or state-of-affairs is (prudentially) preferable seems to depend partly on what has appeal to creatures like us and what can reasonably be thought as enhancing the quality of a human life. Whether an act is maleficent partly depends on data about what is harmful to vulnerable creatures like homo sapiens. Whether a person is courageous cannot be determined independent of empirical information about what counts for weak, but pretty smart creatures like humans as reckless behavior and what counts as cowardly behavior, having it that courage is defined as the middle between recklessness and cowardness. It so seems that answering the question whether the attribution of an evaluative property is appropriate requires some reference to characteristics of human creatures to which we have empirical access. This is clearly visible in the semantics of so called ‘thick’ evaluative concepts like ‘courageous’ or ‘generous’ that have a descriptive and a prescriptive side to them. But if the essence of all evaluative and deontic properties is such that their instantiation is somehow dependent on typically human sensibilities and concerns of the evaluator, then there is an empirical side to all normative evaluation. The question that interests me is whether and in what sense our moral judgments would change if the empirically determinable human characteristics on which the correctness of these judgments hinges were to change. Let us make the comparison between colour-judgments and moral judgments.

Comparing colour and value: the issue of rigidification A response-dependence account of colour and value has to make explicit our intuitions about the way attributions of colour and value are relativized to the kind of creature that attributes. In the words of Frank Jackson:

Conceptual Analysis, Normativity and the Empirical

143

The relativity to kinds of creatures arises from the fact that which properties of the world around us stand in the right relation to certain experiences for those experiences to count as presentations of the properties is, in part, a matter of how the creatures having the experiences are, just as which kinds of intruders a burglar alarm latches onto is in part a matter of how the alarm is made, and which weather conditions a barometer records is in part a matter of how the barometer is calibrated (Jackson 1998, p. 95). The question is what this relativity to kinds of creatures exactly involves. In what sense are these attributions dependent on the (possibly changing) characteristics of that creature. If our sensitivity for colours were to change radically, so that everything that appeared as yellow came to appear as blue to us (and the other way around), would we say that yellow things have become blue (and blue things yellow) or would we say that what looks blue actually is yellow (and the other way around)? My intuition is that the world of colours would have changed relative to a change in human perception. Perhaps in the first year ‘after the change’ people would still say: “this blueness is what we used to perceive as yellow”, but such references will soon become redundant. I side here with Simon Blackburn: It is not altogether simple to characterize the ‘mind-dependence’ of secondary qualities. But it is plausible to say that these are relative to our perceptions of them in this way: if we were to change so that everything in the world which had appeared blue came to appear red to us, this is what it is for the world to cease to contain blue things, and come to contain only red things (Blackburn 1985, p. 14). But, says Blackburn, moral categorizations are not in this way relative to our sensitivities: The analogue with moral qualities fails dramatically: if everyone comes to think of it as permissible to maltreat animals, this does nothing at all to make it permissible: it just means that everybody had deteriorated (Blackburn 1985, p. 14).

144

Theo van Willigenburg

According to this view the instantiation of moral properties is ‘objective’ (or non-relative) in the sense of being independent of changes in our sensibilities and attitudes. What does that mean? Take the following intriguing example discussed by Peter Railton. Railton presents a science fiction case in which human reproduction through cloning has become a wide spread practice. Many choose to have full-scale adult replications of themselves, instead of having to care for a naturally born child. However, the cloning process does not produce exact replicas. Railton tells us that replica-individual… …lack a special, intrinsic interest in those from whom they are replicated or who are otherwise genetically close to them. Biological relatedness does not seem to matter in its own right to them, even after they have fully “matured” and are successfully integrated into the community, and even when they are well informed and reflective. In virtually all other respects, they are emotionally and affectively our replicas (Railton 1998, p. 72). Because of a change in sensibilities these human clones have lost any sense of the value of having and caring for kin relationships. For us kinship matters in its own right, but for them such agent-relative reasons have no appeal. They only have ‘impartial’ concerns. Now, say that after some time in the future cloning proves so satisfactory overall, that it fully replaces other forms of reproduction and that all humans only have ‘impartial concerns’. Should we say that A. although humans have become insensitive to it, this change in sensibilities would not alter the intrinsic value of relations effected among biological kin, or should we say that B. there is nothing desirable or valuable anymore about having kin relationships, as the agent-relative reasons involved in valuing such relationships have definitely lost any appeal. Railton believes that B is the right way of looking at it, because “what is intrinsically good for humans is not rigidly fixed by actual human responses, but reflects instead potentially evolving or changing human responses” (Railton 1998, p. 77). He contends that “what is intrinsically good for an individual I of kind K depends upon the nature of I and K” (Railton 1998, p. 77).

Conceptual Analysis, Normativity and the Empirical

145

Indeed, there is this relation between what is intrinsically good or right and sensibilities that rest in human nature. But can we suspend of rigidification on actual human sensibilities and concerns? Say that we all turned into sadists. Would that mean that torturing animals or maltreating children would no longer be wrong? Railton counters this kind of objection by pointing out that a change in human sensibilities would not change the intrinsic badness of wanton cruelty toward animals or children or any victim of that conduct, simply because such conduct will remain intrinsically unliked. Because moral evaluation is non-partial, we have to take into account the basic interest of others and this implies avoiding cruel behaviour. Fortunately, the newly cloned humans in Railton’s example are fully sensitive to such impartial reasons, and as impartiality is constitutive of morality we can say that their change in response to the meaning of kinship relations does not affect their moral outlook. So, Railton believes that we should not save morality by rigidifying on actual human responses as we now have them (before turning into sadists), but by rigidifying on what is constitutive of morality on a higher level of evaluative conceptualization, namely impartiality. The problem with such an approach is that it presupposes that it is possible to conceptualize moral value so as to distinguish it radically from prudential value. If we can make such a radical distinction, then there seems to be no problem in contending - as Railton does - that what makes the life of a human being good and worthwhile (prudential value) is closely related to the fluctuating sensibilities and concerns of humans. Kinship relations have no meaning whatsoever for the newly cloned humans, so it seems that we cannot hold that having these relationships and caring for them would contribute to the flourishing of their lives. A change of sensibility must result in a change of prudential value. But can we make such a clear separation between what is prudentially valuable and what is morally valuable, given the fact that also what counts as moral values is in some way dependent on typically human sensibilities and concerns, just as with prudential values? Railton believes that morality is limited to the requirement of impartial consideration of the basic interests of others. But does that mean that there is nothing morally wrong with a mother who abandons her children in order to work 24 hours a day for Oxfam? And does that mean that from a moral point of view the man on the Titanic who

146

Theo van Willigenburg

leaves his own little son behind in order to save the two daughters of the steersman must be praised? Able users of the concept ‘moral’ will not think that agent-relative concerns have no moral import and that the only thing that counts in moral judgment is the impartial consideration of the basic interests of all. Our folk intuition about what morality is all about does not limit moral concerns to the important concern for impartiality. And there is no good reason for a reversionary attempt to replace our ‘folk’ use of the concept ‘moral’ with a more philosophical one. For if we think about why impartiality plays such an important role in our idea of morality, it seems that we are led to consider what makes human being into the kind of creatures that deserve equal consideration of their basic interests. Why do humans deserve equal worth and respect? I am inclined to think that human beings have equal worth because their reflective capacities turn them all equally into sources of normativity. Humans are, given their reflective capacities, capable of asking normative questions about the good and the beautiful, and they are capable of answering those questions. As such they come to value things, or to disvalue them. The human capacity for critical reflection is the source of what is taken as normative. Humans spread value around the world and are as such equally to be valued. But not only this capacity of critical reflection is essential to being human. Equally essential is the fact that we may be captivated by love and partial consideration. Harry Frankfurt speaks about the ‘necessities of love’: we have concerns that we cannot help having. The love of parents for their children is the paradigm of such inescapable concerns (Frankfurt 1999). These concerns are not irrational. They provide for focus and structure to our practical life. They give us the identities or self-conceptions under which we value ourselves. I am not just a human being, I am also a father to my foster son, a scholar, a amateur musician and lover of the French kitchen. There is a dynamics between these self-conceptions and the things in the world that I take to be normative and valuable. Our life of rational reflection derives is impetus and structure from what we love and are concerned about. If moral evaluation has to do with what we owe to each other given what is essential to us as human beings, then we cannot say that love for one’s kin is completely void of moral value. Love for kinship is of moral value,

Conceptual Analysis, Normativity and the Empirical

147

because it says something about who we are. It says something about what is essential to ‘us’ humans. The new (cloned) ‘humans’ in Railton’s example have lost something that seems to be definitive of being a human being. This is why love for kinship cannot just be considered to be a fluctuating nonmoral (‘just’ prudential) concern. Caring for one’s family seems to carry with it moral value. Partial concern is to be included in what we owe to each other. Saving moral value by rigidifying on what is constitutive of morality in terms of impartiality therefore cannot do the whole job. What is essential of being morally valuable and right is also related to what we now understand to be constitutive of being a human creature. We need to rigidify the reference of moral concepts in order to avoid a devastating relativity. We need rigidification in order to fix what is essential for the properties to which moral concepts refer. Compare this to the rigidification involved in fixing the essence of the properties to which natural concepts refer. H2O is the essence of water, a fact that has been established a posteriori (via empirical research). That H2O is what makes water into water means that, if on another planet we were to find that the watery stuff that on our planet has the chemical structure of H2O there has a different chemical identity, then we would not call the watery stuff on that other planet “water” but “something that looks like water”. So the essence of water is rigidly determined by the physical nature of our world. Similarly, what is essential of being morally valuable and right is rigidly determined by human nature: that what turns human beings into human beings. And, we have at least partial empirical access to factors that make a being into a human being. We can to a large extend account for this in psychological and sociological terms. Of course, humans may change. But this change will be within a range of what can genuinely be called ‘human’. Such changes will result in changes of value. But if changes result in humans loosing what makes them human, such changes will not result in a change of value. Rigidifying the reference of moral concepts by relating moral evaluation to what we count as typical human sensibilities and concerns is a way of specifying what we mean if we say that a state-of-affairs or person merits

148

Theo van Willigenburg

a particular moral response. A response is merited if it results from a match between what we rigidly describe as typically human concerns (e.g. sociality, love for kin etc.) and features of acts, states of affairs and persons. The rigidification needed is not directly to specific responses, but to human character as we understand and value it.5 Such a ‘background’ way of rigidification is enough as a ‘stopper’ to prevent that if we were all to change into sadist, torturing children would be morally all right.

References Blackburn, S. (1985): “Errors and the Phenomenology of Value”, in: T. Honderich (ed.): Morality and Objectivity, London, pp. 1-22. Frankfurt, H. G. (1999): „Autonomy, Necessity, and Love“ and „On Caring“, in: Necessity, Volition and Love, Cambridge. Jackson, F. (1998): From Metaphysics to Ethics. A Defence of Conceptual Analysis, Oxford. Johnston, M. (2001): „The Authority of Affect“, in: Philosophy and Phenomenological Research, LXIII, pp. 181-214. Lovibond, S. (2004): “Naturalism and Normativity”, in: Proceedings of the Aristotelian Society CIV, Part 2, pp. 185-201. McDowell, J. (1985): “Values and Secondary Qualities”, in: T. Honderich (ed.): Morality and Objectivity, London, pp. 110-129. Railton, P. (1998): “Red, Bitter, Good”, in: European Review of Philosophy, 3, pp. 73-90.

5

My project could be partly described as ‘bringing anthropology to the Kantian story’. The Kantian influence is visible in the arguments available to prevent counting e.g. racism as a characteristic feature of humans and therefore as something to be valued. Racism has to be countered because it conflicts with the idea of profound respect for humans as equal sources of normativity.

Conceptual Analysis, Normativity and the Empirical

149

Smith, M. (1994): The Moral Problem, Oxford. – (1998): “Response-Dependence Without Reduction”, in: European Review of Philosophy 3, pp. 91-114. Wiggins, D. (1987): “A Sensible Subjectivism?”, in: Needs, Values, Truth, London, pp. 185-214.

Ulvi Doguoglu Naturalism and Rule-Following Practices: Finding Fault with Kripke’s Notion of Objectivity1 The normativity of rule-following is, according to Wittgenstein, bound to practice. For some this makes it quite difficult to see how rule-following can remain objective notwithstanding. An exemplar occupied with such difficulties of reconciling objectivity and rule-following as a practice is Kripke’s influential reading of Wittgenstein. In the following I claim that this reading is based on a specific naturalist stance influenced by Quine that entails a distinct notion of objectivity which undermines any attempt at reconciliation from the very beginning – according to this stance rule-following needs to be naturalized leaving no room for normativity. I contrast this approach with a naturalist stance based mainly on McDowell’s work which takes rule-following practices as natural. In this view normativity can be accounted for as an objective aspect of practices. 1. Normativity is a central issue not only in moral or practical, but also in theoretical philosophy. In all philosophical disciplines, normative issues – e.g. the normativity of meaning in the philosophy of language, accounts of conceptual content in the philosophy of mind, or the normative authority of moral requirements in practical philosophy – are intimately connected to the rule-following problem raised by the later Wittgenstein. This very problem concerns the relation between a rule and the (course of) action that is in accord with it (PI, §§ 198-202):2 how can a rule determine what is in accord with it? On the face of it, two answers to this question suggest themselves: Platonism and interpretationism. The Platonist answer contends that the 1

I am indebted to Jonna Truniger for sedulous corrections and for sprucing up my English; also I owe thanks to Matti Eklund, Johannes Giesinger, Alessandro Maranta, Marc Neumann, Chiara Tabet and Edmundo Ptero for various substantial comments on earlier drafts and for linguistic suggestions. 2 For Philosophical Investigations see Wittgenstein (1958).

152

Ulvi Doguoglu

rule intrinsically determines what is, or is not, in accord with it, while the second answer holds that the accordance between rule and action is determined by interpretation. Wittgenstein, however, argues to the conclusion that both answers paradoxically fail. A rule by itself does not intrinsically determine what accords with it; but neither does an interpretation, since it stands in need of further interpretation thus leading to an infinite regress. 2. According to Wittgenstein, the solution to the paradox is to be found in the claim that rule-following is a practice, which is to say that the rule determines which actions accord with it not all by itself but rather within a practice and that, usually, within such a practice no interpretation is needed; within a practice we are able to follow rules blindly. This solution of Wittgenstein’s includes the idea that rule-following remains objective in the sense that there actually is a difference between seeming right and being right, or between “thinking one was obeying a rule” and “obeying it” (PI, § 202). This much being uncontroversial, what is contentious is concealed by the restriction of the objectivity of rule-following to a practice and hence the question arises whether or not we can make sense of objectivity bound to practices. In order to get clear about these matters, the notion of practice, its role in the solution to the paradox, and the notion of objectivity need to be cashed out. 3. Saul Kripke proposed an influential reading of Wittgenstein on rules and private language in which he pursues a sceptical line concerning the objectivity of rule-following (Kripke 1982).3 According to Kripke, Wittgenstein develops a sceptical paradox concerning meaning and rule-following,4 and thus, since the paradox is unavoidable, proposes a sceptical solution.

3

In what follows I will not discuss the details of Kripke’s argument any further than necessary. I will neither cite critical evaluations concerning the inadequacy of Kripke’s exegesis in terms of a sensible exegesis of textual and historical evidence nor those concerning the systematic matters (for the former issue see Hacker 2001; for the latter see, e.g., the contributions of John McDowell, Colin McGinn, and Paul Boghossian collected in Miller & Wright 2002). 4 Kripke indeed reads the rule-following considerations without further argument as concerning meanings or semantics, a restriction we do not find in Wittgenstein and

Naturalism and Rule-Following Practices

153

The sceptical paradox goes as follows: the basic problem is to find facts about an individual which determine the relation between what she means to say and her actual use of words, that is, facts determining the relation between a rule (meaning) and the actions in accord with it (meaningful uses). In a Platonist framework, things would be quite simple: it is the meaning that determines the use of linguistic expressions. Wittgenstein, however, argues forcefully against such Platonist conceptions and puts forward a conception of meaning as use, within which meaning boils down to uses in the past, hence facing a problem of determining the relation between past uses and according future applications of words. The point is that, since the relation of the individual’s meaning (or of her intention) to her future actions is normative not descriptive, we are in need of some kind of normative machinery or some nomological explanation that would determine future cases of application. Kripke takes Wittgenstein to show that such a machinery does not exist and what is more that, since there are no normative facts, no fact at all will serve for determination. There are no such facts about an individual projecting or determining future cases of application out of past applications. But if there are no objective facts about an individual, what about subjective facts, that is, private experiences of meaning (or intending) something? Such an exit is ruled out by reference to Wittgenstein’s claim that there must be a difference between thinking one was obeying a rule and obeying it; a claim that is argued for, Kripke takes it, in various passages of PI, §§ 138-202. It is for this reason that Kripke claims that “Wittgenstein rejects ‘private language’ as early as § 202” (Kripke 1982, p. 79) and that it “… is his solution … that contains the argument against ‘private language’; for allegedly, the solution will not admit such a language” (Kripke 1982, p. 60). Hence, in this interpretation, the sceptical paradox consists in language use being neither objectively nor subjectively founded. Kripke’s sceptical point concerning the objectivity of rule-following can then be put in the following way: the relation between facts about me and what I mean or how I act is not nomologically explicable, neither by strict universal nor by probabilistic laws; there is no room for normativity in lawful necessitation. This claim is already radical, since it amounts to saying which is not uncontroversial. Yet, I shall not develop this line here and will, as Kripke, not differentiate between issues of rule-following and semantic issues in the following.

154

Ulvi Doguoglu

that there are no facts as to an individual meaning something by using a linguistic expression, and no facts as to her intending something determinate. All the same, Kripke goes even further by denying there being any subjective foundation either and by claiming that, as a matter of fact, the individual does, on her own, not mean or intend anything by her words, “the notion of a rule as guiding the person who adopts it can have no substantive content” (Kripke 1982, p. 89)5. This is Kripke’s sceptical paradox. 4. In Wittgenstein’s name, Kripke puts forward a sceptical solution approaching objectivity as close as possible. Since there is nothing, that is to say, no fact about an individual determining the normative relation between a rule and her according to it, each new application she makes “… is a leap in the dark; any present intention could be interpreted so as to accord with anything we may choose to do” (Kripke 1982, p. 55). As a consequence, Kripke holds Wittgenstein to take recourse to practice as something else than the desired factual and objective underwriting establishing at least some relation between rule and action. Since there are, according to Kripke, no objective facts that determine what I mean by the use of certain words or how I should go on in the future (the sceptical problem), the social practice of language use determines what I mean and how I should go on (the proposed sceptical solution accepting the sceptical problem). This solution presupposes the following notional landscape, consisting of three absolutely distinct notional fields or realms, namely (1) a field of notions like objectivity, factuality, reality and nature, (2) a field with notions like intersubjectivity, community, society, and practice, and (3) a field of notions like subjectivity, individuality, and privacy. These three fields are strictly separate and do not overlap, which is to say that what is objective can neither be social nor individual, what is intersubjective can neither be strictly real nor private. Since rule-following is to be conceived as a practice, it can, within this notional landscape, neither belong to the notional field (1) of objectivity, nor to the realm (3) of subjectivity. An answer can only be sought within the notional field (2), that is, within a social notion of practice. Therefore also the notion of normativity belongs to (2).6 5

Without further notice, all emphases within quotes are in the original. Bloor (1983), Bloor (1997), Kusch (2002), Williams (1999), and Pettit (2002) share Kripke’s communalist view of matters in Wittgenstein at least in general.

6

Naturalism and Rule-Following Practices

155

Practice not exactly being the solution Kripke initially was watching out for, it serves merely as a second-rate, or a sceptical solution. As belonging to the notional field of intersubjectivity, it cannot count as objective. Yet at the same time the practice of a community precludes any appeal to privacy or subjectivity. In Kripke, therefore, rule-following practices are merely considered as social practices, the notional fields (1) and (3) then being irrelevant. Wittgenstein’s difference between seeming right and being right in this view reduces to a difference between seeming right to an individual and being right according to the community the individual is integrated into. This sort of intersubjectivity is Kripke’s version of what can be called ‘objective rulefollowing within a practice’, with strict objectivity not being achievable. Rule-following cannot attain objectivity for it cannot be neatly separated from the perspective of a community. In fact, Kripke’s account does not amount to a naturalist view of rule-following, but to a naturalization of rulefollowing which “quines away” the objectivity of such practices.7 The proposed sceptical solution within the notional field (2) thereby functions as some kind of quietening placebo. 5. Kripke compares Wittgenstein’s rule-following considerations with Goodman’s new riddle of induction and with Quine’s arguments against the factuality of semantics or the objectivity of meaning and reference.8 While, indeed, Goodman’s results are comparable to Wittgenstein’s in their spirit and purpose,9 Quine’s way of dealing with philosophical problems and results is fundamentally different. Whenever Wittgenstein in his considerations comes across puzzling results, he turns back to the presuppositions that led to the puzzlement and reviews the base he started out from. Quine on the other hand insists on the puzzles he develops and claims that, as a matter of fact, we are faced with ontological relativity as well as with semantic indeterminacy. This difference in strategy is usually illustrated by the slogan that one philosopher’s 7

The verb “to quine” is due to Daniel Dennett. In the first paragraph of “Quining Qualia” he cites the entry from his The Philosophical Lexicon: “quine, v. To deny resolutely the existence or importance of something real or significant” (see Dennett 1978). 8 See Goodman (1983); Kripke (1982), p. 20, p. 58; further see, e.g., Quine (1960), or the title essay of Quine (1969); Kripke (1982), pp. 55-58, p. 114. 9 It is surprising that there exists no single study devoted at length to both Wittgenstein on rule-following and Goodman on projection.

156

Ulvi Doguoglu

modus ponens is another philosopher’s modus tollens.10 It is, among other things,11 his modus ponens attitude that leads Kripke to misread Wittgenstein in a Quinean spirit and develop a sceptical interpretation. Kripke is well aware of the weighty differences between Wittgenstein’s problem of rule-following and Quine’s discussion of the indeterminacy of meaning, similarities notwithstanding. According to Kripke, the main difference between Quine and Wittgenstein is this: Wittgenstein does not limit himself to the kind of behaviourism Quine favours, since he does not only consider behavioural facts as candidates for a possible solution of the paradox but also a variety of dispositions, qualia, images, mental states or entities (Kripke 1982, pp. 22-54). In the eyes of Kripke, this fact renders Wittgenstein’s scepticism even more forceful than Quine’s (Kripke 1982, p. 14). But despite this, going beyond Quine and considering mental facts (or states or objects) as aspirants to a solution, all the candidates Kripke deems worthy presuppose a naturalist notion of fact (or states or objects). In taking Wittgenstein to promote a sceptical paradox, Kripke is convinced right from the beginning that whichever behavioural or mental facts about me we are consulting they cannot, as such, determine what I am to do in order to act according to a certain rule. Therefore none of the facts Kripke examines and ponders would do as a normative fact. 6. The fundamental claim of Quine’s is – in his own wording – the following: “Naturalism looks only to natural science, however fallible, for an account of what there is and what what there is does” (Quine 1992, p. 9). Sellars, for his part, puts this commitment into a neat Protagorean formula that would surely meet Quine’s approval: “… in the dimension of describing and explaining the world, science is the measure of all things, of what is that it is, and of what is not that it is not” (Sellars 1997, p. 83).12 Since philosophical questions in general are to be pursued and answered in such a scientific spirit, 10

See, for example, Putnam (1994a), p. 280. See Fn. 4 above. 12 See, e.g., Plato (1973), 152a, for a record of the Protagorean formula, and the subsequent pages for a lengthy critical discussion of it by Socrates. 11

Naturalism and Rule-Following Practices

157

epistemology, as the project of explaining how we as “physical denizens of the physical world” (Quine 1995, p. 16) manage to develop fruitful scientific theories out of mere stimuli at our nerve endings, is turned into a chapter of empirical psychology (Quine 1969, p. 83) or cognitive science. In his last book From Stimulus to Science, Quine claims that the traditional epistemological quest for knowledge about knowledge had already lost its sense by the times of Aristotle: “Knowledge itself … outpaced knowledge about knowledge” (Quine 1995, p. 2). The tremendous progress of knowledge acquisition and its successful application have ruled out epistemological doubts concerning the reliability of these knowledge claims. Scientific realist he is, Quine simply takes for granted the ever evolving stock of scientific (and only scientific) knowledge, as well as the ways we attain it. Still, one might detect a certain tension in Quine’s writings. While access to the structure of nature is secured by scientific means and methods, nature itself ultimately remains undisclosed. In rejecting the two dogmas of empiricism (Quine 1980) and adopting a holistic web of belief, Quine commits himself to a negative concept of nature as an independent effective cause, which cannot become part of the web itself but merely impinges on it from the outside, staying, so to say, incognito, or, to use Quine’s term, inscrutable. Science therefore cannot achieve objectivity in a strict sense, but merely intersubjectivity (Quine 1995, pp. 44). Eventually, Quine claims that all we can grasp of nature is just its structural traits, a claim he puts into a neat slogan that in turn might find Protagoras’ approval: “Save the structure and you save all” (Quine 1992, p. 8). Hence, Quine is committed to a concept of nature which remains a conceptually unattainable “Ding an sich”, a “noumenon” in the negative understanding, as Kant called it (Kant 1990, B294-315). Hilary Putnam’s charge, argued for in various of his writings (see, e.g., Putnam 1994b), is that Quine is committed to metaphysical realism. The tension I mentioned above results from a combination of this notion of nature remaining undisclosed with Quine’s declared scientific realism. It is resolved by a closer look at Quine’s notion of “scientific objectivity”, which amounts to intersubjectivity only:

158

Ulvi Doguoglu

Science ventures its tentative answers in man-made concepts, perforce, couched in man-made language, but we can ask no better. The very notion of objects, or of one and many, is indeed as parochially human as the parts of speech; to ask what reality is really like, however, apart from human categories, is self-stultifying. … (Quine 1992, p. 9; see also Quine 1995 pp. 44f.) It seems, then, that the appropriate characterization of nature in Quine’s global structuralism (Quine 1992) would, in allusion to Wittgenstein’s thought experiment of the beetle in a box (PI, § 293), be the following: Quine’s nature is a nature concealed in a box nobody has access to; it thereby just cancels out. On the other hand, since in Quine’s naturalism all that belongs to the notional field (3) is reduced to overt behaviour, subjectivity cancels out as well. – If this rough sketch of Quine’s naturalism is correct, it is devastating, since it seems that the concept of science Quine favours faces the same problem as the notion of normativity of rule-following: the only place being left over for human (including scientific) practices is the notional field (2). In his discussion of various failing strategies for naturalizing rule-following and in arguing to the conclusion that objectivity cannot be attained if rule-following is a practice, Kripke draws on a notional landscape, I claimed, that separates into three distinct notional fields, the realm of the objective, or natural, the realm of the subjective, or private, and a realm in between, the realm of social intersubjectivity. This notional trichotomy is due to forms of naturalism of a broadly scientistic brand à la Quine in which objectivity is taken to be absolute and separate from human practices, let alone individual subjects. In effect, it is just such a Quinean naturalist conception of objectivity and factuality which is implicit in Kripke’s discussion of possible facts determining the relation between past and future action. 7. Still, from a pragmatist point of view, naturalism suggests itself as an attractive position for various reasons. A characterization of naturalisms may take a negative and a positive form:13 negatively described it aims at avoiding various problematic “–isms”, turning against epistemological scepticism, 13

For informative overviews and discussion as well as further references see, e.g., Kitcher (1992), Rea (2002), Rosenberg (1996), or Shook (2003).

Naturalism and Rule-Following Practices

159

transcendentalism or apriorism and various forms of Platonism, dualism or supernaturalism (McDowell 1996, p. 77)14; what is more, it sets itself against various forms of relativism, culturalism or subjectivism. Positively described, the core idea of naturalism is to take nature to be the only and ultimate ground for objectivity in which recourse to that same nature amounts to appealing to an independent authority that decides between correct or incorrect. Furthermore, nature, or its structure, is claimed to be accessible by certain means and methods of inquiry. A third distinctive feature of naturalism is that the results of such inquiry as well as its methods are in principle subject to error and that they might need to be revised in view of recalcitrant record. In sum, what is distinctive for naturalist positions of all sorts is – in positive and negative terms – that they (i) propagate one continuous system of nature or reality or objectivity (in contrast to dualistic or pluralistic views of various provenances), (ii) uphold a pragmatic realism concerning epistemological access to this system of nature by certain means and methods (against sceptical doubts), and (iii) maintain that all knowledge claims as well as the ways to attain knowledge are principally fallible or answerable to nature or reality (against apriorism); everything might be subject to reconsideration; “no statement is immune to revision” (Quine 1980, p. 43). As regards rule-following, talk of “naturalism” wavers ambiguously between naturalism and naturalization. Restricting (i)-(iii) above, as Quine does, to natural science leaves no room for a substantial notion of normativity. Normativity in Quinean naturalism calls for naturalization, that is, for some explanatory reduction of normative phenomena to natural facts, laws and processes. But, as Kripke has shown, within such a naturalism any project of naturalization of normativity is doomed to fail. We might, though, develop a view which is properly called naturalism regarding rule-following practices by rejecting Quine’s limitation of (i)(iii) to science and holding a position that takes (at least certain paradigmatic forms of) rule-following practices to be objective natural phenomena. In such a view, the notion of objectivity will not fit Kripke’s notional trichotomy mooted above. To the contrary, since practices are not separated 14

McDowell characterizes his own sort of naturalism as “naturalized Platonism” (McDowell 1996, pp. 91-95).

160

Ulvi Doguoglu

from objectivity but are themselves taken to be paradigmatically objective from the very beginning, Kripke’s trichotomy collapses. This, then, makes room for a different notional landscape. 8. There are many reasons to dismiss Kripke’s sceptical naturalization of Wittgenstein’s rule-following considerations and his sceptical solution which accepts the sketched notional landscape. In effect, Kripke’s so-called sceptical solution amounts to just this: an eliminative account of (the objectivity of) rule-following, thereby evoking sceptical and relativistic worries concerning a fundamental part of our lives. This might lead us to reject the presupposed Quinean scientistic naturalism with the intention to save the objectivity of rule-following. Indeed, I take Kripke’s sceptical reading of the later Wittgenstein on rules as self-defeating since it fails to provide for the objectivity of rule-following practices.15 Cancelling out the objectivity and even the subjectivity of rule-following practices overturns any position that leads to this conclusion.16 As noted above, Quine’s medication against the “melancholy” of the sceptic concerning scientific knowledge is to invoke the fact that “knowledge itself … outpaced knowledge about knowledge” (Quine 1995, pp. 1f). In the same vein and with the same right one could suggest extending the line and acknowledge the fact that our knowledge concerning rule-following practices easily outpaces sceptical worries regarding it. Fundamental scepticism concerning rule-following, as fundamental epistemological scepticism, is some sort of melancholy (or, maybe more precisely, scientistic infatuation) to be overcome. At this point we might now ask ourselves in more detail what sort of naturalism would do for substantial objectivity in rule-following practices and how the notional landscape in this case has to be drawn. Let me therefore contrast these considerations of naturalizing rule-following practices with naturalism concerning these practices.17 9. A promising strategy in regard to securing the objectivity of rule-following 15

For an argument along these lines see Boghossian (2002), p. 185. 16 Hilary Putnam argues in just this vein against any position that leads to ontological relativity (see, e.g., Putnam 1994b). 17 There are, of course, non-naturalist alternatives in the field claiming, for example, that there is a fundamental cleavage between the scientific explanation of nature and the

Naturalism and Rule-Following Practices

161

is to reject the scientistic naturalist’s notional topology and her scientism, that is, to question her restriction of objective viz. natural knowledge to scientific knowledge as well as her restriction of possible access to nature to scientific means and methods. On the other hand, this strategy would hold that we should stick to the attractive sides of naturalism, namely fallibilism, and the various anti-metaphysical and anti-sceptical stances. According to the very short do-it-yourselfer’s guide to naturalism we developed above, we are to specify (i) the realm of the natural and objective, by determining (ii) the paradigmatic knowledge claims and the paradigmatic means and methods of inquiry we want to admit, and (iii) the ways in which these claims and methods might become subject to revision. First and foremost, we should take the objectivity of (at least some) rulefollowing practices for granted. As suggested above, this includes as well a rejection of Kripke’s notional trichotomy as a different view at Wittgenstein’s paradox – lines of argument pursued by authors like Hilary Putnam, John McDowell or John Haugeland.18 Albeit these philosophers do not deny the social character of language and complex rule-following practices, they claim, however, that these practices do not solely depend on us being

understanding of human society, culture and mind. In such a view the cleavage between “Erklären” and “Verstehen” stands for two projects wide apart, science on the one hand, humanities and social sciences on the other hand (“Naturwissenschaften” vs. “Geistes-” and “Sozialwissenschaften”). It contends that scientific explanation in the field of human culture and mind is doomed to fail because it misses the essential point of individuality or subjectivity. To argue thus, however, is not to challenge the Quinean naturalist, but in fact to accept her notional framework. Hence, she would accept the claim with a smile. If answers to, say, semantic facts or moral questions do not count independently from any subjective point of view but are merely relative to a particular culture and its subjects, then no wonder they seem not very coercive to her. All the same, the naturalist will not be able to resist taking the enticing further step and eliminate individuality. – It is, by the way, striking that McDowell, in his reading of Gadamer (see McDowell 1996, esp. Lect. 6), does not mention or reflect upon the opposition between “Erklären” and “Verstehen” the ensuing opposition of “Natur-” and “Geistes- und Sozialwissenschaften” which is held by Gadamer at least in the form of an opposition of truth and method. Therefore Gadamer first and foremost focuses on the notion of “Verstehen” in “Geisteswissenschaften”; still, he claims “Verstehen” to have self-standing validity also in science which cannot be reduced to method (Gadamer 1990, p. 1). 18 See, e.g., Putnam (2004); McDowell (1998b), esp. chap. 13 and 14-16; McDowell (1998c), esp. chap. 11-14, Haugeland (1998), esp. essays 1, 9, 10, 13.

162

Ulvi Doguoglu

members of a society but as well on us living with physical bodies in a certain physical environment. In a different social and physical environment we could therefore not develop the same rule-following practices. As regards language use, this is exactly the insight of semantic externalism (Putnam 1975). As a consequence, the community cannot be rock bottom where the spade turns;19 objectivity would collapse if it were merely reduced to a Kripkean substitute like “being accepted by a community (or by most members of it)”. Note that by reference to rock bottom Wittgenstein does not give a dogmatic answer but rather rejects sceptical and unsatisfying questions leading to it. This becomes obvious in PI § 217, when Wittgenstein speaks of “architectonic demands”, saying that sometimes the demanded explanation is comparable to a “Scheingesims” (ornamental coping) fulfilling no static function. In this reading, the upshot of Wittgenstein in developing a paradox is not sceptical as regards rule-following as a practice, but concerning the notional trichotomy making it impossible to see how there can be an objective difference between seeming right and being right. Another aspect that is usually neglected is the aspect of the participant subject. So, in § 217, Wittgenstein says: “This is simply how I act”20, thereby not speaking of a community (he does not say ‘This is simply how we act’), but of an individual actor. In Kripke’s account, the individual does not play any substantial role; she frisks about blindly in the dark as long as there is no community: “All we can say, if we consider a single person in isolation, is that our ordinary practice licenses him to apply the rule in the way it strikes him” (Kripke 1982, p. 88). According to Kripke, the social practice does not determine what to do, if one is on his or her own, since it is only within the community and its sanctioning the actions of a single person that we can speak of obeying a rule. Again, “… considered in isolation, the notion of a 19

“Habe ich die Begründungen erschöpft, so bin ich nun auf dem harten Felsen angelangt, und mein Spaten biegt sich zurück. Ich bin dann geneigt, zu sagen: ‘So handle ich eben.’ (Erinnere dich, dass wir manchmal Erklärungen fordern nicht ihres Inhalts wegen, sondern der Form der Erklärung wegen. Unsere Forderung ist eine architektonische; die Erklärung eine Art Scheingesims, das nichts trägt.)” (PI, § 217). 20 Anscombe translates “So handle ich eben” as “This is simply what I do”.

Naturalism and Rule-Following Practices

163

rule as guiding the person who adopts it can have no substantive content” (Kripke 1982, p. 89). Therefore not only objectivity but also the idea of an acting subject goes overboard in Kripke’s interpretation. In our alternative, though, the notion of practice is disrupting Kripke’s notional corset. Including its environment, physical and social, practice is objective, its notion is neither restricted to mere intersubjectivity nor is it opposed to the notion of objectivity. On the other hand, dispensing with privacy does not mean to rule out the specific role of an individual subject of rule-following either: the individual can be the master of her practice, she is herself engaged in following rules. In this view, Wittgenstein’s paradox is even more radical, requiring the notional landscape to change fundamentally: practice involves both subject (subjectivity) and object (objectivity) and since practice is both public and subjective, in the notional topology of this naturalism, subjectivity does not coincide with privacy. As a consequence, even Kripke’s contention that the impossibility of a private language is proved by § 202 collapses. What a private language would amount to and why it is not possible needs further argument and is not yet included in the paradox of § 201. 10. It seems essential to knowledge resulting from scientific inquiry that it is not telling us how reality should be, but just how it is. Of course, we may expect nature to be one way rather than another because our theories say so. Often, however, nature is, so to say, not “behaving” as it should, running counter to all expectations. Yet, this is not the mistake of nature but simply the error of our theory or inquiry. Nature can never be wrong, only our model of it can; it is our theory that is to blame. If we have a scientific law saying that something is not possible – say travelling faster than light-speed – we rely on this being impossible. However, if we find out that it is possible all the same and this particular law is broken, we are urged to think about revising our theory to some extent. Not so in language, or in moral philosophy. Take any moral requirement – you will find it being broken all over the world all the time. Notwithstanding, this fact will not lead us to a revision of our concepts of, say, human rights. Rather we will argue the other way round and say that it is not possible to act in disaccord with fundamental moral requirements and, still, remain morally

164

Ulvi Doguoglu

decent. Nevertheless, in a certain sense, moral requirements might be used to predict which of various physically possible actions in certain situations would be morally possible actions, that is, actions in accord with the moral requirements. The same applies to rule-following practices in general. Let us take Kripke’s example of addition. Insofar as all the physical or mental facts about an individual do not determine how to go on, the individual has, from a physical point of view, the choice to do whatever she wants; it is up to her. From the point of view of the practice of addition, though, there are strong restrictions concerning the possible answers to the question ‘What is the result of the addition of the numbers two and three?’. From this point of view the only possible (in the sense of correct) answer, is ‘five’, while any other answer is, so to say, not a possible move in the game. But again: wrong answers will not lead us to change the practice of addition. In the same sense, an illegal move of a chess player does not give reason to reconsider the conformity of certain moves. You cannot play chess with me if I do not, in practice, accept the rules of the game and act accordingly.21 The normative practice of playing chess is highly inert as regards recalcitrant experience. It is in just this vein that Putnam argues to the conclusion that his story of Brains in a Vat, while being physically possible (because no physical law is broken), is philosophically impossible (Putnam 1981, Chap. 1). The case of the Brains in Vats raises a sceptical worry as to our representational or semantic access to the world we live in by imagining that our experiences could be simulated by a powerful computer connected to our nerve endings. Hilary Putnam claims that this sceptical position is not open to us since it is unintelligible or conceptually impossible. This he shows by inquiring “… into what is reasonably possible assuming certain general premises …” (Putnam 1981, p. 16), since what is “conceptually impossible” or what is “reasonably possible” is determined not by physics but, as Putnam would state matters in recent years, by our practices, or by how we live our lives (as embodied brains) in an environment. From within our lively practices the 21

We might think of a person not accepting chess rules but playing according to them all the same – let us say, she accepts the rules in practice.

Naturalism and Rule-Following Practices

165

idea of us being envatted brains is impossible. In sum, determining the rules which guide normative practices and deciding what accords to these rules and what does not, is definitely not the sort of theorizing we come across in natural science. But does not this difference in the playing of the game of scientific inquiry and of our normative practices pose a problem for the proposed broad naturalist conception of rule-following? One might hold that it casts doubt on the possibility of reconciliation of these different games under one umbrella termed naturalism. The rejoinder to this charge is that the difference between normative and descriptive relation to the world is not as crucial as it presents itself. Both relations are, in fact, normative; the difference depends only on what Haugeland calls the “direction of fit”. The descriptive relation stands for a “world-to-rule direction of fit”, whereas the normative relation stands for a “rule-to-world direction of fit” (Haugeland 1998, pp. 305f). 11. It is such a view of normativity going both ways that leads McDowell to claim that, if anything, experience can merely play an exculpatory role in Quine’s structuralism, it cannot justify one structure rather than another (McDowell 1996, Lect. 1). Therefore, knowledge claims cannot face what Quine calls a “tribunal of experience”, since everything Quine can invoke just serves as an excuse, not as a reason. Recalcitrant experience solely causes revisions in the structure, its development being evolutionary, with the fittest structure surviving. Hence, science develops, in Quine’s view, blindly; it does not reason, but only guess shrewdly in the dark (Quine 1995, p. 46). Quine would probably not take this as an innuendo. This is obvious in Quine’s picture of knowledge acquisition (Quine 1995, esp. Chap. 2). According to Quine we start out from global “stimuli resulting” from momentary stimulations of our nerve endings, and then proceed by reacting to perceptual similarities between parts of global stimuli. From these effective similarities we go on by forming expectations, by learning and developing habits – that is by what Quine calls “primitive induction”. We then advance by “sharing information”, starting to use language by handling grammar and logic, and simulating science with our “first faltering scientific laws”, that is with “observation categoricals” (Quine 1995, p. 25). Soon bodies, or objects dawn, indistinctly at first, but focused soon. These steps represent the advent

166

Ulvi Doguoglu

of science, thereby, according to Quine, opening up the development of scientific knowledge and complex scientific theories. Susan Haack blames Quine for wavering between two concepts of scientific knowledge, since in some places he refers by this term strictly to knowledge as a result of natural science, while in others he takes it in a less terminological way to take recourse to empirical knowledge in general (Haack 1993). I agree with Haack on there being a tension in his use of these notions, in Quine’s eyes, tough, this obviously is neither tension-carrying nor wavering; to the contrary, he deliberately uses terms like “induction”, “expectation”, “learning” and the like univocally as if they were continuous and (almost) unchanging all the way down to primitive forms of induction, expectation or learning. In order to back up his natural history of our way to science, Quine sensibly renders the different stages of these phenomena more similar than they actually are. This fact prevents him from having to admit that in his description of how science develops onto- and phylogenetically the primitive expectations we develop at an early stage and which we still share with some animals are different in kind from the expectations we develop later on, when forming and checking scientific theories. Holding primitive induction to be the cradle of science and observation categoricals like “When lightning, thunder” to be already “complete … miniature scientific theor[ies]” (Quine 1995, p. 26), Quine disclaims that we might at some further stage develop forms of inquiry or of complex human practices different from scientific inquiry or practice; forms of inquiry that would allow, for example, for moral reasoning. Admitting a tension would then mean for Quine to jeopardize his whole endeavour, since the continuation of primitive notions and complex scientific notions justifies his restriction to scientific inquiry. But even if Quine were right concerning continuation, peril still crouches in another angle. 12. The mark of complex expectations (as opposed to Quine’s primitive expectations) is their being prospective in the sense that there is not only a definite range of possible outcomes, but also a certain range of impossibilia. Haugeland puts it like this: What does it mean for phenomena to “make sense”? It means that they

Naturalism and Rule-Following Practices

167

are accessible not just in their actuality but in their possibility. For the possible, as a (severe) restriction of the actual, can account for the fact that … within the conceivable, the actual has the determinate character it has, rather than various others (namely, others that are conceivable but ruled out as impossible) (Haugeland 1998, pp. 352f). Expecting things to be this or that way is the characteristic feature of our rational relation to objectivity or objective reality, and if our expectation proves wrong, we reject either the grounds for our expectation – e. g. by claiming that our theory was wrong or not fine-grained enough to account for the experience – or else we reject part of our experience, e.g. by appealing to extraordinary circumstances of perception like hallucination or a foggy sight. Such a relation to the world which has an objective, and determinate character is different from Quine’s “primitive” relation to an environment. Again, Haugeland, comparing human reaction to recalcitrant experience with a dog’s reaction, explores the matter: Consider … how the members of a family are perceivable (on a corporeal level): each has his or her own characteristic visual appearance, sound of voice, odor, way of moving, and so on; and, of course, their various parts stay attached in the same way. But suppose, one day, all these aspects started permuting: what looks like Sister sounds like Father, moves like Grandma, and smells like Kid Brother. Even the parts could mix up: Mother’s head (but Father’s hair) on Uncle’s torso with Baby’s limbs – or just two heads with no limbs or torso at all (sounding like a truck …). And moments later, they switch again, with new divisions and new participants. What would you say? Surely something like: “Egad! Am I going crazy? Am I being tricked or drugged? I can’t really be seeing this – it’s impossible”. That is, you would reject what you seemed to perceive, you would not accept them as objects. Now suppose that, instead of you, it were the family dog who came home to this. We can’t ask what it would say, because dogs can’t talk; and, of course, any estimate of its reaction at all is bound to be largely conjecture and prejudice. But, by way of counterpoint to sharpen the main point, I’ll express my own prejudice: I think the dog would bark.

168

Ulvi Doguoglu

I expect it would be disoriented and distressed, maybe even frightened. But I can’t imagine any part of a dog’s reaction amounting to a rejection of the scene, a discounting of its reality, on the grounds that it’s impossible. Though Fido can tell Sister from Brother, and humans from cats, I don’t think he can distinguish in any sense between possible and impossible (Haugeland 1998, pp. 261f). Human perception is objective in the sense that the way things can be is constitutive for our expectations – if perception does not match up to them, we tend to react like the dog and bark “Egad!”. But soon we call our perception or other mental capacities into question (“Am I going crazy?”), take the circumstances, in which we perceive these strange things, into consideration (“Am I being tricked or drugged?”), and reject what we see as impossible or incoherent. Of course, unless things alter or cease to happen again and again, we will probably widen the circle of circumstances to be reconsidered, as in an exemplary case in which we imagine physicists finding out that some fundamental constant of nature, say Planck’s constant, is staggering a little every now and then, causing intense and strange effects on things, which eventually result in weird objective perceptions of higher animals. If our everyday objectivity already involves such stepping back from our experiences, such questioning and inquiring into their reliability, even more so does complex scientific research. Scientists are, most of the time at least, way beyond barking around if something does not match their expectations. Usually they have quite clear ideas about which outcomes are possible and which are not. McDowell and Haugeland argue to the conclusion that views like Quine’s, that is views which are not able to encompass possibilia and impossibilia, or which, in Haugeland’s terms, are not “letting entities be” (Haugeland 1998, pp. 325-354), fatally fail to attain objectivity. In their eyes, Quine does not see that even for a world-to-rule (or scientific) relation of fit we ought to let things have normative authority, since “to understand normative authority of objects … is to understand their objectivity” (Haugeland 1998, p. 339).22 Only if we conceive of our experience of things as not only causally but also normatively constrained by reality, we can, so to say, submit cases to the 22

Haugeland continues that “to pursue this understanding … it will be necessary to make an apparent, and perhaps surprising, digression into subjectivity”.

Naturalism and Rule-Following Practices

169

tribunal of experience for confirmation. Normativity and objectivity end up as just two sides of the same coin. 13. In “Two Sorts of Naturalism” John McDowell discusses a provocative thought-experiment in which we are asked to imagine rational wolves (McDowell 1998a). A rational wolf, McDowell holds, would be able to step back, become conscious of what he does, roam his mind over the (physical) possibilities there are, and think or reason about how to go on. Whereas the common wolf cannot help being part of its pack and take part in, for example, hunting, the rational wolf would be in the position to see that he might shirk as well and thus, instead of partaking in the ever exhausting hunt, he could simply decide not to show up until “dinner is ready”. In other words, while common wolves cannot, at least as far as we know, reflect upon what they are doing, for they just do blindly what they always did (be it instinctive or trained), a rational wolf could consider the reasons for possible actions. With rationality, McDowell therefore claims, the wolf attains freedom. Against this background, our presumed rational wolf, reflecting on possible actions, might face the following quandary: What reasons should motivate him to exhaust himself in hunting activities although he could be content with keeping the fire and waiting for the prey to be delivered? In such a case, it is reasonable, McDowell claims, to argue within the bounds of what is good for wolves by drawing on natural facts as, for instance, joining forces in the hunt which is a matter of survival for wolves (even though, wolves being rational, they could, in principle, decide to organize their hunting patterns differently). Yet, for a rational wolf such reasons might be convincing but still not motivating; they might well record an objective requirement, but why should he act accordingly? And what does “act accordingly” mean, anyway? The upshot of McDowell’s argument is that by developing our capacities to step back from what we were doing implicitly and by developing and cultivating reasoning we acquire freedom and the capacity to see unexpected possibilities, new ways of doing things and of living our lives. Nevertheless, we should not loose our sensitivity to reasons that base on the very possibility of our practices. These are the two essentials of what McDowell, following Gadamer, calls “Bildung” or, following Aristotle, “second nature”:

170

Ulvi Doguoglu

acquisition of freedom and consciousness in our doings on the one hand without loosing the sensitivity for how these practices work on the other. Or, in McDowells wording, “rational necessitation is not just compatible with freedom but constitutive of it” (McDowell 1996, p. 5). Wittgenstein’s recurrent considerations of children in learning situations serve not only, as Kripke contends, to deconstruct the trust in our ability to follow rules, but also to point out that the learner is in a different situation compared to us who master the practice already well: the child, learning to master a certain practice and to follow some rule within this practice, is not, as we are, in the position to see the various alternative ways to proceed and to choose one of them. This is why we cannot explain to the learner how to carry on according to the rule, since our explanations would be meaningless to her – we can only train her and try to show her how to move on. This is neither to say, that practices are immune from criticism or revision nor that they are, all by themselves, justified, even though human practices have – to some extent at least – their own right stemming from their relevance to our lives. No practice is immune to revision; there might be good reasons to call their relevance into question. Still, if someone, be it a sophist or a sceptic, principally or methodologically casts doubt on the objectivity of well-rehearsed practices out of the blue, she completely drops out of the game.23 Practices of reflection, of which philosophy is the paradigmatic exemplar, draw upon the practices they reflect. It is central to philosophical reflection that it takes its start from a paradox arising in practice.24 Indeed, philosophical reflection cannot absolutely be detached from the practices it deliberately ponders.

23

For an argument as to the incoherence of scepticism on these lines see, e.g., Putnam (2001). 24 As regards philosophy arising out of paradox see Schulthess (2004).

Naturalism and Rule-Following Practices

171

References Bloor, D. (1983): Wittgenstein. A Social Theory of Knowledge, London. – (1997): Wittgenstein, Rules and Institutions, London/New York. Boghossian, P. (2002), “The Rule-Following Considerations”, in: Miller/ Wright (2002), pp. 141-87. Dennett, D.C. (1978). “Quining Qualia”. http://ase.tufts.edu/cogstud/ papers/quinqual.htm, accessed on June 2, 2004. [Originally published in: A.J. Marcel/E. Bisiach (eds.) (1978): Consciousness in Contemporary Science, Oxford] Gadamer, H. G. (1990): Wahrheit und Methode. Grundzüge einer philosphischen Hermeneutik, Tübingen. Goodman, N. (1983): Fact, Fiction and Forecast, Cambridge (MA)/ London. Haack, S. (1993): “The Two Faces of Quine’s Naturalism”, Synthese, 94 (3), pp. 335-356. Hacker, P. M. S. (2001): “On Misunderstanding Wittgenstein”, in: Wittgenstein. Connections and Controversies, Oxford, pp. 268-309. Haugeland, J. (1998): Having Thought. Essays in the Metaphysics of Mind, Cambridge (MA)/London. Kant, I. (1990): Kritik der reinen Vernunft, Hamburg. Kitcher, P. (1992): “The Naturalists Return”, Philosophical Review, 101 (1), pp. 53-114. Kripke, S. A. (1982): Wittgenstein on Rules and Private Language. An Elementary Exposition, Oxford. Kusch, M. (2002): Knowledge by Agreement. The Programme of Communitarian Epistemology, Oxford/New York.

172

Ulvi Doguoglu

McDowell, J. (1996): Mind and World, Cambridge (MA)/London. – (1998a): “Two Sorts of Naturalism”, in: McDowell (1998c), pp. 167-197. – (1998b): Meaning, Knowledge and Reality, Cambridge (MA)/London. – (1998c). Mind, Value, and Reality, Cambridge (MA)/London. Miller, A./Wright, C. (eds.) (2002): Rule-Following and Meaning, Chesham. Pettit, P. (2002): Rules, Reasons, and Norms, Oxford/New York. Plato (1973): Theaetetus, transl. by J. McDowell, Oxford. Putnam, H. (1975): “The Meaning of ‘Meaning’”, in: Putnam, H.: Mind, Language, and Reality, Cambridge/New York, pp. 215-271. – (1981): Reason, Truth, and History, Cambridge/New York. – (1994a): “Realism without Absolutes”, in: Putnam (1994c), pp. 279-294. – (1994b): “Model Theory and the Factuality of Semantics”, in: Putnam (1994c), pp. 351-375. – (1994c): Words and Life, Cambridge (MA)/London. – (2001): “Skepticism, Stroud and the Contextuality of Knowledge”, Philosophical Explorations, 4, pp. 2-16. – (1999): The Threefold Cord. Mind, Body, and World, New York. – (2004): Ethics without Ontology, Cambridge (MA)/London. Quine, W. V. (1960): Word and Object, Cambridge (MA). – (1969): Ontological Relativity and Other Essays, New York. – (1980): “Two Dogmas of Empiricism”, in: Quine, W.V.: From a Logical Point of View. 9 Logico-Philosophical Essays, Cambridge (MA)/London, pp. 20-46 – (1992): “Structure and Nature”, Journal of Philosophy, 89, pp. 5-9. – (1995): From Stimulus to Science, Cambridge (MA)/London.

Naturalism and Rule-Following Practices

173

Rea, M. C. (2002): World without Design. The Ontological Consequences of Naturalism, Oxford. Rosenberg, A. (1996): “A Field Guide to Recent Species of Naturalism”, British Journal for the Philosophy of Science, 47 (1), pp. 1-29. Schulthess, P. (2004): “Zum Proprium der scholastischen logica moderna gegenüber Antike und Moderne”, Internationale Zeitschrift für Philosophie (forthcoming). Sellars, W. (1997): Empiricism and the Philosophy of Mind, Cambridge (MA)/London. Shook, J. R. (ed.) (2003): Pragmatic Naturalism & Realism, Amherst. Williams, M. (1999): Wittgenstein, Mind and Meaning. Towards a Social Conception of Mind, London/New York. Wittgenstein, L. (1958): Philosophische Untersuchungen - Philosophical Investigations, transl. by G.E.M. Anscombe, Oxford.

List of Contributors Norbert Anwander is Lecturer in Ethics at the University of Zurich. David Copp is Professor of Philosophy at the University of Florida. Ulvi Doguoglu is Assistant Professor for Philosophy at the University of Zurich Neil Roughley is Assistant Professor for Philosophy at the University of Constance, Germany. Peter Schaber is Associate Professor of Practical Philosophy at the University of Zurich. Thomas Schmidt is Assistant Professor at the Department of Philosophy of the University of Göttingen. Theo van Willigenburg is Professor of Ethics in the Faculty of Philosophy of Erasmus University Rotterdam.

Name Index Achtziger, A. 69 Aristotle 157, 170 Audi, R. 90 Ayer, A. J. 55, 131 Baldwin, Th. 89–90 Blackburn, S. 24, 50, 52, 58, 60, 78, 118, 119, 131, 143 Block, N. 58 Bloor, D. 154 Boghossian, P. 152 Boyd, R. N. 87, 130 Brink, D. 87, 92, 105, 107, 130 Carnap, R. 64 Castañeda, H. 62 Copp, D. 61, 62, 78, 88, 90, 108 Crisp, R. 93, 108 Dancy, J. 25, 93, 98, 102, 114 Danto, A. 52 Darwall, S. 17, 20, 31, 36 Davidson, D. 66 Dennett, D. C. 155 Dworkin, R. 132 Fine, K. 101 Foot, Ph. 14, 20, 29 Frankfurt, H. G 146 Gadamer, H. G. 161, 170 Gaut, B. 113–114 Gerrard, E. 126 Gibbard, A. 10, 11, 24–25, 78, 131 Gollwitzer, P. 69 Goodman, N. 155 Haack, S. 166 Hacker, P. M. S. 152 Hampton, J. 89, 101 Hare, R. M. 7, 24, 54, 58, 60–61, 70, 126, 131 Harman, G. 65 Haugeland, J. 161, 165–168 Hubin, D. C. 33 Hume, D. 59–60, 70

Jackson, F. 11, 58–60, 91, 98, 99, 109, 110, 117, 135–137, 142–143 Johnston, M. 141 Kant, I. 14, 30, 35–36, 49–50, 56–57, 73, 75, 136, 148, 157 Kenny, A. 64 King, J. C. 11 Kitcher, P. 158 Korsgaard, C. M. 17, 30, 36–39, 101–102, 132 Kripke, S. A. 151–170 Kusch, M. 154 Lewin, R. 68 Lewis, D. 58, 70, 135 Lovibond, S. 136 Mackie, J. L. 9, 80, 84, 88, 103, 121–132 McDowell, J. 93, 133, 136, 142, 151–152, 159, 161, 165, 168–170 McNaughton, D. 98, 126 Miller, A. 152 Moore, G. E. 9, 13, 49–50, 53–54, 75, 89–90, 95–96, 105–108, 124. Nagel, T. 36 Papineau 52 Parfit, D. 99–101, 115 Pettit, P. 59, 154 Pinker, S. 91 Plato 21, 49–50, 75, 125, 151–152, 156 Protagoras 156–157 Putnam, H. 58, 105, 123, 156–157, 160–162, 164, 170 Quine, W. V. 52–54, 151, 155–161, 168 Rachels, J. 88 Railton, P. 144–145, 147 Rawling, P. 98 Rea, M. C. 158 Rosenberg, A. 158 Savage–Rumbaugh, S. 68 Sayre–McCord, G. 114 Scanlon, Th. 111–112, 121

Name Index Schaber, P. 80, 87, 102 Schelling, T. C. 40 Schulthess, P. 170 Seebass, G. 64 Sellars, W. 156 Shafer–Landau, R. 87, 92, 110, 117, 129 Shook, J, R. 158 Smith, M. A. 10–11, 17, 34–35, 54, 58–59, 139, 142 Stemmer, P. 65

177

Stevenson, C. L. 24, 131 Sturgeon, N. 11–12, 89 Tresan, J. 10, 18, 43 Tugendhat, E. 70 Wiggins, D. 93, 142 Williams, B. 66, 70–71, 106 Williams, M. 154 Wittgenstein, L. 130, 151–156, 158, 160, 162–163, 170 Wright, C. 152

Subject Index antirealism moral 121–123 authoritative normativity 96 authoritative reasons proposal 20–21, 30–33 behaviouri 58 behaviourism 54, 156 belief 64 buck passing view 111 categorical and hypothetical imperatives 14, 19–21, 29–41 closed question proposal 20–21, 30, 32–33 cognitivist 95 descriptivism 24 disagreement moral 123–124 enabling condition 114 etiquette 14–16, 31, 61 expectations 166–169 explanation 160, 162, 169 explanatory relevance, counterfactual test of 56, 113 expressivism 24–25, 47, 76–82 externalism moral 87, 98 semantic 162 fact moral 87, 121, 123, 129, 131 normative 56, 87, 99, 100–103, 115, 156 social 73 fallibilism 159, 170 functionalism “moral functionalism” 59–61 in philosophy of mind 57–58 generic normativity 14–16, 25–27 standard-based account 26

identity 37, 89, 92 indeterminacy of meaning 155–156 induction, new riddle of 155 inscrutability of reference 157 internalism motivational 17, 125–128, 131 normative 17, 126, 129 interpretationism 151 intersubjectivity 75–78, 154–156 intuitionism 65–68, 124–125 judgement moral 55–56, 76, 79–81, 121, 126–127, 129–131, 133 normative and non-normative 132 knowledge moral 124–126 language private 152–154, 163 social character of 161 learning 165–166 metaphysics 62–65, 87, 93, 100–101, 103, 121, 125, 131–132 mixed worlds 117 motivating reason 112, 169 motivation 16–19, 27 moral 126–127 motivational normativity 14, 16–20, 22, 25–34 myth of Gyges 21–22, 39–40 naturalism, characterization of 156–160 necessity 90, 92, 101–102 non-naturalism ethical 89, 103, 109 noncognitivism 131 normatively relevant fact 115 normative reason 101–102, 112 open question argument 9, 90, 105 optative attitudinising 48, 64–73, 76–83

Subject Index paradox, sceptical 153–154, 156 physicalism 135 practical question 70–71, 73, 82 private language. See language, private property aesthetic 88, 139 deontic 141–142 descriptive 59–62, 79, 109–110, 135–137 evaluative 106, 141–142 moral 9–13, 23–26, 87, 89, 92–93, 95, 100, 135, 139 natural 9, 11–13, 25, 38–39, 53–54, 89–90, 105, 108–109, 135 non-natural 49, 66, 75, 89, 106, 115 secondary quality 139 prudence 14 psychology 51–54, 68–69, 82–83 queerness, argument from 9, 88, 103, 124–126, 128–130 rationality self-grounded 14, 33–34 rational agency 10, 28, 30, 33–34 realism metaphysical 156–157 moral 11–12, 15–16, 23–24, 26, 77–78, 87, 95, 118–119, 121–125 naturalist 118 naturalist and normative 128–131 scientific 157–158 realist-expressivism 24–25

179

reasons 126, 128 authoritative 19–22, 28–32, 34 kinds of 15 moral 132 motivating. See motivating reason normative. See normative reason overriding 21, 40–41 reductionism 10–11, 82, 89, 90–100 relativity, argument from 122–123 response-dependence 140 rule-following normativity of 151 objectivity of 151–154, 157, 159–162, 170 social character of 154–155, 160–162 sanctions 63–64, 73 scepticism 160, 164 semantic externalism. See externalism, semantic shapelessness, the problem of 93–96 social norms 48, 61–65, 72–73, 74–76, 79–80 speech acts 10 standards 48, 61–79 structuralism 157–158, 165 supervenience 57, 60–61, 91, 107 thick terms 106, 142 thin terms 106 values moral 88, 121–126, 128 why be moral? 20–21, 29, 37, 40–41